untitled

ILLINOIS POLLUTION CONTROL BOARD

June 16, 2006

IN THE MATTER OF

)

PROPOSED NEW 35 ILL ADM. CODE) R06-25

225 CONTROL OF EMISSIONS FROM) (Rulemaking - Air)

LARGE COMBUSTION SOURCES )

(MERCURY)

)

TESTIMONY OF DR. GERALD KEELER

PART II

BEFORE MARIE E. TIPSORD

HEARING OFFICER

The testimony of Dr. Gerald Keeler, a

witness called in the rulemaking proceeding before the

Illinois Pollution Control Board taken on June 16, 2006,

at 9:00 a.m., at the offices of the Environmental

Protection Agency, Springfield, Illinois, before Holly

A. Schmid, Notary Public and Certified Shorthand

Reporter, CSR No. 084-98-254587 for the State of

Illinois.

Page1

A P P E A R A N C E S

MEMBERS OF THE ILLINOIS POLLUTION CONTROL BOARD:

Ms. Marie E. Tipsord, Hearing Officer;

Dr. G. Tanner Girard, Board Member;

Ms. Andrea S. Moore, Board Member;

Mr. Anand Rao, Board Staff;

Mr. Thomas Johnson, Board Staff;

Mr. Tim Fox, Board Staff;

Mr. Nicholas Melas, Board Staff;

Ms. Alisa Liu, Board Staff.

COUNSEL FOR THE ILLINOIS

ENVIRONMENTAL PROTECTION AGENCY:

Mr. Charles Matoesian;

Ms. Gina Roccaforte;

Mr. John Kim;

Mr. Richard Ayres;

COUNSEL FROM SCHIFF-HARDEN

Ms. Kathleen Bassi;

Mr. Stephen Bonebrake;

Mr. Sheldon Zabel;

Mr. Jim Ingram, Dynegy, Inc.

COUNSEL FROM JENNER & BLOCK

Mr. Bill Forcade;

Ms. Katherine Rahill.

COUNSEL FROM McGUIRE-WOODS:

Mr. James Harrington;

Mr. David Rieser.

COUNSEL FROM THE CHICAGO LEGAL CLINIC

Mr. Keith I. Harley

COUNSEL FROM MIDWEST GENERATION

Mr. Basil G. Constantelos

Page2

IDENTIFICATION

PG.

Exhibit 29:

Exhibit 30:

Exhibit 31:

Exhibit 32:

130

Page3

MADAM HEARING OFFICER: Good morning,

everyone. Welcome back. This is day five, and we're

making some good progress.

My name is Maria Tipsord. I'm the

hearing officer in this proceeding in No. RO6-25. My

opening will be fairly truncated, since I don't see any

new faces.

I just want to introduce the panel to

you today. Board member Dr. G. Tanner Girard; Board

Member Anand Rao; the far right Nicholas Melas. To my

far left, Tom Johnson, also a board member. Our

technical unit today is represented by Anand Rao. Tim

Fox is with us again today, and Matt Reed is joining us

today with the board staff.

Before we go back to Dr. Keeler's

testimony, Mr. Rieser, I believe you were going to have

some exhibits for me this morning.

MR. RIESER: Why, yes, I do. We have

copies of the preliminary results of Steubenville. I'm

giving one to Mr. Harley, right off the bat, "The

Preliminary Results for Steubenville Mercury Deposition

Source Apportionment Study" from Tim Opelt, April 27,

2005. I don't recall what number you have that at.

MADAM HEARING OFFICER: We're going to

Page4

mark that as Exhibit 29, if there's no objection.

Seeing none, that will be Exhibit 29.

(Exhibit 29 was admitted.)

MR. RIESER: Then the second one is

"The Ambient Measurements to Support Coal Combustion

Emission Research" October 21, Beijing, China presented

by Matthew S. Landis, 2005. Again, we are presenting a

copy to Mr. Harley. We'll mark that as Exhibit No. 30,

if there's no objection. Seeing none, we will mark that

as Exhibit No. 30.

(Exhibit 30 was admitted.)

MS. BASSI: What was 29?

MADAM HEARING OFFICER: 29 is the

preliminary results. And "The Ambient Measurement" is

the title page on the second one. That will be Exhibit

30. Thirty is the Beijing study; 29 is the U.S. EPA

preliminary. Have we got the exhibits straightened out?

Has everybody got 29 and 30? I believe we are ready to

start with Question No. 10 for Dr. Keeler.

DR. KEELER: Question No. 10: "In

using the source-receptor approach to determine the

source of wet deposition mercury, is it accurate that

you determine the source by analyzing for other

constituents associated with those sources? A, does

Page5

that analysis demonstrate the proximity of those

sources? B: Does this analysis differentiate among

different types of coal used?" I assume you're asking

source receptor modeling approach not -- in using the

source-receptor approach.

MR. RIESER: Correct.

DR. KEELER: If you're referring to

the multivarious statistical analysis approach, it does

not demonstrate the proximity of the sources, so the

answer to A is no. If you're asking about the combined

hybrid receptor modeling approach, the answer is yes, so

just to differentiate --

MR. RIESER CONTINUES:

Q. Just for the record I note that the term

source-receptor study is the one that you used in your

testimony. For example, in this second paragraph of the

conclusion, we talked about, on page 5, that says,

"Source receptor studies have recently been completed

that indicate the coal-fired utilities contributed,

approximately, 70 percent of the mercury," so it wasn't

a term I made up. It was something that I thought I

took directly from your testimony.

A. Right. All the methods that we use are

looking for source-receptor relationships, and I was

Page6

just asking a clarifying question. That's all. It's a

correct use of the term.

Q. Thank you.

DR. KEELER: Question No. 11 -- just

so it's clear, B, the answer was no. "In your

testimony, you state that coal combustion was clearly

dominant, in terms of explaining the mercury deposition,

approximately, 70 percent. What is the basis for this

figure?" Do you want me to answer this again?

MR. RIESER: Yes, please.

DR. KEELER: This was the average of

the results of the PMF and the Unmixed models.

MR. RIESER CONTINUES:

Q. Of the two years, correct?

A. For the two years.

Q. 2003 and 2004?

A. Right, for the two-year combined analysis.

Q. There were differences in meteorology

between the two years, were there not?

A. Yes, there was. I said the deposition was

quite different between the two years.

Q. And in 2004, if I recall your testimony,

was dominated by four major precipitation events. Is

that correct?

Page7

A. Actually, I don't recall saying four, but

if that's what I said.

Q. I think you said "several" I'm sorry, but

it was dominated by several large precipitation events?

A. Yes, that's correct.

Q. And in what way was it dominated by those

events?

A. Those were very large deposition totals

relative to the total amount for the entire year.

Q. And what does that mean with respect to

your findings, if any?

A. Well, in one respect what it means is that

for a model, such as a deterministic model, if a missed

accurately calculating the deposition for that one

event, it will significantly alter their source-receptor

relationship and the ultimate approportionment of

whatever came from whatever source.

Q. I'm sorry. What does its mean for the

findings of the receptor study, the one you performed?

A. Not much. I mean, it's just another

example. I was highlighting the fact that it can have a

few single events that can lead to very large

deposition, and it's important in a sense of if one is

trying to compare the results of a CMAQ or TEAM, or that

Page8

type of a model to a receptor modeling that those single

events are very, very important, in terms of the overall

deposition to that location for the entire year, and so

not accurately predicting them is a major error, so that

was the purpose in my presentations and in my comment.

Q. Does it also suggest that source-receptor

done in a year with -- let me ask this, did you consider

-- and I do see in your testimony it's "a few large

precipitation events," I'm sorry. Is it your belief

that these large precipitation events were unusual with

the relationship to the typical weather in Steubenville?

A. They are not unusual in the sense that we

see these large events, typically at every site that we

measure. The sites in Michigan, the sites in Vermont,

the sites in Florida, we'll see a couple of very large

events that contribute a fairly large percentage of the

total deposition in all the sites that we make

measurements at.

Q. Is it accurate that several of these

precipitation events you observed in Steubenville were

results of hurricanes that came up from the Gulf of

Mexico?

A. Two of the events were the result of

cyclones that came up from the Gulf. They were the

Page9

remnants of hurricanes and provided very large amounts

of precipitation.

Q. Had you done the study in 2004, and only

had the 2004 data, would that have skewed the data

higher than the usual data that you would see for

Steubenville because of the large precipitation events?

A. No.

Q. In what way would it not?

A. It turns out that, for those large

precipitation events, the source-receptor model

underestimated the total deposition for those events,

and in fact, underestimates them quite a bit. Extreme

events in statistical analyses tend to be smoothed out,

so those couple big events, in terms of the deposition

-- they were over a microgram per square meter per event

-- were underestimated, and therefore, would not have

overestimated or given more of an emphasis to those

events.

Q. But you stated in your testimony that, at

least, one of those events added 8 percent of the total

mass of mercury to the amount of mercury observed for

the year 2004.

A. That's correct.

Q. And you don't consider that a significant

Page10

additional amount based on that one storm?

A. No. The point of my testimony was that it

does -- 8 percent is a significant additional amount,

but you asked the question whether that had a

significant impact on the results of our source-receptor

modeling, and I said no.

Q. I understand. I see where the four event

comes from. I'm sorry, Madam Hearing Officer. Which

was the Beijing?

MADAM HEARING OFFICER: The Beijing

study is 30?

MR. RIESER CONTINUES:

Q. Thank you. If you turn to Exhibit 30,

there is an "EPA PMF Estimated, versus Measured Mercury

Deposition."

A. Yes, I see that.

Q. And it's a graph comparing mercury

deposition in events, and there are four peaks, at

least, four large peaks that are larger than the other

peaks in that graph.

A. Correct.

Q. Do those represent the significant

precipitation events that you discussed?

A. Those are, yes.

Page11

Q. So would you agree that, in order for the

source-receptor -- excuse me -- receptor modeling to --

strike that. Let's go on to the next question.

DR. KEELER: "What range of values is

represented by this approximate value?" An approximate,

that is. I'm sorry. What did I say?

MR. RIESER: I thought I heard you say

an approximate.

DR. KEELER: "An approximate" it says.

I'm sorry, if I mumbled my words. The PMF and Unmixed

numbers, if I said I think PMF came out to be,

approximately, 70, might have been 69, and Unmix was 74,

and so that's where the number of, approximately, 70.

That wasn't meant to be a mathematically-rounded number,

by any means. It was supposed to be an approximation.

MR. RIESER CONTINUES:

Q. I asked this, I suppose, indeed, but is

there a plus or minus range associated with that?

A. Yeah. As I said yesterday, I believe the

uncertainty, if you do it the same way for both of the

models, is around 15 percent.

Q. Thank you.

DR. KEELER: The question is asking,

"What is the purpose of expressing this as an

Page12

approximate value?" For the purpose of simplicity, and I

have already answered what level of uncertainty to place

on this figure, which is the 15 percent. Question 12:

"You state that `a meteorological analysis corroborates

that a substantial amount of the mercury deposition

found at the Steubenville site was due to local and

regional sources."

MR. RIESER CONTINUES:

Q. Let me stop you. You use the term

"corroborate." What findings do you have, aside from

the meteorologic, that indicate a substantial amount of

mercury was due to local and regional sources?

A. Well, if one takes the emission

inventories for the region for mercury, and looks at the

amount of mercury and other constituents that are there,

one can then see that the largest contributor in the

emissions inventory in that region is actually

coal-fired utilities, so again, that's what I'm saying

is we, again, find that this matches up with what the

EPA emission inventory tells us.

Q. The emissions inventory is of emissions,

not deposition, correct?

A. It is emissions inventory. It's the

actual amount of emissions estimated from the source.

Page13

Q. So the emissions inventory doesn't, in and

of itself, tell you what the deposition is, does it?

A. No, it does not, unless you use a

deterministic model to model from emissions to

deposition?

Q. As I understand it, you did not use a

deterministic model?

A. No, I did not.

Q. And so in what way did you arrive at the

finding through your study that a substantial amount of

mercury deposition found at the Steubenville site was

due to local and regional sources?

A. When one does a source-receptor modeling,

one will often take the data, do the analysis, determine

an estimate of what a local source contributions and

regional source contributions are and then use the local

emissions inventory to what we call -- it's not a

validation. It's a comparison to the results that you

find. It doesn't make sense, in terms of what was

emitted in the region, to what you actually predicted.

This is a very common practice. It's done in a lot of

receptor modeling results. For example, if you estimate

that diesel emissions are an important emissions are

important and go to an area, and there are no diesel

Page14

trucks estimated to have emissions in that area, then

you find your results very unsatisfying.

Q. What is it about the emissions inventory

that allows you to make a qualitative statement with

respect to the deposition in the area, as you say, it's

a substantial amount that was due to local and regional

sources?

A. Are you asking me about the emissions

inventory?

Q. I'm asking you what about the emissions

inventory allows you to make the statement that a

substantial amount of mercury deposition found at the

Steubenville site was due to local and regional sources?

A. We have observational evidence that shows

us that, when we have high concentrations of reactive

mercury in the atmosphere, we would expect that to be

removed fairly close by precipitating systems. Taking

this understanding of the processes, one can then take

and use that understanding when looking at, okay, here

are emissions that are, to a large fraction, greater

than 50 percent as estimated by the utilities to a

common form of reactive mercury. When one goes and

looks at when and where the precipitation fell, one can

then make an estimate that this mercury would have been

Page15

removed through wet deposition in that area.

Q. How do you know what the amount of

reactive gaseous mercury is emitted by the sources

around Steubenville?

A. It's in the emissions inventory.

Q. The amount of reactive gaseous mercury is

in the inventory?

A. That's right. They provide an estimate of

the percentage of reactive mercury, particulate mercury

and elemental mercury that's emitted from each source.

Q. Did you do anything to -- the emissions

inventory is from 1999. Is that correct?

A. That's correct.

Q. Did you do anything to evaluate whether

that emissions inventory was still accurate as of the

time you did your study?

A. We only have the data that is presented to

us, so the answer would be no. There was no follow-up.

We can only use the information that is provided to us,

so the same information that all the modelers use.

Q. Could you have taken steps to -- strike

that. Put it this way, yesterday, when I asked you

about the sources that were nearby, you said that you

didn't know what they were, and you directed us to the

Page16

EPA map.

A. I don't think that's what I said, that I

didn't know what they were. You asked me how many,

specifically, and how much power was produced by the

plants, and I believe I answered I didn't have that

number in front of me.

Q. Did you have --

A. I know exactly where the sources are, not

just the coal-fired utility, but all of the other

sources in the vicinity.

Q. Did you have that number in front of you

when you performed your study?

A. We combine all the analysis tools that we

have including the emissions inventories, the maps

showing the locations of all the sites, how much of the

mercury is emitted from all those, and trying to

understand the results of our modeling, yes.

Q. Now, of course, we are limited in that we

don't actually have a copy of your study, but were there

maps showing the locations of the power plants, the coal

used, and then megawattage included as part or your

study?

A. The publication has a map showing all the

locations of the coal-fired utilities based on the 1990

Page17

information that was provided -- 1999, excuse me. I

misspoke. And I would have to go back and remember if

in the manuscript version that we submitted, whether it

had anything showing the actual amount of mercury that

was emitted from each stack. I don't recall that, but

it does show, geographically, where the plants were

located.

Q. So your report identifies specific power

plants as contributing to the mercury levels that you

observed in Steubenville. Is that correct?

A. No. As I mentioned yesterday, the

receptor modeling approach does not allow you to

identify a particular power plant. It identifies source

contributions from coal combustion.

Q. Then what was the point of including a map

of certain power plants in your study?

A. It's very important when you're publishing

any type of paper in international journals to show

people the location of where you are studying and the

major emissions sources in that vicinity. This is

common practice in the scientific literature.

Q. So the point was just to show the

location. Did you have other features, such as items of

historical interest in Steubenville?

Page18

A. This was a scientific peer-reviewed

literature paper. It's not a historical overview of the

area of Steubenville.

Q. I'm trying to understand why, if the point

of a receptor study, such as the one you performed, is

that it's not, A, can't use it to identify specific

sources, and B, it's not important to know what those

sources are, I'm trying to understand why you would

identify specific sources within the study, itself.

A. Again, you have misinterpreted and

misquoted what I said. A, I did not say it was not

important to know are the individual sources are. B, I

didn't say that individual sources are not important,

and I've been working in Eastern Ohio, Western

Pennsylvania air shed understanding the sources of, not

just mercury, but other compounds for over 20 years. I

have driven that area myself more hours than I can tell

you. I have spent an enormous amount of time on the

grown. I have actually gone around and photographed the

sources. Identified that, in fact, they were in the

locations that were given because, in the early days, we

were given locations for plants that actually didn't

coincide with where the plants were located, so we

actually went and verified that they were there, so I

Page19

have a long historical context, in terms of the research

I have been doing in that area. I published my first

source receptor paper on aerosols and their sources in

the Southwestern Pennsylvania area in 1986, so this is

not something that is new to me, and I think, from over

20 years worth of scientific experience, I believe I

have a good sense for what information is important to

provide in a scientific peer-reviewed paper and what's

not, and so I'm not sure why you keep saying I'm

referring to these things as being unimportant.

Q. Well, in your testimony, on page four, in

the paragraph that begins, "Multivariant statistical

receptor models have been successfully used to apportion

sources of mercury deposited in South Florida and

elsewhere." What you say after that in the next

sentence is, "These techniques have the advantage of not

requiring prior measurements of source profiles or

emission inventories." Similarly, in your report, which

is attached as Exhibit B, and I believe it's on page 18,

again, in a paragraph that begins with the statement,

"Regarding multivariant statistical receptor models,"

there's a sentence right before the end of the paragraph

that says, "Again, both techniques referring to

statistical techniques such as Unmix and deposit matrix

Page20

factorization, both techniques have the advantage of not

requiring prior measurements and source profiles or

emission inventories." So since you identify an

important, positive point that it doesn't require -- the

studies you performed don't require emissions

inventories, I'm having trouble understanding what role

they have to play in the findings that you make, and

obviously, I'm hampered that I don't have the actual

report in front of me.

A. I will try to say it again in a different

way, so perhaps it's more clear. Receptor modeling

allows you to estimate the source contributions by

source category for the pollutants that you're looking

at. It does not identify the specific stack or source,

a specific company that emitted the pollutant. You

asked the question about corroboration, and this is how

we got into discussing emissions and emissions

inventories, what we call "source reconciliation." If

one does a receptor modeling study, one that comes to

mind is one Glenn Kass did in the early 70's looking at

the sources in the Los Angeles Basin. He was looking at

the source of L and L carbon (phonetic) and basically,

did a receptor modeling very similar to what we done in

this study, and estimated the contributions to L and L

Page21

carbon and then went, as a corroborating piece of

information, took the inventory that estimated, from a

source perspective, how much carbon came from all the

sources in LA to see if he could get an approximate

balance. What goes up, has to come down. If you

estimate an overabundance of material that you can't

corroborate with observations or with emissions

estimates, then it gives you some idea that the

emissions are not correct or that there are other

sources contributing that perhaps the emission inventory

didn't conclude, so this is a standard practice in the

field, but it does not require -- receptor modeling does

not require the use of emissions data.

Q. Did you quantify the amount of mercury

deposition that you expected to find in Steubenville

based on the emissions inventory?

A. No, sir.

Q. In what way did the review of the

emissions inventory, then, corroborate the finding that

a substantial amount of mercury deposition found at the

Steubenville site was due to local and regional sources?

MADAM HEARING OFFICER: With all due

respect, I think that's where this all started. I think

he answered that question in the beginning.

Page22

MR. RIESER: Well, I'm really limited

here because I don't think he did answer the question.

He said he looked at the inventory and that supported

the finding that a substantial amount of mercury

deposition found at the Steubenville site was due to

local and regional studies because that finding is

corroborated by other information I have to assume

that's a finding, and again, I don't have the report in

front of me.

MADAM HEARING OFFICER: You have made

that point several times this morning, as well.

MR. RIESER: Right, but the point

being says he looked at the emissions inventory. We

don't know what plants he looked at. We don't know what

particular things he found. We don't know whether this

was a qualitative or quantitative finding of his.

MADAM HEARING OFFICER: I understand

that, but he has also repeatedly stated that the

receptor modeling does not look at individual plants.

It looks at a source category. Am I characterizing that

correctly?

DR. KEELER: Correct.

MR. MATOESIAN: Yes, and he's answered

that several times, category of sources you can trace

Page23

back to.

MR. RIESER: Thank you.

DR. KEELER: Just for the record, I

did state that we use the emissions inventory from the

1999 modeling that EPA did, the same inventory that was

used in the CAMR Rule, and we looked at all the

emissions for the entire Midwest, and actually, for all

the entire Eastern United States, so --

MR. RIESER:

Q. What I would like, then, is a list of the

plants that Dr. Keeler identified as being the plants

that he looked at the inventory for to identify the

amount of mercury that corroborated this finding.

A. It's the entire list that's in the CAMR

Rule, so if you want the entire listing, you can get

that on the EPA website.

MADAM HEARING OFFICER: Once again,

Dr. Keeler, when you start telling us to get stuff, that

means the Agency needs to provide it to us, if you have

not already. We need that as an exhibit.

MR. RIESER CONTINUES:

Q. I'm sorry, the entire list of all power

plants in the country?

A. That's the list we used and we looked at

Page24

the plants east of the Mississippi.

Q. Did you look at a specific range of plants

for this study, all the plants in the Mississippi --

A. We did not look at specific plants, as I

said earlier. We looked at the region.

Q. Well, I'm --

A. We didn't identify -- we did not identify

individual plant contributions that are in the receptor

modeling, or in terms of trying to understand the

contribution of a particular plant. That's not in the

scope of the work that we did.

MR. RIESER: I don't think it's unfair

to ask what plants, to get a list of the plants.

MADAM HEARING OFFICER: I have already

asked them.

MR. RIESER: I'm hearing "region" and

"list." I'm not hearing, "I looked at these plants,"

and I'm assuming there is a list.

MR. MATOESIAN: He stated this --

MR. RIESER: I'm not done talking.

I'm assuming there is a list of plants that he looked at

because he describes, "We looked at these RGM's, and we

calculated the RGM's with the" -- blah, blah blah, and

that's what we need to corroborate, the meteorological

Page25

findings, so I would like a list of those plants.

MR. MATOESIAN: I believe you directed

us to provide that and as he stated, it was a list of

plants east of the Mississippi, I believe.

DR. KEELER: Right.

MS. BASSI CONTINUES:

Q. Sorry to jump into this fray, but please

correct me if I'm wrong, and that's my question. I

thought, at one point, you said that, in your

manuscript, or in your published data, or some place,

there was a map on which you identified the plants that

were in the region or in the vicinity or in the local

area, or however you defined it, but it sounded to me as

if you were saying that there were certain power plants,

perhaps, other types of sources that were on a map

around the Steubenville area that is plotted on a map

that you have identified as dots on a map. Is that

correct?

A. That's correct. In the submitted version

of the paper we submitted, a figure, which had a copy --

I mean had a spot showing the location of the

Steubenville site, and then had within -- I would say it

showed the greater Steubenville area, including most of

Ohio, half of Pennsylvania, down to the south of it, up

Page26

to the north. It didn't include a large area. It was

really meant to show where Steubenville is, but it did

include a point showing a major source of mercury from

the `99 inventory plotted on that.

Q. Then somehow that got expanded to the

entire emissions inventory used in the CAMR?

A. Because there are two different questions.

One was asking a question about figures, and one was

asking what we look at when we look at the emissions,

and I was trying to be clear about that.

Q. So when you looked at the emissions to

corroborate your findings from your study, you were

looking at more than just what was plotted on your map.

Is that correct?

A. That is correct.

MR. ZABEL CONTINUES:

Q. Maybe this is obvious, but it seems to me

we have been dancing around the question of why we can't

get a copy of the draft report.

A. It's not a report. It's a manuscript that

was submitted to a scientific journal that's in the

review process. We've already been given reviews, and

we're in the process of finalizing that paper. The

paper will be provided to everyone, once the paper is

Page27

sent for publication. One other thing that is very

difficult today, especially when doing environmental

research, is to maintain the integrity of the scientific

research publication process, and one of the reasons why

we didn't want to make presentations and do briefings in

front of the EPA, until we had our final analysis was

because these results, then, get misinterpreted. People

that don't do scientific research often don't understand

the process that science is -- you learn as you go. You

make corrections and improvements to the scientific

research that you are doing, and that, at times, if you

don't have the data that you would like at hand, you get

that data, and people provide new data. You submit your

information and you refine your analysis, and if

versions of publications are going around prior to you

actually submitting it to the peer-reviewed literature,

then the whole process becomes -- basically, the

integrity of the process disappears.

Q. I'm sorry. My question wasn't answered.

A. I'm saying that this hearing has come at a

time where the paper has not come out in press, yet, so

that's the reason why the paper hasn't been provided.

Q. I perfectly understand that. I want to

know is there some legal reason why we cannot have that

Page28

document as it exists today? They are asking this Board

to rely on what you're saying is in that paper. We have

no chance to judge that. If I were in litigation, I

could get it in Discovery. I could order it produced.

They want this Board to rely on what Dr. Keeler is

telling us. I have no reason to think he's not telling

us the truth, but I can't check it. I don't know of a

legal reason -- if there is one, counsel can tell me --

why we can't have copy of that draft.

MADAM HEARING OFFICER: I actually

believe that yesterday Dr. Keeler invoked intellectual

property.

MR. ZABEL: Then he can have it under

confidentiality. We won't publish it. That's not the

point.

DR. KEELER: I'm not sure how to

answer you. I just know that I know that, in terms of

the scientific peer review process, this is not

appropriate and --

DR. GIRARD: Could I ask a question.

You say it's been reviewed. Have you already submitted

the revised paper to the journal?

DR. KEELER: No, we have not. It's --

right now it's under going, again, another level of

Page29

review at the Agency, so it's not a finalized document,

yet, and that's the main reason why -- if I hadn't

submitted the final version, I would have no problem

with sharing that, but again, this is -- it's one of

these thing that it's just not done this way.

MR. RAO: And based on the comments

you have received from this peer-review committee, does

it change your results or conclusions in any way?

DR. KEELER: No, it does not.

MS. MOORE: Is the peer view committee

looking at it now within the EPA?

DR. KEELER: Yes. It's an internal

review.

MS. MOORE: Given the fact that your

findings are U.S. EPA excuse me -- are a little

different than the direction that the U.S. EPA has

thought to go in regard to mercury, don't you think

there would be some chance that the release of this

paper might be held for some period of time for further

and further and further peer review?

DR. KEELER: I don't believe they,

legally, can hold up the publication. I'm the one that

has to then make a determination that, okay, we are not

going to wait anymore, so we have a deadline of July 1

Page30

to return the final version of the paper to the Journal,

and so that's the drop dead deadline for all of their

discussions and review, no matter what, in my mind.

MS. MOORE: So July 1 it would

actually be available?

DR. KEELER: That should be -- that's

correct. It should be available around the July 1 time

frame.

MS. MOORE: It would be pretty clear

to you then if there were some other reason they didn't

want to release it, not that that kind of thing would

ever happen.

MS. BASSI: Certainly not in Illinois.

DR. KEELER: I think that's correct.

MR. ZABEL CONTINUES:

Q. Is my understanding correct, Doctor,

that's the version you would be willing to release, the

July 1 version?

A. The version that we will submit as a final

version to the Journal will be the one I'm happy to

share.

Q. Then I guess my only inquiry would be

whether we can have Dr. Keeler back after our experts

have had a chance to study that.

Page31

A. It all depends on when.

MR. RIESER: It would be my

suggestion, frankly, that, if it's going to be released

by July 1, it would be my suggestion we just halt the

cross-examination here --

MADAM HEARING OFFICER: We're not

going to halt the cross-examination because, if it

doesn't happen July 1, and we're right here arguing

about whether or not the Steubenville study is going to

be issued. We're -- we'll continue on.

Here's what I'm going to do. I'm

going to direct the Agency to provide a copy of the

Steubenville report no later than July 5. If it can't

be available by July 5, then I will be willing to hear

motions at that point.

MR. ZABEL: I will have a motion at

that point, Madam Hearing Officer.

MADAM HEARING OFFICER: I just want to

put it off, in case we can get to --

MR. ZABEL: I will reserve my motion,

until the end of the Fourth of July holiday.

DR. KEELER: Can I just make one

point? I'm not going to be in town -- I'm actually going

to be in the field, until about July 10, so I'm not

Page32

going to be dealing with this at all. I told you July

1, just so that you would know that that's when we have

to do it, but I'm not going to be in town to be able to

do anything in regards to this, until I get back on

about July 10. I think I get back to town late on July

9, so putting a date of July 5 is not going to work for

me. I'm just telling you. I have other commitments

that I have made previously. This hearing got postponed

from May, to now, and I had to cancel two different

trips that I had planned for June in order to be here

this week, and I can't cancel any more in the future

because it involves field work, and I have contracts

that are obligating me to be in the field.

MR. RIESER: We're just asking for the

report, not you.

DR. KEELER: But if EPA responds back

to me with whatever the final comments and all that kind

of stuff very late in the game -- I'm hoping I get it

before that, so I can get this taken care of, but if

they get it back to me June 30 or something like that,

I'm going to have a really hard time in dealing with

that. I don't know what will happen. That's out of my

hands.

MR. ZABEL: Madam Hearing Officer, I

Page33

will make my motion, regardless. I understand his

difficulty. I'm giving up most of Father's Day with my

children because of this hearing on Monday, so I'm

sympathetic doctor, but at some point, we need the

evidence in the record, and if it isn't there, I will

make my motion.

DR. GIRARD: One problem we have got

is that we have already got partial -- we have partial

information in the record now, which you already said is

not final. It seems to me that really need to talk with

your attorneys to see if it is possible to put a draft

copy, as it is now, in the record. There's no way it's

going to be released for publication. We take trade

secrets all the time in cases, so you need to talk with

your attorneys. I think it would short circuit some of

this questions by three or four hours, if you could

simply look at what methods were used, and a lot of

these questions go to procedures. They don't go to

results at all. So I think you need to go talk with

your attorneys.

(At which point, a break was taken.)

MADAM HEARING OFFICER: Back on the

record.

MR. MATOESIAN: In response to the

Page34

question of -- on the question of Dr. Keeler's report,

we're going to have to talk to the U.S. EPA about

whether it can be released and what portion, so we

cannot today commit to the July 5 date. We are going to

reach out to them, and hopefully, get an answer by early

next week, Monday, perhaps Tuesday. Then give you a

more definitive answer then. That's the best we can do

at this point.

MADAM HEARING OFFICER: Mr. Zabel.

MR. ZABEL: May I add to my request

then that we also, if we receive anything, receive the

200 pages or thereabouts of commentary that Dr. Keeler

indicated he received from U.S. EPA? You are going to

be talking to U.S. EPA, so you might as well talk to

them about both.

MR. MATOESIAN: We will bring that

issue up, as well, to them.

MR. RIESER: One other item, which I

asked for yesterday, was if the LADCO presentation could

be provided and I think Dr. Keeler said he would look

for it.

DR. KEELER: It's being photocopied as

we speak I believe.

MR. RIESER: Great. Thank you.

Page35

MADAM HEARING OFFICER: At this point,

then, we will move on with the cross-examination of

Dr. Keeler, and we will address issues considering the

availability or unavailability of the Steubenville study

published report at a later date.

MR. RIESER: I think we are on 12-A.

DR. KEELER: 12-A: "In your

testimony, how do you define "local" and "regional"

sources?" I believe I answered that previously. We had

an extended discussion about that yesterday, so that's

been answered. B: "Are the sources described in this

state limited to coal-burning sources?" No, they are

not.

DR. RIESER CONTINUES:

Q. What other sources are?

A. I believe I answered this yesterday, as

well. Iron and steel production, municipal waste

incineration, refining, metals production I think were

the sources that I mentioned yesterday. C: "Is the

deposition" --

Q. I'm sorry --

A. "Limited to wet deposition?"

Q. I'm sorry, one question on that. In

identifying where the sources -- were the sources you

Page36

identified as a result of the receptor sampling and

modeling or as a result of the review of the emissions

inventory?

A. Only the receptor modeling.

Q. Thank you.

DR. KEELER: C, "Is the deposition

described here limited to wet deposition?" The source

apportionment results receptor modeling performed were

only including the wet deposition?

MR. HARLEY CONTINUES:

Q. Dr. Keeler, mindful of the fact that the

source apportionment was based on wet deposition, on

days when it didn't rain, when there was only dry

deposition, do you have an opinion about how that might

have effected the results of the work that you did at

Steubenville?

A. As part of this overall project, we will

be looking at, both, wet and dry deposition. Again,

this was the first paper that we attempted to get into

the scientific journal, the wet deposition took

precedence. The dry deposition work is ongoing, and

it's going to be based on the extensive on-site

meteorological measurements, as well as the reactive

gaseous mercury, particulate mercury, and elemental

Page37

mercury measurements that we're actually making on site.

We have, at this time, not modelled the deposition using

the ambient data. However, based on our past projects,

we would assume -- not one would assume. The dry

deposition component should be a significant additional

loading of mercury to this ecosystem. In our studies in

Detroit, where we were in areas with similar source

contributions, or similar source densities, the portions

of dry deposition was -- well, the dry deposition that

was about half of the overall deposition that we saw at

the site at Steubenville was different in terms of the

mix of sources than Detroit is, so we expect to see some

different contributions, and so forth, but dry

deposition is half the story we believe, and in an area

with lots of sources, dry deposition could dominate the

local deposition to this area.

Q. To be clear, that would be added to what

you have already presented in your testimony about

deposition through precipitation?

A. Yeah. The wet deposition -- just to be

clear, the wet deposition that we modelled and think

that 70 percent of which is coming from coal-fired

utilities, is only about half of the total deposition,

so mercury enters the waterbody by wet deposition, and

Page38

then there is a whole other amount of mercury that

enters through mercury that's in the air, basically,

interacting with the surface and falling to that

surface, so at this point, we haven't really apportioned

half of the total deposition, which is again, part of

the goal, but that's further down the road.

Q. Based on previous work that you have done,

would you reasonably expect that the utility

contribution to dry deposition would be present in the

same proportion as it is for wet deposition?

A. The importance of wet, versus dry, can

vary from one spot to another, but based on, again, past

experience, I would expect that the sources that are

within the 50-mile radius of Steubenville will have a

major importance, in terms of the dry deposition and

influence in the ambient levels of mercury in the air at

Steubenville.

CROSS EXAMINATION BY MR. HARRINGTON:

Q. I don't believe you answered the question

as to whether utilities would have the same impact on

dry as wet deposition. Could you comment on that

further, please?

A. I can't give you a quantitative answer to

be specific on that. I would expect them to be a

Page39

significant contributor to the dry deposition. I

cannot, because I have not done that analysis, give you

a quantitative number what percentage of dry deposition

to coal-fired power plants is.

Q. Do things such as steel mills, metal

working refineries have potentially and proportionate

greater impact on dry deposition?

A. We're talking about dry deposition of

mercury, right?

Q. Dry deposition of mercury in Steubenville.

Keep it within the context of --

A. Again, local sources, if they are closer

to the site and have different forms of mercury that

they are emitting, would have a different contribution

than coal-fired utilities would, yes.

Q. Thank you.

MR. RIESER CONTINUES:

Q. And during the 2003-2004 sampling, did you

measure dry deposition in Steubenville?

A. We did not measure continuously for the

2003-2004 time period a direct measurement of dry

deposition that's analogous to what we did for wet

deposition. What we have done and what we plan to do to

provide that estimate is to use ambient concentrations

Page40

of mercury that have been measured, so we have

continuous measurements of reactive gaseous mercury,

elemental mercury, particulate mercury that are

performed on an hourly time basis together with the

onsite meteorological data, and we will model the dry

deposition quantity to the surface at the Steubenville

site. As part of intensive periods, we have direct

measurements of mercury of dry deposition using

surrogate surfaces, and other techniques, which all of

these techniques have uncertainties, so we compare those

measurements together with models to come up with a best

estimate for dry deposition of mercury in Steubenville,

but again, that analysis is not completed.

MR. HARLEY CONTINUES:

Q. In talking about the conduct of your

ongoing research, you testified yesterday that you have

assembled, but not compiled data for wet deposition for

the year 2005. Is that correct?

A. The database is not finalized. We haven't

compiled all of the trace element data and mercury data

and ion data at this point.

Q. Based on your firsthand review of the data

as it now exists, do you have an opinion as to whether

or not the information from 2005 is consistent with what

Page41

you saw in 2004, 2003, or is it just too early?

A. We haven't done any source apportionment

modeling on the 2005 data at all because the data is not

completed. The total amount of mercury deposition in

Steubenville in 2005 which is the data that is done I

don't have the exact number in front of me, but it's

greater than the average. It's more than 2003, and I

think less than 2004, so it's in the same vicinity.

MR. BONEBRAKE CONTINUES:

Q. Is the relative proportion of different

species of mercury deposited different in wet

deposition, as opposed to dry deposition?

A. The answer to that question -- I'm going

to try to simplify it. The reactive gaseous mercury

forms and the particulate mercury forms are the ones

that will most readily go into precipitation. When we

analyze our samples for our precipitation samples for

these species we see a very wide range in how much

reactive particulate is there and part of that is

because the reactions occur in the precipitation sample,

itself, giving us somewhat an uncertain answer as to how

much particulate was there, in the first place.

Reactive mercury is also going to dominate the dry

deposition because of its properties. It has a very

Page42

high deposition velocity, and so similar to wet

deposition, reactive mercury is going to dominate in wet

deposition, and also going to dominate in dry

deposition, so in the absence of precipitation, you

would expect do see a fairly rapid depreciation of the

reactive mercury to the surface similarly as if it was

going into cloud water, so I would expect reactive

mercury to dominate, both, the wet and the dry. The one

part, the dry deposition, that is less well understood,

and is not included in the models, and the models that I

have reviewed in the current literature is the uptake of

elemental mercury by force canopies and into the plant

materials. This form of mercury and this deposition

loss is something that's not characterized well, but

it's a significant amount of mercury on an annual basis.

It can be as much as three times the wet deposition to a

forested ecosystem. It's not a simple answer.

Q. The deposition -- the mercury uptake in

the forest canopy that you just mentioned, would you

anticipate that that ultimately ends up in the

environment, and if so, what portion of the environment,

aside from the trees, themselves?

A. Well, what we've quantified thus far is

the mercury that's taken up in deciduous leaves and

Page43

confer needles and so forth in the forest canopy. That

material, at least, for the deciduous trees -- every

fall, that material drops from the trees and falls to

the forest floor, so mercury gets bound up in the

organic matter and that material is bound to decompose

like the other plant material and start working through

the process of weathering and so forth, go into the

soils, and some of it will run off, depending on the

ecosystem structure, so it's a delayed signal, in terms

of deposition to the forest canopy uptake through the

growing season, deposition and then a slow process where

it will then slowly make its way to the ecosystem.

MR. ZABEL CONTINUES:

Q. Just for my benefit, Doctor, you mentioned

in discussing dry deposition the use of surrogate

surfaces. Explain to me what that is or what that

means.

A. Sure. One of difficulties in making

measurements of the deposition of any pollutant, whether

it's mercury or sulfur or any pollutant that's in the

atmosphere, is that the surface of the earth is complex.

Trees present a very complex surface in an urban area,

buildings and roads and sidewalks, all those surfaces

are very complex, in terms of the surface area, the

Page44

composition, how they move, how the wind and meteorology

influences those, and so we can't make a replicate of a

building and stick it up and then somehow wash the

building down to get how much pollutant was deposited

into that building, so we come up with a surrogate,

something what we hope will mimic the surface, so we use

a variety of different surfaces, and when we do this, we

try to provide surfaces that have properties that won't

artificially enhance the deposition, so there are other

dynamic surfaces. Sometimes we will use a water surface

if we are trying to look at and estimate the amount of

mercury that's deposited into the Great Lakes, for

example. If we're trying to assimilate the deposition

to a specific type of surface, we would put that

material on a small aerodynamic surface and actually

measure how much deposits to that, and then to verify

that we're getting reasonable results, we actually then

use models to estimate how much mercury deposited to

that small surface, and then use an understanding of the

landscape and some of the information that we can get

from a remote sense to tell us about how the landscapes

change and the proportion of different types of trees

and vegetation and plants and so forth to come up with a

larger estimate of the deposition.

Page45

Q. So in doing that, you use multiple

surrogates, I take it, for different types of surfaces?

A. We use more than one type of surrogate

surface, that's correct.

MADAM HEARING OFFICER: Anything else?

Moving on to D, I believe.

DR. KEELER: D: "What meteorological

analysis was performed to demonstrate this statement?"

We used a combination of meteorological analysis tools

to perform the meteorological interpretation for this

project. This included detailed plotting analysis of

surface and upper air meteorological maps, backward air

mass trajectories, NEX-RAD precipitation data sets,

together with statistical techniques, such as cluster

analysis to institute meteorological influences on the

deposition.

MR. RIESER CONTINUES:

Q. If you look at Exhibit 29, which is the

Beijing, China, Landis Report, there are -- there's a

page after the page we looked at before called "Hy-split

Back Trajectories."

A. Yes. That's what I was referring to in my

statement.

Q. So these are two examples of the back

Page46

trajectory?

MADAM HEARING OFFICER: Point of

clarification, the Beijing study is Exhibit 30. 29 is

the preliminary results. That's right. 30 is the

Beijing study.

MR. RIESER CONTINUES:

Q. I had two identical exhibits. Sorry about

that. Exhibit 30. Doctor, you're there. "Hy-split

Back Trajectories."

A. Yes.

Q. So these are examples of the type of work

that you would do?

A. These are examples of one of the tools

that I listed there. These are the -- as I said, the

backward air mass trajectories.

Q. What information do using these as

examples -- what information do these provide to you?

A. If you look at these plots, these plots,

basically, the trajectories are calculated to correspond

to the maximum hour of the precipitation for a couple of

the precipitation results we received in Steubenville,

and these black lines represent the center of

probability of mass that air would have had thought

going backwards from Steubenville and where the air mass

Page47

would have come from moving backwards in time and I

think it's six-hour increments, and that's what those

little stars are.

Q. The stars are six-hour increments?

A. I believe so, yes. That's going back 72

hours. These are three-day back trajectories.

Q. Then there are graphs below these that are

on the side that says "HPA." Would you tell us what

these rep?

A. Yes. The bottom plots HPA is a unit of

pressure, and so this gives us a vertical slice looking

up in the atmosphere of the height that the air mass was

predicted to be at, so again, if you start at

Steubenville, or to the right, looking at the first here

moving backwards in time, it shows that our site was at

the surface, and then the air mass actually came moving

backwards in time, and went up into the atmosphere a

little bit into the higher into the mix layer and then

up as it went backwards and slowly came down again.

Q. Did these examples -- again, do these also

represent the difficulty in identifying the proximity of

regional source because the three days' time travelled

put you pretty far away from Steubenville? Can you be

more specific about the difficulties you had to -- why

Page48

do you select three days? Let me ask that.

A. For the purposes of this demonstration,

three days was selected because, when you start going

back further, you actually have missing data in the data

streams, and you wind up actually having the model

condition calculate very often, so we use three days

because, as part of that the original analysis we did

back in the late 80's looking at some of the utility

acid precipitation data, we found that three days

represented some synoptic meteorological conditions very

well, and that going further -- actually, the

uncertainty in the trajectory calculation was back much

more than three days which was so great to make them

unreliable.

Q. Looking up, again, at the specific

question of 12, "Meteorologic analysis corroborates that

a substantial amount of the mercury deposition found at

the Steubenville site was due to local and regional

sources," would the local and regional sources be

sources that contribute mercury emissions to the

atmosphere or all along the lines of these back

trajectories?

A. The line here, again, is meant to

represent the most probable path, so as you move

Page49

backwards in time, the path that actually the air mass

could have taken actually grows wider, so if you were to

take and draw an increasing, like, wedge of air moving

backwards in time, it actually encompasses a larger area

than what's shown by just this line. This is just a

representation, and so what we do then is to use this as

a guide, and then take the surface and upper air

meteorological maps together with the NEX-RAD to help

this in terms of guiding us in terms of the timing for

how quickly the air mass has moved, where the

precipitation fell along this trajectory, and what

sources could have fallen into this kind of large area

that surrounds this line that you see here, and so you

put that all together and together with the known source

locations to be able to then say, "Here's the sphere

that we think has contributed to the deposition at that

point."

Q. When you -- tell us what "NEX-RAD" is, I'm

sorry.

A. NEX-RAD is the weather radar that you see

presented on the television. It's the nice way to show

the storms for Springfield, show the nice, heavy

thunderstorms that come through in different colors, so

it's next-generation radar is what it stands for.

Page50

Q. So going back to your description of this

you describe sort of a -- since I'm used to ground water

what I will call an inverse plume where it spreads out

from the back and focuses in at the front end. Is that

an apt description?

A. I'm not a ground water person, but plume

is okay.

Q. So back to that question, would it be in

the way you did the study that all of the sources that

you identify within this area, looking at this first map

on the left, which starts at, essentially, the Rio

Grand, all the sources within that area leading up to

Steubenville would be included in your analysis of the

mercury emissions that fall at Steubenville?

A. We will utilize going as far back as the

trajectory would tell us for the three days. We

actually look at the meteorological numbers for the

entire country when we are doing this, but we would look

at and investigate any potential influences that would

occur along that entire path.

Q. So when you talk about local -- we had an

extended discussion yesterday I understand about local

and regional sources and things like that, and you were

-- again, correct me if I'm mistaken -- but you were

Page51

concerned that the local and regional as the proximity

issue was problematic because, in your mind, it really

wasn't meteorologic?

A. That's correct.

Q. The meteorology that's described in this

first plot, the one on the left on this Exhibit 30,

would suggest that the mercury that's being contributed

into Steubenville that you measured came from sources

along this line and then spread out as far as the data

shows that the air mass moved. Is that correct?

A. I believe what you're trying to convey is

correct.

Q. Is it, again, looking at this map, is it

the assumption that you're going back three days because

that's an understanding of what's useful data? Is that

the way to say that? After three days it gets mushy?

A. Yes.

Q. It doesn't really tell you anything. Do

you have any assumption, and I asked this yesterday, but

there is a good graphical representation of it. Do you

have any assumptions that any amount of mercury comes

into this back end that you have got at the Rio Grand in

looking at how much mercury is being deposited in

Steubenville?

Page52

A. Again, let me see if I understand what

you're asking me. This line, or whatever, that goes

back toward the Rio Grand here, we don't add up

emissions along that line or suggest that emissions

are -- that's not included in our analysis. There are

no assumptions in terms of the analysis. What we would

do is we would take and look at the entire upwind

history of this precipitation event. One tool to look

at where the air mass came from would be this line, and

along this line we would look at things, such as did it

rain? What was the precipitation rate moving backwards

i time along this line, and what we found is that, by

using precipitation, you can look to see whether the

probability that the pollutant would be removed is

greater in times when precipitation along these back

trajectories is high or not, so we look at the storms,

and in this case, this one that's for May 8, it turned

out that the air mass actually wasn't moving very fast.

If you look at those stars there, there are a few stars

very close to Steubenville, so the air mass was not

moving very fast. In fact, those first four or five

stars there reflect the meteorological situation at the

time, which basically, had a storm which stalled out in

that area, so the storm was a big storm. Things didn't

Page53

move very quickly and the precipitation would have been

washing out all of the reactive mercury, and all of the

other very reactive pollutants from the atmosphere and

cleaning things up for a very long time, so that the air

masses that could have contributed to the loading that

was occurring through that storm couldn't come from very

far, and so this is how we do that type of analysis, and

one can take and model that specific storm and compare

that against using observations because there are no

assumptions in the observations. It's meteorological

data that we have and we don't look at a line underneath

the map. We actually look at the whole area, and we,

basically, say, from this, we know that we had a very

slow-moving air mass that took a long time for it to

actually go anywhere, so we are talking about emissions

that were in the vicinity of the Steubenville area. The

ones that were coming from Southern Ohio that were --

the emissions sources were, basically, emitting up into

that area, and then that storm was staying there, and it

was precipitating out for an extended period of time, so

that the air mass, by the time it got up towards,

Steubenville had already been washed out. The air that

was in that storm had already had a great deal of

removal and cleansing, just to put it in simple terms,

Page54

so the highest probability of where the air masses that

fed the clouds could have come from had to have been in

that general vicinity. Again, I'm saying Southern Ohio,

Northern Kentucky, that general area, based on the

service meteorological data our on site, data as well as

using this as a tool. When you start talking about

meteorological analysis and showing maps and so forth,

everyone turns off and goes to sleep in your talk, but

when you show a trajectory plot, people get the idea

that the air started down somewhere in Texas and that's

the path that it took before it got to Ohio, so it's

simplification for presentation purposes, but by no

means is it what we use as the definitive answer because

there's very complex three-dimensional flows that occur

in the atmosphere, and one has to take those into

account, as well, when one is looking at what potential

source areas could have contributed.

Q. So that description is what tells you that

the mercury deposited in Steubenville came from the

specific area you described, Southern Ohio, Kentucky,

and not from, say, Houston, which is more or less passed

-- is that correct?

A. Yes, in simple terms.

Q. In doing that analysis, do you have to

Page55

take into account the atmospheric chemistry in

transformations of mercury in the atmosphere?

A. We do rely on our understanding of the

reactivity of mercury and the chemistry of mercury in

order to be able to look at what distance scales we

think things could take place on. However, there are no

assumptions made. We use what our best available

understanding is based on, both, our observation data

that we have taken, both, at Steubenville and at our

Michigan sites, together with what is published in the

literature.

Q. And is that understanding -- you use the

term "understanding." Is that correct?

A. I believe that's what I said, yes.

Q. Is that understanding different than the

atmospheric transformation of mercury which are

described in the deterministic models?

A. That would depend upon the specific

deterministic model. This gets to the whole question

that I raised yesterday about the uncertainties in the

models.

Q. Let's use CMAQ to narrow it down?

A. CMAQ -- the version of CMAQ that EPA uses

has a very simplistic atmospheric chemistry for mercury.

Page56

There are lots of issues in that chemistry. It does not

include all the relevant reactions that we understand

today. It doesn't include the most up-to-date reaction

rates, so I would say no. We rely upon more than that.

Q. And have you published or is there another

paper published paper that you rely on that identifies

the atmospheric transformation reactions that you rely

on in doing your work?

A. We have papers. We have a paper that we

have submitted to -- see, we have a paper that we

submitted to a peer-reviewed scientific journal that

details some of our deterministic modeling, which is not

at all what was done here, but that's not really

relevant to the situation, but it incorporates some of

the mercury chemistry that we rely on. I mean, some of

these chemical reactions and updating the atmospheric

chemistry happened very rapidly. I was just in Russia

at a mercury meeting a long-range transport mercury

meeting in Moscow and learned of some new reactions that

had just come out, so we rely upon, both, what our

colleagues tell us at meetings, as well as what the

peer-reviewed literature has. I don't write down, and I

don't have, like, a publication where I have listed all

the atmospheric chemistry that I rely on. I rely on

Page57

what's in the peer-reviewed literature.

Q. I guess what the bottom line of what I'm

trying to get at is what Sheldon would say is, is there

a way that we can see how you describe the atmospheric

transformation reactions? I assume there is some type

of mathematical equation or part of your model or

something that another person can look at and apply in a

similar setting and test.

A. We don't use chemical reactions in our

modeling. I stated yesterday that the receptor modeling

does not use chemical reactions. We only used observed

information.

Q. Then I completely misunderstood because I

thought you said that, in evaluating where mercury came

from, you made certain judgments, decisions about the

atmospheric chemistry that mercury undergoes.

A. What we use as our understanding that

reactive mercury is removed very easily with

precipitation, which we have published and have

observations that show this, and that elemental mercury

is not removed as rapid, so again, this is based on

observation of published work, and in understanding the

information, we have an idea of how long reactive

mercury would last in the atmosphere based on almost

Page58

instantaneous removal in precipitating systems. That's

the atmosphere chemistry that I'm referring to is that

there are different chemical properties of mercury forms

that may get removed at different time scales.

Q. Do you have a numeric rate that you use to

evaluate the removal of reactive gaseous mercury?

A. No, sir.

Q. When you say it's rapidly removed, what's

the quantification of rapidly?

A. Well, we have hourly data, and you will

see within an hour time frame a very rapid, so within an

hour time frame, we see a significant -- more than 50

percent -- removal of the mercury within one hour of the

onset of precipitation.

Q. When you talking about that data, what

you're talking about is the data at your wet deposition

sampling locations, correct?

A. That's correct.

Q. So the storm begins at five o'clock. At

5:15, you have a sample that's got a lot of mercury in

it. At 6:15, there's no mercury, just --

A. The reactive mercury would have dropped,

that's right.

Q. And what does that -- but is it correct

Page59

that you use that data to make decisions as to how much

reactive gaseous mercury is in the air mass that's

moving into Steubenville?

A. No. There's no assumption there.

Q. So how do you know how much mercury is in

the air mass that's moving into Steubenville or where

it's from?

A. We have on-site measurements of the

reactive mercury at the site. There's no assumption.

That is my point.

Q. What does it tell you -- what does that

data tell you about -- well, step back. That's why

you're not in position to identify the proximity of the

sources based on the samples that you do at

Steubenville, correct?

A. We have not done a detailed source

apportionment of the ambient mercury at Steubenville,

yet. What we have done at this point is look to see

specific episodes when we have high reactive and

particulate mercury to see what the meteorological

conditions were like and where the air was coming from

based on on-site meteorological measurements, as well as

other meteorological data, such as the trajectories, to

see if there is a strong association, so whether we see

Page60

repeated pattern of things like high sulfur dioxide,

high reactive mercury, high nitrous oxide, high CO,

whatever the on-site measurements we have, high

particulates, and then we analyze the particulates for

there elemental composition, as well, so we can look for

the same tracers in the air as we see in the

precipitation, and that analysis has not been completed,

yet, because we collect four samples a day. That would

collect them for -- well, we're on our third year now,

and it's thousands of samples that we are in the process

of analyzing, so that analysis and information will

come. We do see strong relationships that winds that

flow from the south, southwest with higher mercury

levels and higher S02 (phonetic) concentrations that are

indicative of transport from fossil-fuel-burning plants

from coal-burning plants.

Q. What tells you that those are fossil

burning plants, again, looking at this first map as an

example in Kentucky, Ohio River Valley and not Houston?

A. Well, we have an idea of the location of

where the fossil-fuel-burning facilities are, where all

the sources are, and then we can look along that swath

that we think where the air could have possibly come

from, and we look at the inventory to tell us what the

Page61

possible sources could have been. It's an aid, not as a

quantitative tool.

Q. Then you have to make -- then you have to

evaluate whether precipitation events have washed out

mercury that could come from other sources along that

line. Is that correct?

A. That's right. It's something I have been

working on for more than 20 years, so in my best expert

judgment, I have done this for multiple pollutants,

including mercury, and I feel like I have a very good

handle on this situation, and on this type of analysis.

Q. If there was an example where there was an

air mass moving into Steubenville, and there was no rain

fall events going on along the line back to Texas, what

would the results of Steubenville look like?

A. You know, I would have to go and look to

see if we had those type of events in order for me to be

able to predict because you can draw the same line

moving from Steubenville to Texas, and depending upon

the winds, how strong the winds were, and at all those

segments, what the atmosphere was like, the stability of

the atmosphere along the trajectory, the temperature

profile, all these different meteorological parameters

would affect disbursion, the chemistry, the transport,

Page62

and the dry deposition that occurred along that

trajectory. One of the things that, if one was to take

and estimate how much mercury you would expect left in

an air mass that was submitted to Texas, by the time it

got to Steubenville, one can take and estimate how much

you would have expected to have loss from dry

deposition, how much you might expected to have lost

from chemical reactions, and just by the air mass being

deluded as you got there and when you get that far back

in time, there's very small contribution that you would

see in Steubenville from sources way back in Texas, but

that's a general comment. I would have to look at

specific information to give you a quantitative answer.

Q. In doing that estimate that you just

described, one of the factors would be the atmosphere

transformation that mercury undergoes --

A. Yes.

Q. -- during that time of travel?

A. Yes, sir.

MR. HARLEY CONTINUES:

Q. The May 8 event that you described on your

hy-split trajectory, is that one of the days during that

year where there was an event that precipitation led to

a significant deposition of mercury in Steubenville?

Page63

A. Yeah. That was one of the four largest

events depositing. Again, I think I've given a figure

of maybe the four largest or the several largest, as

much as 8 percent of the annual total, so these are

significant events that occurred at this site.

Q. In the two-year period of 2003, 2004, how

many total mercury rains have there been on Steubenville

comparable to the event on May 8 where we have

hyperloading of mercury on Steubenville?

A. Gee, I don't have that in front of me. My

memory is that we had in the two-year period about eight

events that were greater than .6 micrograms per meter

squared I believe, which that's a big event. And then a

couple that were greater than one, and again, if you are

looking at 10 to 20 micrograms per square meter at a

year at a site, if you get more than one in one rain

event, that's a significant contribution in that one

day.

Q. There was one other question I had. The

CMAQ model, you gave some reasons why the CMAQ model is

it not as precise as the receptor-based approach that

you used in Steubenville. One question I had about the

CMAQ model is you said the CMAQ model underestimated

mercury loading in Steubenville by comparison to the

Page64

results of the work that you have done. Is that

correct?

A. Yeah. Just to qualify, the CMAQ for 2001

is where they calculated that 43 percent was coming from

coal-fired utility boilers. Our data is from 2003,

2004, we did have the opportunity to compare our

Michigan network sites for the 2001 year, together with

our estimated for Vermont, and the CMAQ model

underestimated the deposition that we measured at our

Michigan sites and at the Vermont site for 2001. It was

off by a factor of two at one site, as I recall, so it

grossly underestimated the total deposition at our sites

that we had measuring in 2001. The way you phrased it,

"CMAQ being less precise" I think the way I would rather

phrase it is that these type of models are more

uncertain, that there are more parameters and processes

that are poorly defined in these models, which make

their uncertainties greater and the conclusions drawn

from those models much greater than I would say that

they are from the receptor modeling. That's a correct

statement.

Q. Is the CMAQ model a model which accounts

for, both, wet and dry deposition of mercury?

A. Yes, it does.

Page65

Q. And is it fair to characterize your

testimony as indicating that, in the Steubenville

situation, your modeling of wet deposition alone was

greater than the CMAQ model of, both, wet and dry

deposition at that site?

A. My memory tells me that CMAQ actually

estimated a greater dry deposition at Steubenville than

it did for wet deposition for the 2001 year, so no, our

wet deposition wasn't greater than the CMAQ's total. I

believe the CMAQ total was around 30 micrograms per

square meter, or 34. It's hard to tell from the

modeling because they do it in colors with ranges, but

that, again, I'm just going from my memory of what Russ

Bullock presented, but I believe that that's true, that

the dry deposition was greater in this area than the wet

deposition was. Using the CMAQ model, not from our

observations.

MR. BONEBRAKE CONTINUES:

Q. On the May 8 back trajectory, a line, is

that, essentially, indicative of a wind direction from

the southwest?

A. Yeah. The line that starts at

Steubenville or the line that starts at the site there

on the border of Pennsylvania and Ohio, if you follow

Page66

that backwards towards the bottom of the left page that

is trying to illustrate the most likely path that an air

mass took before it got to Steubenville.

Q. In this particular case, through Texas and

Louisiana, the air mass was moving from the southwest

and northeast?

A. That's correct, southwest.

Q. Yesterday we talked a little about wind

directions in Illinois. My understanding from your

testimony was that, in the winter season in Illinois,

the wind direction is frequently from the northwest. Is

that correct?

A. Yeah. The Great Lakes region, as a whole,

has a higher frequency of winds from the north,

northwest during the wintertime as a result of synoptic

meteorological conditions.

Q. And during the summer months, are winds in

the state of Illinois most typically from the southwest?

A. From the data I looked at for O'Hare, it

appears as if the south is the dominant winds with

south, southwest being one of the more frequent wind

directions, but knowing that westerly winds are also the

dominant wind areas during the summer.

Q. Would O'Hare's -- would the direction of

Page67

wind at O'Hare be impacted by Lake Michigan?

A. Yes.

Q. So would the data from O'Hare, in your

view, be representative of the rest of the state of

Illinois?

A. No. If you look at wind measurements made

in Southern Illinois, you see a slight variation in that

overall pattern. You might have slightly more winds

from the south than you would at O'Hare, so the winds

will change slightly. The dominant wind direction still

is going to be west with a southerly component being

strongest.

Q. With respect to the northwest portion of

the state of Illinois, would you expect, then, that

during the winter months the predominant wind direction

would be from the northwest and then the summer months

it would be from the southwest?

A. South, southwest. From what I have seen

for -- and again, I haven't looked at a climatological

average, but that's about right.

Q. What would you view, Dr. Keeler, to be the

states which contain upwind sources of mercury with

respect to the state of Illinois?

A. Again, taking what you said in terms of

Page68

the dominant wind directions, you would have the states

of Missouri, Iowa, and farther to the west would be

upwind of Illinois, and then the states to the south,

Kentucky, Tennessee, Louisiana, Alabama.

Q. Would Texas also be in that list?

A. Texas would be in that list, as well.

Q. One other question for you, Dr. Keeler.

You mentioned in response to a question from Mr. Harley

that you had done, as I understand it, a comparison of

CMAQ predictions to Michigan deposition data that you

had available to you. Is that correct?

A. Yes.

Q. Is that comparison in a published article?

A. No, it's not.

Q. Is that comparison publicly available?

A. You know, is it in Dr. Landis' briefing?

MADAM HEARING OFFICER: Exhibit 30,

Beijing Exhibit 30, yes.

DR. KEELER: It says, "Comparison of

CMAQ Model Versus" --

Q. Back toward the end of the document.

MADAM HEARING OFFICER: Yeah,

two-thirds back he said. "Comparison of CMAQ Model

versus UMAQL Measured Mercury Wet Deposition Testimony."

Page69

MR. BONEBRAKE CONTINUES:

Q. Is that correct, Dr. Keeler, that's the

page you're on?

A. Yes, and refers to the left column to

Dexter, Michigan.

Q. Helston, Michigan?

A. That's correct.

Q. Both, the CMAQ predictions and the

deposition data that you were using in the comparison

were both from 2001?

A. Yes, sir. These numbers were provided to

Dr. Landis I believe by Russ Bullock of U.S. EPA, and

these are the actual data that went into the CAMR

modeling.

Q. Is there a discussion of this comparison

in the study that's been at issue today in terms of

what --

A. No. This is an additional analysis. The

Michigan site data is not included in the Ohio paper.

Q. Other than what's in this particular page,

has the comparison or the bases of the comparison

otherwise been made publicly available?

A. All of the Michigan data for 2001 is in

peer-reviewed publications and the CMAQ model results

Page70

are in the public docket that was filed by IEPA.

Q. So --

A. So it's all public.

Q. From your perspective, then, someone could

take the publicly-available information in those

documents and repeat the comparison?

A. Yes, sir.

MR. ZABEL CONTINUES:

Q. Earlier I think you said, in doing the

corroboration using a Michigan inventory, you looked at

sources east of the Mississippi. Is that correct?.

A. I said that was the area that we focused

on, yes.

Q. You didn't use, for a particular

corroboration, also sources east of the Mississippi?

A. Again, for trying to speak generally to be

inclusive of all the work that we did for a particular

event, we would use our understanding of the

meteorologic and transport for a particular period, such

as the May 8 period, where we would focus on those

sources, and sources that were to the southwest, for

example, in the most highest probable transport area, so

we wouldn't be then taking out the list of sources from

Minnesota and Wisconsin at that time to look at -- in

Page71

order to think about the May 8 event. On another day

where the flow came from those directions, we would look

at another set of sources that would help us interpret

that specific day. We didn't limit our analysis to a

number of sources. We had the entire -- plotted all of

the sources plotted in 1999 EPA database together with a

list of all those sources and the estimated fraction of

mercury emitted from each source in each fraction as a

tool to guide us in terms of our interpretation.

Q. That was sort of the point of my question,

and the May 8 is a good example. There's some subset of

sources east of the Mississippi you would have used in

looking at emission inventories for them?

A. Again, we would have focused on the

sources in the states which we thought had a probability

of contributing on a specific day for a specific event.

Q. Just to make this perfectly clear to me,

there are coal-fired plants in Tampa Bay in Florida.

They would not have been included, would they, in your

May 8 analysis?

A. I would say that I did not look at Tampa

Bay power plants on interpretation of the May 8.

Q. On that May 8 one, both, Louisiana and

Texas I believe are west of the Mississippi, were they

Page72

not?

A. That would be a geographically-correct

answer.

Q. And in fact, Minnesota -- for instance, if

you're looking in the other direction, although I

understand Minnesota has a dispute on which side of the

Mississippi they are in, at least, part of it is west of

the Mississippi, is it not?

A. Yes, it is. In my use of east of the

Mississippi and west of the Mississippi, I was trying to

give a general division line, but as I mentioned

earlier, I was not trying to distinguish an absolute

line of demarcation for any of the analysis.

Q. Well, then in the May 8 case, would you

have looked at cases in Louisiana and Texas?

A. We would have looked at sources along that

trajectory, yes, or in that area that would be indicated

by that air mass trajectory, yes.

Q. In answer to one of the questions I

understand that cone gets wider the far away from

Steubenville you get?

A. That's right. So sources all the way up

from Kansas to Atlanta, Atlanta area, Georgia would also

have been considered in looking in the backward upwind

Page73

path.

Q. Most of Texas, I take it, at that point?

A. Yes.

Q. Thank you.

MADAM HEARING OFFICER: Anything

further? I believe 12-E --

MR. RIESER CONTINUES:

Q. This will actually address some of the

questions, but since Mr. Harley brought up the

comparison of CMAQ and the receptor modeling, it's

correct, isn't it, that CMAQ in your type of receptors

study are designed to achieve two different results. Is

that correct? It's designed for two different purposes.

Is that correct?

A. How a model is used is defined by the user

and so in my case, which I can attest to, we were using

receptor models to calculate how much mercury was

deposited via wet deposition from the major source

categories for the period of 2003-2004. So that

apportionment to source categories was our objective,

and that's where we did the modeling. Why EPA did the

CMAQ modeling for 2001? As I understand, it was to

estimate the contributions from all the major sources to

the deposition of mercury across the entire United

Page74

States, so by definition, it has a broader purpose, and

was not looking at one site. It was not looking at just

wet deposition. It looked at wet, dry, and ambient, so

yes, they have -- they have different purposes, but to

say CMAQ model has one purpose and the type modeling we

did had only one purpose, it really depends on what the

user was using that model for.

Q. The CMAQ model, among its utilities, is

that it can be used for predictions. Isn't that

correct?

A. Yes, sir.

Q. So you can use it to decide what would

happen, as an example reflected in Exhibit 30, "Utility

zero out," of what mercury deposition would look like if

there were no utility emissions?

A. Yeah. I believe, like you said, in

Exhibit 30, I believe that map included a couple special

plots. Maybe I'm wrong.

Q. You're correct. There is a CMAQ simulated

total mercury deposition for 2001 and underneath it

says, "Utility zero out," and prior to that, there's a

base case CMAQ simulated total mercury for 2001 base

case, so as we said, you can use CMAQ to, as an example,

take out all of the utility emissions and see what

Page75

things look like?

A. Correct.

Q. Can you use your model to do that?

A. The receptor model, by definition, does

not have a predictive capability.

Q. So if I -- well, let me ask you, if you

assumed that the coal-fired power plants, within a 50

kilometer range of Steubenville, ceased to operate,

would you have any conclusion as to what that would do

to the mercury deposition in Steubenville?

A. Based on our work, what I would say is

that if the coal-fired utilities, regardless of

location, stopped emitting, so they were zeroed out, we

would see close to a 70 percent reduction in the mercury

deposition that we measured at Steubenville.

Q. Could that conclusion, in making that

statement, what you were referring to are not the

coal-fired utilities within a 50 kilometer distance from

Steubenville, but all coal-fired utilities in the United

States?

A. We -- again, based on our analysis, our

meteorological analysis, we will consider sources that

were beyond regional, so very long range sources, so

ones that are the western part of the United States

Page76

would contribute very little to that total and part of

that is backed up by just looking at the deposition data

that's in the Western United States, which is very low.

It's for around four micrograms per square meter, so I

would say that it's the sources in the eastern United

States coal-fired utilities in the Eastern United States

which are contributing that 70 percent, and if those

were reduced, we would see that commensurate drop in the

mercury deposition at that site for those years that we

modelled.

Q. But you wouldn't be able to use your

receptor model to determine what would happen at

Steubenville if, for example, all the power plants in

the state of Ohio reduced mercury emissions by 70

percent?

A. No. The model is not capable of

predicting calculations.

MADAM HEARING OFFICER: Are we ready,

then -- I believe we have answered 12-E and I think

perhaps 12-F already. In the discussion we had this

morning, E is, "Does the analysis differentiate between

sources located at different distances?" and we have had

substantial discussion about --

MR. RIESER: Correct, yes.

Page77

MADAM HEARING OFFICER: F is, "Have

you quantified the substantial amount as used in this

statement?"

MR. RIESER: I was looking at the

wrong number, yes, correct.

MADAM HEARING OFFICER: And we

answered 13-A and B yesterday, so 13-C.

DR. KEELER: I actually believe I

answered this one as well. "Are they fired with

bituminous or sub-bituminous coal?" Again, in the

complete listing of the emissions sources and the EPA

inventory, it lists the type of coal burned. I believe

there is a mix of plants that burn, both, bituminous and

sub-bituminous. Some are all sub-bituminous and some

are -- a few that are bituminous, so there's a

combination. D: "Does the chemical profile of

bituminous as opposed to sub-bituminous affect the

amount and type of mercury emitted by a coal-burning

generating unit?" Again, I believe I answered this

yesterday in saying that I'm not an expert on control,

and so perhaps, an expert for the State will provide --

later will talk about how various control will affect

the emissions, but clearly, the type of coal burned will

have an influence on the type of type and form of

Page78

mercury, in respect to whether it's reactive mercury or

gaseous mercury or particulate mercury and we talked

about the importance of chlorine and fly ash content and

iron and others things in the coal that would cause

those differences. E: "Would you expect a different

result at Steubenville if the surrounding units burned

sub-bituminous coal?" I don't have the ability to

provide an answer to that question. I don't understand

exactly what you are asking. I assume that the power

plants in the area are not -- from an inventory, some do

burn sub-bituminous coal, so I'm not sure if you are

asking me if all the plants in the Steubenville area

burn sub-bituminous but I see --

Q. I am asking, if all the power plants

surrounding Steubenville burn sub-bituminous coal, would

you expect a different result?

A. Again, I guess I'm not prepared to answer

that question. I think I would have to have more

understanding of how changing to a different coal type

for the various types of controls that are used in that

area would affect speciation and I'm not prepared to

answer that.

MADAM HEARING OFFICER: And I believe

you have answered 13-F, as well, whether airborne

Page79

sources of mercury are located within 50 miles.

DR. KEELER: So G: "In what way, if

any, are the conditions at Steubenville analogous to the

conditions in Illinois?"

MR. RIESER: To be honest, there was

some movement in that answer, so if we get too

repetitive, cut us off, but I would like to go through

this because I heard two different things, to be honest.

MR. KIM: Could you clarify what you

mean because Dr. Keeler did describe the distinctions in

terms of meteorological conditions and so forth and I

recall him going into something about topography. Were

you wanting something beyond that, something different?

MR. RIESER: What we're looking for is

whether the findings with respect to Steubenville are

analogous to Illinois, and some of that discussion has

to do with things that were discussed and things weren't

discussed. I heard, on the one side, Steubenville was

typical, but in some ways, it's obviously atypical due

to the amount of power plants that are around, so I

think we need to walk through each of the features and

talk about them.

MR. KIM: There again, I'm just for my

sake just trying to get it clear. You said, "What we

Page80

want to know is whether the finding at Steubenville are

analogous to Illinois." When you say the findings at

Steubenville are analogous, you have to be comparing

that to some findings in Illinois, I assume, so are you

referring to something specific, or are you saying can

the findings in Steubenville be transferred to Illinois?

MR. RIESER: I'm assuming this

testimony he has presented had some meaning to the Board

as far as what the study means and the decisions they

should make with respect to Illinois, so I do think we

have to decide how applicable these findings are to

Illinois.

MR. KIM: I agree. The only reason

I'm saying that is he's already provided testimony that

he believed the findings could be applied, not only to

Illinois, but to other locales, once you take individual

variances into account. With this question on its own

is read, it seems like you are referring to something

specific, some kind of fact-specific situation or --

MR. RIESER: I'm happy to break it

down if the question, itself, is not clear. I'm happy

to break it down and go through each of the issues that

I had in mind and maybe that will take care.

MR. KIM: Let's do that.

Page81

MR. AYRES: I thought we had a

discussion that went for, at least, a half an hour.

MR. KIM: I think, if you break it

down, my guess is some of that stuff has already been

answered, but if you want to break it down, that's fine.

MR. AYRES: There was a discussion

that was very fact specific, and then we talked about

the transferability of the learning, if you will, from

Steubenville. Do you recall that?

MR. RIESER: I recall that and my

problem is my recollection is -- my recollection is that

the answer wasn't clear. There are, obviously, some

informational issues that are portable and Steubenville

was a very specific place that has the conditions that

are represented because of conditions that are specific

to Steubenville.

MR. KIM: Sure.

MR. RIESER: Seriously, I think we

could have got through this pretty quickly.

MR. KIM: I wanted to be clear.

MR. RIESER CONTINUES:

Q. I honestly don't mean to belabor the

point, but I just don't think it was clear on the

record. Let me ask, Steubenville is in a river valley.

Page82

Is that correct?

A. The city of Steubenville has part of its

boundaries that fall in the Ohio River Valley, that's

correct.

Q. Was the sampling location in the part of

Steubenville that was in the River Valley?

A. You get down to, I think, definitions. I

think, technically, that whole area is in the Ohio River

Valley, but just to be clear, in terms of topography,

the site was not down on the river in the valley. It

was actually up on top of the area that then flattens

out going to the west in Ohio, so it was not down in the

valley where local sources would have inundated the data

that we were getting. It was actually up on top, and I

think, like, 400 feet above the river, to be clear.

Q. Is the meteorology of the Ohio River

Valley, as you have described it, different than the

meteorology one would typically see in Springfield,

Illinois?

MR. KIM: I think this has been

answered. I think he went through an extended

discussion about meteorological distinctions between

Steubenville and other parts of any other part of the

country. I think his specific testimony was you are not

Page83

going to find that replicated anywhere, but

Steubenville.

MR. RIESER CONTINUES:

Q. Is that your testimony?

A. I think what I said yesterday was that the

conditions at Steubenville, in terms of the overall

source-receptor relationships, which meteorology is a

part of, are specific, to a point, and that those would

be unique to Steubenville. The question is whether the

conditions in Steubenville make it anomalous, or somehow

unique where the results would not be transferable to

somewhere else in Ohio, somewhere else in the Great

Lakes or in Illinois and I would suggest that the

conditions are not unique or anomalous to make them so

they are not usable or transferable to conditions that

we would have in Illinois.

Q. What are the conditions that are not

anomalous?

A. The weather in Ohio, just as it is in

Indiana, Illinois, Michigan, Minnesota, Wisconsin, are

all controlled by synoptic meteorology, which is the

large scale movement of the highest and low pressure

systems across the Great Lakes, so unlike some places

where they have some dominant feature, such as Bermuda

Page84

high, which may dominate the weather in a certain

location, that might make it different than another one.

The Great Lakes, themselves, are dominated by synoptic

meteorological transport, and therefore, as long as

you're taking into account the specific meteorology that

occurs at that spot, it's not going to be all that much

different. The controlling factors are not that much

different for Illinois than Ohio.

Q. Do the presence of a large number of power

plants surrounding Steubenville, does that factor make

it anomalous to other locations?

A. Again, I think I said this before, but I

believe that the high density of power plants in the

Steubenville area is a prime reason why power plants are

a prime contributor to the extremely high deposition

that we see there. However, in another place which

would have a similar density, I believe you would see a