U.S. ENVIRONMENTAL PROTECTION AGENCY

SCIENCE ADVISORY BOARD STAFF OFFICE

CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE (CASAC) OXIDES OF NITROGEN
(NOx)PRIMARY REVIEW PANEL

PUBLIC MEETING

MARRIOTT AT RESEARCH TRIANGLE PARK

4700 Guardian Drive

Durham, North Carolina 27703

MAY 2, 2008



2

1	U.S. ENVIRONMENTAL PROTECTION AGENCY

2	SCIENCE ADVISORY BOARD STAFF OFFICE

3		CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE (CASAC)

4		OXIDES OF NITROGEN (NOx)PRIMARY REVIEW PANEL

5	May 2, 2008

6	DR. HENDERSON:	Welcome, everybody, back

7  for our second day.  We think we're in pretty good

8  shape time-wise, we finished our air quality discussion

9  for the REA yesterday, so I think we can, excuse me,

10   start with a discussion of the exposure analysis

11   section.  It looks like we may have time, as we go

12   through, we want to discuss this, all of this in great

13   depth, but later in the morning we may have time to

14   start writing up some of our consensus replies to the

15   Administrator, so you might keep that in mind.

16	But let's start off with our discussion

17   of the exposure analysis section and Pat Kinney, who

18   just joined us, will lead, hi, Pat, will lead us off.

19	DR. KINNEY:	Okay, get right into it

20   here.  I wasn't sure how far you'd gotten yesterday,

21   'cause I was teaching yesterday afternoon. I'm sorry I

22   wasn't able to be here in person yesterday or even on

23   the phone for very long.  So my understanding, my

24   assumption is that we're talking about Chapter 7 of the	

4

1  checked out some rates for Philadelphia, I think the

2  rates that were used, which came out to be around 17

3  percent for childhood asthma, is a pretty good number

4  from what I could tell, but it raised that sort of

5  general generic issue of how specific, how

6  geographically specific the data are that are used in

7  the exposure assessment.  I think it applies to other

8  aspects of it as well, not just, not just the asthma

9  question.

10	You know, I'm thinking about sort of the

11   fine field geography of a place like Philadelphia and I

12   know that the economic and racial patterns are pretty

13   distinct there like they are in many other cities, and

14   I just wanted to make sure that the staff who are doing

15   the risk assessment keep in mind potential unique

16   aspects of the inner city that may impact on their

17   exposures and health risks due to NO2.  Not only, you

18   know, are asthma prevalence rates greater in some inner

19   city neighborhoods but also there might tend to be

20   simultaneously higher than average exposures to traffic

21   emissions on roadways.  I know from my experience in

22   New York that people that live in the city spend a lot

23   more time commuting by foot and along roadways, and to

24   the extent that those sidewalk commutes are not

25   accounted for in the exposure assessment, that is a



3

1	DR. HENDERSON:	It's surely Chapters 5

2  and 7, and I think answering the charge questions under

3  the--

4	DR. KINNEY:	Okay, so I'll go through

5  my chart the way  I've answered those charge questions.

6  Overall, as I say in my written comments, I thought

7  this was a really good effort by, by staff at EPA.

8  Very comprehensive, technically sound, the writing is

9  very, is very clear, and the interpretation of the

10   results I think are quite good.  There are of course,

11   you know, some areas where any critical reviewer will

12   find areas for improvement and I certainly did as well.

13   One of the fundamental issues in terms of the, I guess

14   this comment mainly focuses on Chapter 7, and the order

15   of my comments was sort of structured around the charge

16   questions rather than sort of being sequentially going

17   through the document.

18	So the issue of focusing on asthmatics

19   as a sensitive subpopulation I thought was very

20   reasonable, given what we know about health effects.

21   In my initial review I was not clear about whether the

22   asthma prevalence that was used for the exposure

23   assessment work was based on sort of Philadelphia inner

24   city asthma rates, I went and when I dug into the

25   technical support document and also went online and	

5

1  source of potential underestimate of some of the

2  extreme exposures.

3	So if it can be done, it should be, and

4  if it can't be done it should just be discussed as an

5  uncertainty that, you know, goes at the end of the

6  document.

7	Another thing, along those lines is air

8  conditioning prevalence,  which I wasn't clear from the

9  discussion whether sort of average regular air

10   conditioning were used or whether, you know, sort of

11   urban specific rates, inner city specific and income

12   specific rates, where you just, 'cause that'll impact

13   the indoor/outdoor relationship, and it may be the

14   indoor/outdoor relationship would be somewhat stronger

15   in the inner city if air conditioning is not as

16   prevalent.

17	So that sort of stuff, the kind of

18   questions where we try to really understand all these

19   exposure determinants and also risk determinants on a

20   finer SES, racial scale in a city I think is important

21   to think about, either to directly model it or else

22   just discuss it as an uncertainty.

23	The question of modeling of the air

24   pollution concentrations based on stationary and mobile

25   sources I thought was very well done for the most part



6

1  and appropriate, using the best available models.  I

2  think that I was a little, I had a question about the

3  reliance for the mobile source emissions only on major

4  roadways as opposed to all roadways, and I'd like to at

5  least know in the, in the main document to what extent

6  that, the omission of the non-major roads results in

7  potentially any downward bias or exposure for the,

8  again, in the urban corridor where there may be lots of

9  roads where people drive but maybe they're not, they're

10   not called major roadways, so I just want to make sure

11   those are accounted for.  Or if they're not, give us

12   some discussion about, you know, how much, how big of

13   an issue that might be.

14	In terms of microenvironments that were

15   chosen for the apex modeling, I think they make good

16   sense. I was unclear about and I think in looking more

17   carefully at the technical support document, I think it

18   became more clear that pedestrian, a pedestrian

19   microenvironment I don't think was explicitly modeled,

20   and again, going back to what I was saying before, the,

21   you know, I'm thinking about kids walking from their

22   apartment to their school or taking the bus to their

23   school, I don't think it's fair to necessarily just

24   assume that the exposures they receive during those

25   commuting times are the same as the home exposures,	

8

1  7 or section 7 and the same is true I think at the

2  beginning of section 5  and section 6, that we need to

3  have, particularly section 6, we need to have sort of

4  a, more of a background, a conceptual discussion about

5  what, what the context is for the analysis that's going

6  to be presented.  It sort of jumps right into

7  methodology without, without really helping the reader

8  understand how these, these two major analyses, the one

9  in chapter 6 and the one in section 7, how they relate

10   to the overall question of assessing risks of NO2.  So

11   we need the context, the rationale for the particular

12   approach and then in the objectives of that and then go

13   into the method, which is, which is very well

14   described.

15	So I think that's it for now.  The rest

16   are just detailed comments.

17	DR. HENDERSON:    Thank you, Pat.  Is Doug

18   Crawford-Brown on the phone today?

19	DR. CRAWFORD-BROWN:    Yes. I'm here.

20	DR. HENDERSON:    Great, Doug.  It's time

21   for your comments on the exposure analysis chapters.

22	DR. CRAWFORD-BROWN:    Okay, give me just

23   a second here, I got to open up that file, I closed it

24   down a second ago.  The first thing, while I'm getting

25   it open here, the first thing I'll say is that I



7

1  both because they're at lower elevation and they're

2  also more likely to be along roadways, where cars are

3  emitting NO2, NOx and NO2.

4	I think the uncertainty assessment was

5  quite, quite good as a first effort, I'm sure that that

6  section will expand as, as more and more comments are

7  received, but I thought it was a really good first

8  effort to think about comprehensively all the

9  uncertainties that are present in the analysis.

10	I concurred with the staff decision to

11   base the risk assessment work on the chamber, human

12   chamber study results for choosing the benchmarks.  I

13   thought there was a good discussion about the

14   epidemiologic work. I tend to agree that for this kind

15   of analysis and the, because of the problems with

16   co-pollutant confounding in the epidemiologic studies

17   that it's really hard to figure out how to isolate an

18   NO2 effect from those studies.  And so although I'm

19   certainly interested in hearing what others think about

20   that, but I, so I concur with the decision that was

21   made to, to stick to the chamber results in this

22   particular part of the document.

23	Then I think beyond that it's largely a

24   series of specific comments, there is one general

25   comment that I think occurs at the beginning of chapter	

9

1  actually found it to be a well described plan if you

2  have knowledge of what is a relevant ambient

3  concentration to use.  So I think what I'll focus on

4  here is that if we know the ambient concentration, then

5  it's possible through the plan that they've laid out to

6  get a good estimate of the individual exposure and the

7  whole population exposure.  The problem that obviously

8  arises is, and I don't know how they're going to deal

9  with this, the difficulty of figuring out exactly what

10   that ambient level is.  I saw two ways of doing it in

11   there.

12	One way is the, the use of the

13   monitoring results, the other way is the use of this

14   sort of first principle approach, where you actually

15   have source streams, inventories, and dispersion and so

16   forth.  And when I originally read through, I was, as I

17   mention in my comments, I was originally thinking that

18   the monitoring results became a kind of input into the

19   more person specific variability analysis that was

20   done, it was pretty clear to me after about two-thirds

21   of the document that I was mistaken about that, that

22   the monitoring results are an alternative way, as far

23   as I can tell, an alternative way of getting at the

24   exposures of the population based solely on ambient

25   exposures.



10

1	So one thing I'll ask the EPA is just

2  to, if they can just make sure that I'm absolutely

3  right about that, that the monitoring results are one

4  way of doing it , and the second way, which does the

5  individual exposures

6  is instead based entirely on first principles, but in

7  any event, I had sort of thought that those two might

8  have been at least compared against each other, or the

9  recommendation was that they would be compared against

10   each other and perhaps the comparison might be used in

11   a kind of uncertainty analysis.

12	DR. HENDERSON:    Would you like for the

13   EPA to --

14	DR. CRAWFORD-BROWN:    And I didn't see

15   that.

16	DR. HENDERSON:    Doug, would you like for

17   the...

18	DR. CRAWFORD-BROWN:	Let me stop with

19   that and just make sure, first, that I've understood

20   correctly that those are two alternative approaches.

21	DR. HENDERSON:	Okay, Doug, they, they,

22   they will tell you.

23	DR. GRAHAM:    Thanks, Doug.  Yes, I could

24   say they were two alternative approaches, but there was

25   overlap with the monitor information, the monitors that	

12

1  correct that there literally are in the end two sets of

2  results that will come out, one will be a kind of

3  natural exposure scenario based on the monitoring

4  results and the other will be for the test cities or

5  whatever we would call them, a much more detailed

6  inter-subject variability distribution?

7	DR. GRAHAM:    Are you referring to the

8  comparison between the air quality analysis and the

9  exposure analysis?

10	DR. CRAWFORD-BROWN:    Yes.  Yeah, I was,

11   I wasn't -- my understanding was that, when I read it,

12   it sounded like the air quality analysis was an

13   alternative way of asking the kind of risk-based

14   question, right?  In the, in the air quality analysis,

15   I think what's being proposed is that there's a kind of

16   margin of exposure approach being used, where you look

17   at what fraction of the population is above some

18   ambient air concentration.

19	DR. GRAHAM:    Right, but in the air

20   quality analysis, there is no relationship between the

21   population that are, say, surrounding that particular

22   monitor, at least with the data that are presented

23   right now.

24	DR. CRAWFORD-BROWN:	Then I guess then

25   I didn't understand in the decision making process how



11

1  were within the Philadelphia county for the exposure

2  analysis, the more complex analysis, those monitor data

3  were used to adjust for local concentrations in the

4  dispersion model, air, air concentrations.

5	DR. CRAWFORD-BROWN:    In the sense of

6  being, of having the model results normalized against

7  the monitoring results?

8	DR. GRAHAM:    Right, that's correct.

9	DR. CRAWFORD-BROWN:    Okay, and, and is

10   there reasonable comfort that the monitoring results

11   are good estimates of ambient exposures for people,

12   because I thought we had this issue of the monitors not

13   necessarily being located at places that were

14   representative of actual exposures to individuals.  Oh,

15   I guess if you're going to do the normalization, it

16   doesn't matter whether the monitoring results are

17   representative or not, you're just mo -- you're, you're

18   adjusting the model to fit the monitoring results, but

19   you're not using those locations as the places where

20   exposures are getting estimated.  Is that right?

21	DR. GRAHAM:	That's right, but if there

22   were individuals that resided in that centroid, the

23   census block centroid, then those, those would be used.

24	DR. CRAWFORD-BROWN:	Okay.  Okay, so

25   I'm relatively comfortable with that.  Am I also	

13

1  the chapter, what was it, I don't have it in front of

2  me here, how the chapter on air quality monitoring

3  results were going to be used to compare against the

4  inter-subject variability result.

5	It's, my understanding was that the

6  inter-subject variability result was going to be used

7  to ask what fraction of the population has a personal

8  exposure above some threshold value we'll call it, or

9  some sort of benchmark value. And, but I was reading it

10   and thinking that a similar thing was being proposed

11   for the air quality one, but instead of comparing it

12   against an actual personal exposure, there was going to

13   be a comparison against an ambient exposure level. But

14   maybe I'm wrong about that.

15	MR. RICHMOND:    This is Harvey Richmond.

16   The air quality analysis is not intended to address the

17   population -- it's not an estimate of population

18   exposure given of course all the time spent indoors,

19   given indoor sources, you know, given the different

20   location of people away from the monitors or on the

21   road, so I don't, you don't think we view those as

22   alternatives that should be compared.  I think there

23   was some discussion whether or not you wanted to

24   evaluate parts of the exposure model just in terms of

25   the air quality characterization, against the air



14

1  quality characterization, you know, in those two

2  elements, that's one thing, but we're not trying to

3  compare or say they should be compared, they shouldn't,

4  between the air quality characteri

5	DR. CRAWFORD-BROWN:	Oh, okay, I mean,

6  it may just be my reading of the document that at some

7  point I began to think that the, that there really were

8  two different ways in which you were going to go about

9  assessing the, the, the need for further controls, the

10   degree of exposure to the general population, and that

11   one was the use of the air quality results, the

12   monitoring results and one was this sort of first

13   principle,

14   but what I'm starting to hear now is that really the

15   air quality monitoring results are playing the role of

16   normalizing the more detailed stuff that goes on in the

17   chapter that I reviewed, which is the inter-subject

18   variability chapter.

19	MR. RICHMOND:    Right, well, they're also

20   looking at whether or not just at the monitors alone do

21   you have exceedances of these benchmarks, not answering

22   the question about what population, but it helps us

23   focus on what cities might have the potential just at

24   the monitors alone.

25	DR. CRAWFORD-BROWN:    Okay, well, I'm not	

16

1	So I'm not going to push that any

2  further, I will say that there obviously is a point at

3  which you've shot past the reasonableness of the

4  resolution and this, as laudable as it is, is sort of

5  pushing the envelope on that.  And that's all I would

6  say.

7	DR. HENDERSON:    Thank you, Doug.  Dale?

8	DR. HATTIS:	Yeah, I have considerably

9  more difficulty with several aspects of the analysis.

10   One, and perhaps the most serious one was dealt with

11   yesterday with this adjustment for the changing of the

12   outdoor concentrations, the ambient concentrations,

13   while holding constant the effective amounts of the

14   indoor source-related exposures.   That appears to have

15   been dealt with, although it was certainly inadequately

16   described in the document.  That having been said, you

17   know, I still need to see that really fully fleshed

18   out, you know, in an altered write-up.

19	The second important problem relates to

20   the adjustment of the source plus dispersion model

21   predictions to correspond to the observed data from the

22   air quality monitors, which is what was, we started to

23   talk about yesterday.  The current comparison is based

24   on only three different locations where there were the

25   three available sets of monitoring data, but it was



15

1  going to push this anymore, I guess what I came at it

2  from the angle of a decision that was being made, I

3  just wasn't clear what role those two kinds of pieces

4  of information were playing in the decision, but I

5  guess what you're saying is you will provide that to

6  the policymakers, the decision makers, and it's up to

7  them to decide how to weight those two pieces of

8  information in in setting a sort of final standard

9  there.  So that makes me comfortable then.

10	The last thing I'll say is that, in a

11   remark that's similar to something I've made in some

12   other contexts on CASAC, I begin to wonder whether the

13   level of detail on inter-subject variability

14   distributions, as laudable as it is and as much as I

15   like running these sorts of models, shoots beyond our

16   ability to really make good estimates of these

17   inter-subject variability distributions.  So there's,

18   there's one part of me that says, well, yeah, just do

19   absolutely the best job you possibly can and generate a

20   whole inter-subject variability distribution, but

21   there's another part of me that wonders whether we've

22   shot past the level of resolution of inter-subject

23   variability by doing all of this, you know, census

24   block, census tract kinds of centroid analyses and then

25   putting in inter-subject variability distributions.	

17

1  based only on long term averages, and that's just not,

2  because you're, you're really trying to model the

3  hourly concentration distribution and the hourly

4  ambient outside levels do drive, although they don't

5  correspond completely, but they do drive the indoor

6  concentrations through the air exchange rates.   It's

7  important that the distribution of hourly outdoor

8  concentrations after adjustment, you know, correspond

9  reasonably, so you, with the, with the predictions for

10   those monitors.  So you really must make the prediction

11   on the basis of the distributions of hourly

12   concentrations at the, the three monitors for which you

13   have observations and show that comparison, otherwise I

14   don't think that anybody should believe that you've got

15   the hourly ambient drivers correct.

16	Secondly, I'm pretty sure that a simple

17   additive adjustment will not be right.  I mean, I

18   expect that because conditions like, you know, the

19   turbulence of the atmosphere and things of that sort

20   likely to be exerting multiplicative influences on

21   concentration.  I bet you that you're going to be

22   better off with a multiplicative mode of adjustment

23   rather than an additive mode of adjustment of the

24   difference between the, whatever remaining difference

25   there is.  Now an additive mode of adjustment, you



18

1  know, could be expected if you, if you just, just, if

2  you're just adding sources essentially that you hadn't

3  modeled, but I expect that in fact you're going to need

4  to multiply this, this overall distribution.  Anyhow,

5  you'll have to see that by making the comparison,

6  showing the comparison, and showing that after you've

7  made your adjustment, the distribution that you get

8  corresponds to what's at the actual monitors.

9	If there's a huge difference in the

10   variance between your predicted hourly ambient levels

11   and your observed hourly ambient levels, then you need

12   to do something more fancy.  Now I'm not sure exactly

13   what that will be, 'cause I haven't got the comparison

14   in front of me, but clearly you're much more likely to

15   come to grief comparing the two distributions than you

16   are comparing the means, and there's, as we discussed

17   yesterday, there's some, there's some signals of

18   trouble in just comparing the means, because the

19   difference among the monitors is, as predicted by the

20   models is much different than the difference among the

21   models, monitors that's actually observed.

22	Okay, so that's the second issue.  The

23   third issue relates to the same correction in another

24   way, and that is basically when you're adjusting the

25   model predictions to the outdoor ambient	

20

1  height adjustment comparing the expected concentrations

2  at the heights of the human receptors versus the height

3  of the monitors that you're using to calibrate your,

4  your exposure model.  So, so that's a, the third

5  problem essentially.

6	The fourth issue is the representation

7  of a few key sources of variability in the apex

8  exposure modeling, which inadvertently I think

9  artificially truncate likely distributions of

10   variability and certainly uncertainty. Some examples of

11   this, the first of the air exchange distributions are

12   contingent on temperature in the presence or absence of

13   air conditioning. Overall I don't have any objection to

14   the idea of using log normal distributions with very

15   broad limits in this case, .1 to 10 air changes per

16   hour.  However, the detailed results seem to show

17   different patterns with temperature in arbitrarily

18   blocked into a few ranges and I think you, there

19   doesn't seem to be any great consistency in the

20   observations that I saw , or a theory to say what it

21   should be like.  So a better description of the data as

22   a whole I expect might be produced by a more extensive

23   regression study using temperature and some transformer

24   of temperature as a continuous variable, either in a

25   fixed effects or a mixed effects modeling of



19

1  concentrations, you're essentially using the heights of

2  the outdoor monitors to adjust everything to, and what

3  you want to do is to make predictions at the heights of

4  the human receptors, which you have said, I think

5  reasonably are 1.8 meters above the ground.

6	So by not, by, by essentially correcting

7  the model predictions to the, to the monitor

8  observations, you haven't got the height adjustment

9  right.  So I would have you make the height, add in the

10   height adjustment for the particular monitors that

11   you've got, and that means you can't completely dismiss

12   the, I mean, the, the, currently the discussion in the

13   document, you know, makes a comparison that says, well,

14   not, most of the monitors aren't 15 meters, they're

15   mostly four or five meters and that's not such a big

16   difference.  I think you need to model that difference.

17   I think you need to model, and basically there's two

18   kinds of models that I, I try to use in preliminary

19   calculations, I try to use either a simple exponential

20   decline with height or more like a gaussian decline

21   with height, and, you know, your model itself probably

22   has vertical distance in it, so you might even have a

23   better way of making that adjustment from your aeromod

24   than I would just from first principles.

25	So anyhow, you do need to make that	

21

1  differences among cities for the air conditioner

2  predictor variable.

3	Secondly, you've got a, NO2 removal rate

4  distribution, and I just completely object to the

5  narrow fixed limits used this removal rate

6  distribution, based on 6 values from a study in 1993 by

7  Schleisser.  The abstract of Schlesser's paper makes it

8  clear that all 6 observations were made in a single

9  house, and that there are additional complications from

10   the presence of HONO, an apparently lived NOX species.

11   So you've got to, basically, you just can't do a

12   uniform - between the limits - of those 6 observations

13   collected from a single house, you've got to expand

14   that distribution beyond that, even with a log normal

15   or more, or a broader distribution, depending on some

16   expert opinions- as to how broad, likely, the overall

17   distribution among real houses is likely to be.  Houses

18   are likely different in their contents, and they are

19   going to absorb at NO2 at different rates, and that

20   variability is totally missed in your current

21   representation of it.

22	The uniform distribution with its fixed

23   boundaries is particularly inappropriate, or it's

24   limited, as you can see.  I also strongly object to the

25   use of the uniform distributions of the concentrations



22

1  of NO2 from gas stoves.

2	The very breadth of the bounds were

3  derived coordinate, 188 parts per billion, argues

4  against the uniform distribution in favor of something

5  much more skewed, like a log normal.  Log normal

6  guarantees a positive contribution, but it doesn't have

7  the unfortunate property of implying a zero chance that

8  the things are outside those limits.  Also, if you are

9  in fact doing a, some sort of a mass balance type

10   model, then you should be entering this contribution

11   not in the form of a concentration, but in the form of

12   mass units. So you should have a certain number of

13   molls of Nitrogen Oxide produced per your cooking

14   events.  Because the concentration implications will

15   differ according to variability of your house,

16   characteristics and your air exchange rates and all

17   that, and that's not, you know, necessarily represented

18   by this range that you've got here.

19	So, that's, and finally, the assumption

20   that all the cooking events, contributed to the indoor

21   NO2, last exactly one hour, which appears to be what

22   you assume, I think, also, artificially limits the

23   variability of the NO2 inputs, and therefore, exposures

24   that are represented in the model.  So, if you're going

25   to do this presence type calculation, which I,	

24

1  was excluded because it was incomplete, it was out of

2  the calculation.  And yet, the exceedances are going to

3  depend quite a bit on the monitor siting.  And, so I

4  just wasn't sure we could make enough sense out of

5  that.  Another problem, which I think you stated pretty

6  clearly earlier, is that the monitor distribution

7  doesn't really represent the population distribution.

8  And so, the counts don't, necessarily, align that well

9  with the population.  And, so I have a lot of trouble

10   with that whole calculation, and I'd like to see

11   something done that, at least, directly acknowledges

12   that the major differences in monitor siting

13   characteristics in counting exceedances, and how that

14   aligns with the population in some kind of a gross

15   sense.

16	With the apex modeling section, you

17   know, you all have been working with that model for a

18   long time, and you've made a lot of progress with it,

19   and I think that's really great.  There's obviously a

20   lot of assumptions in that, and the devil's in the

21   details, but it does look quite thorough, and it looks

22   like you've thought about a lot of the important

23   sources of variability, and the incorporation of the

24   day-to-day correlation of activities with individuals

25   over time, I think, is a really good new enhancement.



23

1  actually, counseled against in favor of a concentration

2  response function the last time we met, but if you're

3  going to do this, it's clear that the, you've got to

4  try to model the extremely high concentration tail of

5  the distribution as well as you possibly can.  And you

6  mustn't artificially limit the variability that's built

7  into the model in the ways that you've done so far.

8  That having been said, the adjustments are not so

9  difficult as to use some alternative distributions as,

10   you know, might first be implied.  I mean, so, 'cause a

11   lot of these things can be just changed as parameters

12   of the model, but you do need to rerun the model with

13   some altered sets of parameters described in these

14   variability distributions for the indoor exposures.

15	DR. HENDERSON:   Okay, thank you, Dale,

16   and Lianne?

17	DR. SHEPPARD:   So, as I mentioned

18   yesterday, my biggest concern is with variability,

19   because you're counting exceedances.  In, was it,

20   chapter five, the direct air quality analysis section,

21   where you're looking at monitors.  It appeared to me

22   that the monitors were treated as completely, or the

23   monitor years were treated as completely exchangeable,

24   so it didn't really matter what year, or whether the

25   monitor was near a road or not.  If, for instance, it	

25

1  And, particularly important, when you're counting

2  exceedances and wanting to look within person.  So,

3  that's a good addition.

4	I had problems with the air mod

5  predictions.  Like Dale, I think that the simple

6  comparison of the means with the monitor means is

7  completely inadequate.  Here we're calcula-, we're

8  looking at the exceedances and relying on the

9  distribution, and then particularly, the tails of the

10   distribution, and getting the means to align is not,

11   necessarily, going to get the tails right at all, and

12   in particular, predictions tend to be a whole lot

13   smoother than the data in general.  And so, there's a

14   very high chance that you're under-predicting the

15   variability, and therefore, getting the tails wrong.

16   So, and then, we heard yesterday that several important

17   sources of variability of NO2 were left out,

18   suggesting, again, the variability's not being

19   captured.  So, I think the, that it's, there's a very

20   good chance we're undercounting because of the way the

21   modeling's being done, and that needs to be addressed

22   head on.

23	I had a couple other less important

24   comments.  I wasn't as clear about the diurnal cooking

25   pattern description.  And I was a little bit worried



26

1  that that smooths the exposure too much.  Let's see.

2  And this is, I think for clarity, I think it's

3  reasonable to say that exposures for asthmatics and

4  non-asthmatics are the same, state that up front, and

5  then, a lot of the language in 7.9 will be clearer, I

6  think, assuming that that's an assumption that people

7  feel comfortable with, stating explicitly.  So, I'd be

8  interested in your comments about the variability,

9  'cause I think that is crucial to counting exceedances.

10	DR. HENDERSON:   Okay, let's let everyone

11   make, anybody else who wants to comment on - -

12	DR. SHEPPARD:   Can we hear from, can we

13   hear from EPA first?

14	DR. HENDERSON:   That's, okay, you want to

15   hear that now, okay.

16	DR. GRAHAM:   And you mean specifically on

17   the air mod, the modification with the monitors and the

18   adjustment, is that what you mean?

19	DR. SHEPPARD:   I think that's a big piece

20   of it.  I'm also worried about the air quality analysis

21   as well with treating the monitors exchangeably, and

22   not representing population exposure, but the, so I

23   think that that, the variability of those monitors does

24   not necessarily align at all with the population

25   exposure, and that could be a problem.  And, also, from	

28

1	DR. SHEPPARD:   But within those time

2  periods, there's not really a distinction about whether

3  how many monitors are included or excluded based on

4  siting characteristics.  So, I guess, at a minimum, I'd

5  like to see that distinction be made, because I think

6  that that matters.  The ones near roads are going to

7  have more exceedances than the ones that are not near

8  roads.  And in some ways, it's captured by the time

9  periods, because there were different, the sitings were

10   different in the different time periods.

11	DR. GRAHAM:   Yeah, in the TSD, I did

12   provide a table as an example for Boston, where we had,

13   whether the monitor was operating at a given year, as

14   well as whether it met the completeness criteria.  And

15   you can see that it is variable over time.  And

16   interestingly, I was looking at it the other day, and I

17   saw that, in fact, more monitors were determined

18   incomplete in the more recent time period than

19   previously, but consistently, there was probably about

20   thirteen monitors within any given year.

21	DR. HENDERSON:   So, Lianne, do you want

22   some more answers?

23	DR. SHEPPARD:   I think we need to hear

24   about the air mod, too.

25	DR. RUSSELL:   Yeah, particularly about



27

1  year to year, and city to city, which monitors are

2  sited near roads will matter in terms of what the

3  counts look like.  And that's all buried in this

4  analysis in a way that it's not clear.  But, yeah, but

5  the key question about the variability is really the

6  air mod.

7	DR. GRAHAM:   Right, but as far as the air

8  quality, I recognize that we did have, well, what we

9  did, I'll explain that first.  I mean, we broke it up

10   into these two five-year periods, recognizing that, of

11   course, there are going to be monitors that are

12   operating, or, perhaps, they didn't meet that valid

13   year criteria, so they would be, perhaps, in one year

14   and out the next year, which is why we grouped them

15   into these five-year periods, so we could, at least,

16   capture, or represent the area fairly well.  And that's

17   why the site years were considered the unit, rather

18   than, okay, well, this year, we had a different set of

19   monitors.  And then, we had this many exceedances, and

20   then the next year, we had possibly a different set of

21   monitors.  So, it was, really, to try and characterize

22   a time period in the region better than, you know,

23   doing it sequentially, because then we'd have less

24   monitors for each given year.  So, is that sufficient

25   for the air quality?	

29

1  the variability.

2	DR. GRAHAM:   Yeah, and if you could

3  reframe the question, please.

4	DR. SHEPPARD:   So, when, in general, when

5  you predict, your predictions are smoother than the

6  data.  And you made an adjustment based on annual

7  average.  And your counting exceedances, which is

8  focusing on the tail of the distribution.  So, if you

9  don't get the whole distribution right, you're not

10   going to count the exceedances right, because you're

11   not, because the exceedances are up in the upper tail.

12   And so, the table 26, that compares the means, is not

13   targeting the aspect, I mean, it's an important aspect

14   of the distribution, but it's completely insufficient

15   for getting at the question about whether your

16   predictions are capturing the quantity of interest,

17   which is being able to calculate exceedances.

18	DR. ROSENBAUM:   Well, actually, we did

19   look a little bit at the comparing the variances, and

20   because we didn't model all of the sources, we only

21   modeled the on-road, the major roads and some of the

22   point sources, actually, the model variance is higher

23   than the monitor variance, because we applied a

24   temporal profile to the on-road emissions.  So, we get

25   this variant, we have an implied hour-to-hour variance



30

1  on the on-road emissions.  So, it, actually, it turned

2  out that, like I said, that the variance was higher

3  with the model results than the monitor results before

4  we made the adjustment.  But your point is well taken.

5  I mean, several people have said that, have pointed out

6  that it was a very simple adjustment we made for the

7  sources that weren't modeled.  And we should probably

8  go back and look at that, and we'll go look at that and

9  see if we can make a better adjustment that's

10   stratified in some way, either regionally or temporally

11   stratified to reflect that better, and to try to match

12   the variance better.

13	DR. SHEPPARD:   Yeah, I think it's

14   important to, when the summaries that show that you've

15   done a reasonable job need to talk about more features

16   of the distribution than just the mean, and

17   particularly, need to focus on that upper tail that's

18   so important.

19	DR. KINNEY:   One simple thing, this is

20   Pat.  One simple thing you can do, of course, is just

21   look at the full, you know, plot histograms of

22   concentrations for the observed and the predicted for

23   the given site over a year, that, of course, masks

24   what's really contributing to the variability.  So,

25   another thing you can do, in addition, is do some	

32

1  buses to school.  We had personal monitors on them.

2  And we were monitoring fine particles, not NO2, but we

3  also were monitoring NO2 at the central, at the van

4  right there at their school.  And, again, what you see

5  is, and we were also monitoring traffic at the highway

6  right next to the, on the Major Deegan.  And when the

7  traffic was the greatest in the morning, we saw the

8  highest NO2 levels.  In the afternoon, we saw nearly as

9  high traffic levels, but not as high NO2.

10	Really, the highest levels of NO2 and

11   fine particles are in the morning, and that make sense,

12   because the wind speeds are lower, and the mixing

13   height is much lower in the morning.  And so that

14   morning, going to school in the morning is when these

15   kids get a big exposure.  And I think the model should

16   reflect that.  The second, I mean, in terms of ambient

17   exposures, that's a big time period for them.

18	The second thing is, you know, I just

19   want to step back from this an listen.  I've been

20   listening to Dale's comments and the others' comments,

21   and you know, I know that we were told that it's too

22   hard to do the epidemiology, but it seems the

23   challenges in doing these exposure estimates are far

24   beyond the challenges involved in the, doing, applying

25   epidemiology.  And then, of course, you're using data



31

1  representative diurnal plots of the model data and the

2  observed data to give the reader a sense of, you know,

3  how they, how similar they look in amplitude.  And

4  then, you know, people also do Fourier analyses, where

5  they disentangle the daily variance, the weekly

6  variance, and the annual variance, and thing, you know,

7  and are they rough, you know, you might do really well

8  on one part of the variance, but not on the other.  You

9  know, if they all do pretty well, that's very

10   reassuring.

11	DR. HENDERSON:   Okay, any more on that

12   point?  Now, are there any other comments on the

13   exposure analysis, yeah, George?

14	DR. THURSTON:   Well, a couple of things.

15   First of all, I wanted to agree with what Patrick

16   Kinney was saying about the morning exposures of

17   children.  And I think the document contained some

18   results that we published from our Bronx studies

19   showing the differential between the, our van at ground

20   level the measurements at the top of the school

21   buildings when we were doing those studies.

22	But, one of the things we observed, and

23   I'm sure this is documented elsewhere, that some of the

24   biggest exposures the children got was right around the

25   time they're going to school, walking and riding on	

33

1  that don't consider the interactions with particles and

2  the real world, and you're not considering the broad

3  breadth of the population, which you consider in the

4  epidemiology.

5	The epidemiology has so many strengths

6  that it makes it just the most obvious.  And after

7  listening in the last hour, I just don't understand the

8  decision by the staff to not consider the epidemiology,

9  and to do risk assessments.  And, yes, you have

10   co-pollutant models, and, you know, I really feel

11   strongly that the single pollutant models are the way

12   to go for this.  You're supposed to do a conservative

13   analysis.

14	That's a conservative analysis, and you

15   have this, the co-pollutant models that show you that

16   it's robust to the inclusion of those, so it's not some

17   artifact.  I think it's the way to go, and everything

18   I've heard this morning, tells me that, it makes that

19   conviction even stronger.  Thank you.

20	DR. HENDERSON:   Frank?

21	DR. SPEIZER:   Well, I was going to make,

22   somewhat, the same point that George just made, in the

23   sense that I thought we weren't yet prepared to talk

24   about the health effects side of this.  And only

25   because Pat raised the issue that he agreed with staff



34

1  in doing, in not doing the epidemiology, I would argue

2  that we don't leave this hotel until we convince staff

3  that they must do the epidemiology side of the risk

4  assessment.  So, I would put that on the agenda as

5  saying, I am forcefully objecting to Pat's agreeing

6  with staff.  Now, if you want to talk about it now, we

7  can, or we may want to talk about it a little bit

8  later.

9	DR. HENDERSON:   We'll be sure and talk

10   about it later, but I'm glad you made the point.  Now,

11   is there any more, are there any more comments on the

12   exposure analysis?  Yes, Ed?

13	DR. AVOL:   I have two small comments,

14   really, questions for staff.  And that is, with regard

15   to the emission strengths and functions that were

16   listed and specified in the document, there's one case

17   with regard to the selection of the upper air stations,

18   and the choices that were made.  So, for Philadelphia,

19   you used Washington Dulles data; for Phoenix you used

20   Tucson data; for Los Angeles, you used San Diego data;

21   and I guess there's sort of a question because of the

22   distances as to how relevant those selections are.  And

23   I understand, and maybe, it's because that's the only

24   data that was available, but I, so, it makes me worry

25   or wonder, sort of, how good that relationship is over	

36

1  What was the other question, I'm sorry?

2	DR. AVOL:   Freeway speeds.

3	DR. ROSENBAUM:   Oh, the speed, yeah.

4  Actually, we looked at that, and we thought they seemed

5  high, too.  And, in fact, we talked to the people in

6  Philadelphia that we got the data from, and they, also,

7  thought they looked a little high.  But, and that was

8  what the, you know, that came out of the model.  And it

9  was, actually, verified somewhat against a small amount

10   of monitored data of, you know, of traffic.  So, it's

11   just, you know, it's something that's not a

12   well-characterized variable in the model.  But that's

13   what came out of the model, and so that's what we used.

14	DR. HENDERSON:   That doesn't sound good

15   to me because I think the model should reflect reality.

16	DR. GRAHAM:   Well, it's something that

17   could be considered in additional sensitivity analyses

18   to see if it is an important parameter.  You're so old

19   fashion, Rogene.

20	DR. HENDERSON:   I know.  Call me an

21   experimentalist.  Okay, now, are there other things

22   that people want to bring up about the exposure?  I

23   would like to know from the EPA staff over here if we

24   have addressed the questions, the charge questions

25   you've listed here.  There are six of them.  Have we



35

1  those distances.

2	And then, secondly, on table 20, about

3  average calculated speeds that are used in the

4  modeling, coming from Los Angeles, it's hard to imagine

5  that there actually are average speeds of 63 to 65

6  miles an hour on the freeway.  But, that's what's used

7  here, and it, that seems to me to be high.  And so, I

8  just have a question about how and where that comes

9  from.

10	DR. HENDERSON:   Do you want to answer

11   that?

12	DR. ROSENBAUM:   Well, as far as the upper

13   air data, that's like what's available.  So, I mean,

14   that was the clo-, those were the closest available

15   data on the upper air.  I think that, Christian might

16   be able to speak to this better, but I think that the

17   upper air, it's not as variable as the surface data so

18   that it's not, you know, it's better to use something

19   closer, but it's not as serious to use something

20   farther away.  Do agree with that?

21	DR. SEIGNEUR:   Yes, well, I guess, you

22   know, since it's the only thing you have available,

23   there is not much choice then.  I think that's the

24   answer.

25	DR. ROSENBAUM:   Yeah, right, of course.	

37

1  addressed those, are we giving you the information you

2  need?  Is there something that you wanted advice on

3  that we haven't spoken to?

4	DR. ROSENBAUM:   Just one point of

5  clarification, as far as, a couple of people mentioned

6  about the pedestrian commuting.  There is a, we do have

7  a micro-environment, outdoors near roadway, that covers

8  when people are walking.  And that is, that's one of

9  the activities that's listed in the CHAD database.  So,

10   it is covered to the extent that the activity patterns

11   in the CHAD database show people outdoors on the

12   sidewalk.  Now, it's not inner-city specific, or you

13   know, whatever.  It's what, just whatever is in the

14   CHAD database.

15	DR. GRAHAM:   And also, along the same

16   lines, it was the commuting.  It may have been

17   misleading, or shall I say misstated in the document.

18   We do, if, of course, an individual is on a bus, a

19   child, then that person is modeled implicitly in that

20   micro-environment.

21	DR. HENDERSON:   Okay, of the people who

22   are on the phone, are you silent because you don't have

23   questions, or do you, anybody on the phone need to,

24   want to ask questions?

25	DR. BALMES:   This is John Balmes.  I



38

1  don't want to ask a question.  I just want to say that

2  I totally agree with George and Frank that we need to

3  talk about the use of the epidemiology.  And I was just

4  holding my fire till later.

5	DR. HENDERSON:   I understand, okay.

6	DR. GORDON:   And this is John.  I don't

7  have anything to say right now, but I'll join in the

8  discussion later with interest.

9	DR. HENDERSON:   Okay, well, then can we

10   move on to, oops, I'm sorry.

11	DR. AVOL:   Will you have the resources to

12   model all five cities?

13	DR. GRAHAM:   I don't think I can answer

14   that right now.  Resources, I think the limiting factor

15   here is the time, and I could probably safely say, I

16   don't think there's time to do all five cities.

17	DR. WYZGA:   So, do you want some advice

18   as to the cities that you should choose?

19	DR. GRAHAM:   Right, we did ask that, I

20   think.

21	DR. HENDERSON:   Asked for that, yes.

22	DR. GRAHAM:   As far as prioritizing,

23   which would be the next cities to, perhaps, look at,

24   and I think I heard a vote for Los Angeles and Atlanta

25   yesterday.	

40

1  commute in.  And I'm gathering that's not included in

2  the analysis.  You're only analyzing residents, or

3  maybe if you include the outlying counties, then it

4  would be inclusive, but I don't know if you're going to

5  do that in Atlanta.  I just raise that concern about

6  the commuters, because they get some of the biggest

7  exposures.  There's probably a lot of people with

8  asthma who drive in to Atlanta, and so, that's

9  something to keep in mind when choosing Atlanta, and

10   when analyzing Atlanta.

11	DR. RUSSELL:   And if I might, actually,

12   for Los Angeles, as well as Atlanta, as well as

13   Detroit, how much of the urban area are you expecting

14   to get?  I mean, let's start with Los Angeles.  Is it

15   Los Angeles County, and that's it?  And for Atlanta, is

16   it just Fulton County, or not even that, or do you have

17   a real feel for what fraction of the population you're

18   going to capture?

19	DR. GRAHAM:   Right, and I guess this

20   opens up an additional discussion, because time is the

21   limiting factor here.  It becomes a matter of how many

22   receptors are we actually going to model.  I mentioned

23   earlier that Philadelphia, we looked at 17,000

24   receptors, and we're in the process of investigating

25   ways or methods to reduce the number of receptors that



39

1	DR. WYZGA:   I guess one of the questions

2  of Atlanta, Detroit, and Phoenix, is there a difference

3  in the terms of the quantity of air quality data for

4  the three cities?

5	DR. GRAHAM:   Well, I think Los Angeles

6  would have more air quality.

7	DR. WYZGA:   Right, of the three, yeah.

8	DR. GRAHAM:   Right, and Atlanta and

9  Detroit?

10	DR. WYZGA:   And Phoenix, I think were the

11   other three candidates you had.

12	DR. GRAHAM:   Phoenix is probably the

13   least, I think.

14	DR. WYZGA:   The least, okay.

15	DR. GRAHAM:   Yeah, so probably, Detroit

16   and Atlanta.  I think Detroit is fairly limited, too.

17   I don't have the numbers in front of me, but I guess,

18   that would be a good criteria for selection, too.

19	DR. HENDERSON:   Any more?  Yes, George?

20	DR. THURSTON:   Well, I'll just raise

21   again the issue with Atlanta.  I know you're just doing

22   the residents, but I think you'll give the impression

23   that you're doing an analysis for the city of Atlanta,

24   when 60 percent of the people who are there during the

25   day when the NOX levels are highest are commuters who	

41

1  are modeled, while not losing too much information.

2  So, then, in reducing the number of receptors, we can

3  expand the actual modeling domain.  So, for something

4  like Los Angeles, I mean, it would be extremely

5  important, as well as Atlanta, if we are going to

6  extend it outside of a given county, we're going to

7  have to reduce the number of receptors that are

8  actually modeled.

9	DR. ROSENBAUM:   And the other

10   consideration is the, since we're doing the major

11   roadways, is how many roadway links, because that's the

12   number of sources.  So, and then it gets into how you

13   define what's a major roadway.  For Philadelphia, we

14   defined it as 15,000 annual average daily traffic in

15   one direction.  So, I mean, there's a couple of knobs

16   you can play with to adjust the number of sources, the

17   number of receptors.  And then that plays into how big

18   a modeling domain you can model.  So, there's just a

19   lot of considerations.

20	DR. HENDERSON:   Go ahead, Rob.

21	DR. WYZGA:   Please, if you do Atlanta, I

22   may have some data that would of some help to you,

23   including some new personal exposure data.

24	DR. GRAHAM:   That's great, thanks.

25	DR. HENDERSON:   Okay, very good, Lianne?



42

1	DR. SHEPPARD:   So, my recommendation on

2  how to move forward, given you have limited time, is to

3  first make sure that you can actually, that you're

4  doing the job you think you're doing in Philadelphia

5  with respect to the prediction of ambient.  And, maybe,

6  that, if you can't be reassured that you can capture

7  that variability well enough, maybe that's enough to

8  suggest that, for the purposes of estimating

9  exceedances, this is not the way to go.  And then shift

10   the energy into the epi.

11	DR. HENDERSON:   Okay, well, with that,

12   let's proceed and, at least, start on our discussion of

13   the - -

14	DR. LARSON:   Rogene?

15	DR. HENDERSON:   Yes.

16	DR. LARSON:   Can you hear me?  This is

17   Tim Larson calling here.

18	DR. HENDERSON:   Yes, I can hear you.

19	DR. LARSON:   Yeah, thanks.  So, in the

20   exposure assessment, the, it seemed like, one of the

21   biggest chunks of the exposure above 300 ppb was these

22   near-road exposures, not the on-road, but the near-road

23   exposures.  And I was just curious.  Is this due to

24   like a one-hour encounter near the road of a few

25   people, or is this, or is this just, I mean, of a	

44

1  discussion.

2	So, first, with regard to the chapters

3  in the document, and beginning with the chapter three,

4  the at-risk populations, yesterday, we talked and had a

5  discussion about vulnerability and susceptibility, and

6  talked about the distinction and usage in both the ISA

7  and in this current document.  Although, in the context

8  of public health, it might be of more important just,

9  to just be counted, rather than to argue whether it's

10   susceptible or vulnerable, because both are potentially

11   at risk here.

12	Yesterday, Steve Kleeberger recommended

13   that an expanded table of at-risk populations be

14   identified, and I think that's appropriate here,

15   because it's more than just the three categories that

16   are discussed here.  In here, you talk about disease,

17   which operationally means airway responsive hyper, I

18   mean, airway hyper-responsive asthmatics.  Then you

19   talk about age, sort of, in children and older adults.

20   And you talk about proximity to roadways, but even with

21   regard to proximity to roadways, for example, there's a

22   whole host of people that, I think, would expand that

23   population that include residents that live there,

24   schoolchildren that attend school at, near busy

25   roadways, a range of occupational exposures that you



43

1  number of people, or is this just a few people that are

2  being exposed continuously over many hours near the

3  road?  Do you know?

4	DR. GRAHAM:   Without having the actual

5  data in front of me, I can suggest that, of course, we

6  do have less individuals that are exposed to that

7  highest concentration.  So, it is suggesting that, of

8  those people that are receiving those high

9  concentrations, there are very few of them, and that's

10   where they would occur.  I can't respond other than

11   that way, at this point, without the data in front of

12   me.

13	DR. LARSON:   It seems like an important

14   micro-environment, as has been mentioned, and just to

15   reemphasize that point, because it seems to be driving

16   the highest exposures in your analysis.

17	DR. HENDERSON:   Thank you, Tim.  Are

18   there any other comments before we move to health?

19   Well, we have some strong feelings on the health we've

20   already heard.  We'll led Ed Avol is going to lead off

21   on this one, and we may not finish it before the break,

22   but we'll get a good start.

23	DR. AVOL:   Thank you.  So, I have a

24   couple comments on the chapters, and then I'll come

25   back to the five questions to sort of start the	

45

1  brought up.  But also, in addition to the highway

2  patrol, and the toll collectors, and the bus drivers,

3  you talked about there's truck drivers.

4	There's rail workers.  There's off-road

5  construction operators.  There's the port and dock

6  workers in other communities.  There's commuters, both

7  workers and children on school buses.  And then,

8  there's a whole SES component, in terms of proximity to

9  roadways that raises, sort of, these issues.

10	And on the disease side, of course,

11   there's obesity, and we talked about, in cardiovascular

12   disease, and COPD, and diabetes.  But there are other

13   risk categories as well, in terms of the genetics.

14   The, again, SES itself may be a category, and smokers

15   in terms of increased NOX exposure.  So, I think

16   there's a large, a larger, a much larger universe of

17   at-risk populations that need to be, sort of,

18   considered here.

19	With regard to the chapters on health

20   risks, which here, sort of, operationally is just

21   airway  responsiveness, and I guess, that flows from, I

22   understand that flows from the ISA, and we had that

23   discussion yesterday, but, and it's sort of started

24   already this morning.  I think that there's going to

25   be, you're going to get a couple of comments about the



46

1  fact that it's more than just airway responsiveness,

2  and I think - -

3	DR. BALMES:   Could you repeat that?  Ed,

4  we couldn't hear for a second.

5	DR. AVOL:   I'm sorry.  I think that, in

6  terms of the health, there's more than just airway

7  responsiveness, and you know, in some ways, I think,

8  with apologies to physicists, I think there's sort of a

9  grand unification theory at work here.

10	And I think that, you know, if you look

11   at lung function growth decrements, and respiratory

12   illness, and epi,  the epidemiological effects, and

13   lung host defenses, and immunity, and airway

14   inflammation, respiratory emergency admissions, and

15   even cardiovascular disease, which currently is weak,

16   and mortality, I think there's a number of things.  And

17   yesterday, Frank Speizer suggested tallying up pluses

18   and minuses to look at this, sort of whole thing.  And

19   I think if you look at, sort of, make a list of, sort

20   of, short term NO2 effects, and look at adverse

21   respiratory effects, which you categorize as

22   sufficient, and list a whole host of them, and

23   cardiovascular effects, which you list as inadequate.

24	Mortality is suggestive.  You sort of

25   have a balance in one direction, and in long-term NO2,	

48

1  speak.  The animal data shows biological plausibility.

2  But it's the epidemiology that shows, sort of, the real

3  world perspective.  And I agree, it's not always clean.

4  It's not simple.  It's not a single exposure.  But,

5  it's what really happens.  And so, I think, that's,

6  sort of, something that has to be taken into account

7  here.

8	In the range of benchmark values

9  discussion that you have, I think the range is

10   reasonable, but the question I have is, given the

11   uncertainties associated with the measurement area, and

12   the misclassification, that even when you adjust and

13   take them into account, the improvements in whatever

14   modeling differences you're going to come up with, I

15   wonder if that's just not overwhelmed, but those are

16   small errors compared to the uncertainty that you have

17   around the whole aspect, and whether it's worth

18   worrying a lot about refining there, when you have a

19   much bigger, overall uncertainty about it.

20	So, with regard to the five questions

21   that were charged to us in terms of the

22   characterization of health risks, in my judgment, the

23   question, is the documentation clear and appropriately

24   balanced.  I think it's not always the case and not

25   completely.  And so, I think it needs a little bit of



47

1  you look at morbidity, which ends up being suggestive,

2  based on both our work in children's health study, and

3  Mike Brower's Netherlands birth cohort.

4	And even though there are some

5  inconsistencies in cross studies, there's still a

6  biological plausibility from animal toxicology here.

7  And there's some discussion about the inadequacy of,

8  the current inadequacy of mortality data and the

9  long-term data.  But still, I think in all of this,

10   there's a, sort of, a general preponderance of

11   information that sort of sways it in a direction.

12	And there's a discussion here about an

13   effects threshold.  And sort of an interest in it in

14   the document.  And it's not clear to me, I mean,

15   there's not a focus, per se, in the epidemiological

16   work on an effects threshold, necessarily, but at

17   least, to an issue that I'll come back to in a moment.

18	And so, again, in this context, that

19   sort of grand unification, you have, sort of, clinical

20   research.  You have the toxicology.  And you have the

21   epi work.  And each of them, I think, provides some

22   components here.  The controlled exposures have, and

23   the data, which you sort of weight on, sort of shows

24   hyper responsive airways as being, sort of, the

25   information that you, sort of, hang your hat on, so to	

49

1  work there.

2	With regard to questions two and three,

3  on the range of effects and so forth, I think because

4  of the uncertainty issues, I'm just not sure where the

5  benchmark discussions leave us.  So, I think the values

6  chosen are reasonable, but I think you need to do a

7  better job about establishing what's meaningful and

8  worthwhile in the document to be more convincing.

9	I think discussion yesterday on question

10   four about the data input for Philadelphia, about the

11   model output, about adjustments, about moving, you

12   know, alternate, all, or other communities that could

13   be tested, such as Los Angeles or Atlanta, sort of, I

14   think make moving to a richer data area like Los

15   Angeles may be more scientifically interesting and

16   appealing to do this, in terms of better emissions,

17   inventory data, better monitoring data, more traffic.

18   But I think the decisions you, the discussions you

19   raised yesterday about modeling domains and commuter

20   activity into and out of the area, and downwind

21   photochemistry, I think, are going to make that a

22   challenging undertaking as well.  And so, all in all, I

23   think question four is, sort of, a work in progress

24   that sort of be continued.

25	And then, finally, with question five, I



50

1  think that there have been a number of concerns

2  expressed about it.  But in my opinion, you know, and

3  sort of by way of analogy, reality tv is sort of

4  currently in vogue.  And so, I think epi may not be

5  unilaterally decisive, but it's the real world.  And

6  so, I think, you need to give it more measure and

7  stature here in the document.

8	DR. HENDERSON:   Thank you, Ed.  And,

9  John, are you on the phone?

10	DR. BALMES:   I am.

11	DR. HENDERSON:   Okay, it's your turn.

12	DR. BALMES:   My son is calling me 'cause

13   he just woke up, but I'll try to ignore him.  So, I

14   have a lot of agreement with Ed's comments.  I'll just

15   go through the questions.  With regard to question one,

16   I think that the staff did a fairly good job of

17   characterizing the health evidence concerning NO2

18   exposures.  In fact, I sort of, I like the fact that

19   this a much shorter version of what's in the ISA, so

20   it's easier to get one's mind around.  But, I guess I

21   have the same concerns that I have with the ISA, that I

22   think the long term exposure to NO2, and its effects on

23   growth of lung function in children is, I find the data

24   more compelling than the staff did in the ISA, and

25   that's, of course, reflected here as well.	

52

1  California Air Resources Board, and I always thing

2  about trying to explain to my non-physician colleagues

3  on the Board a risk assessment like this.  And to try

4  to explain to them, you know, a little delta and airway

5  responsiveness in asthmatics, as opposed to, you know,

6  sort of real world morbidity endpoints like going to

7  the hospital, or going to the emergency department.  I

8  think it's going to be really problematic, especially

9  given, well, I'll just leave that alone.  I was going

10   to make a politically inappropriate comment.  I think

11   the benchmark values are reasonable.

12	And then, the question four, to what

13   extent is the assessment, interpretation, and

14   presentation of initial health risk result technically

15   sound, clearly communicated, appropriately

16   characterized.  I think, again, the staff has done a

17   good job with the exception that I would have really

18   preferred to have morbidity endpoints to be used.

19	And, you know, again, well, turning to

20   question five, while I understand the rationale for the

21   staff judgment that's presented in section 4.233, I'm

22   just not persuaded that the judgment is the correct

23   one.  I think that, although many of the epi studies on

24   the effects of short term exposure to NO2 have been

25   conducted outside of the United States, and there's a



51

1	And I do agree with Ed that, sort of,

2  the overall gestalt, which I think staff shares, is

3  that NO2 does cause short term respiratory health

4  effects, but it's, when you go through everything

5  that's written down, and you end up using airway

6  responsiveness as, sort of, major health endpoint for

7  the risk assessment, it seems like it waters down the,

8  sort of, large body of evidence, much of which is

9  epidemiological.  I, actually, find the epi, I do

10   controlled human exposure studies, and I, actually,

11   find the epi evidence more convincing about health

12   effects with regard to NO2 exposures, than I do the

13   controlled human exposure data.  That's in part because

14   my lab has generated some relatively negative data with

15   regard to NO2 exposures, to show my bias.

16	With regard to question two, again, I

17   understand why the staff decided to use the

18   experimental data on airways responsiveness in

19   asthmatic adults to identify potential health benchmark

20   values to characterize risk, but I would have

21   preferred, as I've already said, to see asthma epi data

22   used.  I find asthma exacerbation data, you know,

23   hospital admission, emergency department admission,

24   much more compelling, and much more, much easier for

25   policy audiences to understand.  I mean, I'm now on the	

53

1  relatively small number of U. S. studies that have

2  been, U. S. cities that have been studied, you know, I

3  think the entire body of epi literature is consistent.

4  And so, I think you could develop a concentration

5  response.

6	Now, I know the staff is also concerned

7  that this is problematic because of the, of

8  co-pollutant effects, and I agree with George Thurston

9  strongly that, if you're trying to public health

10   protective, then you can use, you could use single

11   pollutant NO2, you could derive a single pollutant NO2

12   concentration response.  It might be not entirely

13   reflective of the independent NO2 effect.  I'm sure it

14   isn't, but it still would be public health protective.

15	And I just have to say for the nth time

16   that trying to tease out an independent effective NO2

17   isn't possible.  We're never going to be able to do it.

18   And, just like we can never tease out independent ozone

19   effects, or independent PM effects, yet we were able

20   to, sort of, use, we use epi data for ozone and PM in

21   risk assessment work.

22	I guess my last comment would be

23   regarding the controlled human exposure data.  It is

24   experimental, but it's, there's not that much of it.

25	Even using, or even basing a lot of the



54

1  assessment on the meta-analysis that Larry Folensbee

2  did for the '93 effort, you know, there's a fair amount

3  of variability in those results.  The subsequent

4  results are also variable, with regard to allergens

5  responsiveness, and it's not even, it's probably even

6  an underestimate of the effects, since the people that

7  participate in controlled human exposure studies are

8  adults, to be adult asthmatics, not children, and also

9  they're relatively mild asthmatics.

10	Now, I know the staff knows this.  So,

11   hanging our hat on something that's hard for lay folks

12   to understand isn't exactly the greatest, doesn't

13   really have a large end, even pooling studies, and is

14   only representative of a sub-fraction of the asthmatics

15   that are potentially at risk, I think is, isn't real

16   world, and the epi data, I think, would be a better way

17   to approach this.

18	DR. HENDERSON:    Thank you John. James?

19	DR. CRAPO:    I'm going to agree with the

20   previous speakers, I think that the staff needs to step

21   back in terms of the health risk assessment on this and

22   look at another way to do it. Let me also go through

23   the questions and sort of tell you why I think that.

24   The first question is in terms of was the assessment of

25   the overall health evidence balanced and the answer is	

56

1  happened.

2	But the laboratory can see a small shift

3  in a spirometry value. But the subject feels healthy,

4  they don't require any medication, they don't realize

5  you did anything to them and most of them would tell

6  you nothing happened that day. These are people who

7  have twitchy airways and occasionally are around an

8  animal or a cat or a dusty environment and they have a

9  little reaction of airways and it's a very common event

10   in their lives.

11	They treat these as insignificant

12   events. I think that taking that and then- when you put

13   that together with these studies that were used to do

14   this, there were 3 of them done in the same laboratory,

15   and not reproduced by any other laboratory worldwide,

16   really raises criteria question of- A- Is it real? And

17   then B- Is it significant?

18	I think it forms a very poor basis to do

19   a quantitative risk assessment around. So my answer to

20   question #1 is it was not a balanced assessment of that

21   and the wrong endpoints were chosen. The second

22   question was whether or not the, well it's the same

23   question, it's based on this particular airway

24   responsiveness, so that's another issue. I think the

25   benchmark values are appropriate in the sense that they



55

1  no, it was not. The choice by the group was to look at

2  the human health clinical studies in an uncritical

3  fashion and accept it without applying rigorous

4  criteria to it's value. And so what was done was to

5  choose the effect that had a quantitative number

6  associated with it, with the lowest observed effect,

7  but without asking if it was reproducible and

8  consistent and clinically significant.

9	Therefore, doing the exact opposite of

10   the Hill criteria. I mean, it fails every Hill

11   criteria. I want to say that I agree that the

12   epidemiology is the strongest evidence and does, I

13   think, meet those criteria, provide a basis for going

14   forward. I think there's also a tendency not to

15   understand exactly what's being measured.

16	When you measure airway and hyper

17   reactivity in a subject like this, you have to remember

18   that asthmatics have twitchy airways and they're

19   sensitive to lots of things- to temperature, to

20   moisture, to common allergens or dust. In this case,

21   the ability to stimulate a reaction measured by a

22   usually very tiny change in FPP-1 on a spirometer in

23   response to some challenge is a test in which the

24   subject is unaware of any effect. In most of these

25   cases, the subject is totally unaware that anything	

57

1  represent the high end of what's clinically- what's

2  observed in their environment. I mean you couldn't go,

3  you couldn't take them much higher because there's no

4  values up there. So you're looking at a reasonable set

5  of values, but it's being designed around the question

6  of perhaps 0.26 ppb, well ppm, causes airway hyper

7  sensitivity.

8	Therefore, by finding an exceedence of

9  that, maybe once a year, you've said that there's one

10   day one child might have a spirometry measure....I mean

11   you're asking how often that happens.

12	What I'm trying to say is it's am

13   insignificant event but you're actually asking the

14   question, how many times in a year would a certain

15   number of people have the potential of this event

16   happening and that's probably the wrong question given

17   that your best data is epidemiology. The epidemiology

18   says that you can get a change in respiratory

19   morbidity, they say ER admissions, by a 20 ppb drop in

20   the 24 hour daily average. Perhaps you're endpoints

21   should be designed around that.

22	You ought to be asking how many areas

23   have a exposure pattern where a different standard

24   could drop the, if you assume you met it, how many

25   people would have a 20 ppb drop in their daily 24 hour



58

1  exposure. How many times would that occur to a

2  population? So I'm trying to suggest you do as an

3  endpoint that's related to the epidemiology and the

4  only one you've got there is a 20 ppb drop. So I think

5  it's possible to ask how many days have enough

6  elevation and enough change that you could cut those

7  out and how many days would be vulnerable to that kind

8  of thing.

9	Therefore, direct it right at the

10   epidemiology instead of the clinical science. The

11   assessment of the health risks, because the endpoints

12   are not chosen to be ones that I can accept as likely

13   to be clinically significant or statistically valid, I

14   think you can't really assess it. I mean there coming

15   from a negative assessment for question 4 so you can't

16   do that because the endpoint chosen was wrong. It would

17   be,  I think you could use much of the same technology

18   though in the way I just described and apply the

19   epidemiology very effectively.

20	The last question really relates to the

21   same question again. I strongly concur with those that

22   are arguing we ought to stay here today until we agree

23   to do the study focused on the epidemiology as the

24   critical endpoint that we can hang our hat on.

25	DR. HENDERSON:    Thank you James. Is	

60

1  argument for it. My take on the epi results for NO2 is

2  that NO2 central site monitoring is a good marker for

3  day to day variability or spatial variability in air

4  pollution.

5	In particular, related to motor vehicle

6  exhaust emissions. So it's a nice marker of those

7  exposures. So it's perfectly fine to use those studies,

8  as long as the results are presented as NO2 as a

9  measure, as a proxy for general day to day or spatial

10   variations in air pollutions. But, I don't think it's

11   fair to interpret those results where we see effects at

12   very low levels of NO2 in the ambient atmosphere as

13   really representing NO2 per se effects.

14	So as long as we're clear about it, I

15   think it's fine to use that data in the risk assessment

16   but it has to be couched properly. One important aspect

17   of this is that indoor NO2 is the dominant exposure for

18   people who have gas stoves which is very common in many

19   cities. So really understanding the relationship

20   between central siting and personal exposures to NO2 is

21   something that creates a great deal of uncertainty and

22   again sort of argues for the central site data as not

23   just representing NO2. So I'm uncomfortable just taking

24   it at face value and using it in a risk assessment as

25   an NO2 exposure response relationship. That's it for my



59

1  Terry on the phone.  Is Terry... Pat Kinney?

2	DR. KINNEY:    Going down the list of

3  charge questions, the second one which addresses

4  whether the benchmark values are....basically what are

5  the views of us on the benchmark values. I...My concern

6  about the range that was given is that the.....clearly,

7  as John Balmes has mentioned, the subjects that are

8  looked at in clinical studies are not a representative

9  set from the general population and particularly don't

10   have good representation of the most sensitive tail of

11   the general population distribution and even when you

12   look at asthmatics in the laboratory, they're not the

13   most severe asthmatics. So if you're going to use sort

14   of lowest effect values from clinical studies I think

15   it's really important to make some adjustment for this

16   problem. How to do it quantitatively I'm not sure.

17	I wouldn't simply take sort of the

18   lowest effect obtained in laboratory settings and say

19   that's the relevant benchmark in the general

20   population. Of course this does relate to the other

21   issue that's being discussed, which is whether the

22   epidemiologic findings ought to play a larger role,

23   because clearly there you do have the full distribution

24   of people represented in your findings. So it's very

25   tempting to go with that and I think that is an	

61

1  comments.

2	DR. HENDERSON:	Thank you , Pat. I'm

3  going to skip down to Ed Postlethwait because I know he

4  has to leave a little early. So Ed can you give us your

5  comments?

6	DR. POSTLETHWAIT:    Thanks Rogene. I

7  agree with everything that's been said so far. I

8  thought James did a very nice job of going through it

9  question by question.

10	What I picked up from this, and perhaps

11   this is a bit trivial, but in terms of the titling of

12   the chapter with the information that's presented,

13   there seems to be an implicit assumption that exposure

14   as predicted here equates to 100% of a health outcome.

15   And we know there's implicit, inherent variability in

16   populations. And so the question is, are we predicting

17   or are we attempting to predict a worst case scenario

18   here, in terms of if somebody is exposed, then at that

19   event negative outcome will occur.

20	Or are you trying to say, gee, if the

21   asthmatic population we've selected to be the target as

22   receptive to potential health risk from exposure, how

23   many of those people get exposed and then how many

24   exposed people suffer a bronchial spasm event or

25   whatever outcome you want to pick.



62

1	MR. RICHMOND:    This is Harvey Richmond.

2  Clearly, closer to the latter in that we clearly

3  discuss in the chapter that we would not expect all

4  asthmatics exposed to this level to see even the

5  response we're talking about. So we have caveated.

6  That's why we've used the word potential health effect

7  benchmark, not health effect benchmark. It's some

8  unquantified fraction. We're saying the evidence is not

9  there for us to quantify, but some significant fraction

10   we're expecting would, but we're not making any attempt

11   to quantify whether that's 10% or 15% or whatever

12   because we don't think the evidence allows us to

13   quantify what fraction. But there's enough evidence at

14   those different benchmarks and certainly would be an

15   increase, we would expect, at higher levels that they

16   have the potential to experience the responses observed

17   in the clinical studies.

18	DR. POSTLETHWAIT:    Is there a way to put

19   boundaries on the numbers?

20	MR. RICHMOND:    As to what fraction? I

21   haven't seen, out of the clinical data, our ability to

22   do that so far. If others on the committee think

23   there's data, I'd like to see what specifics they point

24   to. The meta analysis that was done by phones took all

25   the individual data from those studies but it just	

64

1  up....

2	DR. HENDERSON:    Okay, Kent do you want

3  to go?

4	DR. PINKERTON:    Well, thank you, and I

5  certainly agree with all the comments that have been

6  made and will add perhaps just a repetitive comments on

7  some of these issues.

8	I think that it is important to look at

9  the full breadth of potential health effects that are

10   there.  I know that airway hyper-responsiveness seems

11   to be the easiest measure to use, but I think it's also

12   important to consider all other indicators as well as

13   potential things that you would want to examine.

14	I think that the selection for using a

15   one hour average makes a lot of sense, but I'm somewhat

16   concerned because of some of the points that have been

17   made about long term effects of No2, so we assume that

18   it's an acute exposure to a relatively high level, as

19   high as we would see in ambient conditions for No2 that

20   may elicit these kinds of responses under special

21   conditions for those who may be challenged with an

22   allergen or something like that.

23	So again it's really a specialized

24   population but the question really comes up is some of

25   the issues that we're dealing with in the Southern



63

1  looked at if there was a positive or negative response

2  without, and I think 76% of the subjects within the .2

3  to .3 range fell in as having a positive response. But

4  it didn't deal with the quantification of the magnitude

5  of that response. Therefore, not all the response might

6  be clinically significant as James Crapo was talking

7  about. So we don't feel comfortable quantifying that

8  percentage in terms of a degree of response that would

9  be judged as either adverse or clinically significant.

10	DR. POSTLETHWAIT:    Well, perhaps

11   something to think about then is to either bring those

12   caveats to the forefront a little more....

13	MR. RICHMOND:    Right, we clearly could

14   communicate it better....

15	DR. POSTLETHWAIT:    ...and again,

16   depending upon who the target audience is you're really

17   trying to write thing, as opposed to the folks in this

18   room, is the presentation of the data that looks like

19   you're predicting numbers of people are going to suffer

20   an adverse event as opposed to a potential adverse

21   event. That's really all I have to add.

22	DR. HENDERSON:    Thank you Ed. Now is

23   Steve on the phone?

24	DR. NUGENT:    He was planning to be on

25   the phone but maybe because we moved this discussion	

65

1  California children's health study which seemed to

2  indicate very clearly that there are reductions in lung

3  growth function in children and that that  has been

4  continually associated with No2, even when we adjust

5  for the co-pollutants so I think that really is an

6  important finding.

7	I'm also thinking about the benchmarks

8  that you've used, they seem to be as appropriate as you

9  could possibly do, I think that seems very reasonable

10   but I'm also concerned that for some of the studies

11   that have come out of Australia, the children there

12   where they actually have shown differences based upon

13   levels of No2 that may be present indoors.  I realize

14   that you're dealing with an outdoor situation, but

15   again those differences were well below the hundred

16   parts per billion where they saw those distinct

17   differences.

18	And again, I would just echo the

19   feelings that you've heard repeatedly this morning, and

20   that is that I think to dismiss the epidemiological

21   studies are...would be a mistake.  I think that there

22   are things that you can do to try to make the most of

23   those, the best of that, and also the modeling that

24   you're taking about doing in these cities, it seems as

25   though you have lots of plans that are there that seem



66

1  fairly reasonable but again, it would be wise to use

2  those resources in the best way possible whether it be

3  that or further analysis of epidemiological studies or

4  human clinical studies.  Thank you.

5	DR. HENDERSON:    Thank you Kent.  John

6  Samet are you there?

7	DR. SAMET:    I am here, I've been

8  listening with interest.  So I'm getting more persuaded

9  that the epidemiological data, some of it should be

10   used, perhaps in addition to some of the clinical

11   studies as an endpoint.  The....I think John Balmes

12   said that well, you know we can we look at the effects

13   of ozone and we look at the effects of PM and we can

14   identify the effects of No2 and epidemiological

15   studies, I do think that there has to be much more

16   caution around interpreting the quantitative estimates

17   from epidemiological studies on No2 and again I think

18   that's because of the particular relationship of No2 to

19   other mixture of components, formation of secondary

20   particles, in some places a contribution to ozone and

21   so on and I think those are what was in that figure

22   that is now in Chapter One of the ISA, some of those

23   relationships that are complicated and really can't be

24   teased out with multi-variable models.

25	The evidence for short term effects,	

68

1  done.  If all the thinking were done following on

2  Frank's comment, no one would ever leave the hotel room

3  because it would just take too long.  There's a lot of

4  difficult issues here, so I guess that's where I sit at

5  the moment.

6	DR. HENDERSON;  Thank you John.  Rich

7  Schlesinger are you still.

8	DR. SCHLESINGER:    I'm here can you hear

9  me?

10	DR. HENDERSON:    Yes, very well Rich.

11	DR. SCHLESINGER:    Okay, well being

12   towards the end of the list, almost everybody has

13   covered all the issues that I had written down and

14   maybe then some, but just some the issues.  One was

15   raised yesterday and Ed Avol raised it again and that

16   needs to be a better separation of the definition of

17   vulnerable versus susceptible.   As I said yesterday

18   and apparently others have felt there was confusion in

19   terms of the definition used or they gave a definition

20   and then made it confusing by putting everything both

21   definitions.

22	The... I would like a better

23   justification for the benchmark that they did use, both

24   in terms of what Kent said in terms of some of the

25   lower levels found in indoors causing an effect, and



67

1  some of it does come from the Australian intervention

2  study which is useful.  There's complimentary data from

3  asthma, asthma panels, and other kinds of studies, so I

4  think to focus on asthma there again with support from

5  the toxicological studies, is reasonable and probably

6  if that's to be the case, then the risk assessment

7  document needs to do a better job in sort of making

8  that case and putting together these two bodies of

9  data.

10	I think the probably the most difficult

11   aspect of moving to use the epidemiological data is how

12   to interpret.  There's two things, I mean one is the

13   quantitative estimate of effect and then the

14   interpretation of that because I think given some of

15   the limitations of the data available, these

16   complicated relationships among variables, the fact

17   that we truly are dealing with a mix looking at one

18   component of it, I think the interpretation of these

19   kinds of numbers in a causal framework, that is to say

20   this is the burden of disease attributable to No2

21   specifically,  has to be carefully and cautiously made.

22	I think we have to think about what

23   these estimates mean in terms of potentially bounding

24   estimates or are they under estimates I mean I think

25   there's some difficult thinking that would need to be	

69

1  even if they ignored that, I still needed a better

2  justification, based on the data set that they did use

3  which was presumably the clinical studies.

4	The other issue, in spite of my

5  toxicology background, I do feel that and given the

6  caveats and the use of epidemiology that we all know

7  about, that it should not have just been discounted in

8  terms of the development here and that it needs to be

9  considered and based on that consideration, EPA may

10   decide that there are valid reasons why they can't use

11   this, or can't use that, but this to me looked like

12   just a outright decision not to consider it without

13   necessarily evaluating it to the extent that it should

14   be evaluated and that's my two cents, given everything

15   else that was said.

16	DR. HENDERSON:    Okay Rich well thank

17   you.  And now we have Frank Speizer.  It's your turn

18   to....

19	DR. SPEIZER:    Well my bias is on the

20   table and most everybody has said most of the things

21   that I was going to bring up.

22	I was really quite shocked that with

23   fifty studies that has been done, that we were just

24   going to cast it aside.  Now one of the concerns is

25   that is how to do this and I do want to go home today,



70

1  and it may be that we need to sort of have a more... a

2  greater discussion with Harvey and his group as to how

3  we can help him if if that's necessary.   I mean it

4  seems to me that if we don't do this, we're going to be

5  here five years from now, some of us will be here five

6  years from now, going through the same argument,

7  because the studies that are going to be done are going

8  to provide similar kind of information.  Well it may be

9  a little more biological markers and stuff but we're

10   going to have the same problem of the mixture.  So I

11   think we need to sort of at least figure out, if I mean

12   certainly do the risk assessment, and see what the

13   flaws are and hopefully those flaws can lead to a

14   better direction to where things will go in the future.

15	The...I think that's an important

16   component of doing the analysis using the epi data.  I

17   would propose that it might be appropriate for a

18   sub-committee of this group to be involved before we

19   see the next draft as Harvey does some of these

20   analyses, certainly it would be better for us to look

21   at it than for OMB to look at it, and I am concerned

22   that OMB will look at it and I think it might be better

23   to have a sub-group.  Now I don't know whether that's

24   feasible or doable, but I'm sure there are people who

25   have some expertise around the table that might be able	

72

1  look at particle interaction

2	DR. THURSTON:    Oh, you know me so well,

3  no I'll be brief because I think all the points have

4  been made really.  I first of all I want to say that I

5  think we are in good shape, I think we're in good shape

6  with the documents that we have as a stepping off point

7  for doing what needs to be done.  And I think that the

8  ISA has got to connect the epi and the controlled

9  exposure assessments better and give the basis for you

10   know...  that's what it should do because that's the

11   gap there.   And I think that will give us the basis

12   for doing the risk assessment. And I think at this

13   point I would say to take the Philadelphia analysis and

14   dot all the Is and cross all the Ts, listen to all the

15   advice and that's enough of that kind of analysis

16   because I think it tells us what it needs to tell,

17   doing more cities.

18	I mean just listening to the discussions

19   of trying to do L.A. and Atlanta, and I don't think

20   they have time to do all those and when they're done

21   with it you're going to be talking about such a small

22   percentage of America being considered and such a

23   narrow population base within those cities, that it's

24   not going to have any kind of....I mean it's a useful

25   exercise that do once at Philadelphia and to learn



71

1  to at least react to what Harvey will do with these

2  data and maybe we should have a discussion about that.

3	DR. HENDERSON:    Yes, I agree, when I

4  came to this meeting I thought that really in depth

5  questions that faces us here is this multi-pollutant

6  problem, and as we have all said, it's not easy, it's

7  something that no one has a magic answer for but  I

8  think you're right within this group, there can be wise

9  advice as to how to handle this.  This is something...I

10   would like to ask the folks in the air office, in what

11   way can we help you with this major problem.  The

12   agencies moving sort of multi-pollutant approach but

13   you're not there yet I mean and it's not something we

14   just say say you do that, so maybe Karen can tell us

15   how we can help her.

16	DR. MARTIN:    What I'm going to ask you

17   back first is if we are about at the point of taking a

18   mid-morning break, it would give us an opportunity to

19   collect our thoughts, with regard to the issues you

20   have raised and I'd be happy to say some words on this

21   point, perhaps right after the break, that would give

22   you instantaneous reaction.

23	DR. HENDERSON:    That sounds good now we

24   had two more people listed here, I don't want to

25   neglect them, George is going to tell us so we need to	

73

1  about that, but I think at this point what we really

2  need to do is do the next stage of analysis, which is

3  to consider the entire US, or at least all the major

4  cities in the US, and do an epi analysis and yes there

5  are uncertainties but we can document those

6  uncertainties.  And that's when you can look at the

7  multi-pollutant analysis and get an idea of what the

8  uncertainties of the estimates from the single

9  pollutants might be, and document the uncertainties but

10   do the analysis, do what will tell the decision makers

11   what they need to know and will give us some more a

12   broad breath of information and I think that's more

13   doable than trying to do several more cities of this

14   intense modeling, with its own huge uncertainties as

15   well that have to be documented and I think Frank's

16   idea of the sub-committee is a wonderful one.   I'd be

17   glad to participate in such a sub-committee and try and

18   work with the staff to choose how to go about getting

19   the co-efficience to use where, if they need the kind

20   of help because I know they have their own experts so

21   that's about it.

22	DR. HENDERSON:    Thank you and Ron did

23   you want to add something?

24	DR. WYZGA:    Sure, let me say I think

25   it's obvious that we have to go out and try use the epi



74

1  data.   I think we shouldn't ignore the clinical data,

2  I think you've done a lot of work along those lines,

3  and I think you need to refine along some of the

4  suggestions that have been made, but I will say that in

5  using the epi data is going to be a challenge, you know

6  as Frank said, the devil's going to be in the details.

7	You've got a lot of questions with

8  respect to which study to choose, etcetera but one of

9  the major problems is the fact that when you look at

10   the ISA and sort of  the emphasis of the discussions,

11   we've spoken about how important some of the peak

12   exposures are, exposures near roadways, etcetera.

13   When you look at the epi studies, they tend to do use

14   twenty four hour averages, and so you know you're

15   looking at something quite different from the hourly

16   average and that's going to be a challenge, and I think

17   that we need to reconcile these two and to think about

18   how we're going to reconcile it, and you know that's

19   something that I think we all need to put our thinking

20   caps on, and think about how to do that.

21	DR. HENDERSON:	Thank you Ron, and

22   let's take Karen's suggestion and we'll take a  fifteen

23   minute break, come back about 10:35 or 40, and then

24   have a really good discussion on how we can attack this

25   problem.	

76

1  with from our perspective and hopefully to provide some

2  context in terms of from our perspective again what we

3  think might be most productive in terms of the

4  discussions you're now about to have.

5	And I would like to -- I would to raise

6  two general points here.  The first point relates as

7  much to process as substance, but I think it's relevant

8  here.

9	As I -- as I am sure you all know, the

10   process that we are working -- doing these NAAQS

11   reviews now call for us to produce different kinds of

12   documents than we did before and we now have an

13   integrative science assessment, we now have a

14   quantitative risk assessment document, and that  both

15   of those documents are to be followed by an advanced

16   notice of proposed rule making, which captures a policy

17   assessment and historically that's been in a staff

18   paper, which you would be seeing now at the same time

19   that you're seeing a quantitative risk assessment

20   document and that document has historically provided a

21   broader perspective from a policy perspective for

22   thinking about quantitative risk assessment, but we

23   don't have that now because our process has changed.

24	And all that is salient to the issue

25   about what we are or are not doing with regard to



75

1	DR. LARSON:    Rogene, when people are

2  putting their thinking caps on, Tim Larson...

3	DR. HENDERSON:    Oh Tim...

4	DR. LARSON:    Yes, this seems like an

5  opportunity to rethink where we put and how we design

6  No2 networks, because I mean if we had the data  near

7  the roads as well as away from the roads right now, we

8  could probably do a better job of combining these two

9  approaches but right now we're got at least one hand

10   behind our back so it's something to think about as an

11   opportunity going in the future.

12	DR. HENDERSON:    Okay thank you.  Tim

13   we're going to take a break.

14   (WHEREUPON  , a brief recess was taken.)

15	DR. HENDERSON:    Let's move on or expand

16   our knowledge of the problem and I really appreciate

17   your doing this because it is a problem.  We all want

18   to discuss it.  It's the biggest problem we address and

19   so, Karen, the floor is yours and we'll appreciate your

20   expanding our knowledge in this area.

21	DR. MARTIN:    Okay.  I'll start by saying

22   I'm sure I didn't promise to solve the problem and my

23   purpose in the remarks I would like to offer aren't --

24   aren't so much expanding knowledge, but to try to

25   provide some context for what you have been struggling	

77

1  considering epi studies.  I've heard the phrase a

2  number of times today that we are dismissing the epi

3  studies and I know our quantitative risk assessment

4  doesn't present any quantitative risk assessment based

5  on information from epi studies, but I would suggest to

6  you that it most surely doesn't dismiss them and it

7  most importantly doesn't dismiss them from a subsequent

8  policy consideration context.

9	There are lots of ways to think about

10   what epi evidence tells us.  One way to think about it

11   is to extract concentration response functions and do

12   quantitative risk assessment.

13	Another way to think about them is to

14   ask questions like what were the air quality levels and

15   patterns in areas where we have signals from epi

16   studies that tell us something about health effects?

17	Were those areas in areas that exceed

18   current standards or that attain current standards?

19	Those kind of questions have some

20   quantitative component to them, but they don't have

21   quantitative risk assessment associated with them, but

22   they still tell us a whole lot about things

23   policymakers need to know in assessing the adequacy of

24   the current standard and potential alternative

25   standards.



78

1	And what I would like to say here is

2  that we fully intend to think about the epi evidence in

3  those contexts in the next document that we produce so

4  that I just want to make real sure you understand that,

5  because this draft quantitative risk assessment didn't

6  draw from epi CR functions to play out quantitative

7  risk assessment -- assessments per se, there's nothing

8  about this document that was intended to convey the

9  notion that somehow we're dismissing that evidence from

10   the broader qualitative policy considerations.

11	So that was -- that was just one general

12   point I wanted to make and that would have been made in

13   the kind of document we used to produce.

14	The second point I wanted to make was,

15   to the extent that we are open to thinking about how we

16   might incorporate information from epi studies into the

17   next draft of a quantitative risk assessment, we have

18   pondered all the kinds of challenges that you have

19   raised.

20	We are not oblivious to those challenges

21   and it's those challenges that caused us to take the

22   position that we took in this document.

23	If we are to move off that position, we

24   would very much appreciate from this group, perhaps in

25   the discussions you now have, somewhat more specificity	

80

1  that we saw a pretty high potential that, even if we

2  went down the road of selecting specific CR functions

3  from specific studies and applying them in specific

4  example cities, in the end, if all we end up with are

5  numbers that say, well, these are numbers that show

6  that if you reduced air pollution in general, you would

7  achieve a benefit in public health protection, but if

8  they aren't numbers that say, if you reduce NO2 in

9  particular by this degree you will achieve benefits in

10   public health protection, then the question I would

11   pose back to you is would we have produced numbers that

12   meaningfully inform the review of an NO2 standard, not

13   an air pollution standard?

14	So those are clearly the kinds of

15   questions we need to grapple with and in our grappling

16   with them, and in our grappling with the notion of also

17   the averaging time associated with effects, the very

18   short-term peaks that we can see in clinical studies

19   that we can tease out in the exposure assessment that

20   we've done, how well represented is that evidence in

21   epi studies that use 24-hour averages or long-term

22   averages?

23	How well -- how well does one body of

24   evidence help to provide plausibility for the other

25   body of evidence when we have significantly disparate



79

1  as to how we might go about surmounting some of those

2  challenges in more concrete ways and I just make a

3  couple more observations.

4	In past reviews of PM and ozone, one of

5  the things that greatly helped us move forward in

6  bringing epi evidence into more quantitative assessment

7  was the fact that we had the ability to look at

8  multi-city studies, multi-pollutant analyses within

9  multi-city studies that helped us better reach

10   inferences about the extent to which that information

11   informed judgments about PM per se, ozone per se, as

12   opposed to just the broad mix of air pollution and I

13   think the observations you were making, the same as we

14   have observed, is that, in the case of NO2, that's

15   going to be much harder to do.

16	It's going to be much harder to do just

17   because of the atmospheric chemistry involved and the

18   correlations among these pollutants.

19	It's also going to be harder to do

20   because we don't have the breadth and depth of

21   multi-city epi studies that focused on NO2 and that

22   found independent NO2 effects, not just in single

23   pollutant models, but in multi-pollutant models.

24	So those are the kind of constraints

25   that we are very well aware of and caused us to say	

81

1  averaging times?

2	There's more I could say, but I'm not

3  sure it's worth my taking more of your time and mine

4  talking about it, but I did want to say that we fully

5  recognize all of these challenges and when we looked at

6  it, our sense was it wasn't clear to us that the

7  numbers we would get out of an epi-driven quantitative

8  risk assessment would do more than tell us that we have

9  a clear air pollution effect as opposed to tell us

10   something about quantitative levels of an NO2 standard

11   beyond what we would get from looking at the epi

12   evidence in a more qualitative way, which we fully

13   intend to do.

14	DR. HENDERSON:    James, go ahead.

15	DR. CRAPO:    Karen, I want to say that I

16   hear you, and I agree with you, and I sympathize with

17   you completely.  I think you hit it right on the money.

18	I would say that I am far less

19   impassioned about stricter standards for NO2 than I was

20   about particulates in ozone for exactly the reasons you

21   just outlined.

22	Having said that, I think our first

23   decision, and your decision and our decision to

24   evaluate it, is do we think that our country should

25   have a short-term standard for NO2?



82

1	I think, looking at all the data, again

2  I'm less impassioned, but I think the answer is yes to

3  that based on primarily the epidemiology, but that's --

4  it starts there because if you're not -- you have to

5  decide I think it should be a standard.

6	Then -- then my next issue is, if we are

7  going to recommend a standard, what would we recommend

8  and how would we defend it?

9	And so I sort of look at the very end

10   point, if that's where -- if that's where we're going

11   to possibly end up.  When I get to that end point, I

12   know that I can't defend by saying that we've done this

13   analysis and we think 50,000 asthmatics will have at

14   least one day a year in which they have an undetectable

15   clinically insignificant airway constriction that would

16   only be made with a spirometer.   We're not going to

17   sell that to anybody, including ourselves.

18	I think the only way we could defend it,

19   if we do it right there, is by saying that we believe

20   that setting the standard at this level will result in

21   this -- a reduction in hospital admissions or morbidity

22   of some kind -- not necessarily mortality, but

23   morbidity, exacerbations or hospital admissions let's

24   say, ER visits for asthma, and we think there -- we

25   think that this level would result in this many	

84

1  suggesting we need to get at, which I completely agree

2  with you, there are two things epi can tell us.

3	One, epi can tell us the estimate of how

4  many hospital admissions we think might be reduced in

5  going from a standard of this to this and coming up

6  with a number which would have, I'm sure we all agree,

7  huge uncertainty bounds around it.

8	The other thing to do is ask the

9  question what were the air quality levels and air

10   quality patterns and profiles in the areas where we saw

11   effects and what level of a standard of a given

12   averaging time would we need to cause reductions in NO2

13   levels to occur in that area?

14	That's another way to look at epi

15   information as sort of a quasi quantitative way, if you

16   will.  I would suggest that, if you look back at the

17   history of decision making on PM and ozone, that is

18   more the way we looked at the epi evidence even in

19   those reviews than by decisions driven on the specific

20   quantitative estimates of risk with large uncertainty

21   bands around them that came out of the quantitative

22   risk assessments we did in those reviews, so that we

23   don't need to do a quantitative risk assessment to use

24   epi to answer the kind of question you're postulating.

25	We could do both, of course, and have a



83

1  reductions in that important parameter.

2	That's the only one I can come up with

3  that I think that I could defend long term and so what

4  I'm really suggesting is that we -- at least as we

5  explore this, we look for a way to do the risk analysis

6  that could give us that answer, tell us if there is a

7  standard, so that -- but that means we've got to do the

8  analysis around a parameter that's going to give back

9  results and right now the only number that I see from

10   this is maybe a 20 PPB -- how many people in the risk

11   category at a certain standard would have a drop of 20

12   PPB -- this is a 24-hour average, I understand that --

13   to figure out what the average time was and figure out

14   what period of time, but you've got to do something

15   with that so that it relates back to our epidemiology

16   and that's sort of where I am.

17	I'm not sure how to do it, but I -- but

18   I can envision that as the only answer that I could end

19   up defending of changing the standard.

20	DR. MARTIN:    If I could just...

21	DR. HENDERSON:    Go ahead.

22	DR. MARTIN:    If I could just react to

23   that as a way to help the thinking that goes into your

24   subsequent discussion, there are -- taking your

25   scenario of what you're trying -- what you're	

85

1  broader characterization of the information, but I'm

2  just saying it isn't -- the quantitative risk

3  assessment alone isn't the only way to use that

4  evidence.

5	DR. HENDERSON:    I'm interested, Karen,

6  in -- you know, you said that this type of analysis

7  that we're looking for would be part of a policy

8  assessment.  You know our problems with the last policy

9  assessment, which was not a policy assessment.

10	In your mind, will there truly be a real

11   policy assessment document developed?

12	DR. MARTIN:    The best way I know how to

13   answer that is comments they made at the last meeting

14   we had and that is it is our intent to produce a policy

15   assessment in the context of the next document we

16   produce, which will be the ANPR, according to our

17   current process, and to do the kind of assessment that

18   we would have historically done in staff papers.  That

19   is our intent.

20	I can't offer guarantees because no one

21   of us alone determines the final outcome of these

22   things, but that is clearly our intent.

23	DR. HENDERSON:    Okay.  I want to hear

24   other people's thoughts.  I mean what Karen has pointed

25   out is the new process has prevented us from seeing the



86

1  kind of data we want as early as we would like and

2  that's not a good thing.

3	So, as part of the problem we're facing

4  here, we would like to see what the agency can do with

5  the epi data, not to do a quantitative risk assessment,

6  but to do the type of thing to answer the question

7  where we see it and epi studies where we see effects

8  associated with NO2 levels, you know, how can we lower

9  the level to avoid that I guess is what you were

10   saying.  I think I said that right.  George?

11	DR. THURSTON:    Well, I guess I'm not

12   ready to roll over on doing a quantitative assessment

13   because I think that's what needs to be done and I

14   would say the uncertainties in the process that was

15   just described regarding the writing of the ANPR are

16   far larger than the uncertainties involved in doing a

17   quantitative risk assessment using epidemiology, so

18   that I think that we should do our work and not rely on

19   the kindness of strangers.  So -- and we should do the

20   best work we can and make it as complete as possible.

21	Now, in reference to the multi-city

22   study question, first of all, I don't think that they

23   are the gold standard they're made out to be sometimes,

24   but I think they are important and they provide

25   information, but I would say my own assessment looking	

88

1	DR. HENDERSON:    Great.  You know, I

2  don't think there's anybody arguing that we don't need

3  to use the epi data as we set standards.  It's just

4  that there's this peculiar mind set that you can't give

5  CASAC a staff paper.

6	It might contaminate it in some way and

7  I don't know what -- I shouldn't say it quite that way,

8  but we need that analysis in order to make -- to meet

9  our congressional mandate to advise and recommend a

10   setting of the standards to the agency and without that

11   staff analysis, we're left to do just what -- you know,

12   to do -- just at the present time to look at a

13   quantitative risk assessment that will have huge

14   variabilities associated with it, but you're saying you

15   want that done.

16	DR. THURSTON:    Yes, I am and I think

17   others are saying that.

18	DR. HENDERSON:    Well, I want to hear...

19	DR. THURSTON:    Yes, ask around.

20	DR. LARSON:    Regine, this is Tim Larson

21   calling.  Yes, I was intrigued by James' idea about the

22   daily metric and the hospital admissions because, if

23   you go to a risk assessment model that predicts daily

24   averages rather than hourly values, I think some of

25   this variability and queasiness about the particular



87

1  at the multi-city studies is that they have confirmed

2  what we have seen in doing meta-analysis of multiple

3  individual level studies, at least the conclusions of

4  those.

5	And so -- and we do have a multi-city

6  study for NO2 that shows its resilience against the

7  inclusion of other pollutants and that is the Canadian

8  work by Burnett using mortality.

9	Now, you know, that's another decision

10   whether we want to do it for mortality as well as

11   hospital admissions, but I do think that answers the --

12   you know, that question has been asked and answered.

13	We've done -- someone has done the

14   multi-city study and has shown -- and shown that

15   principle proven, okay, for mortality.

16	Now, it hasn't been done for hospital

17   admissions to your knowledge, but I don't think it

18   needs to be.  I think we have enough studies.

19	We can do a meta-analysis, come up with

20   an estimate, move forward, and come up with

21   quantitative analyses and we have time to do that and

22   we have time to give the American people an

23   understanding of the width and breadth of the problem

24   that is associated -- the health risks that are

25   associated with NO2.	

89

1  risk exposure model will be diminished just because

2  you're averaging and the fact that the 24-hour average

3  is a much more robust measure.

4	So I -- there's an appeal in that if

5  that's there the health data is taking us.  I think it

6  also diminishes some of the uncertainty in the exposure

7  analysis exercises.  That's just a thought.

8	DR. HENDERSON:    Any response?  Harvey?

9	MR. RICHMOND:    Well, I would like to

10   hear from the committee.

11	DR. HENDERSON:    Yes.

12	MR. RICHMOND:    I thought we heard a lot

13   in the fall in the discussion of ISA about concern

14   about that it was really the clinical evidence and tox

15   evidence supported concern about peak hourly type

16   exposures.

17	The exposure situations, analysis,

18   certainly that Steven has produced, you know, talks

19   about peak exposures and on road and we've heard a lot

20   of concern about that in the study suggesting proximity

21   to roadways where people don't sit for 24 hours.

22	So I would ask back for the other health

23   experts on this committee, when James Crapo said that

24   we probably need a short-term standard, are we talking

25   about 24-hour standards or are we talking about



90

1  one-hour standards?

2	Because one of the things we have to do

3  also is decide what alternative standards -- it could

4  be a pretty broad range, but we will have to decide

5  what alternative standards we're going to analyze in

6  this next phase and we don't have a lot of time, so I

7  would be interested in feedback as well on this

8  averaging time issue to sharpen it a bit.

9	DR. CRAPO:    I would argue that the

10   really important thing is that we establish a

11   short-term standard, the precedence of looking at that

12   and having one, and the form of it is less important

13   than that you form one and I would -- I would pick the

14   form that you can best defend as the first one.

15	It's going to go on for years, but your

16   question is what can you defend as you go to this very

17   important change?

18	DR. LARSON:    Well, again you've already

19   got the hourly predictions, so you could easily come up

20   with an aggregation of 24-hour averages.  So that's not

21   really a big deal.  I'm not saying you said it was, but

22   I think that's -- that's a fairly straightforward

23   analysis.

24	DR. BALMES:    This is John Balmes.  I

25   agree with what Tim is proposing as something	

92

1  or a comment.

2	Isn't the Federal Reference Method now a

3  continuous one hour or continuous and then an

4  accumulation of the one-hour measurements to make the

5  24-hour average?  Is that correct?  If that's

6  correct...

7	DR. THURSTON:    They don't use bubblers

8  anymore.

9	DR. SPEIZER:    Pardon?

10	DR. THURSTON:    They don't use bubblers

11   anymore.

12	DR. SPEIZER:    Right.  That's the old

13   fashioned way, but if that is correct and you've got

14   the data in the Federal Reference Method at hourly

15   levels, there are a considerable amount of work that's

16   been done on modeling the 24-hour averages and, indeed,

17   you could have -- you could have a peak one-hour

18   measure as your -- as your standard and you get all the

19   information you need.

20	DR. HENDERSON:    Well, okay.  Pat?

21	DR. KINNEY:    One way that we -- I think

22   that's emerging from this discussion, one way that you

23   can link the exposure and risk assessment to the

24   epidemiologic findings in a semi-quantitative way would

25   be to do a couple of exposure -- practical exposure



91

1  worthwhile to consider and I agree with James' context

2  that getting any kind of short-term standard would be

3  sort of a major advance and I think actually the epi

4  data, in terms of asthma and exacerbation are -- fit a

5  24-hour averaging time better than an hourly averaging

6  time.

7	DR. HENDERSON:    So you're recommending,

8  because you have more data at 24 hours, to have the

9  short-term standard be for 24 hours?

10	DR. BALMES:    I'm recommending

11   considering that at this point.

12	DR. KLEEBERGER:    Yes, this is Steve

13   Kleeberger.  I would agree with John and all the

14   discussions going on.

15	You know, I recently found transitional

16   epi studies that support a role for short-term

17   high-peak exposures related to death in infants

18   attributed to SIDS and these are three replicated

19   studies in Canada and two in the United States I

20   believe and they all associate with daily peaks of NO2

21   and averaging across 24 hours I think would give you --

22   at least in terms of these kinds of studies would be

23   the most appropriate way to go.

24	DR. HENDERSON:    Okay, Frank?

25	DR. SPEIZER:    I have sort of a question	

93

1  things.

2	I mean one clear thing is, in Section 6,

3  do some work on 24-hour averages, as well as for the

4  peak levels, using the ambient data and then, in

5  Section 7, likewise, you know, why not -- why not look

6  at the 24-hour average personal exposures that you're

7  modeling and importantly add an analysis of how, in the

8  apex modeling process, you end up -- what you end up

9  seeing in the relationship between the central site

10   modeled 24-hour average levels are in the apex analysis

11   and the distribution of personal exposures.

12	What is the relationship between those

13   two quantities?   Because the epidemiologic results

14   obviously are based on the central site type monitoring

15   and it would be very I think informative to our

16   thinking to understand, based on a careful analysis of

17   exposure activities and spatial distributions and

18   everything else, how the sort of typical personal

19   exposures relate to that central site data.

20	DR. HENDERSON:    Thank you, Pat.  Harvey,

21   did that sound -- are we telling you anything here?

22	MR. RICHMOND:    I guess I heard certainly

23   that the data is more available, the epidemiological

24   data, on a 24-hour basis, but I guess I'm asking this

25   committee's expert judgment in a sense.



94

1	Is that really where you believe the

2  health effects -- sort of the source of the health

3  effects, that's it's the 24-hour average or one-hour

4  average?

5	Because again there's a huge difference

6  between looking at alternative peak one-hour standards

7  and looking at 24-hour average standards and the epi is

8  at much lower levels.

9	As Pat Kinney said, is it a signal for

10   on a road as opposed to truly those low NO2 24-hour

11   averages?  Do we really mean to imply that?

12	DR. HENDERSON:    George?

13	DR. THURSTON:    Well, I just have a, you

14   know, clarification question.

15	You say there's a huge difference

16   between looking at the 24-hour and the one-hour.  In

17   what context?  In what sense do you mean that?

18	MR. RICHMOND:    I expect an analysis of

19   analyzing short-term hourly NO2 standards and again we

20   haven't talked levels.

21	Say if it was at our benchmarks in the

22   .2 to .3 range, you know, what kind of 24-hour

23   standards would we be looking at?

24	The epi doesn't point so far because we

25   haven't done that analysis of what their quality levels	

96

1  the epi supports going down at 24-hour averages?

2	DR. THURSTON:    Oh, I can't give you a

3  number.

4	MR. RICHMOND:    Well, that's the problem.

5	DR. CRAPO:    I will.

6	DR. THURSTON:    Well, I would have to

7  think about it.

8	SPEAKER:    In a nutshell.

9	DR. CRAPO:    No, seriously, the way I

10   would answer that is I want to know what level we would

11   have to set the standard at that would result in a 20

12   part per billion decrease for a significant segment of

13   our population.

14	SPEAKER:    For what average, 24 hours?

15	DR. CRAPO:    For a 24-hour average

16   because that's what the epidemiologic data...

17	MR. RICHMOND:    But a 20 PPB decrease

18   from as-is levels, the current...

19	DR. CRAPO:    Yes.

20	MR. RICHMOND:    ...air quality levels?

21	DR. CRAPO:    Yes, yes.  I mean there's...

22	MR. RICHMOND:    Because that 20 PPB was

23   chosen just as a standardized increment to...

24	DR. CRAPO:    I understand that.  I'm just

25   saying that -- saying that that overall exposure on a



95

1  -- other than to say it's below the annual standard.

2	The epi effect estimates are for a

3  standardized change, but there's no -- they're linear.

4  They're, you know, generally linear in nature down to

5  very low levels.

6	So the epi -- I would still like to see

7  how that informs a choice of alternative 24-hour

8  standards if that's what we're talking about.

9	DR. THURSTON:    Well, I would just say I

10   agree with James Crapo's advice, which is to look at --

11   I think the impetus is to get a short-term standard to

12   protect public health and you just -- you have to work

13   with what you've got and I hope that, in the next five

14   years, that people will, you know, look more deeply at

15   these questions that are identified in this process,

16   but you've got to work with what you've got and go with

17   an averaging period where you've got the strongest

18   epidemiological data to support it.

19	So that's the way I would choose between

20   the one-hour and the 24-hour.  You've got most of the

21   studies and you're going to base your risk analysis

22   mostly on a 24-hour average and, you know, that's the

23   way I would go.

24	MR. RICHMOND:    In your interpretation of

25   the epi, what are the appropriate range that you think	

97

1  24-hour average...

2	MR. RICHMOND:    Right.

3	DR. CRAPO:    ...is -- is associated with

4  those health benefits.  Why -- I just used -- because

5  that's your data.  I would use my gold standard and I

6  would say, well, then I've modeled -- if I set the

7  standard at this level for a 24-hour average, how many

8  -- and then I -- I would find a level at which -- at

9  which it didn't change anything in the country.  So you

10   can set it at this level.  Everybody is...

11	MR. RICHMOND:    Well...

12	DR. CRAPO:    ...below it and so it

13   doesn't have any impact.

14	MR. RICHMOND:    Well, the...

15	DR. CRAPO:    Then I would try it down in

16   increments until I started to see how many million

17   people would benefit with a 20 PPB benefit.

18	MR. RICHMOND:    You're going to see

19   current levels are variable.  They're clearly below the

20   current standard, they're in .03, but there's not a

21   uniformity.  There's going to be a variety.  L.A. has

22   higher levels than I would suspect Atlanta and smaller

23   cities.  Applying the epi, you're going to see a

24   continuous decrease no matter, you know, what levels

25   look at based on...



98

1	DR. CRAPO:    But that's exactly what I'm

2  saying and I'm arguing that you want to draw me a

3  curve, so let's -- I don't care.  Let's center it at

4  point -- for 24 hours, what's a reasonable one, .01 or

5  something?  I don't know.

6	Anyway, pick some number at which there

7  is no exceedance by anyplace in the entire United

8  States and then sequentially drop it and calculate how

9  many people in the United States would have a 20 PPB

10   decrease and give me a curve of that so we can see how

11   far you have to drop it before you start to get a

12   significant portion of the population with the benefit

13   and then you can start to analyze your data.  So it

14   will be -- it's a curve function, so draw us the curve.

15	MR. RICHMOND:	I think Ron wants to...

16	DR. HENDERSON:    Ron has got his hand up.

17	DR. WYZGA:    I think that there are a few

18   studies in the epidemiologic literature that look at

19   peak NO2 exposures and I think, you know, if you could

20   look at them systematically and see if they tell you

21   something any different and then at least my

22   understanding from reading the ISA is you basically

23   imply that 24-hour averages are similar to annual

24   averages, is that correct?

25	DR. GRAHAM:    We haven't done that	

100

1  distributed in the atmosphere and how it's measured.

2  It may or may not be related to the health effects and

3  you've done it for ozone, so you can do it for NO2.

4	MR. RICHMOND:    Can I respond to that?

5	DR. SHEPPARD:    Uh-huh.  (Indicating

6  affirmatively)

7	MR. RICHMOND:    A few points where I do

8  think they are different and would like again the

9  committee's advice.  I don't think she's saying whether

10   we are or not.  We have not agreed to pursue a

11   quantitative assessment yet.  We'll have to take that

12   -- you know, your comments and recommendations into

13   account, but I want to point out a few things and get

14   the committee's reaction.

15	In ozone, as in PM, we applied those

16   functions and got baseline incidence data and we

17   plotted it in the cities where we had the health

18   studies.

19	For some of these end points like

20   hospital admissions in the U.S., we have some studies

21   in Atlanta, we have Ito in New York City as an example.

22	I'm hearing Dr. Thurston recommend doing

23   a meta-analysis and implying as though we should apply

24   this in a much broader urban scale than those

25   individual cities, which would presuppose that the



99

1  analysis.

2	DR. WYZGA:    Okay, but then you may also

3  look at the association between hourly peaks and

4  24-hour average and see whether or not the association

5  is relatively consistent across, you know, several

6  cities as well.

7	SPEAKER:    Well, the average point...

8	DR. GRAHAM:    Well, I suspect that

9  relationship is going to vary obviously from monitor to

10   monitor based on its proximity to roadways and other

11   factors that may influence the concentrations.

12	DR. HENDERSON:    Leanne, you had your

13   hand up several times.  I don't know what...

14	DR. SHEPPARD:    Yes, I wanted to say that

15   using the epi evidence I don't really see that the

16   questions are significantly different than what was

17   done for ozone for the quantitative risk assessment

18   based on the time series studies and I would use that

19   as the model and the evidence for NO2 may be a little

20   bit weaker, but it's all a continuum and the questions

21   are all the same in all these epi studies and time

22   series studies where there is PM, or ozone, or NO2.

23	So they may be -- the reservations may

24   be a bit stronger for NO2 than for the other

25   pollutants, but that could be a feature of how it's	

101

1  monitoring networks are comparable, presuppose that the

2  baseline incidence rates are comparable, what baseline

3  incidence rates were used for much larger areas, just

4  huge uncertainties, which were not the way we did it in

5  ozone or PM.  They were much more city specific.

6	I'm hearing a lot of comments about

7  refining the exposure analysis to get into a lot more

8  detail in a much more refined level than we're asking

9  the epidemiology, just to wave our hands and apply

10   functions that we're going to get criticized do not

11   really necessarily -- we haven't shown how different

12   would they be for those cities because we don't have

13   the data for those cities.

14	Collecting baseline incidence data is

15   also time consuming and resource.  You know, we don't

16   have that on a national basis.  We have to go out and

17   get that individually and it's taken a lot -- you know,

18   it takes a lot more time than the two months we have to

19   turn around this document.

20	So I would like to know -- when the

21   committee envisions talking about doing a quantitative

22   assessment, I need a lot more concrete advice as to the

23   nature and assumptions that you're talking about making

24   in such an assessment and what is the degree of

25   uncertainty?



102

1	You're gonna be asking us back, if we --

2  if I pick certain assumptions, well how much difference

3  did it make in the assessment and we're not going to be

4  able to answer that.

5	DR. HENDERSON:    I think a basic problem

6  is there haven't been that many studies with NO2 like

7  there is with ozone, is that correct?  I mean you don't

8  have...

9	MR. RICHMOND:    Particularly in the U.S.

10   there have not been as many studies focused on NO2 as a

11   pollutant, yes.

12	DR. HENDERSON:    And so the data is not

13   there.  Go ahead, Ron.

14	DR. WYZGA:    Do you have the same

15   deadline as you have for the ISA?

16	MR. RICHMOND:    We have -- about a month

17   later -- you have the meeting scheduled for the end of

18   July, right?   That's SO2, sorry.  NO2.

19	SPEAKER:    September.

20	MR. RICHMOND:    September, which means --

21   early September, which means a month before that, so by

22   early August effectively we have to have the document

23   out the door under the current...

24	SPEAKER:    Is it a court ordered deadline

25   as well?	

104

1  the proposed -- is that the early September...

2	DR. JENKINS:	For the ANPR it's

3  -- the ANPR is December, this December.

4	SPEAKER:    She asked for a proposal date.

5	DR. JENKINS:    The proposal is a year

6  after that.

7	DR. HENDERSON:    Oh, okay.  Yes, George.

8	DR. THURSTON:    Well, you know, I

9  acknowledge that, for hospital admissions, we don't

10   have national healthcare, so we don't have national

11   information about that, but we do have it for people 65

12   years and older, who have been identified as a

13   susceptible population through Medicare and that's

14   available nationwide for every city in the United

15   States.

16	So, you know, I think a conservative

17   analysis would be to apply the relative risks that you

18   get from the studies to that subpopulation and then

19   that can be done -- you have the underlying hospital

20   admission risks and it's clearly an underestimate, but

21   that is one avenue that could be pursued.

22	Another is I know people in New York.  I

23   mean I think -- I know there is a central clearinghouse

24   for asthma admissions and, you know, it's a reportable

25   event, emergency room visits and hospital -- and of



103

1	MR. RICHMOND:    It's a consent decree,

2  basically...

3	SPEAKER:    Okay.

4	MR. RICHMOND:    ...the same as a court

5  ordered deadline.

6	SPEAKER:    Okay.

7	MR. RICHMOND:    So we have the rest of

8  May, June, and July.  We have less than three months to

9  carry out this analysis, so I would ask you to consider

10   that as you're making your comments as to what's

11   feasible and doable.

12	Definitely -- at the same time, the same

13   people are also doing the SO2 assessment and planning

14   for the PM assessment.  So I know that's not something

15   you may want to hear, but it's the same group of people

16   doing the work.

17	DR. MARTIN:    Just a -- just a

18   clarification.  The consent to preschedule we're

19   working under has a date for the final ISA.  It has a

20   date for our proposed decision.  It doesn't have

21   interim dates embedded for when the second draft of the

22   risk assessment has to get out, but if we don't keep on

23   the schedule, we can't meet the next consent to create

24   for a proposed decision.

25	DR. HENDERSON:    And what is the date for	

105

1  course hospital admissions, but emergency room for

2  asthma in New York, so an analysis could certainly be

3  done for New York.

4	There may be other cities where that

5  could be done, but, you know, the question of time is

6  pressing, it's true, but what about the Medicare data?

7	MR. RICHMOND:    Are the studies, though,

8  that are applying those functions, are they -- do they

9  match up?  Are they broken out by those same age

10   breakouts?

11	DR. THURSTON:    I would have to look, but

12   I don't think so.

13	DR. BALMES:    This is John Balmes.

14	DR. THURSTON:    You know, I don't think

15   they match up exactly, but they -- you know what I'm

16   saying is there are mortality studies that have looked

17   at that and based on -- you know, the document is

18   saying that this isn't especially susceptible.

19	So applying that risk would be a

20   conservative thing to do. Applying the general

21   population relative risk to that population would be

22   conservative, right?

23	DR. BALMES:    This is John Balmes.  You

24   know, I'm part of the CDC funded environmental public

25   health tracking network and we've chosen hospital



106

1  admissions for -- as the health end point with regard

2  to air pollution and asthma because of its availability

3  in virtually every state of the country.   You have

4  hospital admission data.

5	With regard to Emergency Department

6  visit data, that's less widely available, but

7  California and New York both have that, for example,

8  two of the most populated states in the country, but

9  for hospital admission data, you should be able to get

10   that for just about every state in the country.

11	I guess, you know, it does become a

12   resource in terms of time issue.  They have to go to

13   each state at this point, but they are available.

14	DR. HENDERSON:    Ken, when is the policy

15   assessment document due that's going to take into

16   account the epi studies?

17	SPEAKER:    For the ANPR.

18	DR. JENKINS:    Well, I mean what we've

19   been talking about doing and what Karen had mentioned

20   and what I think I talked about a little bit yesterday

21   and just touched on briefly was, for example, as we

22   identify potential alternative standards, we can take

23   the -- look at the epi studies and look at the air

24   quality levels in those -- in the locations where epi

25   studies were done and where associations were seen and	

108

1  than what we included in the risk assessment, but we do

2  have to make some decisions about what are the

3  appropriate range of standards that should be included

4  in the next phase of this assessment.

5	This standard -- this only looked at

6  meeting the current annual standard, which we meet

7  everywhere, and the as-is situation, which is lower.

8  We have not looked at any alternative standards.

9	So I don't want to phrase it a little

10   bit differently, but it's what alternative standards

11   should be looked at in subsequent exposure and/or risk

12   assessments?

13	DR. HENDERSON:    And that really is based

14   on the health effects.  It seems to me that there's --

15   the health effects associated with the peak exposures

16   are of interest.  I mean I -- and I haven't heard

17   anybody say that that's

18   -- that we should be considering that.  Go ahead,

19   Leanne.

20	DR. SHEPPARD:    Well, I would agree that

21   the peak exposures are of interest, but the data that

22   we have is based on 24-hour average and time series

23   study design.

24	So I think, if there's going to be a

25   quantitative risk assessment, it has to be focused on



107

1  compare levels in those locations against the potential

2  alternative standards that are under consideration.

3	So by doing this we can make a -- take

4  -- make some sort of a qualitative judgment and can

5  form an ultimate decision about what level of public

6  health protection is provided by the different

7  alternative standards under consideration through

8  comparison to the levels that have been -- the levels

9  present in locations where epi studies have been

10   conducted.

11	DR. HENDERSON:    Okay and that's -- the

12   policy assessment will be written when?  I mean when do

13   we get see it?

14	DR. JENKINS:    Well, the date on that is

15   December, this coming December, December of '08.

16	DR. HENDERSON:    Okay.  Well, I don't

17   think we've really answered the question Harvey posed

18   about do you want to -- do health effects suggest we

19   have a one-hour standard or a 24-hour standard or both?

20   I mean how have we tried to answer that?

21	MR. RICHMOND:    Or at least which

22   standards would you, you know, focus on?

23	I'm not saying about the ultimate

24   decision because, as Karen has said, we're going to

25   look at the policy assessment as well that's broader	

109

1  that because that's where the data are and looking back

2  at the ozone staff paper for what was done there,

3  there's three pieces to that.

4	There's the air quality piece, the

5  concentration response piece, and then the estimates of

6  the city-specific health effects incidence rates and

7  population data and you have a number of cities where

8  you have that because you've already done it and so I

9  think the first question is -- that would make it ideal

10   is do you have any studies for NO2 that are in those

11   cities and use that as a way of limiting the workload

12   and, if you can't do that, personally I think the

13   concentration response function can be applied to other

14   cities where it wasn't estimated.  One could argue

15   about that, but that is something that I think is

16   acceptable.

17	So in terms of actually doing a

18   quantitative risk assessment based on the epi studies,

19   I would focus on the areas where you've already done a

20   tremendous amount of work and then see what additional

21   assumptions you have to make in order to make it

22   relevant for NO2.

23	DR. HENDERSON:    And George?

24	DR. THURSTON:    Yes, I just concur that

25   you can apply the relative risk across cities, you



110

1  know, especially where you've done a meta-analysis of

2  multiple cities that use that.  The key there is to

3  have the baseline risk that Harvey has brought up that

4  you're applying the relative risk to.

5	You know, what is -- what is the

6  present?  If you don't have that, then you can't really

7  do a city-specific estimate.

8	But I think, you know, based on what

9  John Balmes is saying and what I was saying about the

10   Medicare data, yes, there are data out there to do many

11   cities and to get at the width and breadth of this

12   problem.

13	MR. RICHMOND:    I'll also ask a question

14   specific on air quality.

15	If you were to apply these effect

16   estimates from the epi to cities, in PM we averaged --

17   because most of the monitors are population --

18   population-oriented monitors -- would you restrict

19   which monitors?

20	Is it applying the air quality data to

21   the average of all the available monitors across the

22   city each day, which is what we did in PM excluding the

23   few limited -- if we found a source-oriented monitor,

24   it was excluded, as it was in the epi.

25	In this case, would you average across	

112

1  studies where they've analyzed the data using the

2  central site monitors and then you are using central

3  site monitors to estimate the risks.

4	So that's, you know, the link and that's

5  an important one and it gives you a lot of -- reduces

6  uncertainty quite a bit.  If they had used personal

7  monitors on every one and done the estimates, then you

8  would have to go out and get personal monitors and

9  then, you know, backtrack to the central-site monitors

10   because that's what's going to be used for the

11   regulation, but the epidemiology has already done all

12   that work for you and linked the exposures and the

13   health effects to the central-site monitors, but I --

14   you know, you'll have to look at each study, maybe

15   contact the authors and say did you exclude from -- you

16   know, I think most people average all the sites and I

17   think they do tend to exclude the hot spot monitors,

18   but I don't know for sure.

19	MR. RICHMOND:    I could ask Ron.  You

20   have one of the co-authors.  Are you not a co-author on

21   the area studies?   You're certainly familiar with

22   that.

23	DR. WYZGA:    No, what we -- but I think

24   that -- I guess a couple of things.  I was going to

25   say, first of all, I think that in some cases you could



111

1  all the monitors or would you exclude roadway --

2  near-roadway monitors and what would be the criteria

3  for such?

4	DR. HENDERSON:    Leanne?

5	DR. SHEPPARD:    For the time series study

6  design, I think you would need to focus on the

7  population-oriented monitors and therefore you would

8  need to exclude the near-road monitors.

9	You know, we are -- for NO2, we know

10   that there's important near-road effects that need to

11   be taken into account, but I don't think this study

12   design or this evidence actually is where you're going

13   to be able to ask those questions.  So that's where the

14   qualitative discussion has to come in.

15	The quantitative risk assessment can

16   only address a piece of it because that's the only

17   numbers that we have.  The rest of it needs to be done

18   differently until the research is better, is more

19   complete in those areas.

20	DR. HENDERSON:    George?

21	DR. THURSTON:    Well, I would find out

22   what was done in the epidemiology.  I mean one of the

23   strengths of the analysis here is that you're applying

24   central site monitoring data.

25	You know, you're using risks from	

113

1  be limited using the 24-hour basis, but you can try and

2  sort of look at some of the associations, you know,

3  within a specific city between one-hour maxes and

4  24-hours to sort of try and milk as much as you can

5  about that whole one-hour versus 24-hour issue.

6	I'll say that in Atlanta there was a

7  part of that study that Helen Su collected some data

8  where she has daily personal data of NO2 exposures I

9  think measured continuously.

10	She measured them outside people's homes

11   and we also have central monitoring data and, you know,

12   I think -- I think that, you know, that's one city

13   where those data may be of help.

14	MR. RICHMOND:    Are those published and

15   in the ISA?

16	DR. WYZGA:    I believe that the paper has

17   been accepted for publication, but I don't think it's

18   been published yet, but she has, you know, the raw air

19   quality data and I don't know if...

20	MR. RICHMOND:    Yes.  That's obviously a

21   consideration for us.  It would have to get into the --

22   into the finalized state for us to use that kind of

23   data and I appreciate knowing about that data certainly

24   and seeing it, but that would certainly encourage -- if

25   you would encourage her to get it to -- to get it to



114

1  Mary Ross and the folks in NCEA...

2	DR. WYZGA:    Okay.

3	MR. RICHMOND:    ...for us to look at and

4  consider.

5	DR. WYZGA:    But I guess one of the

6  things is one's a study and the other is sort of the

7  air quality data, which, you know, are not going to be

8  published as part of the study, but they were used in

9  the study.

10	MR. RICHMOND:    Right, but my question

11   was are you familiar with whether for the Tolbert, the

12   Paige studies, did they use all the monitors including

13   ones that were near roadways or did they -- did they

14   just focus on population?  Do you know that?

15	DR. WYZGA:    They focused on population.

16	MR. RICHMOND:    Oriented monitors?

17	DR. WYZGA:    Yes.

18	MR. RICHMOND:    Okay.

19	DR. HENDERSON:    I'm wondering if some of

20   the people on the phone are wanting to say something.

21   John Samet?  You are an epidemiologist.  What are your

22   views?  Are you there?

23	DR. SAMET:    I'm here.  I've been

24   listening with interest.  You know, it's like when

25   you're in class and you hope not to get called on to	

116

1  know, the right thing to do is to go in and do a quick

2  sort of multi-pollutant analysis and have an NO2

3  coefficient.  That makes me nervous frankly.

4	So I think if the question could be

5  sharpened enough around what kind of an office would be

6  helpful and why.  I think the question of what is the

7  right time is a -- and this goes back over to what

8  James was saying, is it a one-hour, a, you know,

9  three-hour, a 24-hour?

10	To me it would go as much back to what

11   we know from the toxicology.  I mean the

12   epidemiologists -- we study the time domains that are

13   sort of handed to us by those who monitor.  We don't

14   know exactly what the sort of time response

15   relationships are without help from toxicology.

16	So I guess those are my thoughts now.  I

17   mean I think your question is how can we -- you know,

18   how can we move forward in the remaining few hours to

19   try and come to some sense of what will be helpful?

20   That's it.

21	DR. HENDERSON:    Okay, John.  You've laid

22   out the dilemma.  What analysis would help standard

23   setting?  I gather it's what would normally be in the

24   staff paper, if that -- which you've explained to us,

25   Karen.  Can we have that now?



115

1  answer the question.

2	You know, I guess what's all mixed up

3  are complicated matters of process and science and as

4  I've heard the discussion play out, I guess the

5  question that's come to my mind repeatedly is what --

6  what analysis in the end would benefit setting a

7  standard?

8	And I'm not -- I'm not sure I've heard a

9  sharp answer to that.  I mean I guess you could

10   question the existing risk assessment.  Is that a -- is

11   that a benefit based around these short terms?

12	So we are really I think in a quandary

13   around sort of all elements of a potential new NAAQS

14   related to, you know -- well, indicative we've got, but

15   form and level I think are difficult.

16	So it seems to me that if we just could

17   -- putting aside all the difficult side questions, I

18   guess one point I'll make is -- I mean that I don't --

19   I don't think this is the time to consider new

20   epidemiological analysis.

21	I've gotten the sense that some are

22   suggesting that that might be the case, but, you know,

23   we, for example, have all the Medicare data up and

24   available.

25	I would be reluctant to say that, you	

117

1	DR. MARTIN:    Well, the types of analyses

2  that I alluded to before, that Scott alluded to, and a

3  number of folks have pointed back to the ozone as an

4  analogy and, of course, in ozone we did exposure

5  assessment, quantitative risk assessment driven by epi

6  and clinical inputs.

7	We did the more qualitative

8  evidence-based consideration, simply asking the

9  question at what exposure levels did we see effects in

10   clinical studies, at what air quality levels did we see

11   effects in epi studies, and if you look back at the

12   basis for the decisions there, you see a lot of

13   emphasis being placed on the more semi-quantitative

14   evidence-driven kinds of questions, at what exposure

15   level were effects seen as opposed to putting a lot of

16   weight on the quantitative numbers of exactly how many

17   additional hospital admissions in City X or Y primarily

18   because of the big uncertainty ranges around those

19   numbers.

20	They prove to be less meaningful as a

21   basis for decision making than the more general

22   messages that these different kinds of evidence provide

23   for us and so, looking back at the history of decision

24   making is what's helping inform our thinking on this

25   and so it's -- it's -- there are different views as to



118

1  what kinds of information ought to be the primary basis

2  for decision making and we all may ave different views

3  on that.

4	But the question here is, in the context

5  of this quantitative risk assessment document, are

6  there ways to grapple with all the quantitative

7  challenges that allow us to do something meaningful in

8  a quantitative risk assessment way with the epi that

9  would be one more piece of information that could be

10   added to the other kinds of information that we will be

11   pulling together in the next document we do?

12	So I think that still would be the most

13   useful-to-us way to focus your continuing discussion

14   here and that's what you've been talking about and I

15   think that's still the most useful thing.

16	DR. HENDERSON:    Okay.  Yes?  Who wants

17   to -- oh, is it Dale?

18	DR. HATTIS:    Yes.  I want to respond to

19   Harvey just a little bit.  I do think that something

20   meaningful is possible with concentration response

21   functions.

22	The issue that requires some

23   toxicological judgment is on the time scale or which

24   the averaging time for the proposed standards and there

25   I think what you need to ask is for the kind of --	

120

1  familiar with the likely causal pathways and their

2  reversal rates.

3	DR. HENDERSON:    Thank you, Dale.  It is

4  important to know how reversible and how quickly they

5  are reversible.

6	Do any people want to -- have additional

7  comments on whether the 24-hour -- I'm wondering if the

8  more sensitive causality end point is this effect on

9  the immune system.  Is that true or -- I'm asking.

10	DR. HATTIS:    I don't know the data well

11   enough to be -- but at least that leads -- that gives

12   me at least a plausible pathway to the kind of

13   impairment of long-term lung function growth that seems

14   to be, you know, among the more interesting -- or even

15   the hospital admissions.

16	DR. HENDERSON:    I see that.  So I think

17   it's important that the immune response is going to be

18   harder to put your finger on, but...

19	DR. HATTIS:    Sure.

20	DR. HENDERSON:    ...it's certainly

21   something that people would be concerned about.  I mean

22   that's -- but lung growth would be a different

23   population.  I mean it would...

24	DR. HATTIS:    Lung growth is...

25	DR. HENDERSON:    ...be the children.



119

1  assuming NO2 does causally do something, over what time

2  scale would the effects likely be reversed that are

3  leading to the effects of interest?

4	So I'm going to put out a hypothesis

5  that, in fact, what matters is the change in the immune

6  response functions.  Now, we could have other

7  hypotheses and the issue is -- and I think you can do

8  -- to some extent this sort of diagnosis leads me to an

9  idea that I'm more likely to have responses

10   accumulating over a day-- that exposures accumulating

11   -- the effects of exposures are likely to accumulate

12   over at least a day rather than that they're reversed

13   so quickly as the FEV1 immediate response, but even the

14   FEV1 immediate response from the ozone analogy, you

15   know, seemed to take quite a few hours to be fully

16   manifest.

17	So I mean I would, as my original -- as

18   an initial response to Harvey, think more closely about

19   daily averaging times than -- than -- than just

20   assuming that if you get a high excursion in one hour,

21   that would be irrelevant and completely reversed by

22   three hours later.

23	I would expect not, but that's, you

24   know, one person's judgment and I think you should

25   probably seek the judgments of people who are much more	

121

1	DR. HATTIS:    ...one thing and the

2  hospital admissions are another, but they both -- you

3  know, so the question is, okay, what's the most

4  plausible, you know, mechanistic pathway and if that's

5  contingent on that, what is the -- what are the

6  response times that are likely to be -- and reversal

7  times that are likely to be relevant?

8	DR. CRAPO:    I would like to say one

9  thing in response to Dale's comment.  You know, I've

10   studied previously really high levels of NO2 and some

11   fairly low levels in biological experiments and

12   clearly, if you go to really high levels, you know, 10

13   PPM or something up really high, you get very acute,

14   severe events that have a repair cycle that we can see

15   and measure.

16	The levels we're talking about are

17   really low by that comparison and the -- and I, in my

18   -- I can't prove this, but my gut instinct from sort of

19   a general understanding of the biology is that the --

20   you're not going to see the levels we're talking about

21   creating these kinds of a -- of an acute repair event

22   that comes and goes in an hour, but, in fact, you're

23   looking at some accumulative oxidative or nitrated

24   stress that's going on that's affecting signals in

25   there because you're looking at quite low levels and it



122

1  -- I'm much less worried about the one-hour averaging

2  time versus the 24 because I don't think you're looking

3  at acute inflammatory injury caused by any peak that

4  we're talking about.

5	We're probably 100 times or at least ten

6  times below that level, but I think that would be the

7  process driving it.  This is all just kind of instinct

8  thinking about how the biology is going.

9	So I -- I don't have a concern which

10   averaging time you pick.  In terms of what I think

11   about the biology, I'm more concerned that you pick --

12   that we look at a short-term average that's probably

13   going to -- and I'm going to -- I'm going to bet that

14   eventually, when people do more critical studies on

15   this, that you're going to find that they all are

16   surrogates for each other.  That's what I think is

17   going to happen.

18	DR. HATTIS:    Well, I agree that they are

19   going to be surrogates for each other, but they're

20   going to be imperfect surrogates.

21	So that -- and that has an implication

22   that the coefficients, effect coefficients, that the

23   epidemiologists observed by, say, taking the 24-hour

24   averages or the highest 24-hour average in a year or

25   whatever -- whatever dosemetric they have that they've	

124

1  seasonal differences.  I know that that may be a very

2  important factor, but I don't know if anyone has ever

3  looked at that.

4	For example, with RSV, typically in

5  children we have -- really our greatest problem is in

6  January and February, at least in California, but can

7  those be related to winter or seasonal differences in

8  terms of NO2 concentrations?

9	And I realize that there are a lot of

10   confounding factors if PM levels are also fluctuating

11   as a function of season.

12	DR. HENDERSON:    And Leanne?

13	DR. SHEPPARD:    So I wanted to try to

14   bring this back to see if we could address your

15   questions in the context of our objections.

16	So it seems to me that the basic

17   question is how do we argue for a short-term standard

18   and is the work that's being done so far or could be

19   done, how do we get there?

20	And the information that's in the

21   current document is I think limited because it's

22   focused on exceedances which have big problems with

23   modeling because you have to get at the tails of the

24   distribution and big problems from a health

25   plausibility point of view as we heard from Dr. Crapo.



123

1  used as their independent variable are going to be

2  imperfect proxies for that causal parameter.

3	Given that, there's likely going to be

4  some tendency to underestimate the effect size.  At the

5  same time, there will be some overestimation of the

6  effect size because of the possibility that NO2 is --

7  you know, is correlated with some other causal

8  parameter that also causes the effect.

9	So there are -- there are potential --

10   there are likely areas in both underestimation and

11   overestimation and that has to be said in terms of a

12   fair communication of the uncertainties.

13	That having been said, I think one can

14   make estimates and attach some, you know, what-if

15   calculations.  So, I mean, I would say that, you know,

16   any estimates that are made along those lines, you

17   know, will conservatively have order of magnitude

18   uncertainty associated with them.

19	That doesn't mean that they necessarily

20   should not be used for regulatory choices, but I think

21   that it's fair to communicate that those --

22   uncertainties of that order are likely there.

23	DR. HENDERSON:    Kent?

24	DR. PINKERTON:    If infection or viral

25   susceptibility is a factor, one might also consider	

125

1	The epi data provide alternate evidence

2  that could be used from a quantitative point of view

3  and the only thing that can be done feasibly in the

4  order of two months is to rely on previous work you've

5  done and the published literature to do a quantitative

6  risk assessment in epi.

7	And I think the question is can -- will

8  that move this whole agenda forward?

9	Will this help to answer the question or

10   are we going to have to rely exclusively on the

11   qualitative -- the more qualitative or

12   semi-quantitative evaluation?

13	I think that's what it comes down to.  I

14   think the quantitative assessment right now is

15   problematic and I'm wrestling with how do we help you

16   move forward in terms of a different quantitative

17   assessment and I guess, without having all the --

18   looking at all the details, which is where the devil

19   is, is in the details, of course, I would say you can

20   do it, but you've done enough work in ozone and there

21   are concentration response functions out there that

22   could be applied and that you do need to restrict to

23   population-oriented monitors because they're all time

24   serious study designs and that is the relevant feature

25   of -- that's used in those designs in terms of exposure



126

1  variation.

2	And a lot of things are going to be

3  ignored that are important as well, the one-hour

4  exposures, the near-road exposures.  Many things will

5  just have to be addressed qualitatively.  They can't be

6  addressed quantitatively and certainly not in two

7  months.

8	MR. RICHMOND:    Just as we asked for a

9  committee response and prioritization, again, if we

10   were to do any quantitative using the epi in a

11   quantitative risk assessment, we've heard a lot of

12   discussion that's bounced around for a variety of

13   different health end points, everything from mortality,

14   to symptoms, to hospital admissions, to Emergency

15   Department visits.

16	Again, knowing what you know about the

17   available evidence in the U.S. in particular, which of

18   these end points would you give the highest priority to

19   including in an assessment?

20	DR. HENDERSON:    Well, you notice from

21   the silence that we all don't have a magic answer.

22   Frank, do you want...

23	DR. SPEIZER:    There's no magic answer,

24   but Table 53 -- Figure 531 lays out what the

25   respiratory effects are and you've got both mortality	

128

1  utilization, just a whole host of differences.

2	It's one thing to use those to conclude

3  the evidence to support whether NO2 may have an effect.

4  It's a very different thing to take that and apply it

5  in the U.S. for purposes of quantifying how much of an

6  effect.

7	DR. SPEIZER:    Well, the thing is within

8  each city you control for all those things.  I mean in

9  a time series analysis those are basically internal

10   controls.  So it's the deltas that you're looking at

11   within cities, isn't it?

12	DR. LARSON:    This is Tim Larson.  The

13   deltas would be different probably just because most of

14   the European monitors are right next to the roadway.

15	DR. SPEIZER:    Well, you would have to

16   obviously exclude those anyway.

17	DR. LARSON:    But they're all like that.

18   It's a totally different network than the U.S.

19	DR. BALMES:    This is John.  Can I

20   comment on this?  I mean we just -- we've done this

21   FENa analysis, which is a combined time series analysis

22   of the European and North American studies and for PM

23   and ozone at least the European/U.S. coefficients are

24   quite similar for what that -- this is for mortality

25   for what that's worth.



127

1  and ED visits, hospital admissions sort of sitting out

2  there with lots of points and to the degree that you

3  can identify cities that relate to those points, I

4  think you go with any of those as a group, which I

5  think what you want to do is maximize the potential

6  diversity of sites that are available in these data

7  and, you know, I can't tell you right now whether

8  that's going to end up being respiratory and mortality

9  or whether it's going to end up being hospital

10   admissions, but I think you're going to get a pretty

11   good smattering across the country from those data.

12	I also think you probably might want to

13   include some of the overseas data, particularly the

14   European data where I think you've got perhaps a little

15   more spread in what the exposure might be because I

16   suspect that most of the NO2 measurements are going to

17   be somewhat higher, so I think that might help some.

18	MR. RICHMOND:    Do other -- I would like

19   to know specific on that last point because I would be

20   inclined not -- other than in a qualitative support way

21   of taking anything quantitative from the European

22   studies and applying it to a U.S. city given the

23   differences in the copollutants, given the differences

24   in the monitoring relationships, differences in health

25   incidents and hospital -- you know, hospital	

129

1	DR. HENDERSON:    Do other people have

2  comments on whether the European cities should be

3  looked at?

4	DR. LARSON:    Well, the PM and ozone

5  measurements right next to the road -- PM right next to

6  the road, it's not that big of a contribution compared

7  to NO2s or NAAQS, so I would suspect that you wouldn't

8  see a big difference in the PM relationships.  Ozone

9  might be a little odd, but NO2 is quite different right

10   next to the road, as we know.

11	DR. HENDERSON:    Yes.

12	DR. LARSON:    So it's just a micro-scale

13   philosophy there as you know.  So I would -- I would

14   probably say that, you know, you would -- you would end

15   up underestimating the coefficient just because you

16   have systematically higher values in the network

17   compared to where they would -- what they would be at

18   sites comparable to the U.S. sites.

19	DR. HENDERSON:    Thank you.  Anybody else

20   have comments on use of European cities?  George, do

21   you want to say something?

22	DR. THURSTON:    Well, I mean I share

23   Harvey's concern and I think that that should be a

24   factor in selecting what health end point to do this

25   analysis, but I do think, you know, we're looking at



130

1  airway hyperreactivity and the benchmark analyses, so

2  we ought to try and stay with something related to that

3  if we can, so obviously respiratory and asthma, either

4  ER visits or hospital admissions and I think you would

5  have to sort of make up a table of the things you would

6  like to have and try and figure out, well, this one, we

7  have more U.S. studies to go on.

8	That's a good thing, but then which ones

9  do you have?  Which ones can you get the data out there

10   for?

11	You know, you know you can get elderly,

12   you know, older adults.  I don't like elderly anymore.

13   I'm getting too close to 65.  Older adult hospital

14   admissions you can get nationwide and so, you know, I

15   think it's sort of you've got to look.

16	I don't have a table like that, but

17   these are all considerations in choosing, but I do

18   think you want to stay -- also the documents -- we

19   don't want to do mortality because mortality

20   -- I mean that's -- when we get to the end of the ISA,

21   we're not saying that we really have a firm conviction

22   that mortality is related to NO2, so we want to stay

23   with what we have our convictions about at this point

24   in terms of the causality, so that would be, you know,

25   hospital admissions or ER visits.	

132

1	DR. SPEIZER:    My concern about the

2  respiratory symptoms would be that you've got a mixture

3  of prevalence and it may not be acute response that

4  you're looking at there.  Some of that may be chronic

5  exposure.

6	DR. BALMES:    This is John Balmes.  I

7  guess I was somewhat reassured about the use of the

8  European study data by the comment that Tim Larson just

9  made that, if anything, it would be an underestimate of

10   the concentration response and that could be, you know,

11   clearly stated in the discussion about assumptions that

12   are -- the underlined assumptions that are made about

13   the use of those data.

14	DR. HENDERSON:    I would like to ask

15   Harvey, if you do a type of quantitative risk

16   assessment, whatever you do, would this analysis help

17   you in standard setting?  I mean I'm going back to...

18	MR. RICHMOND:    That's a tough issue as

19   Karen has already addressed that because we've got a

20   much more extensive quantitative analysis than we're

21   possibly going to be able to do in the next few months

22   for ozone and PM and it was used in a more qualitative

23   fashion in the most recent decisions.	So I

24   really can't answer that question at this point, but,

25   you know, we'll have to do -- you know, if we do an



131

1	So that narrows it down to that and

2  then, you know, I think you have to look at these

3  concerns, you know, which places do we have more U.S.

4  studies and which ones do we have the actual underlying

5  data to apply the risk coefficients in the U.S.

6	MR. RICHMOND:    The one other category

7  you didn't speak to was actually several of the

8  multi-city studies are the respiratory symptoms for

9  asthmatics, the inner-city asthma studies.

10	Where would you put that?  They're not

11   ER visits or hospital admissions.  They're symptoms.

12	DR. LARSON:    Is that Harvey who just

13   said that?

14	DR. THURSTON:    I think you're going to

15   have trouble getting...

16	DR. HENDERSON:    Yes...

17	DR. THURSTON:    ...data.   I don't know.

18	DR. HENDERSON:    ...Harvey said that.

19	DR. THURSTON:    Maybe you can -- maybe

20   you do have -- but...

21	MR. RICHMOND:    That's what's emphasized

22   in the ISA.

23	DR. THURSTON:    Yes, but I think the

24   hospital admissions and the ER visits are the ones.

25   That's where I think we should put our effort into.	

133

1  analysis, we will have to appropriately caveat it and

2  combine it with the evidence-based approach and what

3  the policymakers and which policymakers are in office

4  at that time will have to deal with it.

5	I can't possibly address that.

6  Obviously the staff -- if you're saying, you know,

7  under the current regime and the kind of decision

8  making that's been made most recently, the reason why

9  we didn't put forward and do the assessment is we did

10   not feel, given the benefits and costs so to speak,

11   that it would offer that much given the resources and

12   time and what we would get out of it.

13	We clearly made the recommendation and

14   pursued the analysis the way we did given our best

15   judgment of what value it would have given what we know

16   right now.

17	DR. HENDERSON:    Yes.

18	DR. SPEIZER:    But we do know at the time

19   this decision will be made there will be a different

20   administration, so therefore I don't think that should

21   be part of the consideration.

22	DR. HENDERSON:    I'm thinking about it

23   for any administration, if we know that a quantitative

24   risk assessment will have a lot of variability of -- a

25   lot of uncertainty I should say associated with it and



134

1  we know we can use it in a qualitative fashion, right?

2	I mean and so I'm just wondering if,

3  with the amount of data we have, does it -- is it

4  beneficial to do the quantitative analysis now or wait

5  until you get more data?

6	And these are the thoughts that are

7  going around in my mind.  I'm not recommending one or

8  the other.  I'm just

9  saying...

10	MR. RICHMOND:    And we will have to go

11   back and consider it along with other priorities in the

12   office, resources, other pollutants that Ross is

13   dealing with.  It all has to be in the mix, but I'm not

14   going to give an answer to that today.

15	DR. HENDERSON:    Yes.  Well, I -- well,

16   we have had a good discussion I think.  As we've said

17   all along, it's a very difficult problem.  NO2 is not

18   -- we don't have as much data as we would like for NO2

19   and we have the problem of wanting to use a

20   multi-pollutant approach, but we're not quite there

21   yet.

22	The Clean Air Act is still approaching

23   -- you know, we're still operating under one pollutant

24   at a time, so it makes for difficulties, but I think

25   this has been a productive discussion.	

136

1	So while we could give you a preview of

2  what you're going to see when the document comes out,

3  it's really beyond the point where meaningful input

4  would be -- would be able to be taken into account.

5	DR. HENDERSON:    So this is a first draft

6  and the next draft is -- we would see at what -- at

7  what date?

8	DR. JENKINS:    Yes, we're going to send

9  it to you in probably early to mid August and then we

10   have a CASAC meeting and I'm not sure, Angela, if

11   that's been definitively scheduled for early September.

12	DR. NUGENT:    September 9th and 10th.

13	DR. HENDERSON:    Okay.

14	DR. MARTIN:    When we originally set out,

15   we were expecting that it would be at the end of

16   September and as the scheduling difficulties arose, the

17   date kept creeping earlier and earlier and that just

18   further and further constrains the time we have to do

19   this and that just is what it is.

20	DR. HENDERSON:    But we will make

21   comments on it in early September, but that will be the

22   final draft at that time.

23	MR. RICHMOND:    The second draft we then

24   have in our schedule to take into account your comments

25   to produce a final report.



135

1	DR. SPEIZER:    Regine?

2	DR. HENDERSON:    Yes.

3	DR. SPEIZER:    I earlier raised the

4  question of a subcommittee working with the staff and I

5  understand that that's not possible and I can

6  appreciate why it might not be, but given that they

7  have their own consultants and they can get their own

8  advice, I'm comfortable with that.

9	However, I would like to ask Karen and

10   Harvey whether or not we could hear a little more about

11   this, perhaps as a -- as part of a FACA meeting, which

12   the next one that comes up is the SO2 one in July and

13   whether or not they could make some progress to that

14   point and provide a sort of interim report to us at

15   that point.  Is that a fair request?

16	DR. HENDERSON:    Well, I'll ask them if

17   that's a fair request.

18	You know, ideally I would like to have a

19   staff paper and review it twice, but in the absence of

20   this, would there be anything to report at the SOx

21   meeting?

22	DR. MARTIN:    Well, that's occurring at

23   the end of July, and as we talked about, this next

24   document needs to go out the middle of August at the

25   latest.	

137

1	DR. HENDERSON:    Okay.

2	MR. RICHMOND:    Just like the ISA.

3  You've just reviewed the second and they'll make

4  changes and do a finalized...

5	DR. HENDERSON:    So we will see it in

6  early September, Frank, and we will have an opportunity

7  to comment on it.

8	DR. SPEIZER:    Yes.  Would it help the

9  staff if we saw it -- saw it even earlier and gave them

10   a little more time for their final draft knowing that

11   they're going to get some comments from us in August?

12   You know, why not get them in July?

13	DR. HENDERSON:    Well, that would be up

14   -- I think we respond to your request for help, so you

15   let us help as much as we can and we'll respond.

16	So let us know if we can, if it would be

17   helpful for us to say something at our July meeting.

18   We will be together in July.  Yes, George?

19	DR. THURSTON:    You know, I never got any

20   feedback from the staff about Frank's suggestion of a

21   subcommittee that would work with them.  Huh?  Did I --

22   did I miss it?  I just missed it.

23	DR. SPEIZER:    Well, on a side

24   conversation I understood that we can't do it.

25	DR. THURSTON:    Oh, I'm sorry.



138

1	DR. NUGENT:    If that conversation was

2  with me, whatever interaction members have with the

3  agency that's providing advice, it just needs to be in

4  a FACA context.  So it needs public meetings and...

5	DR. THURSTON:    Yes, that's okay.

6	DR. NUGENT:    So a subgroup could have --

7  it's possible.  If the...

8	DR. THURSTON:    Or a conference call or

9  something?

10	DR. NUGENT:    A conference call, yes.

11	DR. THURSTON:    That would...

12	DR. NUGENT:    I mean if that's what OAQPS

13   would like it's a possibility.  We don't have to

14   resolve it here.

15	DR. MARTIN:    The suggestion is if, as we

16   go away from this meeting and we grapple with the

17   issues raised and think about how we think we can best

18   make our way forward, if we've identified some

19   particular salient questions that we think further

20   discussion with you would be helpful, there's the

21   potential that we could work with Angela to schedule a

22   publically noticed public teleconference perhaps that

23   would focus on specific questions that we would T-up

24   for you and the best I can say at this point is that

25   may or may not be helpful to us depending on how our	

140

1  written input from the leads, the lead discussants, at

2  least by next Friday.

3	DR. HENDERSON:    Or today if you have it.

4	DR. NUGENT:    Right and we're hoping that

5  goes to both Regine and me and then we would turn --

6  integrate the document and our goal is to have it to

7  you a week later, so that would be by May 16th I guess

8  that would be and then there would be a short

9  turnaround time for the committee and we're hoping that

10   both reports, the one on the ISA and the risk and

11   exposure assessment documents, would be posted on the

12   website soon after that for the public meeting on --

13   public teleconference on June 11th, and I'll send the

14   times for that to all the members of this panel in case

15   you would like to participate.

16	DR. HENDERSON:    Yes, and I anticipate

17   that will be a very short call.  Yes, Karen?

18	DR. MARTIN:    If I can just raise a

19   question...

20	DR. HENDERSON:    Yes.

21	DR. MARTIN:    You just said you

22   anticipate that June 11th would be a very short call.

23   That might be an opportunity, if we have identified

24   some questions on which we would like further feedback,

25   perhaps to extend that call to include some



139

1  thinking evolves in the days following today.

2	And so I appreciate the committee's

3  willingness to interact with us in that way and

4  Angela's willingness to help facilitate that and let us

5  go away and think about that and see -- see the extent

6  to which we think that will be helpful.

7	DR. HENDERSON:    Thank you, Karen.  I

8  think that's the best way -- that's the way we can

9  handle it.  We are now at lunch time and I presume

10   there is lunch, but I suggest that people -- we had

11   said we would end at 2:00.

12	We're a bit early, but if people want to

13   spend the extra time working on their paragraphs for

14   the letters to the administrator, now is a good time.

15	I mean after you eat, but otherwise do

16   we have any other business to attend to, Angela?

17	DR. NUGENT:    No, I don't think so,

18   Regine, unless we want to just touch again on the

19   schedule for following up from this meeting?

20	DR. HENDERSON:    Yes, and everybody has

21   that schedule and that's what I was suggesting.  You

22   might want to start working on those paragraphs.

23	Do you have that schedule?  I don't know

24   where -- well, people have it in their folders.

25	DR. NUGENT:    We had asked for the	

141

1  consultative advice on specific questions.

2	I'm not saying I know today that's what

3  we want to do, but I was just raising that as an option

4  to see if you think that's a viable option.

5	DR. HENDERSON:    Well, I think it's a

6  great idea, Karen.  If you do want -- and when you

7  think about it and you need -- that would be a good

8  chance for you to interact with the panel again and, as

9  I say, otherwise it's a very short call because we will

10   have all agreed.  Outside the public view we will have

11   come to consensus on the letter and this would just to

12   be doing it in public, but that's a good idea, Karen.

13	Okay.  Okay, well, lunch awaits us and

14   we thank -- I want to thank the Air Office people for

15   being so -- you know, for working with us so carefully.

16	I think we've had a good dialogue at

17   this meeting and we've discussed a lot of options and

18   we appreciate your taking the time to come be with us

19   and I appreciate all the members of the panel taking

20   the time to come here and work on this important

21   problem and I think we'll see the -- we'll be talking

22   to you June 11th and we'll see you the end of July for

23   the SOx, okay?  Thank you.

24   (WHEREUPON,   the SESSION   was concluded at 12:08 p. m.)

25



142

1	CAPTION

2

3

4  The foregoing matter was taken on the date, and at the

5  time and place set out on the Title page hereof.

6  It was requested that the matter be taken by the

7  reporter and that the same be reduced to typewritten

8  form.

9  Further, as relates to depositions, it was agreed by

10   and between counsel and the parties that the reading

11   and signing of the transcript, be and the same is

12   hereby waived.

13

14

15

16

17

18

19

20

21

22

23

24

25

	

143

1	CERTIFICATE OF REPORTER

2  COMMONWEALTH OF VIRGINIA

3  AT LARGE:

4  I do hereby certify that the witness in the foregoing

5  transcript was taken on the date, and at the time and

6  place set out on the Title page hereof by me after

7  first being duly sworn to testify the truth, the whole

8  truth, and nothing but the truth; and that the said

9  matter was recorded stenographically and mechanically

10   by me and then reduced to typewritten form under my

11   direction, and constitutes a true record of the

12   transcript as taken, all to the best of my skill and

13   ability.

14   I further certify that the inspection, reading and

15   signing of said deposition were waived by counsel for

16   the respective parties and by the witness.

17   I certify that I am not a relative or employee of

18   either counsel, and that I am in no way interested

19   financially, directly or indirectly, in this action.

20

21

22

23

24   CHARLES DAVID HOFFMAN, COURT REPORTER / NOTARY

25   SUBMITTED ON MAY 2, 2008

	

0

0.26 57:6

01 98:4

03 97:20

113:1,5 116:9

120:7 122:23,24

24-hours 113:4

26 29:12

135:19

absolutely 10:2

15:19

absorb 21:19

08 107:15

 		abstract 21:7



1

1 20:15 56:20

 	3

3 56:14 63:3 94:22

300 42:21

accept 55:3 58:12 acceptable 109:16 accepted 113:17

1.8 19:5

 		according 22:15

10 20:15 121:12

10% 62:11

10:35 74:23

100 122:5

100% 61:14

10th 136:12

11th 140:13,22

141:22

12:08 141:24

15 19:14

15% 62:11

15,000 41:14

16th 140:7

17 4:2

17,000 40:23

188 22:3

1993 21:6

2

2 2:5 63:2 94:22

2:00 139:11

20 35:2 57:19,25

58:4 83:10,11

96:11,17,22

97:17 98:9

2008 2:5

24 57:20,25 89:21

91:8,9,21 96:14

98:4 122:2

24-hour 80:21 83:12

89:2,25 90:20 91:5

92:5,16

93:3,6,10,24

94:3,7,10,16,22

95:7,20,22 96:1,15

97:1,7 98:23

99:4 107:19 108:22

4

4 58:15

4.233 52:21

40 74:23

 	

5

5 3:1 8:2

50,000 82:13

53 126:24

531 126:24

6

6 8:2,3,9 21:6,8,12

93:2

60 39:24

63 35:5

65 35:5 104:11

130:13

7

7 2:24 3:2,14 8:1,9

93:5

7.9 26:5

76% 63:2

9

93 54:2

9th 136:12

A

ability 15:16 55:21

62:21 79:7

able 2:22 29:17

35:16 53:17,19

70:25 102:4

106:9 111:13

132:21 136:4

absence 20:12

85:16

account 48:6,13

100:13 106:16

111:11 136:4,24

accounted 4:25 6:11 accumulate 119:11 accumulating 119:10 accumulation
92:4 accumulative 121:23 achieve 80:7,9 acknowledge 104:9 acknowledges
24:11 across 91:21 99:5

109:25 110:21,25

127:11

Act 134:22

activities 24:24

37:9 93:17

activity 37:10

49:20

actual 11:14

13:12 18:8 41:3

43:4 131:4

actually 9:1,14

18:21 23:1

29:18,22 30:1 35:5

36:4,9 40:11,22

41:8 42:3

51:9,10 57:13

65:12 91:3

109:17 111:12

131:7

acute 64:18

121:13,21 122:3

132:3

add 19:9 63:21 64:6

73:23 93:7



added 118:10

adding 18:2

addition 25:3 30:25

45:1 66:10

additional 21:9

36:17 40:20 109:20

117:17 120:6

additive

17:17,23,25

address 13:16 75:18

111:16 124:14

133:5

addressed 25:21

36:24 37:1 126:5,6

132:19

addresses 59:3 adequacy 77:23 adjust 11:3 19:2

41:16 48:12 65:4

adjusting 11:18

18:24

adjustment 16:11,20

17:8,17,22,23,25

18:7 19:8,10,23

20:1 26:18 29:6

30:4,6,9 59:15

adjustments 23:8

49:11

administration

133:20,23

administrator

2:15 139:14

admission 51:23

104:20 106:4,9

admissions 46:14

57:19 82:21,23

84:4 87:11,17

88:22 100:20

104:9,24 105:1

106:1 117:17

120:15 121:2

126:14 127:1,10

130:4,14,25

131:11,24

adult 54:8 130:13

adults 44:19

51:19 54:8 130:12

advance 91:3

advanced 76:15

adverse 46:20

63:9,20

advice 37:2 38:17

71:9 72:15 95:10

100:9 101:22 135:8

138:3 141:1

advise 88:9

ADVISORY 2:2,3 aeromod 19:23 affecting 121:24 affirmatively 100:6
afternoon 2:21 32:8 against 10:8,9 11:6

13:3,12,13,25 22:4

23:1 36:9 87:6

107:1

age 44:19 105:9 agencies 71:12 agency 2:1 86:4

88:10 138:3

agenda 34:4 125:8 aggregation 90:20 ago 8:24

agreed 33:25 100:10

141:10

agreeing 34:5 agreement 50:14 ahead 41:20 81:14

83:21 102:13

108:18

air 2:3,8

5:7,9,15,23 11:4

12:8,12,14,18,19

13:2,11,16,25

14:4,11,15 16:22

17:6 20:11,13,15

21:1 22:16 23:20

25:4 26:17,20

27:6,7,25 28:24

34:17

35:13,15,17 39:3,6

52:1 60:3,10 71:10

77:14 79:12

80:6,13 81:9

84:9 96:20

106:2,23 109:4

110:14,20 113:18

114:7 117:10

134:22 141:14

airway 44:17,18

45:21 46:1,6,13

51:5 52:4 55:16

56:23 57:6 64:10

82:15 130:1

airways 47:24 51:18

55:18 56:7,9

align 24:8 25:10

26:24

aligns 24:14 allergen 64:22 allergens 54:4

55:20

allow 118:7 allows 62:12 alluded 117:2

alone 14:20,24 52:9

85:3,21

already 43:20 45:24

51:21 90:18

109:8,19 112:11

132:19

altered 16:18 23:13

alternate 49:12

125:1

alternative 9:22,23

10:20,24 12:13

23:9 77:24

90:3,5 94:6 95:7

106:22 107:2,7

108:8,10

alternatives 13:22

am 11:25 34:5 50:10

57:12 66:7 70:21

76:9 81:18 83:16

88:16

ambient 9:2,4,10,24

11:11 12:18

13:13 16:12

17:4,15

18:10,11,25

32:16 42:5 60:12

64:19 93:4





America 72:22

American 87:22

128:22

among 18:19,20

21:1,17 67:16

79:18 120:14

amount 36:9 54:2

92:15 109:20 134:3

amounts 16:13 amplitude 31:3 analogy 50:3

117:4 119:14

analyses 8:8

15:24 31:4 36:17

70:20 79:8 87:21

117:1 130:1

analysis 2:10,17

7:9,15 8:5,21 9:19

10:11 11:2

12:8,9,12,14,20

13:16 16:9 23:20

26:20 27:4 31:13

33:13,14 34:12

39:23 40:2 43:16

62:24 66:3 70:16

72:13,15

73:2,4,7,10

82:13 83:5,8

85:6 88:8,11

89:7,17 90:23

93:7,10,16

94:18,25 95:21

99:1 101:7 103:9

104:17 105:2

111:23 115:6,20

116:2,22

128:9,21 129:25

132:16,20 133:1,14

134:4

analyze 90:5 98:13 analyzed 112:1 analyzing 40:2,10

94:19

and/or 108:11

Angela 136:10

138:21 139:16

Angela's 139:4

Angeles 34:20

35:4 38:24 39:5

40:12,14,15 41:4

49:13,15

angle 15:2

animal 47:6 48:1

56:8

annual 29:6 31:6

41:14 95:1 98:23

108:6

ANPR 85:16 86:15

104:2,3 106:17

answer 35:10,24

38:13 54:25

56:19 71:7 82:2

83:6,18 84:24

85:13 86:6 96:10

102:4 107:20

115:1,9 125:9

126:21,23 132:24

134:14

answered 3:5

87:12 107:17

answering 3:2 14:21 answers 28:22 87:11 anticipate

140:16,22

anybody 17:14 26:11

37:23 82:17 88:2

108:17 129:19

anyhow 18:4 19:25

anymore 15:1

92:8,11 130:12

anyone 124:2 anyplace 98:7 anything 38:7 55:25

56:5 93:21 97:9

127:21 132:9

135:20

anyway 98:6 128:16

apartment 6:22

apex 6:15 20:7

24:16 93:8,10

apologies 46:8

apparently 21:10

68:18

appeal 89:4

appealing 49:16

appeared 23:21 appears 16:14 22:21 applied 29:23

100:15 109:13

125:22

applies 4:7

apply 58:18

100:23 101:9

104:17 109:25

110:15 128:4 131:5

applying 32:24 55:3

80:3 97:23

105:8,19,20

110:4,20 111:23

127:22

appreciate 75:16,19

78:24 113:23 135:6

139:2 141:18,19

approach 8:12

9:14 12:16 54:17

71:12 133:2 134:20

approaches 10:20,24

75:9

approaching 134:22

appropriate 6:1

44:14 56:25 65:8

70:17 91:23

95:25 108:3

appropriately 48:23

52:15 133:1

arbitrarily 20:17

are....basically

59:4

are...would 65:21

area 27:16 40:13

48:11 49:14,20

75:20 84:13 112:21

areas 3:11,12 57:22

77:15,17 84:10

101:3 109:19

111:19 123:10

aren't 19:14

75:23,24 80:8

argue 34:1 44:9

90:9 109:14 124:17

argues 22:3 60:22





arguing 58:22

88:2 98:2

argument 60:1 70:6

arises 9:8 arose 136:16 artifact 33:17

artificially 20:9

22:22 23:6

aside 69:24 115:17 as-is 96:18 108:7 aspect 29:13

48:17 60:16 67:11

aspects 4:8,16 16:9 assess 58:14 assessing 8:10 14:9

77:23

assessment 2:25

3:23 4:7,15,25

7:4,11 34:4

42:20 51:7 52:3,13

53:21 54:1,21,24

56:19,20

58:11,15

60:15,24 67:6

70:12 72:12

76:13,14,17,19,22

77:3,4,12,21

78:5,7,17 79:6

80:19 81:8 84:23

85:3,8,9,11,15,17

86:5,12,17,25

88:13,23 92:23

99:17 100:11

101:22,24 102:3

103:13,14,22

106:15 107:12,25

108:1,4,25

109:18 111:15

115:10 117:5

118:5,8

125:6,14,17

126:11,19 132:16

133:9,24 140:11

assessments 33:9

72:9 78:7 84:22

108:12

associate 91:20

associated 48:11

55:6 65:4 77:21

80:17 86:8

87:24,25 88:14

97:3 108:15 123:18

133:25

association 99:3,4

associations 106:25

113:2

assume 6:24 22:22

57:24 64:17

assuming 26:6

119:1,20

assumption 2:24

22:19 26:6 61:13

assumptions 24:20

101:23 102:2

109:21 132:11,12

asthma 3:22,24

4:3,8,18 40:8

51:21,22 67:3,4

82:24 91:4

104:24 105:2 106:2

130:3 131:9

asthmatic 51:19

61:21

asthmatics 3:18

26:3 44:18 52:5

54:8,9,14 55:18

59:12,13 62:4

82:13 131:9

Atlanta 38:24

39:2,8,16,21,23

40:5,8,9,10,12,15

41:5,21 49:13

72:19 97:22 100:21

113:6

atmosphere 17:19

60:12 100:1

atmospheric 79:17

at-risk 44:4,13

45:17

attach 123:14 attack 74:24 attain 77:18 attempt 62:10 attempting 61:17

attend 44:24 139:16

attributable 67:20 attributed 91:18 audience 63:16 audiences 51:25

August 102:22

135:24 136:9

137:11

Australia 65:11

Australian 67:1 authors 112:15 availability 106:2 available 6:1 16:25

34:24

35:13,14,22

67:15 93:23 104:14

106:6,13 110:21

115:24 126:17

127:6

ave 118:2

avenue 104:21

average 4:20 5:9

29:7 35:3,5

41:14 57:20

64:15 74:16

83:12,13 89:2 92:5

93:6,10 94:3,4,7

95:22 96:14,15

97:1,7 99:4,7

108:22 110:21,25

112:16 122:12,24

averaged 110:16

averages 17:1 74:14

80:21,22 88:24

90:20 92:16 93:3

94:11 96:1

98:23,24 122:24

averaging 80:17

81:1 84:12 89:2

90:8 91:5,21 95:17

118:24 119:19

122:1,10

avoid 86:9

Avol 34:13 36:2

38:11 43:20,23

46:5 68:15

awaits 141:13





aware 79:25

away 13:20 35:20

75:7 138:16 139:5

B

background 8:4 69:5 backtrack 112:9 balance 22:9 46:25 balanced 48:24

54:25 56:20

Balmes 37:25 46:3

50:10,12 59:7

66:11 90:24

91:10 105:13,23

110:9 128:19 132:6

bands 84:21

base 7:11 72:23

95:21

based 3:23 5:24

9:24 10:6 12:3

16:23 17:1 21:6

28:3 29:6 47:2

56:23 65:12 69:2,9

77:4 82:3 93:14,16

97:25 99:10,18

105:17 108:13,22

109:18 110:8

115:11

baseline 100:16

101:2,14 110:3

basic 102:5 124:16

basically 18:24

19:17 21:11

98:22 103:2 128:9

basing 53:25

basis 17:11 55:13

56:18 72:9,11

93:24 101:16 113:1

117:12,21 118:1

became 6:18 9:18 become 106:11 becomes 40:21 begin 15:12 beginning 7:25

8:2 44:3

behind 75:10

believe 17:14 82:19

91:20 94:1 113:16

benchmark 13:9 48:8

49:5 51:19 52:11

56:25 59:4,5,19

62:7 68:23 130:1

benchmarks 7:12

14:21 62:14 65:7

94:21

beneficial 134:4

benefit 80:7

97:17 98:12

115:6,11

benefits 80:9

97:4 133:10

best 6:1 15:19

57:17 65:23 66:2

85:12 86:20

90:14 133:14

138:17,24 139:8

bet 17:21 122:13

better 17:22

19:23 20:21

27:22 30:9,11,12

35:16,18

49:7,16,17 54:16

63:14 67:7

68:16,22 69:1

70:14,20,22 72:9

75:8 79:9 91:5

111:18

beyond 7:23 15:15

21:14 32:24

81:11 136:3

bias 6:7 51:15

69:19

bigger 48:19

biggest 23:18 31:24

40:6 42:21 75:18

billion 22:3

65:16 96:12

biological 47:6

48:1 70:9 121:11

biology 121:19

122:8,11

birth 47:3

bit 24:3 25:25

29:19 34:7 48:25

61:11 90:8

99:20,24 106:20

108:10 112:6

118:19 139:12

block 11:23 15:24

blocked 20:18

Board 2:2 52:1,3

bodies 67:8

body 51:8 53:3

80:23,25

Boston 28:12 bounced 126:12 boundaries 21:23

62:19

bounding 67:23 bounds 22:2 84:7 breadth 22:2 33:3

64:9 79:20 87:23

110:11

break 43:21

71:18,21 74:23

75:13

breakouts 105:10 breath 73:12 brief 72:3 75:14 briefly 106:21

bring 36:22 63:11

69:21 124:14

bringing 79:6

broad 20:15 21:16

33:2 73:12 79:12

90:4

broader 21:15 76:21

78:10 85:1

100:24 107:25

broke 27:9 broken 105:9 bronchial 61:24

Bronx 31:18

brought 45:1 110:3

Brower's 47:3 bubblers 92:7,10 buildings 31:21 built 23:6

burden 67:20

buried 27:3





Burnett 87:8

bus 6:22 37:18 45:2 buses 32:1 45:7 business 139:16

busy 44:24

C

calcula 25:7

calculate 29:17

98:8

calculated 35:3

calculation 22:25

24:2,10

calculations

19:19 123:15

calibrate 20:3

California 52:1

65:1 106:7 124:6

Canada 91:19

Canadian 87:7 candidates 39:11 caps 74:20 75:2 capture 27:16 40:18

42:6

captured 25:19 28:8 captures 76:16 capturing 29:16 cardiovascular

45:11 46:15,23

care 98:3 careful 93:16 carefully 6:17

67:21 141:15

carry 103:9

cars 7:2

CASAC 2:3 15:12

88:5 136:10

case 20:15 34:16

48:24 55:20

61:17 67:6,8 79:14

110:25 115:22

140:14

cases 55:25 112:25

cast 69:24

cat 56:8

categories 44:15

45:13

categorize 46:21

category 45:14

83:11 131:6

causal 67:19

120:1 123:2,7

causality 120:8

130:24

causally 119:1

cause 2:21 5:12

18:13 23:10 26:9

50:12 51:3 84:12

caused 78:21

79:25 122:3

causes 57:6 123:8 causing 68:25 caution 66:16 cautiously 67:21 caveat
133:1 caveated 62:5 caveats 63:12 69:6

CDC 105:24

census 11:23

15:23,24

center 98:3

central 32:3

60:2,20,22

93:9,14,19

104:23 111:24

112:2 113:11

central-site

112:9,13

centroid 11:22,23

15:24

cents 69:14

certain 22:12 57:14

83:11 102:2

certainly 3:12 7:19

16:15 20:10

62:14 64:5

70:12,20 89:18

93:22 105:2 112:21

113:23,24 120:20

126:6

CHAD 37:9,11,14

challenge 55:23

74:5,16

challenged 64:21

challenges 32:23,24

78:18,20,21 79:2

81:5 118:7

challenging 49:22 chamber 7:11,12,21 chance 22:7

25:14,20 141:8

change 55:22

57:18 58:6 90:17

95:3 97:9 119:5

changed 23:11 76:23 changes 20:15 137:4 changing 16:11

83:19

chapter 2:24 3:14

7:25 8:9 13:1,2

14:17,18 23:20

44:3 61:12 62:3

66:22

chapters 3:1 8:21

43:24 44:2 45:19

characteri 14:4

characteristics

22:16 24:13 28:4

characterization

13:25 14:1 48:22

85:1

characterize

27:21 51:20

characterized 52:16

characterizing

50:17

charge 3:2,5,15

36:24 59:3

charged 48:21 chart 3:5 checked 4:1 chemistry 79:17

child 37:19 57:10 childhood 4:3 children 31:17,24

44:19 45:7 50:23

54:8 65:3,11

120:25 124:5





children's 47:2

65:1

choice 35:23 55:1

95:7

choices 34:18

123:20

choose 38:18 55:5

73:18 74:8 95:19

choosing 7:12

40:9 130:17

chosen 6:15 49:6

56:21 58:12,16

96:23 105:25

Christian 35:15 chronic 132:4 chunks 42:21 cities 4:13 12:4

14:23 21:1

38:12,16,18,23

39:4 53:2 60:19

65:24 72:17,23

73:4,13 80:4 97:23

99:6 100:17,25

101:12,13 105:4

109:7,11,14,25

110:2,11,16

127:3 128:11

129:2,20

city 3:24

4:16,19,22

5:11,15,20 27:1

39:23 100:21 101:5

104:14 110:22

113:3,12 117:17

127:22 128:8

city-specific 109:6

110:7

clarification

37:5 94:14 103:18

clarity 26:2 class 114:25 clean 2:3 48:3

134:22

clear 3:9,21 5:8

6:18 9:20 15:3

21:8 23:3 25:24

27:4 47:14 48:23

60:14 81:6,9 93:2

clearer 26:5

clearinghouse

104:23

clearly 18:14

24:6 52:15 59:23

62:2 63:13 65:2

80:14 85:22

97:19 104:20

121:12 132:11

133:13

clinical 47:19 55:2

58:10 59:8,14

62:17,21 66:4,10

69:3 74:1 80:18

89:14 117:6,10

clinically 55:8

57:1 58:13

63:6,9 82:15

clo 35:14 close 130:13 closed 8:23

closely 119:18 closer 35:19 62:2 closest 35:14

co-author 112:20

co-authors 112:20

co-efficience 73:19

coefficient 116:3

129:15

coefficients 122:22

128:23 131:5

cohort 47:3 colleagues 52:2 collect 71:19 collected 21:13

113:7

Collecting 101:14 collectors 45:2 combine 133:2 combined 128:21
combining 75:8 comes 35:8 64:24

121:22 125:13

135:12 136:2

comfort 11:10

comfortable 11:25

15:9 26:7 63:7

135:8

coming 35:4 58:14

84:5 107:15

comment 3:14 7:25

26:11 52:10

53:22 68:2 92:1

121:9 128:20 132:8

137:7

comments 3:6,15

7:6,24 8:16,21

9:17 25:24 26:8

31:12 32:20

34:11,13

43:18,24 45:25

50:14 61:1,5

64:5,6 85:13

100:12 101:6

103:10 120:7

129:2,20 136:21,24

137:11

committee 2:3 62:22

89:10,23 101:21

126:9 140:9

committee's 93:25

100:9,14 139:2

common 55:20 56:9

60:18

communicate 63:14

123:21

communicated 52:15

communication

123:12

communities 45:6

49:12

commute 40:1 commuter 49:19 commuters 39:25

40:6 45:6

commutes 4:24

commuting 4:23 6:25

37:6,16

comparable

101:1,2 129:18

compare 13:3 14:3

107:1





compared 10:8,9

13:22 14:3 48:16

129:6,17

compares 29:12

comparing 13:11

18:15,16,18 20:1

29:19

comparison 10:10

12:8 13:13 16:23

17:13 18:5,6,13

19:13 25:6 107:8

121:17

compelling 50:24

51:24

complete 86:20

111:19

completely 17:5

19:11 21:4

23:22,23 25:7

29:14 48:25

81:17 84:1 119:21

completeness 28:14 complex 11:2 complicated 66:23

67:16 115:3

complications 21:9 complimentary 67:2 component 45:8

67:18 70:16 77:20

components 47:22

66:19

comprehensive 3:8 comprehensively 7:8 concentration 9:3,4

12:18 17:3,21

22:11,14 23:1,4

43:7 53:4,12 77:11

109:5,13 118:20

125:21 132:10

concentrations 5:24

11:3,4 16:12

17:6,8,12 19:1

20:1 21:25 30:22

43:9 99:11 124:8

conceptual 8:4

concern 23:18

40:5 59:5

89:13,15,20

122:9 129:23 132:1

concerned 53:6

64:16 65:10

70:21 120:21

122:11

concerning 50:17

concerns 50:1,21

69:24 131:3

conclude 128:2 concluded 141:24 conclusions 87:3 concrete 79:2

101:22

concur 7:20 58:21

109:24

concurred 7:10 conditioner 21:1 conditioning

5:8,10,15 20:13

conditions 17:18

64:19,21

conducted 52:25

107:10

conference 138:8,10 confirmed 87:1 confounding 7:16

124:10

confusing 68:20 confusion 68:18 congressional 88:9 connect 72:8
consensus 2:14

141:11

consent 103:1,18,23

conservative

33:12,14 104:16

105:20,22

conservatively

123:17

consider 33:1,3,8

64:12 69:12 73:3

91:1 103:9 114:4

115:19 123:25

134:11

considerable 92:15

considerably 16:8

consideration 41:10

69:9 77:8

107:2,7 113:21

117:8 133:21

considerations

41:19 78:10 130:17

considered 27:17

36:17 45:18 69:9

72:22

considering 33:2

77:1 91:11 108:18

consistency 20:19

consistent 53:3

55:8 99:5

consistently 28:19 constant 16:13 constrains 136:18 constraints 79:24
constriction 82:15 construction 45:5 consultants 135:7 consultative
141:1 consuming 101:15 contact 112:15 contained 31:17 contaminate 88:6
contents 21:18 context 8:5,11 44:7

47:18 75:25 76:2

77:8 85:15 91:1

94:17 118:4 124:15

138:4

contexts 15:12 78:3

contingent 20:12

121:5

continually 65:4 continued 49:24 continuing 118:13 continuous 20:24

92:3 97:24

continuously 43:2

113:9

continuum 99:20 contributed 22:20 contributing 30:24





contribution

22:6,10 66:20

129:6

control 128:8

controlled 47:22

51:10,13 53:23

54:7 72:8

controls 14:9

128:10

conversation 137:24

138:1

counting 23:19

24:13 25:1 26:9

29:7

country 81:24

97:9 106:3,8,10

127:11

counts 24:8 27:3

county 11:1

40:15,16 41:6

couple 25:23

31:14 37:5 41:15

curious 42:23

current 16:23 21:20

44:7 47:8 77:18,24

85:17 96:18

97:19,20 102:23

108:6 124:21 133:7

currently 19:12

46:15 50:4

curve 98:3,10,14

cut 58:6

cycle 121:14

convey 78:8

43:24 45:25 79:3	 	

conviction 33:19

130:21

convictions 130:23 convince 34:2 convincing 49:8

51:11

cooking 22:13,20

25:24

coordinate 22:3

COPD 45:12

co-pollutant 7:16

33:10,15 53:8

copollutants 127:23 co-pollutants 65:5 correct 11:8 12:1

17:15 52:22

92:5,6,13 98:24

102:7

correcting 19:6 correction 18:23 correctly 10:20 correlated 123:7
correlation 24:24 correlations 79:18 correspond 16:21

17:5,8

corresponds 18:8 corridor 6:8 costs 133:10 couched 60:16 counseled 23:1
count 29:10 counted 44:9 counties 40:3

92:25 112:24

course 3:10 13:18

27:11 30:20,23

32:25 35:25

37:18 43:5 45:10

50:25 59:20

84:25 105:1

117:4 125:19

court 102:24 103:4 covered 37:10 68:13 covers 37:7

CR 78:6 80:2

Crapo 54:19 63:6

81:15 89:23 90:9

96:5,9,15,19,21,24

97:3,12,15 98:1

121:8 124:25

Crapo's 95:10

Crawford-Brown

8:18,19,22

10:14,18 11:5,9,24

12:10,24 14:5,25

create 103:23 creates 60:21 creating 121:21 creeping 136:17 criteria
27:13

28:14 39:18

55:4,10,11,13

56:16 111:2

critical 3:11 58:24

122:14

criticized 101:10 cross 47:5 72:14 crucial 26:9

D

daily 31:5 41:14

57:20,25

88:22,23 91:20

113:8 119:19

Dale 16:7 23:15

25:5 118:17 120:3

Dale's 32:20 121:9

data 4:6 11:2 12:22

16:21,25 20:21

25:13 29:6

31:1,2 32:25

34:19,20,24

35:13,15,17

36:6,10 39:3

41:22,23 43:5,11

47:8,9,23 48:1

49:10,14,17

50:23

51:13,14,18,21,22

53:20,23 54:16

57:17 60:15,22

62:21,23,25

63:18 66:9

67:2,9,11,15

69:2 70:16 71:2

74:1,5 75:6 82:1

86:1,5 88:3 89:5

91:4,8 92:14

93:4,19,23,24

95:18 96:16 97:5

98:13 100:16

101:13,14 102:12

105:6 106:4,6,9

108:21 109:1,7





110:10,20 111:24

112:1

113:7,8,11,13,19,2

3 114:7 115:23

120:10 125:1

127:6,11,13,14

130:9 131:5,17

132:8,13

134:3,5,18

database 37:9,11,14

date 103:19,20,25

104:4 107:14

136:7,17

dates 103:21

day 2:7 28:16 39:25

56:6 57:10

60:3,9 82:14

110:22 119:10,12

days 58:5,7 139:1 day-to-day 24:24 deadline

102:15,24 103:5

deal 9:8 60:21 63:4

90:21 133:4

dealing 64:25 65:14

67:17 134:13

dealt 16:10,15

death 91:17

December 104:3

107:15

decide 15:7 69:10

82:5 90:3,4

decided 51:17

decision 7:10,20

12:25 15:2,4,6

33:8 69:12 73:10

81:23 84:17 87:9

103:20,24 107:5,24

117:21,23 118:2

133:7,19

decisions 49:18

84:19 108:2 117:12

132:23

decisive 50:5 decline 19:20 decrease 96:12,17

97:24 98:10

decree 103:1

decrements 46:11

Deegan 32:6

deeply 95:14

defend 82:8,12,18

83:3 90:14,16

defending 83:19 defenses 46:13 define 41:13 defined 41:14

Definitely 103:12 definition 68:16,19 definitions 68:21 definitively
136:11 degree 14:10 63:8

80:9 101:24 127:2

delta 52:4

deltas 128:10,13

department 51:23

52:7 106:5 126:15

depend 24:3

depending 21:15

63:16 138:25

depth 2:13 71:4

79:20

derive 53:11 derived 22:3 described 8:14

9:1 16:16 23:13

58:18 86:15

description 20:21

25:25

design 75:5

108:23 111:6,12

designed 57:5,21 designs 125:24,25 detail 15:13 101:8 detailed 8:16

12:5 14:16 20:16

details 24:21

74:6 125:18,19

determinants 5:19 determined 28:17 determines 85:21

Detroit

39:2,9,15,16 40:13

develop 53:4

developed 85:11 development 69:8 devil 125:18 devil's 24:20 74:6
diabetes 45:12 diagnosis 119:8 dialogue 141:16

Diego 34:20 differ 22:15 difference 17:24

18:9,19,20 19:16

39:2 94:5,15 102:2

129:8

differences 21:1

24:12 48:14

65:12,15,17

124:1,7

127:23,24 128:1

different 13:19

14:8 16:24 18:20

20:17 21:18,19

27:18,20 28:9,10

57:23 62:14

74:15 76:11

98:21 99:16

100:8 101:11 107:6

117:22,25 118:2

120:22 125:16

126:13 128:4,13,18

129:9 133:19

differential 31:19

differently

108:10 111:18

difficult 23:9

67:10,25 68:4

115:15,17 134:17

difficulties 134:24

136:16

difficulty 9:9 16:9 dilemma 116:22 diminished 89:1 diminishes 89:6
direct 23:20 58:9 direction 41:15

46:25 47:11 70:14

directly 5:21 24:11





discounted 69:7

discuss 2:12 5:22

62:3 75:18

discussants 140:1

discussed 5:4 18:16

44:16 59:21 141:17

discussion

2:8,10,16 5:9 6:12

7:13 8:4 13:23

19:12 38:8 40:20

42:12 44:1,5 45:23

47:7,12 48:9

49:9 63:25 70:2

71:2 74:24 83:24

89:13 92:22 111:14

115:4 118:13

126:12 132:11

134:16,25 138:20

discussions 49:5,18

72:18 74:10 76:4

78:25 91:14

disease 44:16

45:10,12 46:15

67:20

disentangle 31:5

dismiss 19:11 65:20

77:6,7

dismissing 77:2

78:9

disparate 80:25

dispersion 9:15

11:4 16:20

distance 19:22

distances 34:22

35:1

distinct 4:13 65:16

distinction

28:2,5 44:6

distributed 100:1

distribution 12:6

15:20 17:3,7

18:4,7

21:4,6,14,15,17,22

22:4 23:5 24:6,7

25:9,10

29:8,9,14 30:16

59:11,23 93:11

124:24

distributions

15:14,17,25

17:11 18:15

20:9,11,14 21:25

23:9,14 93:17

diurnal 25:24 31:1 diversity 127:6 doable 70:24

73:13 103:11

dock 45:5

document 2:25

3:17,25 5:6 6:5,17

7:22 9:21 14:6

16:16 19:13

31:17 34:16

37:17 44:3,7 47:14

49:8 50:7 67:7

73:5,9 76:14,20

78:3,8,13,22

85:11,15 101:19

102:22 105:17

106:15 118:5,11

124:21 135:24

136:2 140:6

documentation 48:23

documented 31:23

73:15

documents 72:6

76:12,15 130:18

140:11

domain 41:3,18

domains 49:19

116:12

dominant 60:17

done 5:3,4,25

9:20 23:7 24:11

25:21 30:15

52:16 55:4 56:14

62:24 68:1 69:23

70:7 72:7,20

74:2 80:20 82:12

85:18 86:13

87:13,16 88:15

92:16 94:25

98:25 99:17

100:3 104:19

105:3,5 106:25

109:2,8,19 110:1

111:17,22 112:7,11

124:18,19

125:3,5,20 128:20

door 102:23 dosemetric 122:25 dot 72:14

Doug 8:17,20

10:16,21,23 16:7

downward 6:7

downwind 49:20

Dr 2:6,19 3:1,4

8:17,19,20,22

10:12,14,16,18,21,

23

11:5,8,9,21,24

12:7,10,19,24

14:5,25 16:7,8

23:15,17

26:10,12,14,16,19

27:7

28:1,11,21,23,25

29:2,4,18 30:13,19

31:11,14

33:20,21 34:9,13

35:10,12,21,25

36:2,3,14,16,20

37:4,15,21,25

38:5,6,9,11,13,17,

19,21,22

39:1,5,7,8,10,12,1

4,15,19,20

40:11,19

41:9,20,21,24,25

42:1,11,14,15,16,1

8,19 43:4,13,17,23

46:3,5

50:8,10,11,12

54:18,19 58:25

59:2 61:2,6

62:18

63:10,15,22,24

64:2,4 66:5,7

68:6,8,10,11

69:16,19

71:3,16,23 72:2

73:22,24 74:21





75:1,3,4,12,15,21

81:14,15

83:20,21,22

85:5,12,23 86:11

88:1,16,18,19,20

89:8,11 90:9,18,24

91:7,10,12,24,25

92:7,9,10,12,20,21

93:20 94:12,13

95:9

96:2,5,6,9,15,19,2

1,24 97:3,12,15

98:1,16,17,25

99:2,8,12,14

100:5,22

102:5,12,14

103:17,25

104:2,5,7,8

105:11,13,14,23

106:14,18

107:11,14,16

108:13,20

109:23,24

111:4,5,20,21

112:23 113:16

114:2,5,15,17,19,2

3 116:21 117:1

118:16,18

120:3,10,16,19,20,

24,25 121:1,8

122:18 123:23,24

124:12,13,25

126:20,23

128:7,12,15,17,19

129:1,4,11,12,19,2

2

131:12,14,16,17,18

,19,23

132:1,6,14

133:17,18,22

134:15

135:1,2,3,16,22

136:5,8,12,13,14,2

0

137:1,5,8,13,19,23

,25

138:1,5,6,8,10,11,

12,15

139:7,17,20,25

140:3,4,16,18,20,2

1 141:5

draft 70:19 78:5,17

103:21

136:5,6,22,23

137:10

draw 78:6 98:2,14

drive 6:9 17:4,5

40:8

driven 84:19 117:5

drivers 17:15

45:2,3

driving 43:15 122:7

drop 57:19,24,25

58:4 83:11 98:8,11

due 4:17 42:23

106:15

dug 3:24

Dulles 34:19 during 6:24 39:24 dust 55:20

dusty 56:8

E

earlier 24:6

40:23 135:3 136:17

137:9

early 61:4 86:1

102:21,22 104:1

136:9,11,21

137:6 139:12

easier 50:20 51:24 easiest 64:11 easily 90:19

easy 71:6 eat 139:15 echo 65:18

economic 4:12

Ed 34:12 43:20 46:3

50:8 51:1 61:3,4

63:22 68:15 127:1

Ed's 50:14

effect 7:18 53:13

55:5,6,24 59:14,18

62:6,7 67:13 68:25

81:9 95:2 110:15

120:8 122:22

123:4,6,8 128:3,6

effective 16:13

53:16

effectively 58:19

102:22

effects 3:20

20:25 33:24

46:12,20,21,23

47:13,16 49:3

50:22 51:4,12

52:24 53:8,19 54:6

60:11,13 64:9,17

66:12,13,14,25

77:16 79:22

80:17 84:11 86:7

94:2,3 100:2

107:18 108:14,15

109:6 111:10

112:13 117:9,11,15

119:2,3,11 126:25

effort 3:7 7:5,8

54:2 131:25

either 5:21 19:19

20:24 30:10

63:9,11 130:3

elderly 130:11,12

elements 14:2

115:13

elevation 7:1 58:6

elicit 64:20

else 5:21 26:11

69:15 93:18 129:19

elsewhere 31:23 embedded 103:21 emergency 46:14

51:23 52:7

104:25 105:1 106:5

126:14

emerging 92:22 emission 34:15 emissions 4:21

6:3 29:24 30:1

49:16 60:6

emitting 7:3

emphasis 74:10

117:13





emphasized 131:21

encounter 42:24 encourage 113:24,25 endpoint 51:6

58:3,16,24 66:11

endpoints 52:6,18

56:21 57:20 58:11

energy 42:10 enhancement 24:25 entering 22:10 entire 53:3 73:3

98:7

entirely 10:6 53:12 envelope 16:5 environment 56:8

57:2

environmental 2:1

105:24

envision 83:18

envisions 101:21

EPA 3:7 10:1,13

26:13 36:23 69:9

epi 42:10 46:12

47:21 50:4

51:9,11,21 52:23

53:3,20 54:16 60:1

70:16 72:8 73:4,25

74:5,13

77:1,2,5,10,15

78:2,6,16

79:6,21 80:21

81:11

84:2,3,14,18,24

86:5,7 88:3

91:3,16 94:7,24

95:2,6,25 96:1

97:23 99:15,21

106:16,23,24 107:9

109:18 110:16,24

117:5,11 118:8

125:1,6 126:10

epidemiologic

7:14,16 59:22

92:24 93:13

96:16 98:18

epidemiological

46:12 47:15 51:9

65:20 66:3,9,14,17

67:11 93:23

95:18 115:20

epidemiologist

114:21

epidemiologists

116:12 122:23

epidemiology

32:22,25

33:4,5,8 34:1,3

38:3 48:2 55:12

57:17

58:3,10,19,23 69:6

82:3 83:15 86:17

101:9 111:22

112:11

epi-driven 81:7

equates 61:14

ER 57:19 82:24

130:4,25 131:11,24

errors 48:16

especially 52:8

105:18 110:1

essentially 18:2

19:1,6 20:5

establish 90:10 establishing 49:7 estimate 9:6

13:17 67:13 84:3

87:20 110:7 112:3

estimated 11:20

109:14

estimates 11:11

15:16 32:23

66:16 67:23,24

73:8 84:20 95:2

109:5 110:16 112:7

123:14,16

estimating 42:8

etcetera 74:8,12

European

127:14,21

128:14,22 129:2,20

132:8

European/U.S 128:23

evaluate 13:24

81:24

evaluated 69:14

evaluating 69:13 evaluation 125:12 event 10:7 56:9

57:13,15

61:19,24

63:20,21 104:25

121:21

events 22:14,20

56:12 121:14

eventually 122:14

everybody 2:6 68:12

69:20 97:10 139:20

everyone 26:10

everything 19:2

33:17 51:4 61:7

68:20 69:14

93:18 126:13

everywhere 108:7

evidence 50:17

51:8,11 54:25

55:12 62:8,12,13

66:25 77:10 78:2,9

79:6 80:20,24,25

81:12 84:18 85:4

89:14,15

99:15,19 111:12

117:22 125:1

126:17 128:3

evidence-based

117:8 133:2

evidence-driven

117:14

evolves 139:1

exacerbation

51:22 91:4

exacerbations 82:23

exact 55:9

exactly 9:9 18:12

22:21 54:12

55:15 81:20 98:1

105:15 116:14

117:16

examine 64:13

example 28:12 44:21

80:4 100:21

106:7,21 115:23





124:4

examples 20:10 exceed 77:17 exceedance 98:7 exceedances 14:21

23:19 24:2,13

25:2,8 26:9

27:19 28:7

29:7,10,11,17 42:9

124:22

exceedence 57:8 exception 52:17 exchange 17:6 20:11

22:16

expert 21:16 93:25

expertise 70:25 experts 73:20 89:23 explain 27:9 52:2,4 explained 116:24
explicitly 6:19

26:7

explore 83:5 exponential 19:19 exposed 43:2,6

61:18,23,24 62:4

exposure 2:10,17,25

3:22 4:7,25 5:19

6:7 8:21 9:6,7

112:12 113:8

119:10,11 126:4

expressed 50:2 extend 41:6 140:25 extensive 20:22

132:20

extent 4:24 6:5

37:10 52:13

69:13 78:15

79:10 119:8 139:5

extra 139:13 extract 77:11 extreme 5:2

extremely 23:4 41:4

exchangeable 23:23

11:1 12:3,9,16	 	

exchangeably 26:21

exclude 111:1,8

112:15,17 128:16

excluded 24:1

28:3 110:24

excluding 110:22 exclusively 125:10 excursion 119:20 excuse 2:9

exercise 72:25 exercises 89:7 exerting 17:20 exhaust 60:6 existing
115:10 expand 7:6 21:13

41:3 44:22 75:15

expanded 44:13 expanding 75:20,24 expect 17:18 18:3

20:22 62:3,15

94:18 119:23

expected 18:1 20:1

expecting 40:13

62:10 136:15

experience 4:21

62:16

experimental

51:18 53:24

experimentalist

36:21

experiments 121:11

13:8,12,13,18,24

14:10 20:4,8

26:1,22,25 31:13

32:15,23 34:12

36:22 41:23

42:20,21 45:15

48:4 50:22

51:10,13 52:24

53:23 54:7 57:23

58:1 60:17,25

61:13,22 64:18

72:9 80:19

89:1,6,17 92:23,25

93:17 96:25

101:7 108:11

117:4,9,14

125:25 127:15

132:5 140:11

exposures 4:17,20

5:2 6:24,25

9:24,25 10:5

11:11,14,20

16:14 22:23

23:14 26:3

31:16,24 32:17

40:7 42:22,23

43:16 44:25

47:22 50:18

51:12,15 60:7,20

74:12 89:16,19

91:17 93:6,11,19

98:19 108:15,21

F

FACA 135:11 138:4

face 60:24

faces 71:5 facilitate 139:4 facing 86:3

fact 18:3 22:9

28:17 36:5 46:1

50:18 67:16 74:9

79:7 89:2 119:5

121:22

factor 38:14

40:21 123:25 124:2

129:24

factors 99:11

124:10

fails 55:10

fair 6:23 54:2

60:11 123:12,21

135:15,17

fairly 27:16

39:16 50:16 66:1

90:22 121:11

fall 89:13

familiar 112:21

114:11 120:1

fancy 18:12 farther 35:20 fashion 36:19

55:3 132:23 134:1

fashioned 92:13





favor 22:4 23:1

feasible 70:24

103:11

feasibly 125:3

feature 99:25

125:24

features 30:15

February 124:6

Federal 92:2,14

feedback 90:7

137:20 140:24

feel 26:7 33:10

40:17 63:7 69:5

133:10

feelings 43:19

65:19

feels 56:3 fell 63:3 felt 68:18

FENa 128:21

FEV1 119:13,14 field 4:11 fifteen 74:22 fifty 69:23

figure 7:17 66:21

70:11 83:13 126:24

130:6

figuring 9:9

file 8:23

final 15:8 85:21

103:19 136:22,25

137:10

finalized 113:22

137:4

finally 22:19 49:25 finding 57:8 65:6 findings 59:22,24

92:24

fine 4:11 32:2,11

60:7,15

finer 5:20 finger 120:18 finish 43:21 finished 2:8 fire 38:4

firm 130:21

first 7:5,7 8:24,25

9:14 10:6,19 14:12

19:24 20:11

23:10 26:13 27:9

31:15 42:3 44:2

54:24 71:17 72:4

76:6 81:22 86:22

90:14 109:9 112:25

136:5

fit 11:18 91:4

five 19:15 23:20

38:12,16 43:25

48:20 49:25

52:20 70:5 95:13

five-year 27:10,15 fixed 20:25 21:5,22 flaws 70:13

fleshed 16:17 floor 75:19 flows 45:21,22

fluctuating 124:10

focus 9:3 14:23

30:17 47:15 67:4

107:22 109:19

111:6 114:14

118:13 138:23

focused 58:23 79:21

102:10 108:25

114:15 124:22

focuses 3:14 focusing 3:18 29:8 folders 139:24

Folensbee 54:1

folks 54:11 63:17

71:10 114:1 117:3

foot 4:23 forcefully 34:5 forefront 63:12 form 22:11

90:12,13,14

107:5 115:15

formation 66:19 forms 56:18 forth 9:16 49:3 forward 42:2

55:14 79:5 87:20

116:18 125:8,16

133:9 138:18

Fourier 31:4

fourth 20:6

FPP-1 55:22

fraction 12:17 13:7

40:17 62:8,9,13,20

framework 67:19

Frank 33:20 38:2

46:17 69:17 74:6

91:24 126:22 137:6

frankly 116:3

Frank's 68:2

73:15 137:20

freeway 35:6 36:2

Friday 140:2

front 13:1 18:14

26:4 39:17 43:5,11

full 30:21 59:23

64:9

fully 16:17 78:2

81:4,12 119:15

Fulton 40:16

function 23:2 46:11

50:23 65:3 98:14

109:13 120:13

124:11

functions 34:15

77:11 78:6 80:2

100:16 101:10

105:8 118:21 119:6

125:21

fundamental 3:13 funded 105:24 future 70:14 75:11

G

gap 72:11

gas 22:1 60:18 gather 116:23 gathering 40:1 gaussian 19:20 gee 61:20

general 4:5 7:24

14:10 25:13 29:4

47:10 59:9,11,19





60:9 76:6 78:11

80:6 105:20 117:21

121:19

generally 95:4 generate 15:19 generated 51:14 generic 4:5 genetics 45:13
geographically 4:6 geography 4:11

George 31:13

33:22 38:2 39:19

53:8 71:25 86:10

94:12 104:7 109:23

111:20 129:20

137:18

gestalt 51:2

gets 41:12

getting 8:24 9:23

11:20 25:10,15

29:15 66:8 73:18

91:2 130:13 131:15

given 3:20 13:18,19

27:24 28:13,20

30:23 41:6 42:2

48:10 52:9 57:16

59:6 67:14 69:5,14

84:11 123:3

127:22,23

133:10,11,14,15

135:6

gives 112:5 120:11

giving 37:1

glad 34:10 73:17

goal 140:6

gold 86:23 97:5

gonna 102:1

GORDON 38:6

gotten 2:20 115:21

GRAHAM 10:23

11:8,21 12:7,19

26:16 27:7 28:11

29:2 36:16 37:15

38:13,19,22

39:5,8,12,15 40:19

41:24 43:4 98:25

99:8

grand 46:9 47:19

grapple 80:15 118:6

138:16

grappling 80:15,16

great 2:12 8:20

20:19 24:19

41:24 60:21 88:1

141:6

greater 4:18 70:2

greatest 32:7 54:12

124:5

greatly 79:5 grief 18:15 gross 24:14

ground 19:5 31:19

group 55:1

70:2,18 71:8 78:24

103:15 127:4

grouped 27:14

growth 46:11

50:23 65:3

120:13,22,24

guarantees 22:6

85:20

guess 3:13 11:15

12:24 15:1,5

28:4 34:21 35:21

39:1,17 40:19

45:21 50:20

53:22 68:4 86:9,11

93:22,24 106:11

112:24 114:5

115:2,4,9,18

116:16 125:17

132:7 140:7

gut 121:18

H

hand 75:9 98:16

99:13

handed 116:13 handle 71:9 139:9 hands 101:9

hang 47:25 58:24 hanging 54:11 happen 122:17 happened 56:1,6

happens 48:5 57:11

happy 71:20

hard 7:17 32:22

35:4 54:11

harder

79:15,16,19 120:18

Harvey 13:15 62:1

70:2,19 71:1

89:8 93:20

107:17 110:3

118:19 119:18

131:12,18 132:15

135:10

Harvey's 129:23

hat 47:25 54:11

58:24

HATTIS 16:8

118:18

120:10,19,24 121:1

122:18

haven't 18:13

19:8 37:3 62:21

94:20,25 98:25

101:11 102:6

108:16

having 11:6 16:16

23:8 43:4 63:3

81:22 90:12 123:13

125:17

head 25:22

health 3:20 4:17

33:24 43:18,19

44:8 45:19 46:6

47:2 48:22 50:17

51:3,6,11,19 52:14

53:9,14 54:21,25

55:2 58:11

61:14,22 62:6,7

64:9 65:1 77:16

80:7,10 87:24

89:5,22 94:2 95:12

97:4 100:2,17

105:25 106:1

107:6,18 108:14,15

109:6 112:13

124:24 126:13

127:24 129:24





healthcare 104:10

healthy 56:3

hear 14:14

26:12,13,15

28:23 42:16,18

46:4 68:8 81:16

85:23 88:18

89:10 103:15

135:10

heard 25:16 33:18

38:24 43:20

65:19 77:1

89:12,19 93:22

108:16 115:4,8

124:25 126:11

hearing 7:19 100:22

101:6

height

19:8,9,10,20,21

20:1,2 32:13

heights 19:1,3 20:2

Helen 113:7

help 41:22 70:3

71:11,15 73:20

80:24 83:23 113:13

116:15,22 125:9,15

127:17 132:16

137:8,14,15 139:4

helped 79:5,9

helpful 116:6,19

137:17 138:20,25

139:6

helping 8:7 117:24

helps 14:22

HENDERSON 2:6 3:1

8:17,20

10:12,16,21 16:7

23:15 26:10,14

28:21 31:11

33:20 34:9 35:10

36:14,20 37:21

38:5,9,21 39:19

41:20,25

42:11,15,18

43:17 50:8,11

54:18 58:25 61:2

63:22 64:2 66:5

68:6,10 69:16

71:3,23 73:22

74:21 75:3,12,15

81:14 83:21

85:5,23 88:1,18

89:8,11 91:7,24

92:20 93:20

94:12 98:16

99:12 102:5,12

103:25 104:7

106:14 107:11,16

108:13 109:23

111:4,20 114:19

116:21 118:16

120:3,16,20,25

123:23 124:12

126:20 129:1,11,19

131:16,18 132:14

133:17,22 134:15

135:2,16

136:5,13,20

137:1,5,13

139:7,20

140:3,16,20 141:5

hi 2:18

high 23:4 25:14

32:9 35:7 36:5,7

43:8 57:1 64:18,19

80:1 119:20

121:10,12,13

higher 4:20 29:22

30:2 57:3 62:15

97:22 127:17

129:16

highest 32:8,10

39:25 43:7,16

122:24 126:18

high-peak 91:17

highway 32:5 45:1

Hill 55:10 histograms 30:21 historically

76:17,20 85:18

history 84:17

117:23

hit 81:17

holding 16:13 38:4

home 6:25 69:25

homes 113:10

HONO 21:10

hope 95:13 114:25

hopefully 70:13

76:1

hoping 140:4,9

hospital 51:23 52:7

82:21,23 84:4

87:11,16 88:22

100:20 104:9,19,25

105:1,25 106:4,9

117:17 120:15

121:2 126:14

127:1,9,25

130:4,13,25

131:11,24

host 44:22 46:13,22

128:1

hot 112:17

hotel 34:2 68:2

hour 20:16 22:21

33:7 35:6 57:20,25

64:15 74:14 92:3

119:20 121:22

hourly 17:3,7,11,15

18:10,11 74:15

88:24 89:15

90:19 91:5 92:14

94:19 99:3

hours 43:2 89:21

91:8,9,21 96:14

98:4 116:18

119:15,22

hour-to-hour 29:25 house 21:9,13 22:15 houses 21:17

huge 18:9 73:14

84:7 88:13 94:5,15

101:4

Huh 137:21

human 7:11 19:4

20:2 51:10,13

53:23 54:7 55:2

66:4

hundred 65:15

hyper 44:17 47:24





55:16 57:6

hyperreactivity

130:1

hyper-responsive

44:18

hyper- responsiveness

64:10

hypotheses 119:7

hypothesis 119:4

I

I...My 59:5

I'd 6:4 24:10

26:7 28:4 62:23

71:20 73:16

idea 20:14

73:7,16 88:21

119:9 141:6,12

ideal 109:9 ideally 135:18 identified 44:14

95:15 104:12

138:18 140:23

identify 51:19

66:14 106:22 127:3

ignore 50:13 74:1

ignored 69:1 126:3

I'll 3:4 8:25 9:3

10:1 15:10 27:9

38:7 39:20 43:24

47:17 50:13,14

52:9 72:3 75:21

110:13 113:6

115:18 135:16

140:13

illness 46:12

I'm 2:21 4:10

6:21 7:5,18

8:19,24 10:2 11:25

13:14 14:14,25

16:1 17:16 18:12

26:20 31:23

34:10 36:1 38:10

40:1 46:5 49:4

51:25 52:21

53:13 54:19

57:12 58:2 59:16

60:23 61:2 64:15

65:7,10 66:8

68:8 70:24 71:16

75:22 81:2 82:2

83:4,17 84:6

85:1,5 86:11 90:21

91:10 93:24

96:24 98:1,2

100:22 101:6

105:15,24 107:23

114:19,23 115:8

119:4,9 120:7,9

122:1,11,13 125:15

130:13 132:17

133:22

134:2,7,8,13 135:8

136:10 137:25

141:2

imagine 35:4 immediate 119:13,14 immune 119:5

120:9,17

immunity 46:13

impact 4:16 5:12

97:13

impairment 120:13

impassioned 81:19

82:2

imperfect 122:20

123:2

impetus 95:11 implication 122:21 implications 22:14 implicit 61:13,15
implicitly 37:19 implied 23:10 29:25 imply 94:11 98:23 implying 22:7

100:23

important 5:20

16:19 17:7 24:22

25:1,16,23 29:13

30:14,18 36:18

41:5 43:13 44:8

59:15 60:16

64:8,12 65:6 70:15

74:11 83:1 86:24

90:10,12,17 111:10

112:5 120:4,17

124:2 126:3 141:20

importantly 77:7

93:7

impression 39:22 improvement 3:12 improvements 48:13 inadequacy 47:7,8
inadequate 25:7

46:23

inadequately 16:15 inadvertently 20:8 inappropriate 21:23

52:10

incidence 100:16

101:2,3,14 109:6

incidents 127:25 inclined 127:20 include 40:3

44:23 127:13

140:25

included 28:3

40:1 108:1,3

including 41:23

82:17 114:12

126:19

inclusion 33:16

87:7

inclusive 40:4 income 5:11 incomplete 24:1

28:18

inconsistencies

47:5

incorporate 78:16 incorporation 24:23 increase 62:15 increased 45:15
increment 96:23 increments 97:16 indeed 92:16 independent

53:13,16,18,19

79:22 123:1





indicate 65:2

Indicating 100:5 indicative 115:14 indicators 64:12 individual 9:6 10:5

37:18 62:25 87:3

100:25

individually 101:17

individuals

11:14,22 24:24

43:6

indoor 13:19

16:14 17:5 22:20

23:14 60:17

indoor/outdoor

5:13,14

indoors 13:18 65:13

68:25

infants 91:17 infection 123:24 inferences 79:10 inflammation 46:14
inflammatory 122:3 influence 99:11 influences 17:20 inform 80:12 117:24
information 10:25

15:4,8 37:1 41:1

47:11,25 61:12

70:8 73:12 77:5

78:16 79:10

84:15 85:1 86:25

92:19 104:11

118:1,9,10 124:20

informative 93:15 informed 79:11 informs 95:7 inherent 61:15 initial
3:21

52:14 119:18

injury 122:3

inner 3:23

4:16,18 5:11,15

inner-city 37:12

131:9

input 9:18 49:10

136:3 140:1

inputs 22:23 117:6

insignificant 56:11

57:13 82:15

instance 23:25 instantaneous 71:22 instead 10:6

13:11 58:10

instinct 121:18

122:7

insufficient 29:14 integrate 140:6 integrative 76:13 intend 78:2 81:13
intended 13:16 78:8 intense 73:14

intent 85:14,19,22

interact 139:3

141:8

interaction 72:1

138:2

interactions 33:1

interest 29:16 38:8

47:13 66:8

108:16,21 114:24

119:3

interested 7:19

26:8 85:5 90:7

interesting 49:15

120:14

interestingly 28:16

interim 103:21

135:14

internal 128:9

interpret 60:11

67:12

interpretation

3:9 52:13 67:14,18

95:24

interpreting 66:16

inter-subject

12:6 13:4,6

14:17

15:13,17,20,22,25

intervention 67:1

intrigued 88:21

inventories 9:15

inventory 49:17 investigating 40:24 involved 32:24

70:18 79:17 86:16

irrelevant 119:21

ISA 44:6 45:22

50:19,21,24

66:22 72:8 74:10

89:13 98:22 102:15

103:19 113:15

130:20 131:22

137:2 140:10

isn't 53:14,17

54:12,15 85:2,3

92:2 105:18 128:11

isolate 7:17

issue 3:18 4:5 6:13

11:12 18:22,23

20:6 33:25 39:21

47:17 56:24

59:21 69:4 76:24

82:6 90:8 106:12

113:5 118:22 119:7

132:18

issues 3:13 45:9

49:4 64:7,25

68:4,13,14 71:19

138:17

Ito 100:21

it's 3:1 6:23

7:17,23 8:20 9:5

13:5,17 15:6

17:6 21:23 23:3

25:19 26:2 27:4

28:8 29:13,14

30:13 32:21

33:16,17 34:23

35:4,17,18,19,22

36:10,11,16

37:12,13 44:9,15

45:23 46:1 47:14

48:2,3,4,5,17,24

50:5,11,20 51:4

52:8 53:24 54:5

55:4 56:9,22,23

57:5,12 58:5





59:15,24

joined 2:18

60:6,7,10,15

62:7 64:11,18,23

69:17 71:6,13

72:23,24 73:25

75:10,18 76:7

78:21 79:16,19

81:3 88:3 90:15

94:3 95:1 98:14

99:20,25 100:1

101:17 103:1,15

104:2,20,24

105:6 108:10

113:17 114:24

116:23 117:25

120:17,20 123:21

124:21 127:9

128:2,4,10,18

129:6,12 130:15

134:17 136:3

138:7,13 141:5,9

I've 3:5 10:19

15:11 32:19

33:18 51:21 66:7

77:1 97:6 114:23

115:4,8,21 121:9

J

James 54:18 58:25

61:8 63:6 81:14

88:21 89:23 91:1

95:10 116:8

January 124:6

JENKINS 104:2,5

106:18 107:14

136:8

job 15:19 30:15

42:4 49:7 50:16

52:17 61:8 67:7

75:8

John 37:25 38:6

50:9 54:18 59:7

66:5,11 68:6 90:24

91:13 105:13,23

110:9 114:21

116:21 128:19

132:6

join 38:7

judged 63:9

judgment 48:22

52:21,22 93:25

107:4 118:23

119:24 133:15

judgments 79:11

119:25

July 102:18 103:8

135:12,23

137:12,17,18

141:22

jumps 8:6

June 103:8

140:13,22 141:22

justification 68:23

69:2

K

Karen 71:14 75:19

81:15 85:5,24

106:19 107:24

116:25 132:19

135:9 139:7 140:17

141:6,12

Karen's 74:22

Ken 106:14

Kent 64:2 66:5

68:24 123:23

key 20:7 27:5 110:2 kids 6:21 32:15 kindness 86:19

kinds 15:3,24 19:18

64:20 67:3,19

76:11 78:18

80:14 91:22

117:14,22 118:1,10

121:21

Kinney 2:17,19

3:4 30:19 31:16

59:1,2 92:21 94:9

Kleeberger 44:12

91:12,13

knobs 41:15

knowledge 9:2

75:16,20,24 87:17

L

L.A 72:19 97:21 lab 51:14 laboratory

56:2,14,15

59:12,18

laid 9:5 116:21 language 26:5 large 45:16 51:8

54:13 84:20

largely 7:23

larger 45:16

59:22 86:16 101:3

Larry 54:1

Larson

42:14,16,17,19

43:13 75:1,2,4

88:20 90:18

128:12,17 129:4,12

131:12 132:8

last 15:10 22:21

23:2 33:7 53:22

58:20 85:8,13

127:19

later 2:13

34:8,10 38:4,8

102:17 119:22

140:7

latest 135:25

latter 62:2

laudable 15:14 16:4

lay 54:11

lays 126:24

lead 2:18 43:20

70:13 140:1

leading 119:3

leads 119:8

120:11 140:1

Leanne 99:12 108:19

111:4 124:12

learn 72:25

least 6:5 10:8

12:22 24:11

27:15 39:13,14

42:12 47:17

70:11 71:1 73:3





75:9 82:14 83:4

87:3 91:22 98:21

107:21 119:12

120:11,12 122:5

124:6 128:23 140:2

leave 34:2 49:5

52:9 61:4 68:2

led 43:20

less 25:23 27:23

43:6 81:18 82:2

90:12 103:8

106:6 117:20 122:1

let's 2:16

26:1,10 40:14

42:12 74:22

75:15 82:23 98:3

letter 141:11 letters 139:14 level 9:10 13:13

15:13,22 31:20

62:4 64:18

82:20,25 84:11

86:9 87:3 96:10

97:7,8,10 101:8

107:5 115:15

117:15 122:6

levels 17:4

18:10,11 32:8,9,10

39:25 60:12

62:15 65:13

68:25 77:14

81:10 84:9,13 86:8

92:15 93:4,10

94:8,20,25 95:5

96:18,20

97:19,22,24 106:24

107:1,8 117:9,10

121:10,11,12,16,20

,25 124:10

Lianne 23:16

28:21 41:25

likely 7:2 17:20

18:14 20:9

21:16,17,18

58:12 119:2,9,11

120:1 121:6,7

123:3,10,22

likewise 93:5

limit 23:6 limitations 67:15 limited 21:24 39:16

42:2 110:23

113:1 124:21

limiting 38:14

40:21 109:11

limits 20:15

21:5,12 22:8,22

linear 95:3,4

lines 5:7 37:16

74:2 123:16

link 92:23 112:4 linked 112:12 links 41:11

list 46:19,22,23

59:2 68:12

listed 34:16

36:25 37:9 71:24

listen 32:19 72:14

listening 32:20

33:7 66:8 72:18

114:24

literally 12:1

literature 53:3

98:18 125:5

little 6:2 25:25

29:19 34:7 36:7

48:25 52:4 56:9

61:4 63:12 70:9

99:19 106:20 108:9

118:19 127:14

129:9 135:10

137:10

live 4:22 44:23 lived 21:10 lives 56:10 local 11:3 located 11:13
location 13:20 locations 11:19

16:24 106:24

107:1,9

log 20:14 21:14

22:5

long 2:23 17:1

24:18 50:22

60:8,14 64:17 68:3

83:3

long-term 46:25

47:9 80:21 120:13

Los 34:20 35:4

38:24 39:5

40:12,14,15 41:4

49:13,14

losing 41:1

lot 4:22 23:11

24:9,18,20,22

25:12 26:5 40:7

41:19 48:18

50:14 53:25

64:15 68:3

74:2,7 77:22

89:12,19 90:6

101:6,7,17,18,22

112:5 117:12,15

124:9 126:2,11

133:24,25 141:17

lots 6:8 55:19

65:25 77:9 127:2

low 60:12 94:10

95:5 121:11,17,25

lower 7:1

32:12,13 68:25

86:8 94:8 108:7

lowest 55:6

59:14,18

lunch 139:9,10

141:13

lung 46:11,13 50:23

65:2 120:13,22,24

M

magic 71:7

126:21,23

magnitude 63:4

123:17

main 6:5

mainly 3:14

major 6:3,10 8:8

24:12 29:21 32:6

41:10,13 51:6





71:11 73:3 74:9

91:3

makers 15:6 73:10 mandate 88:9 manifest 119:16 margin 12:16 marker
60:2,6 markers 70:9

MARTIN 71:16

75:21 83:20,22

85:12 103:17 117:1

135:22 136:14

138:15 140:18,21

Mary 114:1 masks 30:23 mass 22:9,12 match 30:11

105:9,15

matter 11:16

23:24 27:2 40:21

97:24

matters 28:6

115:3 119:5

maxes 113:3 maximize 127:5 may 2:5,11,13

4:16 5:13 6:8 14:6

34:7 37:16 41:22

43:21 45:14

49:15 50:4

64:20,21 65:13

69:9 70:1,8

99:2,11,19,23

100:2 103:8,15

105:4 113:13 118:2

124:1 128:3

132:3,4 138:25

140:7

maybe 6:9 13:14

34:23 40:3

42:5,7 57:9

63:25 68:14

71:2,14 83:10

112:14 131:19

mean 14:5 17:17

19:12 23:10

26:16,18 27:9

29:13 30:5,16

32:16 35:13

40:14 41:4,15

42:25 44:18

47:14 51:25

55:10 57:2,10

58:14

67:12,23,24

70:3,11 71:13

72:18,24 75:6

85:24 93:2

94:11,17 96:21

102:7 104:23

106:18 107:12,20

108:16 111:22

115:9,18 116:11,17

119:17 120:21,23

123:15,19 128:8,20

129:22 130:20

132:17 134:2

138:12 139:15

meaningful 49:7

117:20 118:7,20

136:3

meaningfully 80:12

means 18:16,18

19:11 25:6,10

29:12 44:17 83:7

102:20,21

measure 50:6

55:16 60:9 64:11

89:3 92:18 121:15

measure....I 57:10

measured 55:15,21

100:1 113:9,10

measurement 48:11

measurements

31:20 92:4

127:16 129:5

mechanistic 121:4

Medicare 104:13

105:6 110:10

115:23

medication 56:4

meet 27:12 55:13

88:8 103:23 108:6

meeting 71:4

85:13 102:17 108:6

135:11,21 136:10

137:17 138:16

139:19 140:12

141:17

meetings 138:4

members 138:2

140:14 141:19

mention 9:17

mentioned 23:17

37:5 40:22 43:14

59:7 106:19

messages 117:22

met 23:2 28:14

57:24

meta 62:24

meta-analysis

54:1 87:2,19

100:23 110:1

meters 19:5,14,15 method 8:13 92:2,14 methodology 8:7 methods 40:25

metric 88:22

microenvironment

6:19

micro-environment

37:7,20 43:14

microenvironments

6:14

micro-scale 129:12

mid 136:9

middle 135:24

mid-morning 71:18

Mike 47:3 mild 54:9 miles 35:6 milk 113:4

million 97:16

mind 2:15 4:15 40:9

50:20 85:10 88:4

115:5 134:7

mine 81:3 minimum 28:4 minuses 46:18 minute 74:23





misclassification

48:12

misleading 37:17

miss 137:22

missed 21:20 137:22 misstated 37:17 mistake 65:21 mistaken 9:21

mix 67:17 79:12

134:13

mixed 20:25 115:2

mixing 32:12

mixture 66:19 70:10

132:2

mo 11:17

mobile 5:24 6:3

mod 25:4 26:17 27:6

28:24

mode 17:22,23,25

model 5:21

11:4,6,18 13:24

16:20 17:2 18:25

19:7,16,17,21 20:4

22:10,24 23:4,7,12

24:17 29:20,22

30:3 31:1 32:15

36:8,12,13,15

38:12 40:22

41:18 49:11

88:23 89:1 99:19

modeled 6:19 18:3

29:21 30:7 37:19

41:1,8 93:10 97:6

modeling 5:23

6:15 20:8,25 24:16

35:4 41:3,18 48:14

49:19 65:23

73:14 92:16 93:7,8

124:23

modeling's 25:21

models 6:1 15:15

18:20,21 19:18

33:10,11,15

66:24 79:23

modification 26:17 moisture 55:20 molls 22:13

moment 47:17 68:5

money 81:17

monitor 10:25

11:2 12:22 19:7

23:23,25 24:3,6,12

25:6 28:13 29:23

30:3 99:9,10

110:23 116:13

monitored 36:10

monitoring

9:13,18,22 10:3

11:7,10,16,18 12:3

13:2 14:12,15

16:25 32:2,3,5

49:17 60:2 93:14

101:1 111:24

113:11 127:24

monitors 10:25

11:12 13:20

14:20,24 16:22

17:10,12

18:8,19,21

19:2,10,14 20:3

23:21,22

26:17,21,23

27:1,11,19,21,24

28:3,17,20 32:1

110:17,18,19,21

111:1,2,7,8

112:2,3,7,8,9,13,1

7 114:12,16 125:23

128:14

month 102:16,21

months 101:18 103:8

125:4 126:7 132:21

morbidity 47:1

52:6,18 57:19

82:21,23

morning 2:13

31:16

32:7,11,13,14

33:18 45:24 65:19

mortality

46:16,24 47:8

82:22 87:8,10,15

105:16 126:13,25

127:8 128:24

130:19,22

mostly 19:15 95:22

motor 60:5

move 38:10 42:2

43:18 75:15

78:23 79:5 87:20

116:18 125:8,16

moved 63:25

moving 49:11,14

67:11 71:12

multi-city

79:8,9,21 86:21

87:1,5,14 131:8

multiple 87:2 110:2

multiplicative

17:20,22

multiply 18:4

multi-pollutant

71:5,12 73:7

79:8,23 116:2

134:20

multi-variable

66:24

mustn't 23:6

N

NAAQS 76:10

115:13 129:7

narrow 21:5 72:23 narrows 131:1 national 101:16

104:10

nationwide 104:14

130:14

natural 12:3

nature 95:4 101:23

NCEA 114:1

nearly 32:8

near-road 42:22

111:8,10 126:4

near-roadway 111:2

necessarily 6:23

11:13 22:17 24:8

25:11 26:24

47:16 69:13

82:22 101:11

123:19





necessary 70:3

negative 51:14

58:15 61:19 63:1

neglect 71:25 neighborhoods 4:19 nervous 116:3

Netherlands 47:3

network 105:25

128:18 129:16

networks 75:6 101:1 nice 60:6 61:8 nitrated 121:23

Nitrogen 2:4 22:13

No2 4:17 7:3,18

8:10 21:3,19

22:1,21,23 25:17

32:2,3,8,9,10

46:20,25

50:17,22

51:3,12,15 52:24

53:11,13,16

60:1,2,8,12,13,17,

20,23,25

64:17,19 65:4,13

66:14,17,18

67:20 75:6

79:14,21,22

80:8,12

81:10,19,25

84:12 86:8 87:6,25

91:20 94:10,19

98:19

99:19,22,24

100:3

102:6,10,18

109:10,22 111:9

113:8 116:2

119:1 121:10 123:6

124:8 127:16 128:3

129:9 130:22

134:17,18

NO2s 129:7

non-asthmatics 26:4

non-major 6:6

non-physician 52:2

normal 20:14

21:14 22:5

normalization 11:15

normalized 11:6 normalizing 14:16 normally 116:23

North 128:22

nothing 56:6 78:7 notice 76:16 126:20 noticed 138:22 notion 78:9 80:16

NOx 7:3 21:10 39:25

45:15

NOx)PRIMARY 2:4

nth 53:15

NUGENT 63:24 136:12

138:1,6,10,12

139:17,25 140:4

nutshell 96:8

O

OAQPS 138:12 obesity 45:11 object 21:4,24 objecting 34:5 objection 20:13
objections 124:15 objectives 8:12 oblivious 78:20 observations

17:13 19:8 20:20

21:8,12 79:3,13

obtained 59:18 obvious 33:6 73:25 obviously 9:7

16:2 24:19 93:14

99:9 113:20 128:16

130:3 133:6

occasionally 56:7 occupational 44:25 occur 43:10 58:1

61:19 84:13

occurring 135:22

occurs 7:25 odd 129:9 of....I 72:24

offer 75:23 85:20

133:11

office 2:2 71:10

116:5 133:3 134:12

141:14

off-road 45:4

oh 11:14 14:5

36:3 72:2 75:3

96:2 104:7

118:17 137:25

okay 2:19 3:4

8:22 10:21 11:9,24

14:5,25 18:22

23:15

26:10,14,15

27:18 31:11

36:21 37:21 38:5,9

39:14 41:25

42:11 50:11 64:2

68:11 69:16

75:12,21 85:23

87:15 91:24

92:20 99:2 103:3,6

104:7 107:11,16

114:2,18 116:21

118:16 121:3

136:13 137:1 138:5

141:13,23

old 36:18 92:12

older 44:19

104:12 130:12,13

OMB 70:21,22

omission 6:6

one-hour 42:24 90:1

92:4,17

94:3,6,16 95:20

107:19 113:3,5

116:8 122:1 126:3

ones 28:6,7 58:12

114:13 130:8,9

131:4,24

one's 50:20 114:6

online 3:25

on-road 29:21,24

30:1 42:22

oops 38:10

open 8:23,25 78:15 opens 40:20 operating 27:12





28:13 134:23

operationally 44:17

45:20

operators 45:5 opinion 50:2 opinions 21:16 opportunity 71:18

75:5,11 137:6

140:23

opposed 6:4 52:5

63:17,20 79:12

81:9 94:10 117:15

opposite 55:9 option 141:3,4 options 141:17 order 3:14 88:8

109:21 123:17,22

125:4

ordered 102:24

103:5

Oriented 114:16 original 119:17 originally

9:16,17 136:14

others 7:19 32:20

62:22 68:18 88:17

otherwise 17:13

139:15 141:9

ought 57:22 58:22

59:22 118:1 130:2

ourselves 82:17

outcome 61:14,19,25

85:21

outdoor 16:12

17:7 18:25 19:2

65:14

outdoors 37:7,11 outlined 81:21 outlying 40:3 output 49:11 outright
69:12 outside 17:4 22:8

41:6 52:25

113:10 141:10

overall 3:6 8:10

18:4 20:13 21:16

48:19 51:2 54:25

96:25

overestimation

123:5,11

overlap 10:25 overseas 127:13 overwhelmed 48:15 oxidative 121:23

Oxide 22:13

OXIDES 2:4

ozone 53:18,20

66:13,20 79:4,11

81:20 84:17

99:17,22

100:3,15 101:5

102:7 109:2

117:3,4 119:14

125:20 128:23

129:4,8 132:22

 	P

p.m 141:24

Paige 114:12

panel 2:4 140:14

141:8,19

panels 67:3

paper 21:7 76:18

88:5 109:2

113:16 116:24

135:19

papers 85:18

paragraphs

139:13,22

parameter 36:18

83:1,8 123:2,8

parameters 23:11,13

Pardon 92:9

participate 54:7

73:17 140:15

particle 72:1

particles 32:2,11

33:1 66:20

particular 7:22

8:11 12:21 19:10

25:12 56:23 60:5

66:18 80:9 88:25

126:17 138:19

particularly 8:3

21:23 25:1,9 28:25

30:17 59:9 102:9

127:13

particulates 81:20

past 15:22 16:3

79:4

Pat 2:17,18 8:17

30:20 33:25 59:1

61:2 92:20 93:20

94:9

pathway 120:12

121:4

pathways 120:1

Patrick 31:15

patrol 45:2

Pat's 34:5

pattern 25:25 57:23

patterns 4:12 20:17

37:10 77:15 84:10

peak 74:11 89:15,19

92:17 93:4 94:6

98:19 108:15,21

122:3

peaks 80:18 91:20

99:3

peculiar 88:4

pedestrian 6:18

37:6

people 4:22 6:9

11:11 13:20 26:6

30:5 31:4

36:5,22

37:5,8,11,21 39:24

40:7 42:25

43:1,8 44:22

54:6 56:6 57:15,25

59:24 60:18

61:23,24 63:19

70:24 71:24 75:1

83:10 87:22

89:21 95:14

97:17 98:9

103:13,15

104:11,22 112:16

114:20 119:25

120:6,21 122:14





129:1 139:10,12,24

141:14

people's 85:24

113:10

per 20:15 22:3,13

47:15 60:13

65:16 78:7 79:11

96:12

percent 4:3 39:24

percentage 63:8

72:22

perfectly 60:7

perhaps 10:10 16:10

27:12,13 38:23

57:6,20 61:10

63:10 64:6 66:10

71:21 78:24 127:14

135:11 138:22

140:25

period 27:22

28:18 32:17

83:14 95:17

periods 27:10,15

28:2,9,10

person 2:22 9:19

25:2 37:19

personal 13:7,12

32:1 41:23 60:20

93:6,11,18 112:6,8

113:8

personally 109:12 person's 119:24 perspective 48:3

76:1,2,21

persuaded 52:22

66:8

phase 90:6 108:4

Philadelphia 3:23

4:1,11 11:1

34:18 36:6 40:23

41:13 42:4 49:10

72:13,25

philosophy 129:13

Phoenix 34:19

39:2,10,12

phone 2:23 8:18

37:22,23 50:9 59:1

63:23,25 114:20

phones 62:24

photochemistry

49:21

phrase 77:1 108:9 physicists 46:8 pick 61:25 90:13

98:6 102:2

122:10,11

picked 61:10

piece 26:19 109:4,5

111:16 118:9

pieces 15:3,7 109:3

PINKERTON 64:4

123:24

placed 117:13

places 11:13,19

66:20 131:3

plan 9:1,5

planning 63:24

103:13

plans 65:25

plausibility 47:6

48:1 80:24 124:25

plausible 120:12

121:4

play 41:16 59:22

78:6 115:4

playing 14:15 15:4

plays 41:17

please 29:3 41:21

plot 30:21 plots 31:1 plotted 100:17 plus 16:20 pluses 46:17

PM 53:19,20 66:13

79:4,11 84:17

99:22 100:15 101:5

103:14 110:16,22

124:10 128:22

129:4,5,8 132:22

point 14:7 16:2

29:22 30:4 31:12

33:22 34:10 37:4

43:11,15 62:23

71:17,21 72:6,13

73:1 76:6 78:12,14

82:10,11 91:11

94:24 98:4 99:7

100:13 106:1,13

115:18 120:8

124:25 125:2

127:19 129:24

130:23 132:24

135:14,15 136:3

138:24

pointed 30:5

85:24 117:3

points 64:16 72:3

76:6 100:7,19

126:13,18 127:2,3

policy 51:25

76:16,21 77:8

78:10

85:7,8,9,11,14

106:14 107:12,25

policymakers 15:6

77:23 133:3

politically 52:10

pollutant 33:11

53:11 79:23 102:11

134:23

pollutants 73:9

79:18 87:7 99:25

134:12

pollution 5:24 60:4

79:12 80:6,13 81:9

106:2

pollutions 60:10 pondered 78:18 pooling 54:13 poor 56:18 populated 106:8

population 9:7,24

12:17,21 13:7,17

14:10,22 24:7,9,14

26:22,24 33:3

40:17 44:23 58:2

59:9,11,20 61:21

64:24 72:23

96:13 98:12 104:13

105:21 109:7





110:17 114:14,15

120:23

population-oriented

110:18 111:7

125:23

populations 44:4,13

45:17 61:16

port 45:5 portion 98:12 pose 80:11 posed 107:17

position 78:22,23

positive 22:6

63:1,3

possibility 123:6

138:13

possible 9:5

53:17 58:5 66:2

86:20 118:20 135:5

138:7

possibly 15:19 23:5

27:20 65:9 82:11

132:21 133:5

posted 140:11

Postlethwait 61:3,6

62:18 63:10,15

postulating 84:24

potential 4:15

5:1 14:23 51:19

57:15 61:22

62:6,16 63:20

64:9,13 77:24 80:1

106:22 107:1

115:13 123:9 127:5

138:21

potentially 6:7

44:10 54:15 67:23

ppb 42:21

57:6,19,25 58:4

83:10,12

96:17,22 97:17

98:9

ppm 57:6 121:13 practical 92:25 precedence 90:11 predict 29:5 61:17
predicted

18:10,19 30:22

61:14

predicting 61:16

63:19

prediction 17:10

42:5

predictions 16:21

17:9 18:25

19:3,7 25:5,12

29:5,16 90:19

predictor 21:2 predicts 88:23 preferred 51:21

52:18

preliminary 19:18 prepared 33:23 preponderance 47:10 preschedule 103:18
presence 20:12

21:10 22:25

present 7:9 65:13

77:4 88:12 107:9

110:6

presentation

52:14 63:18

presented 8:6 12:22

52:21 60:8 61:12

pressing 105:6 presumably 69:3 presume 139:9 presuppose 100:25

101:1

pretty 2:7 4:3,12

9:20 17:16 24:5

31:9 80:1 90:4

127:10

prevalence 3:22

4:18 5:8 132:3

prevalent 5:16 prevented 85:25 preview 136:1 previous 54:20

125:4

previously 28:19

121:10

primarily 82:3

117:17

primary 118:1

principle 9:14

14:13 87:15

principles 10:6

19:24

priorities 134:11

prioritization

126:9

prioritizing 38:22 priority 126:18 probably 19:21

28:19 30:7 38:15

39:12,15 40:7 54:5

57:16 67:5,10 75:8

89:24 119:25

122:5,12 127:12

128:13 129:14

136:9

problem 9:7 16:19

20:5 24:5 26:25

59:16 70:10

71:6,11 74:25

75:16,17,18,22

86:3 87:23 96:4

102:5 110:12 124:5

134:17,19 141:21

problematic 52:8

53:7 125:15

problems 7:15

25:4 74:9 85:8

124:22,24

proceed 42:12

process 12:25 40:24

76:7,10,23

85:17,25 86:14

93:8 95:15 115:3

122:7

produce 76:11

78:3,13 85:14,16

136:25

produced 20:22

22:13 80:11 89:18

productive 76:3

134:25

profile 29:24

profiles 84:10





progress 24:18

49:23 135:13

promise 75:22 properly 60:16 property 22:7 proposal 104:4,5 propose
70:17

pursued 104:21

133:14

push 15:1 16:1 pushing 16:5 putting 15:25

67:8 68:20 75:2

115:17 117:15

132:15,20 133:23

134:4

quantitatively

59:16 126:6

quantities 93:13 quantity 29:16 39:3 quasi 84:15

proposed 12:15

 		queasiness 88:25

13:10 76:16

103:20,24 104:1

118:24

proposing 90:25 protect 95:12 protection 2:1

80:7,10 107:6

protective 53:10,14 prove 117:20 121:18 proven 87:15

provide 15:5

28:12 55:13 70:8

75:25 76:1 80:24

86:24 117:22 125:1

135:14

provided 76:20

107:6

provides 47:21 providing 138:3 proxies 123:2 proximity

44:20,21 45:8

89:20 99:10

proxy 60:9

public 44:8 53:9,14

80:7,10 95:12

105:24 107:5

138:4,22 140:12,13

141:10,12

publically 138:22 publication 113:17 published 31:18

113:14,18 114:8

125:5

pulling 118:11 purpose 75:23 purposes 42:8 128:5 pursue 100:10

 	Q

qualitative 78:10

81:12 107:4 111:14

117:7 125:11

127:20 132:22

134:1

qualitatively 126:5

quality 2:8

12:8,12,14,20

13:2,11,16,25

14:1,4,11,15 16:22

23:20 26:20

27:8,25 39:3,6

77:14 84:9,10

94:25 96:20 106:24

109:4 110:14,20

113:19 114:7

117:10

quandary 115:12 quantification 63:4 quantify 62:9,11,13 quantifying 63:7

128:5

quantitative 55:5

56:19 66:16

67:13

76:14,19,22

77:3,4,12,20,21

78:5,6,17 79:6

81:7,10

84:15,20,21,23

85:2 86:5,12,17

87:21 88:13

99:17 100:11

101:21 108:25

109:18 111:15

117:5,16 118:5,6,8

125:2,5,14,16

126:10,11 127:21

question 4:9 5:23

6:2 8:10 12:14

14:22 27:5 29:3,15

34:21 35:8 36:1

38:1 48:10,23

49:9,23,25 50:15

51:16 52:12,20

54:24

56:16,20,22,23

57:5,14,16

58:15,20,21

61:9,16 64:24

80:10 84:9,24

86:6,22 87:12

90:16 91:25

94:14 105:5 107:17

109:9 110:13

114:10

115:1,5,10

116:4,6,17 117:9

118:4 121:3 124:17

125:7,9 132:24

135:4 140:19

questions

3:2,5,16 5:18

34:14 36:24

37:23,24 39:1

43:25 48:20 49:2

50:15 54:23 59:3

71:5 74:7 77:14,19

80:15 95:15

99:16,20 111:13

115:17 117:14

124:15 138:19,23

140:24 141:1

quick 116:1

quickly 119:13

120:4

quite 3:10 7:5





24:3,21 69:22

74:15 88:7 112:6

119:15 121:25

128:24 129:9

134:20

R

racial 4:12 5:20

rail 45:4

raise 39:20 40:5

76:5 140:18

raised 4:4 33:25

49:19 68:15

71:20 78:19

135:3 138:17

raises 45:9 56:16 raising 141:3 range 22:18 44:25

48:8,9 49:3 59:6

63:3 90:4 94:22

95:25 108:3

ranges 20:18 117:18

rate 21:3,5

rates 3:24 4:1,2,18

5:11,12 17:6 21:19

22:16 101:2,3

109:6 120:2

rather 3:16 17:23

27:17 44:9 88:24

119:12

rationale 8:11

52:20

raw 113:18

REA 2:9

reach 79:9

react 71:1 83:22

reaction 55:21 56:9

71:22 100:14

reactivity 55:17 reader 8:7 31:2 reading 13:9 14:6

98:22

ready 86:12

real 21:17 33:2

40:17 48:2 50:5

52:6 54:15 56:16

78:4 85:10

reality 36:15 50:3

realize 56:4

65:13 124:9

really 3:7 5:18

7:7,17 8:7 14:7,14

15:16 16:17

17:2,10 23:24

24:7,19,25 27:5,21

28:2 30:24 31:7

32:10 33:10

34:14 48:5 52:8,17

54:13 56:16

58:14,20 59:15

60:13,19

63:16,21

64:23,24 65:5

66:23 69:22 71:4

72:4 73:1 74:24

75:16 83:4 89:14

90:10,21 94:1,11

99:15 101:11

107:17 108:13

110:6 115:12

121:10,12,13,17

124:5 130:21

132:24 136:3

reason 133:8

reasonable 3:20

11:10 26:3 30:15

48:10 49:6 52:11

57:4 65:9 66:1

67:5 98:4

reasonableness 16:3

reasonably 17:9

19:5

reasons 69:10 81:20

reassured 42:6

132:7

reassuring 31:10 receive 6:24 received 7:7 receiving 43:8 recent 28:18
132:23 recently 91:15

133:8

receptive 61:22

receptors 19:4 20:2

40:22,24,25

41:2,7,17

recess 75:14 recognize 27:8 81:5 recognizing 27:10 recommend 82:7 88:9

100:22

recommendation 10:9

42:1 133:13

recommendations

100:12

recommended 44:12

recommending

91:7,10 134:7

reconcile 74:17,18

reduce 40:25 41:7

80:8

reduced 80:6 84:4 reduces 112:5 reducing 41:2 reduction 82:21 reductions
65:2

83:1 84:12

reemphasize 43:15

reference 86:21

92:2,14

referring 12:7 refine 74:3 refined 101:8 refining 48:18

101:7

reflect 30:11 32:16

36:15

reflected 50:25 reflective 53:13 reframe 29:3 regard 34:14,17

44:2,21 45:19

48:20 49:2 50:15

51:12,15,16 54:4

71:19 76:25

106:1,5

regarding 53:23

86:15

regime 133:7

Regine 88:20





135:1 139:18 140:5

region 27:22 regionally 30:10 regression 20:23 regular 5:9 regulation
112:11 regulatory 123:20 relate 8:9 59:20

93:19 127:3

related 58:3 60:5

91:17 100:2 115:14

124:7 130:2,22

relates 16:19 18:23

58:20 76:6 83:15

relationship

5:13,14 12:20

34:25 60:19,25

66:18 93:9,12 99:9

relationships 66:23

67:16 116:15

127:24 129:8

relative 104:17

105:21 109:25

110:4

relatively 11:25

51:14 53:1 54:9

64:18 99:5

relevant 9:2

34:22 59:19 76:7

109:22 121:7

125:24

reliance 6:3 reluctant 115:25 rely 86:18 125:4,10 relying 25:8 remaining
17:24

116:18

remark 15:11 remarks 75:23 remember 55:17 removal 21:3,5 repair
121:14,21 repeat 46:3 repeatedly 65:19

115:5

repetitive 64:6

replicated 91:18

replies 2:14

report 135:14,20

136:25

reportable 104:24 reports 140:10 represent 24:7

27:16 57:1

representation 20:6

21:21 59:10

representative

11:14,17 31:1

54:14 59:8

represented

22:17,24 59:24

80:20

representing

26:22 60:13,23

reproduced 56:15 reproducible 55:7 request 135:15,17

137:14

require 56:4 requires 118:22 rerun 23:12 research 47:20

111:18

reservations 99:23 resided 11:22 residents 39:22

40:2 44:23

resilience 87:6

resolution 15:22

16:4

resolve 138:14

resource 101:15

106:12

resources

38:11,14 52:1 66:2

133:11 134:12

respect 42:5 74:8

respiratory

46:11,14,21 51:3

57:18 126:25 127:8

130:3 131:8 132:2

respond 43:10 100:4

118:18 137:14,15

response 23:2

53:5,12 55:23

60:25 62:5

63:1,3,5,8 77:11

89:8 109:5,13

116:14 118:20

119:6,13,14,18

120:17 121:6,9

125:21 126:9

132:3,10

responses 62:16

64:20 119:9

responsive 44:17

47:24

responsiveness

45:21 46:1,7

51:6,18 52:5

54:5 56:24

rest 8:15 103:7

111:17

restrict 110:18

125:22

result 13:4,6 52:14

82:20,25 96:11

results 3:10 6:6

7:12,21 9:13,18,22

10:3

11:6,7,10,16,18

12:2,4 13:3

14:11,12,15

20:16 30:3 31:18

54:3,4 60:1,8,11

83:9 93:13

rethink 75:5

reversal 120:2

121:6

reversed

119:2,12,21

reversible 120:4,5

review 2:4 3:21

80:12 135:19

reviewed 14:17

137:3

reviewer 3:11

reviews 76:11

79:4 84:19,22





Rich 68:6,10 69:16

richer 49:14

Richmond 13:15

14:19 62:1,20

63:13 89:9,12

93:22 94:18

95:24

96:4,17,20,22

97:2,11,14,18

98:15 100:4,7

102:9,16,20

103:1,4,7 105:7

107:21 110:13

112:19 113:14,20

114:3,10,16,18

126:8 127:18

131:6,21 132:18

134:10 136:23

137:2

riding 31:25

rigorous 55:3

risk 4:15 5:19 7:11

33:9 34:3 44:11

45:13 51:7,20

52:3,14 53:21

54:15,21 56:19

60:15,24 61:22

67:6 70:12 72:12

76:14,19,22

77:3,4,12,21

78:5,7,17 81:8

83:5,10

84:20,22,23 85:2

86:5,17 88:13,23

89:1 92:23 95:21

99:17 103:22

105:19,21

108:1,11,25

109:18,25

58:11 87:24

104:17,20 111:25

112:3

road 13:21 23:25

42:24 43:3 80:2

89:19 94:10

129:5,6,10

roads 6:6,9 27:2

28:6,8 29:21 75:7

roadway 37:7

41:11,13 111:1

128:14

roadways 4:21,23

6:4,10 7:2 41:11

44:20,21,25 45:9

74:12 89:21

99:10 114:13

Rob 41:20

robust 33:16 89:3

Rogene 36:19

42:14 61:6 75:1

role 14:15 15:3

59:22 91:16

roll 86:12

Ron 73:22 74:21

98:15,16 102:13

112:19

room 63:18 68:2

104:25 105:1

ROSENBAUM 29:18

35:12,25 36:3 37:4

41:9

Ross 114:1 134:12

rough 31:7

RSV 124:4

rule 76:16

running 15:15

RUSSELL 28:25 40:11

28:17 32:7,8 65:16

80:1 84:10 137:9

scale 5:20 100:24

118:23 119:2

scenario 12:3 61:17

83:25

schedule 103:23

136:24 138:21

139:19,21,23

scheduled 102:17

136:11

scheduling 136:16

Schleisser 21:7

Schlesinger

68:7,8,11

Schlesser's 21:7

school 6:22,23

31:20,25 32:1,4,14

44:24 45:7

schoolchildren

44:24

science 2:2 58:10

76:13 115:3

SCIENTIFIC 2:3

scientifically

49:15

Scott 117:2

se 47:15 60:13 78:7

79:11

season 124:11 seasonal 124:1,7 second 2:7

8:23,24 10:4 16:19

18:22 32:16,18

46:4 56:21 59:3

78:14 103:21

136:23 137:3

secondary 66:19

110:3,4 111:15

 		secondly 17:16 21:3

115:10 117:5

118:5,8 125:6

126:11 131:5

132:15 133:24

140:10

risk-based 12:13

risks 4:17 8:10

45:20 48:22

S

safely 38:15

salient 76:24

138:19

Samet 66:6,7

114:21,23

San 34:20

saw 9:10 20:20

35:2

section 2:11,17 7:6

8:1,2,3,9 23:20

24:16 52:21 93:2,5

seeing 76:18,19

85:25 93:9 113:24

seek 119:25

seem 20:16,19





65:8,25

seemed 36:4 42:20

65:1 119:15

seems 32:22 35:7

43:13,15 51:7

61:13 64:10

65:9,24 70:4

75:4 108:14 115:16

120:13 124:16

seen 62:21 87:2

106:25 117:15

segment 96:12

SEIGNEUR 35:21 selected 61:21 selecting 80:2

129:24

selection 34:17

39:18 64:14

selections 34:22

sell 82:17

semi-quantitative

92:24 117:13

125:12

send 136:8 140:13

sense 6:16 11:5

24:4,15 31:2 32:11

33:23 56:25

64:15 81:6 93:25

94:17 115:21

116:19

sensitive 3:19

55:19 59:10 120:8

sensitivity 36:17

57:7

separation 68:16

September

102:19,20,21 104:1

136:11,12,16,21

137:6

sequentially 3:16

27:23 98:8

series 7:24

99:18,22 108:22

111:5 128:9,21

serious 16:10 35:19

125:24

seriously 96:9

SES 5:20 45:8,14

SESSION 141:24

sets 12:1 16:25

23:13

setting 15:8

82:20 88:10

115:6 116:23

132:17

settings 59:18

several 16:9

25:16 30:5 73:13

99:5,13 131:7

severe 59:13 121:14

shape 2:8 72:5 share 129:22 shares 51:2 sharp 115:9 sharpen 90:8
sharpened 116:5

SHEPPARD 23:17

26:12,19 28:1,23

29:4 30:13 42:1

99:14 100:5 108:20

111:5 124:13

she's 100:9 shift 42:9 56:2 shocked 69:22 shoots 15:15

short 46:20 51:3

52:24 66:25 115:11

140:8,17,22 141:9

shorter 50:19

short-term 80:18

81:25 89:24

90:11 91:2,9,16

94:19 95:11 122:12

124:17

shot 15:22 16:3 showing 18:6 31:19 shown 65:12 87:14

101:11

shows 47:23

48:1,2 87:6

sidewalk 4:24 37:12

SIDS 91:18

signal 94:9

signals 18:17 77:15

121:24

significant 55:8

56:17 58:13 62:9

63:6,9 96:12 98:12

significantly 80:25

99:16

silence 126:21

silent 37:22

similar 13:10 15:11

31:3 70:8 98:23

128:24

simple 17:16

19:19 25:5

30:6,19,20 48:4

simply 59:17 117:8 simultaneously 4:20 single 21:8,13

33:11 48:4

53:10,11 73:8

79:22

sit 68:4 89:21

site 27:17 30:23

60:2,22 93:9,14,19

111:24 112:2,3

sited 27:2

sites 112:16

127:6 129:18

siting 24:3,12 28:4

60:20

sitings 28:9 sitting 127:1 situation 65:14

108:7

situations 89:17

six 36:25 size 123:4,6 skewed 22:5 skip 61:3

small 34:13 36:9

48:16 53:1 56:2

72:21

smaller 97:22 smattering 127:11 smokers 45:14 smoother 25:13 29:5





smooths 26:1

SO2 102:18 103:13

135:12

solely 9:24 solve 75:22 somebody 61:18 somehow 78:9 someone 87:13

something...I 71:9

somewhat 5:14 33:22

36:9 64:15 78:25

127:17 132:7

son 50:12

sorry 2:21 36:1

38:10 46:5

102:18 137:25

sort 3:15,16,23

4:4,10 5:9,10,17

8:3,6 9:14 10:7

13:9 14:12 15:8

16:4 17:19 22:9

34:21,25 43:25

44:19

45:9,17,20,23

46:8,18,19,24

47:10,11,13,19,23,

24,25 48:2,6

49:13,23,24

50:3,18 51:1,6,8

52:6 53:20 54:23

59:13,17 60:22

67:7 70:1,11 71:12

74:10 82:9 83:16

84:15 91:3,25

93:18 94:2 107:4

113:2,4 114:6

115:13 116:2,13,14

119:8 121:18 127:1

130:5,15 135:14

sorts 15:15

sound 3:8 36:14

52:15 93:21

sounded 12:12

sounds 71:23

source 5:1 6:3 9:15

16:20 94:2

source-oriented

110:23

source-related

16:14

sources 5:25

13:19 18:2 20:7

24:23 25:17

29:20,22 30:7

41:12,16

Southern 64:25

SOx 135:20 141:23 spasm 61:24 spatial 60:3,9

93:17

speak 35:16 48:1

131:7 133:10

SPEAKER 96:8,14

99:7 102:19,24

103:3,6 104:4

106:17

speakers 54:20 special 64:20 specialized 64:23 species 21:10 specific
4:5,6

5:11,12 7:24

9:19 37:12

80:2,3 84:19 101:5

110:14 113:3

127:19 138:23

141:1

specifically

26:16 67:21

specificity 78:25 specifics 62:23 specified 34:16 speed 36:3

speeds 32:12 35:3,5

36:2

Speizer 33:21 46:17

69:17,19 91:25

92:9,12 126:23

128:7,15 132:1

133:18 135:1,3

137:8,23

spend 4:22 139:13 spent 13:18 spirometer 55:22

82:16

spirometry 56:3

57:10

spite 69:4

spoken 37:3 74:11

spot 112:17 spread 127:15 staff 2:2 3:7

4:14 7:10

33:8,25

34:2,6,14 36:23

50:16,24 51:2,17

52:16,21 53:6

54:10,20 73:18

76:17 85:18

88:5,11 109:2

116:24 133:6

135:4,19 137:9,20

stage 73:2

standard 15:8 57:23

77:24 80:12,13

81:10,25 82:5,7,20

83:7,11,19 84:5,11

86:23 89:24

90:11 91:2,9 92:18

95:1,11 96:11

97:5,7,20 107:19

108:5,6 115:7

116:22 124:17

132:17

standardized 95:3

96:23

standards

77:18,25 81:19

88:3,10 89:25

90:1,3,5

94:6,7,19,23

95:8 106:22

107:2,7,22

108:3,8,10 118:24

start 2:10,14,16

40:14 42:12

43:22,25 75:21

98:11,13 139:22

started 16:22 45:23

97:16

starting 14:14





starts 82:4

state 26:4

106:3,10,13 113:22

stated 24:5 132:11

states 52:25

91:19 98:8,9

104:15 106:8

stating 26:7 stationary 5:24 stations 34:17 statistically 58:13 stature
50:7

stay 58:22

130:2,18,22

step 32:19 54:20

stepping 72:6

Steve 44:12 63:23

91:12

Steven 89:18 stick 7:21 stimulate 55:21 stop 10:18

stoves 22:1 60:18

straightforward

90:22

strangers 86:19 stratified 30:10,11 streams 9:15 strengths 33:5

34:15 111:23

stress 121:24 stricter 81:19 strong 43:19 stronger 5:14 33:19

99:24

strongest 55:12

95:17

strongly 21:24

33:11 53:9 58:21

structured 3:15 struggling 75:25 studied 53:2 121:10 studies 7:16,18

31:18,21 47:5

51:10 52:23 53:1

54:7,13 55:2 56:13

59:8,14 60:7

62:17,25

65:10,21

66:3,4,11,15,17

67:3,5 69:3,23

70:7 74:13

77:1,3,5,16

78:16 79:8,9,21

80:3,18,21 86:7

87:1,3,18

91:16,19,22

95:21 98:18

99:18,21,22

100:18,20 102:6,10

104:18 105:7,16

106:16,23,25 107:9

109:10,18 112:1,21

114:12 117:10,11

122:14 127:22

128:22 130:7

131:4,8,9

stuff 5:17 14:16

70:9

Su 113:7

subcommittee

135:4 137:21

sub-committee 70:18

73:16,17

sub-fraction 54:14

subgroup 138:6

sub-group 70:23

subject 55:17,24,25

56:3

subjects 59:7 63:2

subpopulation

3:19 104:18

subsequent 54:3

77:7 83:24 108:11

substance 76:7 suffer 61:24 63:19 sufficient 27:24

46:22

suggest 42:8 43:5

58:2 77:5 84:16

107:18 139:10

suggested 46:17

suggesting 25:18

43:7 83:4 84:1

89:20 115:22

139:21

suggestion 74:22

137:20 138:15

suggestions 74:4

suggestive 46:24

47:1

summaries 30:14

support 3:25 6:17

67:4 91:16 95:18

127:20 128:3

supported 89:15 supports 96:1 supposed 33:12

sure 2:20 4:14 6:10

7:5 10:2,19

17:16 18:12 24:4

31:23 34:9 42:3

49:4 53:13 59:16

70:24 73:24

75:22 76:9 78:4

81:3 83:17 84:6

112:18 115:8

120:19 136:10

surely 3:1 77:6 surface 35:17 surmounting 79:1 surrogates

122:16,19,20

surrounding 12:21

susceptibility 44:5

123:25

susceptible 44:10

68:17 104:13

105:18

suspect 97:22

99:8 127:16 129:7

sways 47:11 sympathize 81:16 symptoms 126:14

131:8,11 132:2

system 120:9

systematically

98:20 129:16

T





table 28:12 29:12

35:2 44:13 69:20

70:25 126:24

130:5,16

tail 23:4 29:8,11

30:17 59:10

tails 25:9,11,15

124:23

taking 6:22 56:12

60:23 65:24

71:17 81:3 83:24

89:5 122:23 127:21

141:18,19

talk 16:23 30:15

33:23 34:6,7,9

38:3 44:16,19,20

talked 36:5

44:4,6 45:3,11

94:20 106:20

135:23

talking 2:24 62:5

63:6 72:21 81:4

89:24,25 95:8

101:21,23 106:19

118:14 121:16,20

122:4 141:21

talks 89:18 tallying 46:17 target 61:21 63:16 targeting 29:13 teaching
2:21

tease 53:16,18

80:19

teased 66:24 technical 3:25 6:17 technically 3:8

52:14

technology 58:17

teleconference

138:22 140:13

temperature

20:12,17,23,24

55:19

temporal 29:24 temporally 30:10 tempting 59:25 ten 122:5

tend 4:19 7:14

25:12 74:13 112:17

tendency 55:14

123:4

term 17:1 46:20

50:22 51:3 52:24

64:17 66:25 83:3

terms 3:13 6:14

13:24 27:2 32:16

39:3 45:8,13,15

46:6 48:21 49:16

54:21,24

61:11,18 63:8

67:23 68:19,24

69:8 76:2,3

91:4,22 106:12

109:17 115:11

122:10 123:11

124:8 125:16,25

130:24

Terry 59:1

test 12:4 55:23 tested 49:13 thank 8:17 16:7

23:15 33:19

43:17,23 50:8

54:18 58:25 61:2

63:22 64:4

66:4,5 68:6

69:16 73:22

74:21 75:12

93:20 120:3 129:19

139:7 141:14,23

thanks 10:23

41:24 42:19 61:6

that'll 5:12

that's 8:5,15

11:8,21 14:2 15:11

16:5 17:1 18:21,22

19:15 20:4

22:17,19 23:6

24:19 25:3

26:6,14,19 27:3,16

30:9,17 31:9 32:17

33:14 34:23

35:6,13,23

36:11,12,13 37:8,9

40:1,8,15 41:11,24

42:7 43:9 44:14

48:5,15 50:25

51:5,13 52:21

54:11 56:24

57:16 58:3

59:19,21 60:25

61:7,12 62:6,11

63:21 66:18 67:6

68:4 69:14

70:3,15,23

72:10,15

73:6,12,21

74:16,18 76:17

79:14 82:3,10

83:2,8,16 84:14

86:2,13 87:9

89:5,7 90:20,22

92:5,12,15,22 94:3

95:8,19,22 96:4,16

97:5 98:1 102:18

103:14 104:13

106:6,15 107:11,25

108:17 109:1

111:13,16 112:4,10

113:12,20 115:5

116:20 118:14,15

119:23 120:22

121:4,24 122:12,16

124:18,20

125:13,25 126:12

127:8 128:25

130:8,20 131:21,25

132:18 133:8

135:5,17,22 136:11

138:3,5,12

139:8,21

141:2,4,12

the.....clearly

59:6

The....I 66:11

The...I 70:15 theory 20:20 46:9 therefore 22:23

25:15 55:9 57:8

58:9 63:5 111:7

133:20

there's 12:15





15:17,18,21

18:9,16,17 19:17

24:19 25:13,19

28:2 34:16,21

38:16 40:7

41:15,18 44:21

45:3,4,5,6,8,11,16

,24 46:6,8,16

47:5,7,10,12,15

52:25 53:24 54:2

55:14 57:3,9 61:15

62:13,23

67:2,12,25 68:3

78:7 81:2 88:2,4

89:4 94:5,15

95:3 96:21

97:20,21 108:14,24

109:3,4 111:10

123:3 126:23

138:20

they'll 137:3

they're 6:9,11

7:1 9:8 14:19

19:14 31:25 54:9

55:18 59:12

72:20 86:23 95:3,4

97:19,20 119:12

122:19 125:23

128:17 131:10,11

137:11

they've 9:5 112:1

122:25

third 18:23 20:4 thirteen 28:20 thorough 24:21 thoughts 71:19

85:24 116:16 134:6

three-hour 116:9

threshold 13:8

47:13,16

Thurston 31:14

39:20 53:8 72:2

86:11 88:16,19

92:7,10 94:13 95:9

96:2,6 100:22

104:8 105:11,14

109:24 111:21

129:22

131:14,17,19,23

137:19,25

138:5,8,11

till 38:4

Tim 42:17 43:17

75:2,3,12 88:20

90:25 128:12 132:8

time-wise 2:8 tiny 55:22 titling 61:11 today 8:18 58:22

69:25 77:2

134:14 139:1 140:3

141:2

Tolbert 114:11

toll 45:2

top 31:20

totally 21:20

38:2 55:25 128:18

touch 139:18 touched 106:21 tough 132:18 towards 68:12 tox 89:14
toxicological

67:5 118:23

toxicology

47:6,20 69:5

116:11,15

tracking 105:25 tract 15:24 traffic 4:20

32:5,7,9 36:10

41:14 49:17

transformer 20:23 transitional 91:15 treat 56:11

treated 23:22,23 treating 26:21 tremendous 109:20 tried 107:20 trivial
61:11 trouble 18:18

24:9 131:15

truck 45:3

true 8:1 105:6

120:9

truly 67:17 85:10

94:10

truncate 20:9

try 5:18 19:18,19

23:4 27:21 30:11

50:13 52:3 65:22

73:17,25 75:24

97:15 113:1,4

116:19 124:13

130:2,6

trying 14:2 17:2

52:2 53:9,16 57:12

58:2 61:20 63:17

72:19 73:13 83:25

Ts 72:14

TSD 28:11

Tucson 34:20

T-up 138:23 turbulence 17:19 turn 50:11 69:17

101:19 140:5

turnaround 140:9 turned 30:1 turning 52:19

tv 50:3 twenty 74:14 twice 135:19

twitchy 55:18 56:7 two-thirds 9:20 type 22:9,25 85:6

86:6 89:15 93:14

132:15

types 117:1 typical 93:18 typically 124:4

U

U.S 2:1 100:20

102:9 126:17

127:22 128:5,18

129:18 130:7

131:3,5

Uh-huh 100:5

ultimate 107:5,23





unaware 55:24,25

uncertainties 7:9

48:11

73:5,6,8,9,14

86:14,16 101:4

123:12,22

uncertainty

5:5,22 7:4 10:11

20:10 48:16,19

49:4 60:21 84:7,20

89:6 101:25

112:6 117:18

123:18 133:25

unclear 6:16 uncomfortable 60:23 uncritical 55:2 undercounting 25:20
underestimate 5:1

54:6 104:20

123:4 132:9

underestimating

129:15

underestimation

123:10

uniform 21:12,22,25

22:4

uniformity 97:21 unilaterally 50:5 unique 4:15

unit 27:17

United 52:25

91:19 98:7,9

104:14

units 22:12 universe 45:16 unless 139:18 unquantified 62:8 upon 63:16
65:12 upper 29:11 30:17

34:17 35:12,15,17

urban 5:11 6:8

40:13 100:24

usage 44:6

useful 67:2 72:24

118:15

useful-to-us 118:13

usually 55:22

utilization 128:1

61:15 88:25 133:24

variability's 25:18

variable 20:24 21:2

28:15 35:17

36:12 54:4 97:19

123:1

variables 67:16

variance 18:10

29:22,23,25

30:2,12 31:5,6,8

variances 29:19 variant 29:25 variation 126:1 variations 60:10 variety
97:21

126:12

vary 99:9 vehicle 60:5 verified 36:9 version 50:19

versus 20:2 68:17

113:5 122:2

vertical 19:22

viable 141:4

underlined 132:12

 		view 13:21 124:25

underlying 104:19

131:4

under-predicting

25:14

understand 5:18 8:8

12:25 33:7 34:23

38:5 45:22

51:17,25 52:20

54:12 55:15 78:4

83:12 93:16

96:24 135:5

understanding

2:23 12:11 13:5

60:19 87:23

98:22 121:19

understood 10:19

137:24

undertaking 49:22 undetectable 82:14 unfortunate 22:7 unification 46:9

47:19

V

valid 27:12 58:13

69:10

value 13:8,9 55:4

56:3 60:24 133:15

values 21:6 48:8

49:5 51:20 52:11

56:25 57:4,5

59:4,5,14 88:24

129:16

van 31:19 32:3 variabilities 88:14 variability 9:19

12:6 13:4,6

14:18

15:13,17,20,23,25

20:7,10 21:20

22:15,23

23:6,14,18 24:23

25:15,17 26:8,23

27:5 29:1 30:24

42:7 54:3 60:3

125:2 141:10

views 59:5 114:22

117:25 118:2

viral 123:24 virtually 106:3 visit 106:6

visits 82:24 104:25

126:15 127:1

130:4,25 131:11,24

vogue 50:4

vote 38:24 vulnerability 44:5 vulnerable 44:10

58:7 68:17

W

wait 134:4

walking 6:21

31:25 37:8

Washington 34:19

wasn't 2:20,22





5:8 12:11 15:3

24:4 25:24 81:6

109:14

waters 51:7

wave 101:9

ways 9:10 14:8 23:7

28:8 40:25 46:7

77:9 79:2 118:6

weak 46:15 weaker 99:20 website 140:12 we'd 27:23

week 140:7

weekly 31:5

weight 15:7 47:23

117:16

Welcome 2:6

we'll 13:8 30:8

34:9 43:20,22

74:22 75:19 100:11

132:25 137:15

141:21,22

well- characterized

36:12

we're 2:7,24 14:2

25:7,20 40:24

41:6,10 53:17

60:14 62:5,8,10

64:25 70:4,9

72:5 74:18 75:9,13

78:9 82:10,16 85:7

86:3 88:11 90:5

95:8 101:8,10

102:3 103:18

107:24 121:16,20

122:4,5 129:25

130:21 132:20

134:20,23 136:8

139:12 140:4,9

we've 15:21 43:19

61:21 62:6 74:11

80:20 82:12 83:7

87:13 89:19 105:25

106:18 107:17

115:14 126:11

128:20 132:19

134:16 138:18

141:16,17

whatever 12:5 17:24

37:13 48:13

61:25 62:11 122:25

132:16 138:2

what-if 123:14

WHEREUPON 75:14

141:24

whether 3:21 5:9,10

11:16 13:23

14:20 15:12,21

23:24 28:2,13,14

29:15 44:9 48:17

56:22 59:4,21

62:11 66:2 70:23

87:10 99:4 100:9

114:11 120:7

127:7,9 128:3

129:2 135:10,13

whole 9:7 15:20

20:22 24:10

25:12 29:9 44:22

45:8 46:18,22

48:17 77:22

113:5 125:8 128:1

widely 106:6

width 87:23 110:11 willingness 139:3,4 wind 32:12

winter 124:7 wise 66:1 71:8 woke 50:13 wonder 15:12

34:25 48:15

wonderful 73:16

wondering 114:19

120:7 134:2

wonders 15:21

work 3:23 7:11,14

46:9 47:2,16,21

49:1,23 53:21

73:18 74:2

86:18,20 87:8

92:15 93:3

95:12,16 103:16

109:20 112:12

124:18 125:4,20

137:21 138:21

141:20

workers 45:4,6,7

working 24:17 76:10

103:19 135:4

139:13,22 141:15

workload 109:11

world 33:2 48:3

50:5 52:6 54:16

worldwide 56:15

worried 25:25 26:20

122:1

worry 34:24 worrying 48:18 worst 61:17

worth 48:17 81:3

128:25

worthwhile 49:8

91:1

wrestling 125:15

write 63:17

write-up 16:18

writing 2:14 3:8

86:15

written 3:6 51:5

68:13 107:12 140:1

wrong 13:14 25:15

56:21 57:16 58:16

WYZGA 38:17

39:1,7,10,14 41:21

73:24 98:17 99:2

102:14 112:23

113:16

114:2,5,15,17

Y

yesterday

2:9,20,21,22

16:11,23 18:17

23:18 25:16

38:25 44:4,12

45:23 46:17

49:9,19 68:15,17

106:20

yet 24:2 33:23

53:19 71:13 100:11





113:18 134:21

York 4:22 100:21

104:22 105:2,3

106:7

you'll 18:5 39:22

112:14

yours 75:19

you've 16:3 17:14

18:6 19:11

21:3,11,13 22:18

23:3,7 24:18,22

30:14 36:25 57:9

58:4 65:8,19

74:2,7 83:14 90:18

92:13

95:13,16,17,20

100:3 109:8,19

110:1 116:21,24

118:14 125:4,20

126:25 127:14

130:15 132:2 137:3

Z

zero 22:7

 

 

 

 

EPA CASAC MEETING 05/02/08 CCR#15905-2	  PAGE  37 

EPA CASAC MEETING 05/02/08 CCR#15905-2	  PAGE  59 

