13 inordinate result or an excessive result

13 inordinate result or an excessive result. Isn't

14 there some way to smooth or otherwise handle data

15 that might appear to be approaching outlier

16 status, short of just rejecting the data entirely?

17 A. In terms of what kind of statistical

18 analysis?

19 Q. Well, the statistical analysis that you

20 have done and that you've been tasked with doing,

21 not only for the cattail coverage, but for the

22 flow data that you're going to be looking at.

23 A. If I understand your question -- let me

24 rephrase it the way I understand it. If you're

25 trying to form certain statistical procedure, in

JACK BESONER & ASSOCIATES, INC.

184

1 this case, let's say regression analysis, you can

2 screen your data to determine if you have,

3 "outliers". You could perform some sort of

4 procedure such as Appendix "E" to screen your data

5 first. You could fit a regression model and then

6 do a test to determine whether any of the points

7 are an outlier, assuming that your regression

8 model is the correct model.

9 Now, that test is a parametric test.

10 So then your question is are there other ways to

11 identify what is an outlier or other ways to

12 somehow incorporate the information from an

13 observation without discarding it totally?

14 Q. The latter.

15 A. Well, there is, for example, the

16 procedure of instead of doing a regression where

17 you use least squares as the minimization

18 criteria, you could do some sort of robust

19 regression like least median squares procedure.

20 Q. Okay. Did you try any of those

21 mechanisms in your analysis, rather than simply

22 rejecting the four or five observations?

23 A. What I did -- I think I said this

24 earlier when I was describing the table of "P"

25 values that I produced. So now I'm talking about

JACK BESONER & ASSOCIATES, INC.

185

1 the analysis that I did where I took the entry and

2 access period data in Loxahatchee, combined that

3 with the data in '78 to '83 in Appendix "E", and

4 then tried to determine if there was really a

5 difference in the average phosphorus concentration

6 after you adjust your stage.

7 So basically comparing the intercept of

8 those two lines, one way to do that is

9 parametrically. Another way to do that is what I

10 -- one thing that I did was you can try and

11 account for stage by still doing least squares

12 fit, but then instead of just assuming that you

13 have normal distributed areas instead of assuming

14 that. You could, instead, do an analysis where

15 you're looking at residuals and then look at the

16 correlation between those residuals using a

17 nonparametrical approach.

18 Q. Have you received flow data yet for S-7

19 and S-8?

20 A. The flow data that I received back in

21 the period -- it must have been between March and

22 June -- that flow data was in a PK zip file, and I

23 don't know if that only included data for S5-A,

24 S-6 and the S-10 structures. Those are the ones

25 that I recall having. It may have included data

JACK BESONER & ASSOCIATES, INC.

186

1 for other structures.

2 Q. Okay. Is that the PK zip file that's

3 on a disk that we just received recently?

4 A. Yes.

5 Q. Okay. You haven't used that data?

6 A. No. No, I haven't. The only flow data

7 that I used was when I started to an analysis

8 trying to relate load coming into S5-A and S-2

9 with the Loxahatchee concentrations.

10 Q. Let me show you a document. I'm not

11 going to mark it yet. The first page is bate

12 stamped 12010119, dated 12/8/93. See if you

13 recognize that.

14 A. Yes, this is the daily load data that I

15 received from Environmental Services & Permitting

16 that covers the period 11/92 through the end of

17 8/93. And John Davis gave me this spread sheet

18 the actual data that he gave me on floppy disks

19 only included two columns that are included are

20 S5-A plus S5-AW and S-6.

21 Q. So at the end of '90, 12/23/93 --

22 12/8/93 apparently, based on the data in the

23 corner, was this an update at S-5 and S flow data

24 you had gotten prior to our last session?

25 A. No, this is load, not flow.

JACK BESONER & ASSOCIATES, INC.

187

1 Q. I'm sorry. Load. So he had conducted

2 the calculation based on the district's algorithm?

3 A. My understanding is that he used the

4 district algorithm.

5 Q. Okay.

6 A. Or it may be at that point maybe they

7 were taking the same kind of sample everyday by

8 then.

9 Q. Do you know if these are based on

10 district samples or are these based on samples

11 done by ES&P? I mean, who's data base is it?

12 A. I believe it comes from the district.

13 Q. Here's 1210164. See if you can tell me

14 what that is.

15 A. This is a printout of some analyses

16 that I ran where I -- these are the analyses that

17 I ran to produce the table of "P" values that I'm

18 talking about. So just to go through this, the

19 first -- the first part of the output on the top

20 of the page where it says district model, that's

21 the model that you get when you plug in the data

22 from Appendix "E" into this regression.

23 And, so, you can see under the column

24 labelled value, those are the coefficients that --

25 more or less what's in Appendix "E". And then I

JACK BESONER & ASSOCIATES, INC.

188

1 went through and used the same model of average

2 log TP versus stage using only the new data --

3 using the new data without 4/93, using the new

4 data, it says use the new data without outliers.

5 This may be an earlier version where I

6 ran an outliers analysis on the new data before I

7 talked with John Davis and got in the results of

8 him going back and looking at the log back to

9 specifically tell me which points are

10 questionable. So this is the last part at the

11 bottom of the page where it says model based only

12 on new data without outliers.

13 The criteria that I used to call

14 something an outlier was I did an analysis similar

15 to what's in Appendix "E", where I fit a separate

16 affect for stages and date on the field

17 replicates, and then just by that statistical

18 analysis threw out the values that were indicated

19 to be "outliers".

20 Q. What was the purpose in producing this

21 document?

22 A. The purpose was to determine how the

23 entry and access data compares to the district

24 model.

25 Q. And what kind of a fit did you get?

JACK BESONER & ASSOCIATES, INC.

189

1 A. The question is too vague.

2 Q. You ran your data sets as described for

3 each model, but the underlying model you were

4 employing with those data sets was the district

5 model from Appendix "E".

6 A. Right, the assumption that you use

7 average data station as a predictor variable.

8 Q. And after you did that for each of the

9 data sets as identified, some without outliers,

10 some solely on the 12/92 to 11/93 data, is that

11 without 4/39 -- the April '93 data, some without

12 outliers, without 4/93-- you identified a

13 different measure of fit for each of these data

14 systems.

15 A. Well, you ought to look at how the

16 coefficients differed.

17 Q. And how did they differ? What kind of

18 range did you see?

19 A. Well, if I recall -- if you're not

20 going to let me look at that -- the slope of the

21 -- just using the district's -- I'm sorry. Just

22 using the entry and access period data and ESP

23 access data, as I recall the slope -- sorry. The

24 intercept was lower than the district -- I can't

25 recall the slope. I believe it was slightly

JACK BESONER & ASSOCIATES, INC.

190

1 different.

2 But, in fact, when you end up combining

3 the data, as I said before, if you leave out April

4 '93, you don't get a significantly different slope

5 between the two time periods, and you possibly

6 marginally get a different intercept.

7 Q. Did you form any opinions based on

8 utilizing these different data sets with the

9 district model -- any final opinions as a result

10 of this comparison regarding the validity of the

11 Appendix "E" model?

12 A. Nothing that I haven't already stated

13 earlier, and that is there's at least two issues

14 that are of concern, one is the April '93 results,

15 and the second is the Department of Justice data.

16 Q. When you run the Department of Justice

17 data, what result obtains then in comparison to

18 the Appendix "E"?

19 A. As I stated earlier, if you take the

20 Department of Justice data and regression, the

21 fourteen station average of log TP versus average

22 stage, you do not get a significant relationship.

23 The observed correlation is positive, but it's not

24 significantly different from zero.

25 Q. Is that reflected here as well?

JACK BESONER & ASSOCIATES, INC.

191

1 A. I'm not sure whether -- it would be on

2 probably the last page -- if I had gotten to that

3 point yet.

4 Q. Why don't you take a look and see if

5 you can find it.

6 A. Okay. No, the Department of Justice

7 data isn't in here. I just started to work on

8 that.

9 Q. Do you have an equivalent document that

10 reflects the results of the work you did on the

11 DOJ data?

12 A. I have one on my computer now.

13 Q. When you provided the total of five

14 data disks to us as part of the current round of

15 trying to complete your deposition, did you

16 include that file? I assume it's under a file of

17 some sort.

18 A. I'm sorry. You're talking about which

19 floppy disks?

20 Q. Well, you've provided the five floppy

21 disks, two appearing to be data for Loxahatchee

22 and data for WCA-2, which I think you've already

23 described as data sets that you got from John

24 Davis that were modified and adjusted.

25 A. Right.

JACK BESONER & ASSOCIATES, INC.

192

1 Q. You give us three additional disks.

2 One is a PK zip. In fact, they all have zipped

3 data on them I believe.

4 A. I'm not -- you probably don't have the

5 updated version of that output because I just

6 produced that a couple of days ago.

7 Q. Did you print out that updated

8 version? Do you have that with you?

9 A. I don't mean right now at this very

10 moment, but here in Miami. It's on my lap top

11 here in the hotel if I can figure it out how to

12 hook it up to a computer to print it out --

13 MR. FITZGERALD: I think that is fairly

14 within the scope of the Notice of Deposition and

15 should be produced while the witness is still

16 here.

17 MR. BLANK: Off the record.

18 BY MR. FITZGERALD:

19 Q. Is that work complete?

20 MR. BLANK: The witness previously

21 testified that he had wanted to look at the

22 replicates.

23 THE WITNESS: Yes, I don't have the DOJ

24 -- as I understand it was -- and I don't know if I

25 said this earlier or not, but yes, the DOJ data,

JACK BESONER & ASSOCIATES, INC.

193

1 as far as I understand, is actually the result of

2 averages of field replicates. And, so, I would

3 actually need to go back and look at the original

4 data so I could get a feel for the kind of field

5 replicate variability and the DOJ data as well.

6 BY MR. FITZGERALD:

7 Q. What kind of field did you develop for

8 the variability in the field replicates on the

9 ES&P data?

10 A. What kind of field did I develop for

11 it?

12 Q. Yes, they are your words. I mean,

13 don't laugh when I use them.

14 A. I guess -- well, it depends on what

15 you're comparing it to, the variable of the field

16 replicates versus the variables -- the variables

17 in the field replicates compared to variables in

18 what the average at the station is overall, the

19 dates or the variables in the model or --

20 Q. Well, you're the one who wants the

21 data, and you're the one who has suggested that it

22 will tell you some -- it will give you a feel for

23 something. What is it going to give you a feel

24 for? How are you going to apply --

25 A. Okay. Well, one thing that I would do

JACK BESONER & ASSOCIATES, INC.

194

1 -- one task that I would do is look at the field

2 replicate variability in the Department of Justice

3 data and compare that to the field replicate

4 variability in the ES&P data to determine whether

5 there's a similar amount.

6 And then, also, if I have those field

7 replicates, then I have a way of looking at them

8 and trying to determine if there are "potential

9 outliers" that would influence any relationship

10 that I may or may not see between the fourteen

11 station average and the stage using just the

12 Department of Justice data.

13 Q. How are you going to resolve the April

14 '93 data problem with the ES&P data?

15 A. When you --

16 MR. BLANK: When you refer to it as the

17 problem, you're talking about the difference in

18 the concentrations?

19 MR. FITZGERALD: Well, the problem is

20 the word the witness used when he said -- and he

21 also has characterized it as an issue of concern,

22 and I took from that that he felt that it needed

23 to be addressed and resolved in some fashion as an

24 historic event. He has the data.

25 BY MR. FITZGERALD:

JACK BESONER & ASSOCIATES, INC.

195

1 Is it their observation, as Dr. Davis

2 is unable to explain why this -- why is the data

3 in the outlier too high in its concentration or

4 does it appear to be too low? What is the concern

5 with the data? Why is it an issue of concern?

6 MR. BLANK: Let me just object to the

7 form of the question. I think we got multiple

8 questions pending here.

9 MR. FITZGERALD: I think it's all asking

10 for the same answer.

11 MR. BLANK: I think so, too, but I think

12 I think the question is why is the concern --

13 BY MR. FITZGERALD:

14 Q. Well, first, what is it about the ESP

15 data of April '93 that causes you a concern?

16 A. If you look at just the ES&P data, as I

17 stated earlier, if you just look the ES&P data,

18 the entry and access, and assuming you're going to

19 average the field reps, which may be one way to do

20 -- it's not necessarily the way to do it. And

21 then looking at the relationship between the

22 fourteen station average and average stage, if you

23 leave 4/93 in, there's not a significant

24 relationship. If you take it out, there is a

25 significant relationship.

JACK BESONER & ASSOCIATES, INC.

196

1 Q. What is it about the 4/93 data that

2 causes you to focus on that instead of some other

3 month, when you're doing your analysis over a ten

4 or eleven month data set?

5 A. If you look at the data plot you can

6 see that it's very low relative to the other

7 observations.

8 Q. Did Dr. Davis suggest to you that it is

9 actually ridiculously low or unbelievably low?

10 A. No.

11 Q. Did he indicate to you that he felt

12 that it, in fact, may reflect an actual true state

13 of affairs at that time in Loxahatchee at those

14 sampling stations?

15 A. He probably didn't say that it in those

16 words.

17 Q. Well, isn't it fair to surmise that he

18 thinks there's something wrong with it too since

19 he has tried to explain it and says he can't find

20 an explanation?

21 A. No, because I asked him about it.

22 Q. So you're the one who identifies the

23 problem. Did you identify the problem solely

24 because of the affect on the correlation analysis?

25 A. Yes.

JACK BESONER & ASSOCIATES, INC.

197

1 Q. All right. I guess at that point you

2 all said, why the heck are these so low in

3 comparison to anywhere else because they are --

4 A. Correct.

5 Q. Which automatically, to a

6 statistician's eye, makes them suspect and worthy

7 of further inquiry.

8 A. Worthy of further inquiry.

9 Q. You don't like suspect I gather. How

10 are you going resolve this issue of concern? Can

11 you resolve this issue of concern?

12 A. I'm not sure at this point.

13 Q. When do you think you'll have a final

14 opinion then, if ever, since you say you don't

15 know if you can? We may have just wasted a day,

16 but let's hear his answer.

17 A. I think I'll have a final opinion

18 within a couple of weeks, however, the caveat is

19 that my final opinion may be that it can't be

20 resolved.

21 Q. Then we have wasted a day and longer.

22 You've looked at the DOJ data set for the same

23 period.

24 A. Yes.

25 Q. Does the same anomalous result appear

JACK BESONER & ASSOCIATES, INC.

198

1 in the DOJ data set?

2 A. By the same anomalous result, you mean

3 -- do you mean that 4/39 appears lower?

4 Q. Aberrationally lower. Extremely low in

5 comparison to all other values.

6 A. No, but on the other hand, the

7 correlation is the opposite direction to have the

8 DOJ data.

9 MR. FITZGERALD: I understand. Off the

10 record.

11 (Discussion off the record.)

12 BY MR. FITZGERALD:

13 Q. In one of your documents you identify a

14 concern over discrepancy between the SWIM Plan

15 data and the data you were utilizing for the

16 Loxahatchee stations 1-1 through 1-16. Did you

17 successfully resolve that concern -- sort it out?

18 A. The way it was left, Bruce Meyer at

19 Environmental Services & Permitting -- in fact, I

20 believe it's probably in that fax that you're

21 holding. He indicated to me --

22 Q. No.

23 A. No? Okay. Then maybe it's in another

24 document. He indicated to me that the data that I

25 had been supplied by in Environmental Services &

JACK BESONER & ASSOCIATES, INC.

199

1 Permitting matched Walker's data.

2 Now, when he says Walker's data, I'm

3 not sure whether he means the data that appeared

4 in Walker's May report -- May '92 report or some

5 other source. But the way it was left -- that was

6 just the way it was left is that yes, there's a

7 discrepancy between the data as it appears in

8 Appendix "E" and the data that he received from

9 Environmental Services & Permitting.

10 However, that discrepancy doesn't end

11 up changing -- as I recall, doesn't end up

12 changing the fourteen station average. It looks

13 like some values were switched. If you're looking

14 at the matrix data and the replicates and the

15 columns representing the stations, it looks like

16 certain values got switched between stations or am

17 I saying that wrong?

18 Q. Well, I'm taking it that you're

19 suggesting that once you looked at it or Bruce

20 Meyer did, you figured out it was essentially

21 topographical. It does not reflect the result or

22 outcome in your analysis.

23 A. Yes, my conclusion was is that it

24 wouldn't affect any conclusions I had.

25 Q. So which data set did you use, the SWIM

JACK BESONER & ASSOCIATES, INC.

200

1 Plan, or did you use --

2 A. The one from Environmental Services.

3 Q. Okay. Did you ever prepare questions

4 for Bill Walker?

5 A. I was asked to do that, but then the

6 litigation was stayed before I made up a list.

7 Q. Okay. In one of your commentaries in

8 the documents you talk about receiving the BDA

9 data, having to ultimately analyze data pertinent

10 to work order 32. I assume this is the vegetative

11 stuff for the Loxahatchee refuge.

12 And you said that you had gotten an

13 updated version of it after doing some data

14 snooping. What are you referring to there?

15 A. To the best of my recollection, I think

16 what happened is I was given a floppy disk with

17 the data, and I was given a hard copy of the data,

18 and I think maybe there was some mismatch between

19 the hard copy and what was on the floppy disk.

20 Q. How did you resolve the apparent

21 discrepancy?

22 A. I called up Breedlove, Dennis and

23 Associates and/or I faxed them or talked to them

24 over the phone and told them where I thought the

25 discrepancies were and requested them to send me a

JACK BESONER & ASSOCIATES, INC.

201

1 revised set of data.

2 Q. You also mention EAA to EMP study,

3 although this may be put on the back burner if the

4 loading estimate issue gets hot, did you ever

5 conduct that work or reach any conclusions or

6 final opinions regarding EAA loading or flows to

7 Everglades National Park?

8 A. No. As I said earlier, I believe

9 Dennis Lettenmaier and I met with Dick Slyfield in

10 May, and we identified stations that would be

11 important for that study, and I probably got as

12 far as me receiving flows from structures coming

13 out of the EAA, possibly Loxahatchee too, as well,

14 but then that was it. I didn't do anymore work on

15 it.

16 Q. Okay. Have you made any effort to go

17 back before the SWIM Plan period of record to '73

18 through '75 for their early data collections by

19 the district to see how concentration of

20 phosphorus in earlier years demonstrates increase

21 in loss of concentration over periods of time?

22 A. No.

23 Q. Have you seen any data sets for that

24 period of time at all, other than the flow and

25 concentration data that you've received from ES&P,

JACK BESONER & ASSOCIATES, INC.

202

1 which I believe goes back to '72?

2 A. No.

3 Q. Let me show you a single page 1210217

4 and see if you recognize the handwriting on that.

5 A. That's mine.

6 Q. Okay. There are notes in there then in

7 which you mention Loftis.

8 A. Yes.

9 Q. Bill Loftis and various commentary on

10 Walker's report or report by Walker and people for

11 DOJ. When were these notes made?

12 A. To the best of my recollection,

13 sometime in the spring of '93. I believe these

14 are notes from a conversation that I had with Mr.

15 Blank.

16 Q. So you've never spoken with Loftis?

17 A. No.

18 Q. And on what basis was it suggested to

19 you Loftis held that Walker didn't do a good job

20 for accounting for dry years?

21 A. I'm sorry. Repeat the question.

22 Q. One of the notes there says or

23 attributes that Loftis -- according to your

24 writing.

25 A. Uh-huh.

JACK BESONER & ASSOCIATES, INC.

203

1 Q. A comment -- an insight that Walker

2 didn't do a good job of accounting for dry years.

3 A. That was information that was conveyed

4 to me by Mr. Blank.

5 Q. I understood that. The question is:

6 What basis was provided to you that led you to

7 believe that was credible, I mean not that Mr.

8 Blank didn't say it to you? Do you understand

9 what I'm asking?

10 A. No. This is a transcription of a

11 conversation that I had.

12 Q. Was Loftis' basis for that view ever

13 communicated to you many?

14 A. I don't recall.

15 Q. Okay. Your notes claim that -- or your

16 statement that you reviewed Loftis reviewed

17 Walker's report and paper for DOJ, down if that is

18 the same report that you reviewed, the May '91

19 report or is it speaking about a different report

20 or reports?

21 MR. BLANK: By the May '91 report,

22 you're talking about the Loxahatchee report,

23 aren't you?

24 MR. FITZGERALD: Uh-huh.

25 THE WITNESS: I'm sorry. Can you tell

JACK BESONER & ASSOCIATES, INC.

204

1 me where on the page you're referring to?

2 MR. FITZGERALD: Oh, I see. All right.

3 It's not there.

4 BY MR. FITZGERALD:

5 Q. Were you aware that Loftis reviewed

6 Bill Walker's May '91 report for the Department of

7 Justice?

8 A. The May '91?

9 Q. Uh-huh.

10 A. And that report had to do with

11 Loxahatchee or inflows to Everglades?

12 Q. Inflows to Everglades?

13 A. Yes, I think I was aware that he

14 reviewed that report.

15 Q. In your handwriting you attribute a

16 comment back to Loftis that a slight trend was

17 identified in phosphorus over the period of

18 record; is that correct? Is that what that note

19 means?

20 A. Yes.

21 Q. Do you know what that percentage of

22 trend or increase over that period of record was?

23 A. No.

24 Q. Who else was present during the

25 conversation in which you received these -- where

JACK BESONER & ASSOCIATES, INC.

205

1 you received the paper?

2 A. At my end?

3 Q. Yes.

4 A. No one except myself.

5 Q. Oh, this was a telephone conversation?

6 A. Yes.

7 Q. Was it a conference call?

8 A. No.

9 Q. So no one was on but you and Mr. Blank?

10 A. As far as I know he was the only one

11 there. An assumption.

12 Q. As a result of your meeting with Dick

13 Slyfield, what's your understanding of how waters

14 reaches Everglades National Park through the

15 Central and South Florida Flood Control Project in

16 periods of low flood?

17 A. I would have to go back and review my

18 notes from that meeting. All I remember is that

19 the main point that I got was that it was a very

20 complicated system.

21 Q. I think you got, you know, the basics.

22 What is your understanding of the source of

23 inflows to Everglades National Park?

24 A. Again, I'd have to look at my notes,

25 but I recall the S-12 structures and some -- one

JACK BESONER & ASSOCIATES, INC.

206

1 or two others, but I can't remember what their

2 labeling is.

3 Q. All right. Is 343 and 344 your --

4 A. Yes, and if I recall, I think you're

5 supposed to subtract out of what goes through 344.

6 Q. Did you determine from your discussions

7 with Dick Slyfield whether water ever moves

8 directly from Lake Okeechobee to Everglades

9 National Park without going through -- over land

10 flow?

11 A. I don't recall.

12 Q. Did you ever evaluate Dr. Walker's

13 application, the seasonal Kendall test to

14 determine what happens if you leave the '84, '85

15 water year out of the analysis?

16 A. No.

17 Q. What caused you to make that comment in

18 your note, or ask yourself the parenthetical

19 question what happens if you do that?

20 A. It somehow came up in the conversation

21 when Dr. Lettenmaier and Slyfield and I were

22 talking. I don't recall the exact context.

23 Q. Were you aware that, in fact, Dr.

24 Walker did that in his analysis?

25 A. Now that you mention it, I think after

JACK BESONER & ASSOCIATES, INC.

207

1 I wrote down that comment, then I think I later

2 read his report and saw that he did.

3 Q. Do you recall what it did to the trend

4 analysis?

5 A. No, I don't recall.

6 Q. And, in fact, did you excise that spike

7 year -- well, what question was focused on for the

8 '84 year? What was it that Slyfield told you that

9 it that raised that one year?

10 A. I don't recall.

11 Q. You weren't aware that there was a big

12 spike and phosphorus in Everglades National Park

13 that year?

14 A. I probably was at that time.

15 Q. At the time. Off the record.

16 (Discussion off the record.)

17 BY MR. FITZGERALD:

18 Q. Did anybody ever discuss with you the

19 inaccuracies that might enter that any flow load

20 calculation based on the use of historic rating

21 curves for the pumping structures?

22 A. In my conversations with Dr.

23 Lettenmaier, I'm aware that he's investigating

24 something along those lines.

25 Q. Do you understand how he intends to

JACK BESONER & ASSOCIATES, INC.

208

1 cope with that potential problem?

2 A. I don't recall exactly what strategies

3 he's carrying out.

4 Q. You had a notation in your notes after

5 speaking with Dick Slyfield about asking John

6 Davis about, first, the model and second, the

7 validity of flow data, real or estimated. Did you

8 ever do that -- discuss that at inspection with

9 John Davis?

10 A. If I did, I don't recall.

11 Q. Okay. Let me show you a document

12 121210 and 233, a schematic of the Loxahatchee

13 National Wildlife Refuge. Who produced that

14 schematic?

15 A. I don't know who actually produced the

16 document. It was given to me during a meeting

17 that I attended with Dr. McClave.

18 Q. What work have you done with Dr.

19 McClave?

20 A. I've only met with him once about two

21 weeks ago. Dennis Lettenmaier and I and John

22 Davis met with him, and one other person from his

23 company, and I don't recall that gentleman's name.

24 Q. A statistician?

25 A. Yes.

JACK BESONER & ASSOCIATES, INC.

209

1 Q. What was the purpose of that meeting?

2 A. The purpose was to discuss the work

3 that Dr. McClave had done analyzing the Appendix

4 "E" data and also the entry and access data.

5 Q. What analysis of the entry and access

6 data is Dr. McClave doing?

7 A. He earlier had done some work, and he

8 -- I believe he testified to this in his

9 deposition -- where he looked at a model where

10 that instead of taking the fourteen stage average,

11 the single response as the single response

12 variable, he divided Loxahatchee up into these

13 four quadrants and then fit some models allowing

14 for -- I guess, as I recall, allowing for a

15 different line -- different rep in the four

16 quadrants.

17 Q. Did the division into quadrants result

18 in different coefficients to a regression line or

19 fundamentally different correlations?

20 A. I don't recall whether the correlations

21 changed or not.

22 Q. Do you have any documents from that

23 meeting?

24 A. Only what was in my folder, which I

25 gave to him. And we gave copies to you.

JACK BESONER & ASSOCIATES, INC.

210

1 Q. So none of that has been withheld on

2 the claim privilege so far as you know?

3 A. No, from that meeting this and

4 probably some handwritten notes are the only thing

5 that were in that folder.

6 Q. Were the notes dated?

7 A. I don't know.

8 MR. FITZGERALD: Counsel, do you know if

9 those notes were provided because I have seen

10 nothing?

11 MR. BLANK: I don't know. We will give

12 you a privileged list. I think the bulk of that

13 list relates to mediation materials.

14 Are we looking for handwritten notes?

15 THE WITNESS: Yes.

16 MR. BLANK: Would the date of these be

17 1/27/94?

18 THE WITNESS: Yes.

19 MR. BLANK: They will be on your

20 privileged list.

21 MR. FITZGERALD: What is the basis of

22 the privilege claim?

23 MR. BLANK: I suspect it's attorney work

24 product, but we will let you know when we give you

25 the list.

JACK BESONER & ASSOCIATES, INC.

211

1 BY MR. FITZGERALD:

2 Q. Was there an attorney meeting?

3 A. No, we were all surprised.

4 Q. Who scheduled that meeting?

5 A. It just evolved out of a consensus

6 between Dr. McClave and Dr. Lettenmaier and myself

7 after we had a phone conversation.

8 Q. Who set up the phone conversation?

9 A. Mr. Blank, I believe, was the one who

10 directed me to get in touch with Dr. McClave.

11 Q. Are you reviewing Dr. McClave's work?

12 A. No.

13 Q. Then what was your purpose for meeting

14 with Dr. McClave?

15 A. To find out exactly what he has done

16 with regard to the limits in the Loxahatchee, it's

17 analysis, and part of it was to make sure that we

18 weren't duplicating effort.

19 Q. And what is your understanding of what

20 he's doing with respect to the limits in

21 Loxahatchee as opposed to what you're doing?

22 A. As I said, he's looking at a different

23 model where he's divided up Loxahatchee into these

24 four quadrants. Rather than just taking an

25 appendix, he used the model.

JACK BESONER & ASSOCIATES, INC.

212

1 Q. Are these regression, his models?

2 A. Yes, I believe so.

3 Q. Do you know what kind of fit he was

4 getting in the individual sets -- quadrants?

5 A. The main point that I remember is that

6 the variability was a bit higher If you partition

7 it up into different quadrants. So the limit is

8 obviously going to be higher.

9 Q. Do you remember what that kind of

10 numbers you're talking about in parts per billion

11 of those limits?

12 A. No, I can't recall.

13 Q. What was Dr. McClave's expressed

14 rationale, if any, for engaging in this construct

15 in Loxahatchee rather than treating it as a

16 hydrological unit?

17 A. Well, any time you take average, you're

18 going to mask any sort of individual relationship

19 that you might see. So he was concerned about

20 looking to see how the -- as I said, looking to

21 see whether the model -- when you get a similar

22 type model in the four quadrants, whether the

23 limits are similar or what happens.

24 Q. Was there similarity among the

25 quadrants at all?

JACK BESONER & ASSOCIATES, INC.

213

1 A. To my best recollection, I think the

2 relationship with stage was similar, but as I say,

3 the variability increased so that the limits

4 increased.

5 Q. Okay. As I understand it, that's the

6 result you'd expect as a result of dividing the

7 area up because you are increasing the

8 variability.

9 A. You're reducing the sample size on

10 which you're basing an average. If you only -- if

11 you take the average just at the station, say in

12 quadrant 3, there are no longer fourteen stations.

13 It's three stations or whatever, and the variance

14 of an average is -- the general formula is the

15 population variance divided by the sample size.

16 Q. Do you agree that's a valid mechanism

17 for considering or modelling the Loxahatchee

18 refuge?

19 A. I think it's a valid alternative to

20 consider.

21 Q. How would you go about determining

22 whether it is ultimately preferable to treating

23 the refuge as a unit -- a single unit rather than

24 four compartments?

25 A. There are a lot of different ways that

JACK BESONER & ASSOCIATES, INC.

214

1 you can go about doing that. One possible method

2 is to look at site entry and access period, and

3 say -- for example, assume that the limits are --

4 in effect, look at whether you're in compliance or

5 out of compliance based on the limits in Appendix

6 "E" to have the entry and access period.

7 It may be that in cases where you are

8 "out of compliance", and it's actually only due to

9 a couple of stations rather than all these

10 stations. So in that case it may make sense to

11 look at a model where, in instead of averaging

12 everything so that you loose the spatial

13 information, it may be preferable to look at a

14 model like this.

15 Q. In your public policy purposes to

16 achieve a certain level or limit in phosphorus

17 within an area you desire to protect -- isn't it

18 counter-intuitive to subdivide the area in a

19 manner that's guaranteed to suggest higher limits

20 that may not be justifiable on a biological

21 basis?

22 MR. BLANK: Object to the form of the

23 question.

24 THE WITNESS: I don't agree with that.

25 If I tell turns out that, for example, stations in

JACK BESONER & ASSOCIATES, INC.

215

1 this area are almost all the time in compliance,

2 whereas stations in this area happen to be out of

3 the compliance for some reason, then it makes

4 sense to try and figure out why those particular

5 stations are out of compliance.

6 I mean, it's the -- it's similar to --

7 in the case of monitoring hazardous waste sites

8 under RCRA regulations, under those regulations

9 the operator of the hazardous waste facility is

10 instructed to look at each down gradient. Well,

11 the person doesn't take an average of those. The

12 person looks at each of those down gradients as

13 well and compares it to the background.

14 BY MR. FITZGERALD:

15 Q. Based on the work as explained to you

16 by Dr. McClave, which of the quadrants are most

17 likely or under his analysis like to the be out of

18 compliance on a regular basis?

19 A. I don't recall.

20 Q. Did he not, in fact, divide the

21 quadrants up so that the influence of the S-5A and

22 S-6 are each restricted to the smallest possible

23 quadrants?

24 MR. BLANK: You're assuming prism

25 influence from 5A and 6.

JACK BESONER & ASSOCIATES, INC.

216

1 BY MR. FITZGERALD:

2 Q. You're familiar with the source of the

3 surface flowing water, aren't are you, Doctor?

4 A. Yes.

5 Q. They are S5-A and 6, aren't they?

6 A. Yes.

7 Q. The ones you've been looking at in your

8 flow data.

9 A. Yes.

10 Q. Did he not construct his quadrants, in

11 fact, not even characterize them as quadrants

12 based on the chart that you're looking at and

13 that's been constructed to bracket with one

14 quadrant each taking in the surface water inflow

15 points, and that also demonstrates the highest

16 levels of phosphorus in surface water column in

17 Loxahatchee, right? I mean, you've looked at the

18 data.

19 MR. BLANK? What's the question?

20 MR. FITZGERALD: And I think it calls

21 for a yes or no at this point.

22 MR. BLANK: What is the question

23 though?

24 BY MR. FITZGERALD:

25 Q. Is that not the effect of what Dr.

JACK BESONER & ASSOCIATES, INC.

217

1 McClave has done?

2 MR. BLANK: I'm going to object to you

3 asking the witness to testify as to the rationale

4 of what Dr. McClave may have used to devise this

5 or -- if you know.

6 MR. FITZGERALD: I'm not asking the

7 rationale, counsel. I understand that he can't

8 answer as to what Dr. McClave's plans were, but he

9 can say yes, that is the effect, and that is the

10 question that is pending.

11 BY MR. FITZGERALD:

12 Q. Is that not the affect of what Dr.

13 McClave has done?

14 A. Is what the affect?

15 Q. Has not Dr. McClave partitioned the

16 refuge in such a way that the two areas most

17 likely -- the two areas that are under the

18 greatest influence of the two surface water inflow

19 points are also the two areas most likely to be

20 out of compliance based on the historical data

21 from the sampling sites?

22 A. I can't answer your question for two

23 reasons. The first reason is that I can't say

24 that this is the area most under the influence of

25 a structure because I don't know what you mean by

JACK BESONER & ASSOCIATES, INC.

218

1 influence and what's been proved.

2 Q. Do you understand the flow path

3 hydrologically of surface water entering

4 Loxahatchee from S5-A?

5 A. In general, yes.

6 Q. Where does the water come from? From

7 Dick Slyfield, I presume, or do you have some

8 other source for this information --

9 A. The water can either flow through

10 Loxahatchee or can go through the canal.

11 Q. That's the extent of your understanding

12 of the flow pattern?

13 A. Yes.

14 Q. And what is your understanding of the

15 flow pattern of 4S-6?

16 A. Similar. It can go out into

17 Loxahatchee or it can go through the canal.

18 MR. FITZGERALD: All right. Why don't

19 we stop here. We can pick up tomorrow.

20 (Thereupon the deposition was concluded.)

JACK BESONER & ASSOCIATES, INC.

219

1 CERTIFICATE OF OATH

3 STATE OF FLORIDA

4 COUNTY OF DADE

6 I, the undersigned authority, certify

7 that Steven P. Millard, Ph.D., personally appeared

8 before me and was duly sworn.

10 WITNESS my hand and official seal this

11 22nd day of February, 1994.

14 _________________________

15 Lora Lee Knorr

16 Notary Public, State of Florida

17 My Commission Expires:

18 April 30, 1995

JACK BESONER & ASSOCIATES, INC.

220

1 CERTIFICATE OF REPORTER

5 STATE OF FLORIDA:

6 COUNTY OF DADE:

I, Lora Lee Knorr, Certified Shorthand

9 Reporter, certify that I was authorized to and did

stenographically report the foregoing deposition;

10 and that the transcript is a true record of the

testimony given by the witness.

11 I further certify that I am not a

relative, employee, attorney, or counsel of any of

12 the parties nor am I a relative or employee of any

of the parties' attorney or counsel connected with

13 the action, nor am I financially interested in the

action.

14 Dated this 22nd day of February, 1994.

Lora Lee Knorr

16 Certified Shorthand Reporter

ACKNOWLEDGMENT

STATE OF FLORIDA

18 COUNTY OF DADE

The foregoing certificate was

19 acknowledged before me this 22nd day of February,

1994, by Lora Lee Knorr, who is personally known

20 to me.

___________________________

JACK BESONER & ASSOCIATES, INC.

221

1 READING AND SIGNING

4 I have read the above transcript, pages 1

5 through 218 and I find: (MARK ONE)

7 ( ) The transcript is true, correct, and

8 completely accurate.

10 ( ) The transcript is true, correct, and accurate,

11 except as set forth in my List of Corrections

12 attached hereto, citing page and line and reason

13 for the correction realizing that, for this

14 purpose, I am still under oath.

_______________ _________________________

17 (DATE) STEVEN P. MILLARD, Ph.D.

Sworn to and subscribed before

21 me this______day of____________, 1994.

22 ___________________________

Notary Public

23 My Commission expires:

JACK BESONER & ASSOCIATES, INC.

222

TO BE EXECUTED BY THE NOTARY IF THE DEPONENT DOES

2 NOT SIGN:

6 I hereby certify that a letter with reference

7 to reading and signing deposition was mailed to

8 the witness through his attorney, on

9 ___________________, 1994, and that the witness

10 ( ) Witness refused to sign, giving the following

11 reason:

12 ( ) Neither the witness nor his attorney has

13 responded to request to read and sign.

_______________ ___________________________

17 (DATE) Notary Public

18 MY COMMISSION EXPIRES:

JACK BESONER & ASSOCIATES, INC.