142
1 Division of Administrative Hearings
2 Department of Administration, State of Florida
3 SUGAR CANE GROWERS COOPERATIVE )
of FLORIDA; ROTH FARMS, INC.; and, )
4 WEDGEWORTH FARMS, INC., )
Petitioners, )
5 V ) DOAH
SOUTH FLORIDA WATER MANAGEMENT ) Case 92-3038
6 DISTRICT, an agency of the State )
of Florida; et al., )
7 Respondents. )
FLORIDA SUGAR CANE LEAGUE, INC.; )
8 UNITED STATES SUGAR CORPORATION; )
and NEW HOPE SOUTH, INC., )
9 Petitioners, )
V ) DOAH
10 SOUTH FLORIDA WATER MANAGEMENT ) Case 92-3039
DISTRICT, an agency of the State )
11 of Florida; et al., )
Respondents. )
12 FLORIDA FRUIT and VEGETABLE )
ASSOCIATION; LEWIS POPE FARMS; )
13 W.E. SCHLECHTER & SONS, INC., )
and HUNDLEY FARMS, INC., )
14 Petitioners, )
V ) DOAH
15 SOUTH FLORIDA WATER MANAGEMENT ) Case 92-3040
DISTRICT, an agency of the State )
16 of Florida; et al., )
Respondents. )
17
Volume II
18 Deposition of Douglas Robson, Ph.D.
19 Taken before April Y. Sapp, Court Reporter
and Notary Public in and for the State of Florida at
20 large, pursuant to notice of taking deposition filed
by the Petitioners in the above cause.
21 - - -
Wednesday March 31, 1993
22 319 Clematis Street, 5th Floor
West Palm Beach, Florida 33401
23 9:00 a.m. - 12:00 p.m.
1:03 - 2:50 p.m.
24 - - -
143
1 APPEARANCES:
2 On behalf of the Petitioners Florida Sugar
Cane League, Inc., United State Sugar Corp.,
3 and New South Hope, Inc.:
Peeples, Earl & Blank, P.A.
4 One Biscayne Tower, Suite 3636
Two South Biscayne Boulevard
5 Miami, Florida 33131
By: ROBERT H. BLANK, ESQUIRE
6
On behalf of the Respondent SFWMD:
7 Popham, Haik, Schnobrich & Kaufman, Ltd.
4000 International Place
8 100 S.E. Second Street
Miami, Florida 33131
9 By: DANIEL J. MCGRATH, ESQUIRE
10 On behalf of the Intervenor United States of America:
Assistant United States Attorney
11 Department of Justice
155 South Miami Avenue, Suite 627
12 Miami, Florida 33130-1693
By: THOMAS A.W. FITZGERALD, ESQUIRE
13
On behalf of the Sugar Cane Growers Cooperative,
14 Roth Farms, Inc. and Wedgeworth Farms, Inc.:
Hopping, Boyd, Green & Sams
15 123 South Calhoun Street
Tallahassee, Florida 32314
16 By: GARY V. PERKO, ESQUIRE
17 Also Present: Dr. James T. McClave
Dr. James Grimshaw
18 - - -
144
1 - - -
2 I N D E X
3 - - -
4
5 WITNESS: DIRECT CROSS REDIRECT RECROSS
6 Douglas Robson, Ph.D.
7
BY MR. PERKO 146
8 BY MR. BLANK 184
9
†䈠⁙前䕐䭒⁏††††††㐱ശ ‸䈠⁙前䱂乁⁋†††††††††††††ㄠ㐸†ഠ ‹
††䉙⁍刮⁐䕒䭏††††††‱㐶ഊ‸†䉙⁍刮⁂䱁之††††††††††††††ㄸ㐍ਠ†ഊ‹††††††††††††††††††††††††‱㐵ഊഊ
145
1 - - -
2 E X H I B I T S
3 - - -
4
5 NUMBER PAGE
6 EXB. NO. 15 152
7 Calculations of standards for the Park
8 EXB. NO. 16 154
9 Settlement Agreement
10 EXB. NO. 17 157
11 Letter to Federico from Robson 7-9-91
12 EXB. NO. 18 158
13 7-18-91 fax from Walker to Robson with
14 attachments
15 EXB. NO. 19 159
16 Draft paper by Walker 5-15-91
17 EXB. NO. 20 165
18 Draft paper by Walker 5-24-91
19 EXB. NO. 21 171
20 Expert witness designation
21 EXB. NO. 22 182
22 Disk print out
23 EXB. NO. 23 236
24 3-12-90 handwritten notes
25
146
1 P R O C E E D I N G S
2 - - -
3 Thereupon,
4 Douglas Robson, Ph.D.,
5 being by the undersigned Notary Public previously
6 sworn, was examined and testified as follows:
7 CONTINUED DIRECT (Douglas Robson, Ph.D.)
8 BY MR. PERKO:
9 Q. Dr. Robson, just to remind you, you are
10 under oath.
11 A. Right.
12 Q. Do you understand that?
13 I just have a couple of follow up questions
14 from what we talked about yesterday.
15 You mentioned that you drafted Appendix E
16 to the SWIM Plan, is that right?
17 A. Right.
18 Q. How much of Appendix E did you actually
19 write?
20 A. I thought I had provided you a copy of the
21 rough draft that I wrote.
22 Q. Okay.
23 MR. MCGRATH: I believe you did. I believe
24 that was also an exhibit.
25 MR. PERKO: Part of it was. Unfortunately,
147
1 it seems to have been separated in the copies
2 that we received. It's difficult to determine
3 what you actually wrote.
4 BY MR. PERKO:
5 Q. Did you write everything in the draft that
6 you produced to us?
7 A. Yes.
8 Q. Okay. Did you write all of the narrative
9 portion of Appendix E?
10 A. I wrote all the narrative portion of that
11 in that draft, yes.
12 Q. In that draft.
13 How does the final Appendix E differ from
14 your draft?
15 A. Very slightly. I did not write the
16 introduction.
17 Q. Okay.
18 A. And in the first paragraph --
19 Q. Of page E-2?
20 A. -- on page E-2 approximately the first half
21 of the first paragraph was not written by me.
22 Q. Can you identify anything else that was not
23 written by you?
24 A. On page E-10 there are two formulas for the
25 percent exceedance limit.
148
1 Q. I believe you previously testified that one
2 of those formulas was written by Dr. Walker, is that
3 correct?
4 A. Right.
5 Q. You don't recall which one or do you?
6 A. I still don't recall which one, but at
7 least the preceding sentence, preceding first
8 formula, was not written by me.
9 Q. Do you know who wrote it?
10 A. I'm not sure.
11 Q. Do you know who wrote the introduction?
12 A. No.
13 Q. Is there anything else that you did not
14 write in Appendix E?
15 A. I don't believe so.
16 Q. Dr. Robson, for the Taylor Slough and
17 Coastal Basin Phosphorus Limits in Appendix E, I
18 believe a 1984 base period is used, is that correct?
19 A. That's correct.
20 Q. Do you know why the 1984 base period was
21 used?
22 A. My understanding is that we were obliged to
23 use the first year that we had records.
24 Q. Do you know why you were obliged to use the
25 first year you had records?
149
1 A. I'm not sure that I know. I thought it was
2 the OFW requirement.
3 Q. So 1984 was the first year that you had
4 data for, is that correct?
5 A. Right.
6 Q. Dr. Robson, have you discussed the Refuge
7 phosphorus levels or ENP phosphorus limits with
8 anyone from the Florida Department of Environmental
9 Regulation?
10 A. No.
11 Q. Other than the brief discussion you had
12 with Dr. Walker and others during the settlement
13 negotiations regarding choice of methodologies to be
14 used in calculating the limits did you speak directly
15 with Dr. Walker during settlement negotiations?
16 A. I have a problem understanding the
17 question. What period are we referring to as far as
18 settlement negotiations?
19 Q. From February 1991 until July.
20 A. I don't recall any.
21 Q. Did you speak to him at all while you were
22 drafting Appendix E?
23 A. Yes.
24 Q. How many times did you speak to him during
25 that time period?
150
1 A. I would have to guess two or three times.
2 Q. And what did you discuss with Dr. Walker at
3 that time?
4 A. These were discussions concerning
5 discrepancies in the data I was analyzing versus the
6 data that Walker was analyzing.
7 Q. Were those discrepancies in the Refuge data
8 or the Park data?
9 A. I believe they were all in the Park data.
10 Q. Did you ever resolve the discrepancies?
11 A. Yes.
12 Q. How did you resolve them?
13 A. Through George Shih.
14 Q. How did Mr. Shih resolve the discrepancies?
15 A. I don't know.
16 Q. How many discrepancies were there?
17 A. Again, I'm guessing, two or three.
18 Q. Was it in the phosphorus concentration data
19 or the flow data?
20 A. Both.
21 Q. What effect did the discrepancies have on
22 your calculations relative to Dr. Walker's
23 calculations?
24 A. The coefficients in my regression analysis
25 are slightly different from his.
151
1 Q. Are they higher or lower?
2 A. Some would be higher and some would be
3 lower. Other discrepancies is the one we just looked
4 at just a moment ago. We were looking at the formula
5 for percent exceedance.
6 Q. Page E-10?
7 A. Yes.
8 Q. Page E-10?
9 A. Yes.
10 Q. What was the discrepancy in those occasions
11 or formulas?
12 A. It's explained in the sentence above the
13 first formula.
14 Q. Okay. You said that you and Dr. Walker
15 ended up with some different coefficients, is that
16 correct?
17 A. Yes.
18 Q. Did the calculations ultimately used in
19 Appendix E use your coefficients or Dr. Walker's
20 coefficients or a combination?
21 A. Used mine.
22 Q. In all circumstances?
23 A. Yes. Except the one I just pointed out.
24 MR. PERKO: Would you mark this, please.
152
1 (The document was marked Exb. No. 15.)
2 BY MR. PERKO:
3 Q. Dr. Robson, yesterday I believe you
4 testified that Mr. Federico or Dr. Federico expected
5 you to eliminate the two stations in the Loxahatchee
6 analysis on July 5, 1991, is that correct?
7 A. Yes.
8 Q. Why did July 5th stick out in your memory?
9 A. Because that was a long day for me.
10 Q. Why was it a long day for you?
11 A. That was the day that the Refuge and the
12 District met in Miami to try to come to -- to try to
13 negotiate an agreement with respect to the Refuge.
14 Q. You were not present at that meeting?
15 A. No.
16 Q. Do you know who was?
17 A. No.
18 Q. Prior to that meeting were there
19 discussions about which stations to eliminate?
20 A. Yes.
21 Q. Were different combinations of stations
22 being discussed?
23 A. I don't believe so.
24 Q. Was it just the two --
25 A. Yeah.
153
1 Q. -- whether the two should be eliminated?
2 A. Yes.
3 Q. That was the single issue?
4 A. That's my recollection.
5 Q. Was there anything else that, to your
6 knowledge, was discussed at the meeting on July 5th
7 between the Refuge and the District?
8 A. Not to my knowledge.
9 Q. Dr. Robson, direct your attention to what's
10 been marked as Exhibit Number 15. Do you recognize
11 this document?
12 A. Yes.
13 Q. And what is this document?
14 A. Document which I sent to Federico on July
15 5th.
16 Q. What is reflected in the document?
17 A. There are calculations of standards for the
18 Park.
19 Q. Were these the standards that or the
20 calculations that were ultimately included in the
21 settlement agreement?
22 A. Do you have a copy of the settlement
23 agreement?
24 Q. I can give you one.
25 MR. PERKO: I only have one copy, counsel.
154
1 We can go ahead and mark this as 16.
2 (The document was marked Exb. No. 16.)
3 BY MR. PERKO:
4 Q. Dr. Robson, show you what's been marked as
5 Exhibit 16. Do you recognize that document?
6 A. Yes.
7 Q. And what is it?
8 A. Settlement agreement.
9 Q. Does that help refresh your recollection as
10 to whether the calculations reflected in Exhibit
11 Number 15 are the ones ultimately included in the
12 settlement agreement?
13 A. These are not.
14 Q. Is it your understanding that the
15 calculations to be used in developing the Park limits
16 and Loxahatchee levels were agreed upon at the July
17 5th meeting?
18 A. No.
19 Q. No? Do you know when they were agreed
20 upon?
21 A. No.
22 Q. Did Mr. Federico call you during -- on July
23 5th during that meeting?
24 A. This makes me wonder whether I've got the
25 date correct, July 5th.
155
1 Q. Why does it make you wonder?
2 A. Because my recollection is that that dealt
3 exclusively with the Refuge.
4 Q. Did -- when he instructed you to exclude
5 the two stations from the Loxahatchee analysis, was
6 Dr. Federico calling you from a meeting in Miami?
7 A. Yes.
8 Q. Did he call you more than once during that
9 meeting?
10 A. Yes.
11 Q. How many times did he call you?
12 A. I don't recall. Several.
13 Q. And why did he call you several times?
14 A. I was faxing him analyses of the Refuge
15 data with and without the exclusions.
16 Q. How many analyses did you fax him? Do you
17 remember?
18 A. No, I don't.
19 Q. Was it more than five?
20 A. I doubt it.
21 Q. Were there any differences in the analyses
22 other than the stations utilized?
23 A. I don't believe so.
24 Q. What effect did the elimination of the two
25 stations have on the limits or levels?
156
1 A. It would have lowered the limits.
2 Q. Do you know why there were discussions at
3 that meeting regarding the elimination of those two
4 stations? Why was that an issue?
5 A. Are you saying stations? Did you say the
6 two stations?
7 Q. Yes.
8 A. I'm not sure what their discussion was.
9 Q. It's your understanding that they were
10 discussing whether or not to eliminate the two
11 stations?
12 A. Yes.
13 Q. Do you know if the elimination of the two
14 sampling dates was also discussed at that meeting?
15 A. Yes.
16 Q. Was it -- did Mr. Federico also instruct
17 you on -- at that time to eliminate the two sampling
18 dates?
19 A. That's my recollection.
20 Q. What effect did the elimination of those
21 two sampling dates have on the Refuge levels?
22 MR. FITZGERALD: Counsel, you've asked and
23 answered that question four times, at least
24 three times yesterday and once this morning. I
25 doubt the answer is going to change.
157
1 THE WITNESS: It would have lowered them.
2 (The document was marked Exb. No. 17.)
3 BY MR. PERKO:
4 Q. Dr. Robson, are you familiar with what's
5 been marked as Exhibit Number 17?
6 A. Yes.
7 Q. What is that document?
8 A. It's a letter I sent to Mr. Federico on
9 July 9th.
10 Q. And what is it concerning?
11 A. ENP interim and long term standards.
12 Q. I direct your attention to paragraph 2 on
13 the first page.
14 A. Yes.
15 Q. States that, "I am working on a frequency
16 exceedance formula. Since I am not certain how Bill
17 W calculated this in the first place, I have left a
18 message on his answering machine and am proceeding
19 with my own interpretation of what he had in mind.
20 George has sent me the relative data."
21 Did you ever resolve your uncertainty as to
22 how Dr. Walker calculated the frequency?
23 A. Yes.
24 Q. Is that reflected on page E-10 of Appendix
25 E, the two formulas you discussed a few moments ago?
158
1 A. No.
2 Q. No. Okay.
3 How did you resolve the uncertainty
4 regarding the frequency exceedance formula?
5 A. As indicated here, it's through proceeding
6 with my own interpretation and finding I was getting
7 results almost identical to Walker.
8 Q. Did you do any additional work after July
9 9, 1991 on the frequency exceedance formula?
10 A. Probably during the next several days.
11 Q. Were your calculations or Dr. Walker's
12 ultimately included in Appendix E?
13 A. Both.
14 Q. Both?
15 A. As far as frequency exceedance, yes.
16 MR. PERKO: Mark these, please.
17 (The document was marked Exb. No. 18.)
18 BY MR. PERKO:
19 Q. Dr. Robson, I direct your attention to
20 what's been marked as Exhibit Number 18. Do you
21 recognize this document?
22 A. Yes.
23 Q. And what is it?
24 A. It's a short listing of data, flows, S12-D.
25 Q. On the first page there's a note that,
159
1 "Attached differences in data bases may explain
2 differences in frequency calculations."
3 Is Dr. Walker referring to the
4 discrepancies in the data that you discussed a few
5 minutes ago?
6 A. I believe so.
7 Q. Does this help refresh your recollection as
8 to how the discrepancies were ultimately resolved?
9 A. Somewhat.
10 Q. How so?
11 A. I still don't recall how George Shih
12 resolved these differences.
13 Q. On the last page, there's a handwritten
14 note that states, "? treated missing conc. as 0? in
15 calculating flow-weighted mean."
16 How were missing concentrations treated in
17 Appendix E?
18 A. As missing.
19 Q. Were they treated as 0?
20 A. No.
21 (The document was marked Exb. No. 19.)
22 BY MR. PERKO:
23 Q. Dr. Robson, direct your attention to what's
24 been marked as Exhibit Number 19. Do you recognize
25 this document?
160
1 A. I'm not sure that I have seen this before.
2 Q. You don't recall or -- just for the record,
3 this is a report entitled -- draft report entitled,
4 Derivation for Phosphorous Limits for Inflows for
5 Everglades National Park prepared for the Department
6 of Justice by William Walker dated May 15, 1991.
7 Dr. Robson, you don't recall whether this
8 is one of the reports by Dr. Walker that you
9 reviewed?
10 A. I don't remember.
11 Q. Did you review any reports by Dr. Walker on
12 the Derivation for Phosphorus Limits for Everglades
13 National Park?
14 A. Yes.
15 Q. Do you recall how long the document was?
16 A. No.
17 Q. Is it possible that this is the document,
18 you just don't know?
19 A. Possible.
20 MR. MCGRATH: Objection to form. Anything
21 is possible.
22 BY MR. PERKO:
23 Q. Dr. Robson, had you calculated phosphorus
24 limits for the Park prior to reviewing the report on
25 that subject by Dr. Walker?
161
1 A. Well, Walker had done it before I did and I
2 had seen reports of his before I did the
3 calculations.
4 Q. Let me direct your attention to page 7 of
5 Exhibit 19 just bates page 0858974. Last full
6 paragraph, last sentence states that, "Despite the
7 limited period of record for these basins, (six
8 years, providing only three degrees of freedom for
9 the regression) results are similar to those for
10 Shark River Slough."
11 Dr. Robson, as a statistician, would you
12 feel comfortable with placing reliance on a
13 regression with only three degrees of freedom for
14 error?
15 A. Well, the reliability would be represented
16 by the standard errors on the regression
17 coefficients.
18 Q. But are you comfortable with a regression
19 that relies on only three degrees of freedom?
20 A. If that's all that's available, I'm obliged
21 to be.
22 Q. Next paragraph, could you just read the
23 next paragraph, Dr. Robson, to yourself.
24 The last sentence of that paragraph states
25 that, "Because of these properties --" that are
162
1 discussed in the paragraph, "-- computed limits for
2 the flow-weighted mean are much more sensitive to
3 outlier screening than the frequency criteria."
4 Does this paragraph indicate a preference
5 for the frequency exceedance model over the
6 flow-weighted mean model?
7 MR. MCGRATH: Let me just object to the
8 form to the extent that you're asking the
9 witness to speculate as to the intent of the
10 author of the document and what his preference
11 was.
12 MR. PERKO: Just asking for his
13 understanding.
14 THE WITNESS: It appears to be slanted
15 towards that, yes.
16 BY MR. PERKO:
17 Q. Do you agree with this paragraph?
18 A. In part.
19 Q. What part do you agree with?
20 A. "Skewness increases the risk of Type I
21 error" and the last sentence, "because of these
22 properties, computed limits," etc.
23 Q. "Because of these properties computed
24 limits for flow-weighted mean are much more sensitive
25 to the outlier screening than are the frequency
163
1 criteria." Is that what you're referring to?
2 A. I would say more sensitive. I wouldn't put
3 the "much" in.
4 Q. Okay. Do you disagree with any statements
5 in this paragraph?
6 A. I'm not sure about the validity of the
7 claim that it increases Type II error because
8 normally when you increase Type I error, you decrease
9 Type II error.
10 Q. Dr. Robson, yesterday you mentioned that
11 you may be relying on a paper by Hirsch in support of
12 your testimony. I direct your attention to the
13 bottom of bates page 0858975. There's a reference to
14 a paper by Helsel and Hirsch. Is that the paper you
15 were referring to yesterday?
16 A. No.
17 Q. Let me direct your attention to Table 2 on
18 bates page 0858979. It's a table that illustrates
19 Limit Sensitivity to Station, Summary Statistic and
20 Outlier Screening. In particular, I direct your
21 attention to the bottom of that table where it shows
22 the flow-weighted mean for station S332. Without
23 Outlier Screening Limit Ranges are from the low of
24 21.2 to a high of 39.7 and the -- and With Outlier
25 Screening the Limit Ranges from a low of 8.0 to a
164
1 high of 12.7, is that correct?
2 A. Yes.
3 Q. Does this demonstrate sensitivity to
4 outliers?
5 A. Yes.
6 Q. Does this cause you any concern? Does this
7 sensitivity cause you any concern about using the
8 flow-weighted mean methodology to set limits for the
9 Park?
10 A. It causes me concern over rejecting
11 outliers when using a flow-weighted mean.
12 Q. Why does it cause you concern for rejecting
13 outliers?
14 A. Because of the sensitivity.
15 Q. Dr. Robson, I direct your attention to
16 bates page 0858993. This is a table entitled,
17 Calculation of Frequency Criteria, Shark River
18 Slough. At the bottom of the page is a reference to
19 the Shapiro Wilk Test for normality. Are you
20 familiar with the Shapiro Wilk Test?
21 A. Yes.
22 Q. And the next line states that, "W equals
23 .9227 w05 equals .8500 null hypothesis accepted."
24 Is that correct?
25 A. That's what it says, yes.
165
1 Q. Do you agree with the phrase, "null
2 hypothesis accepted" when testing for normality?
3 A. No.
4 Q. No?
5 A. No.
6 Q. Why not?
7 A. Null hypothesis is never accepted on the
8 basis of a statistical test.
9 Q. From a statistical standpoint how powerful
10 is the Shapiro Wilk Test when using only 13
11 observations?
12 A. I don't know.
13 (The document was marked Exb. No. 20.)
14 BY MR. PERKO:
15 Q. Dr. Robson, direct your attention to what's
16 been marked as Exhibit Number 20. Do you recognize
17 this document?
18 A. I recognize some of the figures.
19 Q. Do you -- well, just for the record, this
20 is a draft report entitled, Derivation of Phosphorous
21 Limits for the Loxahatchee National Wildlife Refuge
22 prepared for Environmental and Natural Resources
23 Division, U.S. Department of Justice by William
24 W. Walker.
25 Does that mean that you -- does your
166
1 previous response mean that you do not recognize the
2 narrative portion of this report?
3 A. I don't remember seeing it, no.
4 Q. Do you recall reviewing any reports by
5 Dr. Walker on the Derivation of Phosphorus Limits For
6 the Loxahatchee National Wildlife Refuge?
7 A. I don't recall any.
8 Q. Okay. Let me direct your attention to page
9 2 which bears bates number 0870942. Under item
10 number 5 there's an equation that utilizes the term,
11 "C to the negative .5." Is that indicative of the
12 inverse square root?
13 A. Yes.
14 Q. Is this -- have you ever, in your
15 experience as a statistician, used the inverse of the
16 square root to transform data?
17 A. I don't remember any specific instance
18 where I did.
19 Q. In Appendix E you used the logs to
20 transform the data, is that correct?
21 A. Correct.
22 Q. As opposed to the inverse of the square
23 root, is that correct?
24 A. Correct.
25 Q. Next page bates 0870943. First full
167
1 paragraph states that, "The time term B sub 1 D
2 provides a statistical handle for adjusting the
3 standards to the baseline year (June 1978 through May
4 1979.) I'm sorry. First narrative paragraph, first
5 sentence.
6 What is your understanding of the phrase,
7 "statistical handle" as it applies to the dummy
8 variable to define the base year?
9 A. I would read this to mean that he's using
10 the dummy variable for adjusting standards.
11 Q. To the baseline year?
12 A. Yes.
13 Q. Do you know why -- Appendix E uses the
14 dummy variable, does it not?
15 A. Yes.
16 Q. Why does it utilize the dummy variable?
17 A. Pardon me?
18 Q. Why does it utilize the dummy variable to
19 adjust the standards to the baseline year?
20 A. Because at the top of the page D is defined
21 as dummy variable.
22 Q. Why was the dummy variable utilized, then?
23 A. Because we are obliged to adjust to the
24 baseline year.
25 Q. Is there any other way of adjusting to the
168
1 baseline year?
2 A. Yes.
3 Q. What other ways are there to adjust to the
4 baseline year?
5 A. With what was used in the ENP.
6 Q. What was that --
7 A. That was regressing on time.
8 Q. -- time averaging?
9 A. No. Regressing on time.
10 Q. Why did you regress on time for the Park
11 limits but not for the Loxahatchee levels?
12 A. For the Park we were dealing with annual
13 average so the seasonal effect had been eliminated.
14 Q. Why is the seasonal effect important in
15 determining whether to regress on time?
16 A. Otherwise there's seasonal variation in
17 response variable.
18 Q. Farther down on the same page, paragraph
19 following states that "Observed and predicted values
20 are listed in Table 2. Diagnostic checks indicate
21 that model residuals are independent of season,
22 serially independent and normally distributed." Is
23 that correct?
24 A. That's what it says, yes.
25 Q. Do 14 data points provide -- are 14 data
169
1 points sufficient to provide powerful tests of
2 independence of season, powerful in a statistical
3 sense?
4 A. Power is relative. Power starts at a
5 significance level and then goes all the way up to a
6 hundred percent, so it would have some power. I
7 don't know how much.
8 Q. Okay. Direct your attention to Table 4 on
9 bates page 0870950 labeled, Sensitivity Analysis
10 Phosphorus Limit versus Stage for Various
11 Assumptions.
12 What does this table indicate, Dr. Robson?
13 Is this the sensitivity analysis that you mentioned
14 that Dr. Walker had performed yesterday?
15 A. I don't remember the context in which I
16 referred to Walker's sensitivity.
17 Q. Okay. Do you recall if you reviewed this
18 sensitivity analysis before you drafted Appendix E?
19 MR. MCGRATH: You mean independently from
20 this paper since he has testified he doesn't
21 recall seeing this paper?
22 MR. PERKO: Yes.
23 THE WITNESS: I don't recall having seen
24 this.
25 BY MR. PERKO:
170
1 Q. Did you review any sensitivity analysis
2 performed by Dr. Walker prior to drafting Appendix E?
3 A. On either the Park or the Refuge?
4 Q. Let's start with the Park.
5 A. I don't recall any.
6 Q. What about the Refuge?
7 A. Oh. Pardon me. I meant for the Park I do
8 recall. For the Refuge I don't.
9 Q. Okay. Did you yourself perform any
10 sensitivity analysis for the Refuge?
11 A. No.
12 Q. Let me direct your attention to Table 4
13 again. Could you explain to me what this table
14 reflects?
15 A. The entries in the main body of the table
16 are the limits calculated under the conditions
17 specified by the row heading.
18 Q. What does the R squared column indicate?
19 A. R square is variously known as a
20 coefficient of determination.
21 Q. What significance do you attach to an R
22 square value?
23 A. Depending on degrees of freedom available,
24 R square is the index measuring the goodness of fit
25 of the model.
171
1 Q. What do you mean when you say, "depending
2 on degrees of freedom available"?
3 A. It's necessary to know what degrees of
4 freedom to assess the significance of the R square
5 value.
6 Q. If you had three degrees of freedom could
7 you judge how significant -- the significance of an R
8 squared value?
9 A. I'm trying to visualize a table of
10 significance levels and, no, I couldn't tell you
11 without looking it up in the table.
12 Q. Let me direct your attention to items 7 and
13 8 on the table. Number 7 is limits for the perimeter
14 stations and Number 8 is for the interior stations.
15 For the perimeter stations the R squared value is
16 71.6. For the interior stations the R squared value
17 is 55.1. Does this say anything to you about the
18 importance of station effects?
19 A. I'd have to study this document in order to
20 answer that question.
21 Q. Item Number 12 lists the limits for six
22 stations with the lowest geometric means. It's my
23 understanding that for the long term limit levels for
24 the Refuge Appendix E utilized the "3 clean"
25 stations, is that correct?
172
1 A. Correct.
2 Q. Do you remember any discussion of using six
3 clean stations?
4 A. No, I don't.
5 Q. Dr. Robson, is Dr. Walker a statistician?
6 A. No.
7 Q. What is his area of expertise, to your
8 knowledge?
9 A. My understanding would be he's a water
10 quality modeler.
11 Q. Dr. Robson, yesterday you testified that
12 you attended the deposition of Dr. James McClave, is
13 that correct?
14 A. Yes.
15 Q. Did you prepare any written report or
16 critique as a result of that McClave deposition?
17 A. No.
18 Q. Have you discussed the McClave deposition
19 with anyone at the District?
20 A. No, I don't recall that I have.
21 Q. I believe you also mentioned yesterday that
22 you were aware that Dr. McClave attempted to
23 determine whether there's a statistical correlation
24 between phosphorous inflows to the Refuge and
25 interior marsh concentrations, is that correct?
173
1 A. Yes.
2 Q. Did you review the results of Dr. McClave's
3 analysis?
4 A. I looked over the discovery material that
5 was provided at the deposition.
6 Q. Did you formulate any opinions as a result
7 of that review?
8 MR. MCGRATH: I just need to clarify
9 something for the record. Let me object to the
10 form. The term, "review" in the acronym of
11 research community is a term of art and to the
12 extent that Dr. Robson said he looked at, which
13 may be a common --
14 BY MR. PERKO:
15 Q. Did you develop any opinions as a result of
16 looking at the discovery materials produced by
17 Dr. McClave?
18 A. No. I don't have time enough to study it
19 in detail to develop opinions.
20 Q. Do you intend to do any additional work in
21 reviewing the discovery materials produced by
22 Dr. McClave?
23 A. Material I've already seen?
24 Q. Yes?
25 A. No.
174
1 Q. Do you have any opinion or do you recall
2 Dr. McClave stating that he did not find a
3 correlation between phosphorus and inflows to the
4 Refuge and interior marsh concentrations?
5 A. Yes.
6 Q. Do you have any opinion regarding that
7 conclusion?
8 MR. MCGRATH: I would just need to object
9 to the extent that he's got no basis and has
10 testified he's done really no review of work.
11 The question is calling for a speculative
12 conclusion rather than an opinion.
13 MR. PERKO: I'm asking if he has an
14 opinion.
15 THE WITNESS: I reserve my opinion on that
16 issue because I haven't studied it.
17 BY MR. PERKO:
18 Q. Do you also recall that Dr. McClave
19 questioned the sufficiency of the data for setting
20 phosphorus limits or levels?
21 A. Yes.
22 Q. Do you have an opinion regarding
23 Dr. McClave's concerns in that regard?
24 MR. MCGRATH: I need to object to this line
25 of questioning. You are basically asking one
175
1 expert's opinion of another expert's opinion.
2 There's a plethora of case law to say those are
3 improper questions. Again, also object to the
4 form to the extent that the witness has
5 testified he's done no background so there's
6 been no foundation established for that basis.
7 MR. PERKO: He hasn't testified he's done
8 no background regarding the sufficiency of the
9 data. He said he reviewed documents.
10 MR. MCGRATH: Said he looked at documents.
11 MR. PERKO: That's right. Whatever. I'm
12 asking him if he has any opinion regarding
13 Dr. McClave's conclusion that there was
14 insufficient data to set limits or levels.
15 MR. MCGRATH: Just so the record reflects --
16 MR. PERKO: I understand your objection.
17 MR. MCGRATH: -- lack of foundation.
18 THE WITNESS: My understanding is that we
19 are obliged to set levels based on the available
20 data.
21 BY MR. PERKO:
22 Q. If you were not obliged to use the data
23 available would you consider the data sufficient to
24 develop limits for the Park?
25 MR. FITZGERALD: Sorry. You said for the
176
1 Park?
2 MR. PERKO: Yes.
3 THE WITNESS: Well, I have a problem with
4 the term, "sufficient." A statistician always
5 prefers a larger sample size.
6 BY MR. PERKO:
7 Q. If you were not obliged to use the
8 available data to develop limits, would you consider
9 the data sufficient to derive limits -- reliable
10 limits for the Park?
11 A. I still have a problem with sufficiency.
12 Q. If you were not obliged to set limits based
13 on available data, simply asked if you would set
14 statistically reliable limits based on the data
15 that's available for the Park would you do that?
16 MR. FITZGERALD: Object to the question.
17 You start out telling him not to consider the
18 data that's available and then you say assume
19 that's the data. Could you clarify the
20 question?
21 MR. PERKO: He was not obliged.
22 MR. MCGRATH: I would object to the form of
23 the question as being somewhat incomprehensible.
24 I think if you had it read back you would
25 agree with me.
177
1 MR. PERKO: I'll withdraw the question.
2 (The document was marked Exb. No. 21.)
3 BY MR. PERKO:
4 Q. Dr. Robson, direct your attention to what's
5 been marked as Exhibit Number 21 to this deposition.
6 Do you recognize this document?
7 For the record, I'll represent that this is
8 a page out of the District's expert witness
9 designation in the pending administrative challenge
10 to the Everglades SWIM Plan.
11 A. Yes.
12 Q. Did you draft this portion -- the portion
13 pertaining to you of this document?
14 A. I don't believe so.
15 Q. Did you -- have you reviewed it before
16 today?
17 A. I've seen it.
18 Q. Did you provide any input to the District?
19 MR. MCGRATH: At what point in time,
20 counsel?
21 MR. PERKO: Prior to October 1992. October
22 26, 1992.
23 THE WITNESS: Input with respect to this?
24 MR. PERKO: As to what this contained --
25 what is contained in your expert witness
178
1 designation.
2 MR. FITZGERALD: Counsel, could you clarify
3 something for me? Are you asking did he provide
4 information to draft this narrative or had he
5 provided information to the District on the
6 underlying areas? It's not clear to me what
7 you're asking.
8 BY MR. PERKO:
9 Q. Let's start with the first question.
10 A. Well, I have to say I don't remember.
11 Q. Okay. Did you provide input on the
12 underlying areas that are indicated in the expert --
13 your expert witness designation?
14 A. I don't remember being asked to provide any
15 information like this.
16 Q. Okay. Under the heading, Subject Matter of
17 Expected Testimony, this document states that you're
18 expected to testify regarding "Review of statistical
19 methods utilized to develop phosphorous concentration
20 limits for inflows to the ENP and the Refuge."
21 Is that your understanding of the subject
22 matter of your expected testimony?
23 A. Not my understanding.
24 Q. What is your understanding of the subject
25 matter of your expected testimony?
179
1 A. Review the statistical analyses.
2 Q. Once again, I'm a little bit confused by
3 your distinction of statistical methods and
4 statistical analyses. Why are you making that
5 distinction?
6 MR. MCGRATH: I just point out for the
7 record that this was asked and was answered
8 yesterday.
9 MR. PERKO: I just want to make sure
10 there's no misunderstanding regarding what
11 Dr. Robson is expected to testify.
12 MR. MCGRATH: I understand. I'm just
13 noting it for the record.
14 THE WITNESS: Reviewing methods would be a
15 much broader scope.
16 BY MR. PERKO:
17 Q. How do you differentiate between methods
18 and analyses?
19 A. Analyses is a series of calculations
20 dictated by a method.
21 Q. Just so I understand you correctly, you do
22 not intend to provide testimony as to the
23 appropriateness of the statistical methods utilized
24 in Appendix E, is that correct?
25 A. The correctness of the statistical methods --
180
1 statistical analyses. Sorry.
2 Q. Would you consider the decision to utilize
3 a dummy variable in the equations in Appendix E a
4 decision regarding statistical methods or statistical
5 analysis?
6 A. That would be a decision based on an
7 analysis.
8 Q. Just so I understand you correctly, you
9 may, for example, provide testimony as to the
10 appropriateness of the decision to use a dummy
11 variable in Appendix E?
12 A. Yes.
13 Q. Next category, Dr. Robson, states that
14 under Substance of Facts and Opinions, "Expert will
15 support opinions expressed within and conclusions of
16 the SWIM Plan as to the matters described above."
17 Is that an accurate statement assuming we
18 substitute the phrase "statistical analysis" for
19 "statistical methods"?
20 A. Is there any legal jargon involved in this
21 statement?
22 MR. MCGRATH: Just pointing out, for the
23 record, I mean it's obvious that this is an
24 unknown situation. This was prepared -- my
25 understanding, it was prepared by counsel in
181
1 having to comply with all the various deadlines
2 and discovery, so to the extent that Dr. Robson
3 doesn't understand it, that may be quite
4 natural.
5 BY MR. PERKO:
6 Q. Let me ask you this, Dr. Robson. Is it
7 your understanding that you will be providing
8 testimony that will support opinions expressed in the
9 SWIM Plan?
10 A. Well, you're using the same words that are
11 used here. I still don't know whether this is legal
12 terminology when you say, "support opinions."
13 Q. What -- what opinions do you anticipate
14 providing at the final hearing in this matter --
15 MR. MCGRATH: I've got to object to the
16 form of that question --
17 BY MR. PERKO:
18 Q. -- at this time?
19 MR. MCGRATH: -- because essentially this
20 last day and a half of his testimony has been
21 eliciting the nature of his opinions.
22 Read the transcript. We'd be essentially
23 going over everything that you got from this
24 witness in the last day and a half.
25 BY MR. PERKO:
182
1 Q. Can you answer the question, Dr. Robson?
2 A. I don't really know what to expect.
3 Q. Do you anticipate testifying regarding any
4 matters that we have not discussed previously in this
5 deposition?
6 A. No.
7 (The document was marked Exb. No. 22.)
8 BY MR. PERKO:
9 Q. Dr. Robson, let me direct your attention to
10 what's been marked as Exhibit Number 22. I represent
11 for the record this is a print out of the index to
12 one of the disks that you provided in connection with
13 this deposition and it indicates that either a sector
14 anomaly encountered or cluster anomaly encountered
15 for a number of different files.
16 Did you copy the disks that you provided in
17 connection with this deposition or were they copied
18 by someone at the District?
19 MR. MCGRATH: Gary, just so I understand
20 the question, are you asking in the context of
21 copying for purposes of producing to you?
22 MR. PERKO: Yes.
23 MR. MCGRATH: Okay.
24 THE WITNESS: I made copies on my computer.
25 BY MR. PERKO:
183
1 Q. Do you know if all the disks were readable
2 when you sent them? Did you check to see if they
3 were readable?
4 A. I don't know if I checked them all or not.
5 Q. Okay.
6 MR. PERKO: Counsel, just for the record,
7 this was labeled disk number 11 and we simply
8 can't read it. If there is a copy of the disk
9 that we can read --
10 BY MR. PERKO:
11 Q. Let me ask you this, Dr. Robson. Do you
12 have a copy -- copies of the data provided in those
13 disks that you provided on your hard drive?
14 A. Yes.
15 Q. Okay. So --
16 A. Sorry. I have it on floppies.
17 Q. On floppies, okay.
18 MR. PERKO: Counsel, I just request if we
19 could, to make sure that we get a copy of disk
20 number 11 if, in deed, it is available.
21 MR. MCGRATH: Sure. As soon as this
22 deposition ends I'll take a look into it and see
23 what the story is.
24 MR. PERKO: I have no further questions.
25 Thank you, Dr. Robson.
184
1 (Thereupon, a recess was taken.)
2 CROSS (Douglas Robson, Ph.D.)
3 BY MR. BLANK:
4 Q. Dr. Robson, my name is Bob Blank. I am the
5 attorney for the Florida Sugar Cane League and I'm
6 here today representing the League and U.S. Sugar
7 Corporation and New Hope South, Inc. and, as
8 Mr. Perko stated, if I ask you any questions that you
9 don't understand, would you please just let me know
10 and I'll try and rephrase the question.
11 A. Yes.
12 Q. I'll try not to cover too much of the
13 ground that we've already covered, but I do have a
14 few questions for clarification purposes you've
15 already covered.
16 Going back to Exhibit 21 which is your
17 witness designation, I believe I understand the
18 distinction you're making between statistical methods
19 and statistical analyses.
20 Did you review the statistical methods that
21 were utilized in Appendix E to determine the
22 phosphorus limits for the Park and the Refuge?
23 A. I reviewed those that were used in Appendix
24 E.
25 Q. Did you review any other methods used to
185
1 calculate limits for the Park and the Refuge?
2 A. No. Not to calculate limits.
3 Q. Do you have an opinion as to whether the
4 methods utilized in Appendix E were the most
5 appropriate methods to be utilized in determining
6 limits for the Park and Refuge?
7 MR. MCGRATH: Let me make -- reiterate the
8 form objection when we're using the term, "most
9 appropriate."
10 THE WITNESS: I would have a problem with
11 that word also.
12 BY MR. BLANK:
13 Q. What's your problem with the term?
14 A. There may be many other methods one could
15 examine.
16 Q. Are there other methods that you would
17 consider to be more appropriate than the ones that
18 were utilized?
19 MR. MCGRATH: Did you ask if there -- did
20 he review other methods that are more
21 appropriate?
22 MR. BLANK: Would you read back the
23 question, please?
24 (Thereupon, a portion of the record
25 was read by the reporter.)
186
1 THE WITNESS: I've indicated that my
2 opinion was that the tolerance limit would be
3 more appropriate than the confidence limit.
4 BY MR. BLANK:
5 Q. Would that apply for both the Park and the
6 Refuge?
7 A. Yes.
8 Q. What is the statistical method called or
9 known as the weighted least squares?
10 A. It's a method in which individual
11 observations may be assigned different weights in
12 calculating the residuals from squares that is to be
13 minimized.
14 Q. Does this method account for differences in
15 sample sizes between sampling dates?
16 A. It could if they were random samples on
17 each date, if they were random samples.
18 Q. What do you mean by random samples?
19 A. If they represent observations that are
20 independent and identically distributed.
21 Q. Are you using the word, "random" in a
22 statistical sense?
23 A. Yes.
24 Q. Is the least squares the way -- is the
25 least squares method appropriate to use to fit a
187
1 regression model when the response variable itself is
2 an average?
3 A. It could be depending on how weights are
4 calculated.
5 Q. Did you consider using weighted least
6 squares to calculate limits for the Refuge?
7 A. No.
8 Q. Why not?
9 A. I didn't see any circumstance where that
10 would be appropriate.
11 Q. Have you ever been to the Refuge?
12 A. Once.
13 Q. Where did you go?
14 A. We took an airboat tour. This was back in
15 December '90 I guess it was.
16 Q. Was this part of your original orientation
17 meeting that you testified about --
18 A. Yes.
19 Q. -- before?
20 A. Yes.
21 Q. Did you go to the 16 sampling stations?
22 A. No.
23 Q. Why not?
24 A. I was a passenger on the airboat. I had to
25 go wherever it went.
188
1 Q. Did you ask to go to the 16 sampling
2 stations?
3 A. At that point I wasn't aware of the
4 existence of these stations.
5 Q. You previously testified concerning some
6 discussions you had with Dr. Federico on July 5th
7 concerning deletion of sampling dates and station
8 locations for the Refuge, is that correct?
9 A. Correct.
10 Q. What advice, if any, did you give to
11 Dr. Federico or the District as to whether stations
12 CA1 and CA2 should be eliminated when determining
13 limits for the Refuge?
14 A. I did point out that the sample size at
15 these two stations was substantially less than at the
16 other stations.
17 Q. Did you therefore advise them that those
18 two stations should be eliminated?
19 MR. MCGRATH: Just let me object to the
20 extent that his involvement in that decision
21 making process, there have been questions asked
22 and answers given on that matter, but I mean
23 Dr. Robson can answer the question if he can.
24 THE WITNESS: I wouldn't say I advised
25 them. I pointed out the unequal sample sizes.
189
1 BY MR. BLANK:
2 Q. Did you in any way object or express
3 concern with regard to eliminating those two stations
4 from the data base?
5 A. No, I did not.
6 Q. And what about elimination of the sampling
7 dates, 7903 and 7906, did you express concerns or
8 reservation about eliminating those sampling dates
9 from the Refuge data base?
10 A. No.
11 Q. Can you tell me who developed or selected
12 the methods that were utilized in Appendix E?
13 A. It was a combination of Walker and myself.
14 Q. Can you identify for me in general terms --
15 because I think we have covered this to some extent
16 in your prior testimony -- which methods are
17 attributable to Walker and which are attributable to
18 you?
19 A. Pages 48 on I --
20 Q. Are you referring to Exhibit 8, sir?
21 A. I'm sorry.
22 Well, it's from page E-25 onward I
23 attribute to myself.
24 Q. And the methods prior to page E-25?
25 A. Walker.
190
1 Q. Those methods you attribute to Dr. Walker?
2 A. Yes.
3 Q. I think you previously testified you had
4 two contracts with the District, is that correct?
5 A. Correct.
6 Q. Can you tell me what the terms of those
7 contracts were? Let's start with the first one.
8 MR. MCGRATH: It's been asked and answered.
9 MR. BLANK: Not in the way I'm asking it,
10 counsel.
11 THE WITNESS: By term you mean duration?
12 BY MR. BLANK:
13 Q. No. I mean what were the terms of your
14 employment?
15 A. No response.
16 Q. Would you like me to clarify the question?
17 A. No. I'm trying to visualize those
18 contracts and I'm having trouble. I don't remember.
19 Q. Do you remember by what method you were
20 paid? Were you employed on an hourly basis?
21 A. Yes.
22 Q. With regard to both contracts?
23 A. Yes.
24 Q. What was your hourly rate?
25 A. $75.
191
1 Q. Are you presently under contract with the
2 District?
3 A. Yes.
4 Q. At the same hourly rate?
5 A. Yes.
6 Q. The present contract that you have with the
7 District, can you tell me what the scope of that
8 contract is, what has the District asked you to do?
9 MR. MCGRATH: This also has been asked and
10 answered.
11 THE WITNESS: I'm to attend depositions,
12 prepare for.
13 MR. MCGRATH: Before you continue, just so
14 that I am clear, are you asking just the nature
15 of his understanding of what he was retained for
16 or are you asking the scope as it may have been
17 defined in a written contract?
18 MR. BLANK: Well, I think all he can
19 probably answer to is his understanding. I'm
20 asking what his understanding is of what he's
21 required to do under his existing contract.
22 THE WITNESS: My understanding is I am to
23 serve as an expert witness and to assist
24 attorneys at depositions.
25 BY MR. BLANK:
192
1 Q. Have you been requested to prepare any
2 written reports?
3 A. Not in the contract.
4 Q. When did this second contract commence?
5 A. I'm sorry. There's been an amendment to
6 the second contract, which is fairly recent. Second
7 contract itself came on in around May of '91.
8 Q. Okay.
9 A. Something like that.
10 Q. Since May of '91 have you prepared any
11 written reports concerning this case?
12 A. Yes, I have.
13 Q. And what are those reports?
14 A. I prepared a draft in Appendix E.
15 Q. Let me ask the question a different way.
16 Do they include any reports that we have not already
17 identified here today?
18 A. Not that I know of.
19 Q. In the context of the production of
20 documents in connection with your deposition did you
21 provide copies of all written reports or memoranda
22 that you would have prepared in connection with this
23 case?
24 A. All I could find.
25 Q. Were there some that you know of that you
193
1 didn't find?
2 A. Yes.
3 Q. What would those be?
4 A. The paper that I mentioned concerning
5 Seasonal Kendall Test and --
6 MR. MCGRATH: He's asking you for documents
7 and papers that you have prepared, so to the
8 extent if you're referring to another
9 publication, you've looked at, we're focusing
10 now on your work product and are there documents
11 or, you know, reports of your work product that
12 you know exist but you couldn't find and didn't
13 produce? That's what we're focusing on right
14 now.
15 MR. BLANK: Thank you, counsel.
16 THE WITNESS: No.
17 BY MR. BLANK:
18 Q. Are there any documents or reports that you
19 couldn't find that you will be relying on with regard
20 to your anticipated testimony in this case?
21 A. Again, that I prepared?
22 Q. No. That anyone prepared.
23 A. That published paper that I mentioned
24 earlier.
25 Q. Is that the paper -- excuse me. Which one
194
1 was that?
2 A. That was concerning the impact of serial
3 correlation on the Seasonal Kendall Test.
4 Q. This was the one by Hirsch I believe it
5 was?
6 A. Hirsch was one of the authors.
7 Q. Okay. Can you tell me what the detection
8 limits were for total phosphorus during the period of
9 records for the Everglades National Park?
10 A. Started out at two and ended up at four.
11 Q. Parts per billion?
12 A. Yeah.
13 Q. Do you know when the change occurred?
14 A. I don't recall.
15 Q. Do you recall how values at or below the
16 detection limit were treated in the analysis of
17 phosphorus limits for the Park?
18 A. They were set to four.
19 Q. Who made that decision?
20 A. Walker.
21 Q. Did you agree with that decision?
22 A. I don't have a better alternative.
23 Q. Was there a change in the detection limits
24 for total phosphorus during the period of record for
25 the Refuge?
195
1 A. I don't recall when that change took place.
2 I don't know.
3 Q. Do you recall if any of the samples of
4 phosphorus concentration values and the samples taken
5 for the Refuge were adjusted in your analysis to the
6 detection limit?
7 A. I seem to recall that there were some at
8 four.
9 Q. Did you ever review the field data or the
10 raw data for the Refuge?
11 A. I was asked that question yesterday and my
12 response was I couldn't recall whether I had been
13 provided with raw data or with data that appears in
14 Appendix E.
15 Q. So your answer is you don't recall?
16 A. Yes.
17 Q. Can you tell me what are the sources for
18 phosphorus input into the Refuge?
19 A. I don't feel I'm qualified to identify all
20 sources.
21 Q. What sources do you know of?
22 A. There are the inflow structures.
23 Q. Which structures are you referring to?
24 A. I could make a mistake in identifying these
25 structures. I didn't analyze the inflow data. I'll
196
1 guess S5. I guess there's other ones.
2 Q. Is your problem that you just don't recall
3 the nomenclature used to designate those structures?
4 A. I don't remember the structures themselves.
5 Q. You don't remember their names or you don't
6 remember seeing those structures?
7 A. I probably saw them during the helicopter
8 tour during that orientation.
9 Q. Was it your understanding when you were
10 working on Appendix E that the limits being
11 established for the Refuge were to be utilized to
12 regulate inflows from District structures into the
13 Refuge?
14 A. Yes.
15 Q. Did you ever request from the District or
16 anyone else the inflow -- the data from the inflow
17 structures?
18 A. No.
19 Q. Why not?
20 A. I had the understanding that Walker had
21 examined the data and found no correlation in the
22 particular analyses that he did and, in the time
23 schedule I was working under, I would not have had
24 time myself to pursue that.
25 Q. What was the time schedule you were working
197
1 under?
2 A. The date that is paramount in my mind is
3 July 5th, that my analyses had to be essentially
4 completed by July 5th.
5 Q. And when did your analyses commence?
6 A. Again, you're referring to the Refuge?
7 Q. Yes, sir.
8 A. I'm vague on that. I first worked on the
9 Park and I'm not sure exactly when I started working
10 on the Refuge data.
11 Q. Well, I think you just testified that you
12 were working under some time pressure or abbreviated
13 schedule that had to be completed by July 5th. Can
14 you tell me approximately how much time prior to July
15 5th you commenced this work?
16 A. I would guess approximately a month.
17 Q. And your prior testimony was that you
18 understood that Dr. Walker attempted to determine a
19 correlation between inflows into the Refuge and
20 phosphorus concentration levels at the interior marsh
21 stations?
22 A. I have a vague recollection of that, yes.
23 Q. Did you ever review any of Dr. Walker's
24 work in that regard?
25 A. No, I did not.
198
1 Q. Can you tell me who told you that
2 Dr. Walker had attempted to determine a correlation?
3 A. My recollection is that this came up during
4 one of those meetings that I attended.
5 Q. As a statistician were you concerned or
6 troubled by the fact that Dr. Walker wasn't able to
7 find a correlation knowing that these interior marsh
8 limits were going to be used to regulate inflows?
9 MR. MCGRATH: Let me just object to the
10 form of that question in so far as it may
11 mischaracterize testimony which is of record
12 because I don't recall there being any testimony
13 regarding this witness knowing that Dr. Walker
14 did not or could not correlate inflows.
15 MR. BLANK: I don't understand your
16 objection, counsel.
17 MR. MCGRATH: Well, your question is based
18 on this witness' knowledge that there was not a
19 correlation or that he knew that someone else
20 had determined there was not a correlation and I
21 am not aware of anything on the record that
22 establishes that fact which is the basis for
23 your question.
24 MR. BLANK: Do you understand the question --
25 MR. MCGRATH: I'm objecting to the form.
199
1 THE WITNESS: I think so.
2 MR. BLANK: -- Dr. Robson?
3 THE WITNESS: It is an issue that I would
4 pursue if I had the time.
5 BY MR. BLANK:
6 Q. If you determined that there is no
7 correlation or let me -- I'm going to ask you two
8 questions now.
9 If you would determine that there is no
10 correlation between inflows into the Refuge and the
11 phosphorus concentration levels in the interior marsh
12 would you consider it improper to utilize the interim
13 or the long term limits set in Appendix E as a basis
14 of regulating inflows?
15 MR. MCGRATH: Just for the record, let me
16 object to the form, number one, to the extent
17 that you're asking the witness to speculate.
18 Number two, to whatever extent you used the
19 term, "improper."
20 THE WITNESS: I would have second thoughts
21 including wondering about the quality of the
22 data involved.
23 MR. BLANK: Well, Doctor, that's something
24 we all wonder about.
25 MR. FITZGERALD: Move to strike.
200
1 BY MR. BLANK:
2 Q. When you say, "wondering about the quality
3 of the data involved," are you referring to the data
4 for the Refuge?
5 A. What I had in mind was the data on the
6 inflow.
7 Q. Why would you be concerned about the
8 quality of that data?
9 A. If I could find no correlation, then, well
10 both data would be somewhat suspect.
11 Q. Do you have any concerns regarding the
12 quality of the data used in Appendix E to establish
13 the marsh limits for the Refuge?
14 A. My understanding is that we're obliged to
15 set limits based upon the data that are available.
16 There is nothing we can do now about the quality of
17 the data.
18 Q. I don't think your answer was responsive to
19 my question, Doctor.
20 Do you have concerns regarding the quality
21 of the data?
22 A. Well, I thought I answered your question.
23 I'll have to say no, I don't have concerns.
24 Q. What steps did you take to insure that the
25 data that you analyzed for the Refuge was, in fact,
201
1 good data?
2 A. By good data, you mean no errors in data
3 entries?
4 Q. Yes. Or truly representative of the
5 concentration levels that existed when the samples
6 were taken.
7 A. I would have no way of knowing that.
8 Q. Did you inquire concerning the methodology
9 that was utilized to collect the samples?
10 A. Yes.
11 Q. Who did you inquire of?
12 A. Federico.
13 Q. What did you ask him?
14 A. What platform was used in collecting the
15 samples.
16 Q. Excuse me?
17 A. What platform was used in collecting the
18 samples.
19 Q. What do you mean by platform?
20 A. Boat or airboat or helicopter.
21 Q. What did he tell you?
22 A. Helicopter.
23 Q. What else did Dr. Federico tell you
24 concerning the methodology of collection?
25 A. He explained that a bucket was lowered from
202
1 the helicopter to collect the samples. Is that what
2 you --
3 Q. Yes.
4 Did he tell you anything else?
5 A. He did, but I don't remember.
6 Q. Was this communication you had with
7 Dr. Federico verbal?
8 A. Yes.
9 Q. You never received a written report from
10 the District concerning the methodology of collection
11 for the Refuge?
12 A. No, I did not.
13 Q. Did you ever request one?
14 A. No.
15 Q. Was it your understanding, then, that all
16 samples utilized in Appendix E were collected from a
17 helicopter?
18 A. Not necessarily all samples.
19 Q. Could you identify which ones were not?
20 A. No.
21 MR. MCGRATH: With respect to that
22 question, I just want to clarify. Are you
23 talking about with respect to the Refuge in
24 Appendix E or talking about Appendix E as a
25 whole?
203
1 MR. BLANK: All of my questions now are
2 directed to the Refuge.
3 MR. MCGRATH: Okay.
4 Thank you.
5 THE WITNESS: That was my understanding.
6 BY MR. BLANK:
7 Q. Excuse me. Now I am confused.
8 A. Talking about the Refuge.
9 Q. Was it your understanding that all samples
10 for the Refuge were collected from a helicopter?
11 A. Not necessarily, no.
12 Q. But you cannot identify which ones may not
13 have been collected from a helicopter?
14 A. I don't know for sure, no.
15 Q. Do you know whether the helicopter landed
16 prior to collecting the sample?
17 A. No.
18 Q. I'm talking about landed in the Refuge.
19 A. Yeah. No, I don't know.
20 Q. Do you believe that Dr. Federico told you
21 that the samples were collected by lowering a bucket
22 from the helicopter?
23 A. Yeah.
24 Q. What effect would the problem of rotor wash
25 of the helicopter have on the sample?
204
1 A. I don't know.
2 Q. Wouldn't that tend to contaminate the
3 sample?
4 MR. MCGRATH: Let me object to the form to
5 the extent it's outside --
6 MR. BLANK: He's just testifying of his own
7 knowledge, counsel.
8 MR. MCGRATH: I realize that. But,
9 obviously, any opinion by a statistician on the
10 subject matter of the question would be
11 speculation.
12 THE WITNESS: I don't know.
13 BY MR. BLANK:
14 Q. Was it your understanding that the 16
15 sampling stations were permanent locations?
16 A. Yes.
17 Q. And was it also your understanding the
18 helicopter returned to the same location each time it
19 sampled a station?
20 A. Yes.
21 Q. How was the station located?
22 A. I don't know.
23 Q. Do you know if those stations were marked
24 during the sampling period?
25 A. I don't know.
205
1 Q. If they were not would that introduce more
2 uncertainty or variability into your analysis of the
3 limits for the Refuge?
4 A. It would depend on accuracy with which they
5 did return to the same location and upon the
6 magnitude of the local variation.
7 Q. Didn't the model or the analysis used in
8 Appendix E for the Refuge assume permanence?
9 A. Yes.
10 Q. If they did not return to the same site
11 each time they sampled would that assumption be
12 valid?
13 A. Yes. If the location is that target area,
14 I think you return to the same target area.
15 Q. What do you mean by target area?
16 A. I don't know.
17 Q. Well, I don't understand your answer.
18 Could you explain it a little further?
19 A. How large, what the precision is and their
20 ability to return to exactly the same point.
21 MR. MCGRATH: Let me object to the question
22 as being vague and to the extent that you're
23 referring to returning to the sampling site or
24 not returning to the sampling site, I mean what
25 degree of -- are we talking about missing it by
206
1 a foot, by a mile?
2 MR. BLANK: That's what we're getting ready
3 to get into, counsel.
4 MR. MCGRATH: Okay.
5 MR. BLANK: I think my question was phrased
6 in the term, "precise."
7 MR. MCGRATH: Precise is not necessarily a
8 well defined precise term.
9 BY MR. BLANK:
10 Q. Well, what we're dealing here with, Doctor,
11 are we not, is spatial variability if, in fact, the
12 samples were not taken at precisely the same
13 location?
14 A. Yes.
15 Q. Do you have any opinion as to the distance
16 from the original sampling location that would
17 introduce spatial variability for additional samples?
18 A. No, I don't.
19 Q. How would you determine that?
20 A. I couldn't determine it from the existing
21 data.
22 Are you asking me how I would design an
23 experiment?
24 Q. Yes.
25 A. Well, this would involve taking clusters of
207
1 point samples at a number of locations.
2 Q. In the Refuge?
3 A. Yes.
4 Q. Would it also necessitate being able to
5 determine at this time the precise location of the
6 original 16 sampling stations?
7 A. No. I don't think that's necessary.
8 Q. Why not?
9 A. Because whatever spatial variation there is
10 around the target location will simply go into the
11 error variance and have the effect of increasing the
12 error variance.
13 Q. The error variance of what?
14 A. The residual variance around the fitted
15 model.
16 Q. Are you talking about the existing model?
17 A. Any model.
18 Q. Well, Appendix E establishes 14 compliance
19 stations for the Refuge, does it not?
20 A. Yes.
21 Q. As a statistician would you feel
22 comfortable with establishing compliance stations
23 that cannot be precisely located?
24 A. I would only say that in future the same
25 level of precision should be used in locating
208
1 stations that's been used during the POR.
2 Q. Do you know what that level of precision
3 was?
4 A. No.
5 Q. I think you previously testified that you
6 had never been to the sampling stations.
7 A. Right.
8 Q. Do you know what type of hydrological
9 variability exist between those stations?
10 A. I have seen maps, but I haven't retained
11 that in my mind.
12 Q. Is your answer therefore no or you don't
13 recall?
14 A. I don't recall.
15 Q. What type of maps did you see?
16 A. I believe they were contour maps.
17 Q. Have you seen any maps other than contour
18 maps?
19 A. I don't believe so.
20 Q. Do you know how the regulatory stage for
21 the Refuge is established?
22 A. I don't know. I think I answered that
23 question yesterday and I left it up to the judgment
24 of the subject matter specialists.
25 Q. Would you look at Exhibit 8, Table 6? It's
209
1 on page E-18. Do you see that table I'm talking
2 about?
3 A. I do.
4 Q. Can you tell me why missing values were
5 coded as 0?
6 A. That was a decision I made simply because
7 it was easy for my computer to recognize the cells
8 that contained 0.
9 Q. Would you then look at Table 9 on E-22?
10 A. Yes.
11 Q. Why were missing values coded as 1 on that
12 table?
13 A. Because on calculating geometric means I'm
14 taking a product of the entries across the row.
15 Q. Again, that was a decision that you made?
16 A. Yes.
17 Q. With regard to your 0 coding on Table 6 is
18 it possible that in any of your analyses that a
19 missing value could have been read as 0 instead of as
20 a missing value?
21 A. No.
22 Q. Why not?
23 A. My analyses were all run on data sets using
24 the 1 as the --
25 Q. You did not run any analyses using all 16
210
1 stations and all 16 dates?
2 A. Yes. But in those analyses the 0s were
3 replaced by 1s.
4 Q. You are certain of that?
5 A. Yes.
6 Q. Did you provide us with copies of those
7 analyses?
8 A. They would be in floppies that I provided
9 and any print outs I could find would have been
10 included.
11 Q. Okay.
12 MR. BLANK: Why don't we take a lunch
13 break.
14 (Thereupon, a recess was taken.)
15 MR. BLANK: Could you read back the last
16 question and answer?
17 (Thereupon, a portion of the record
18 was read by the reporter.)
19 BY MR. BLANK:
20 Q. Dr. Robson, would you take a look back at
21 Table 9 on page E-22?
22 MR. MCGRATH: Exhibit Number 8?
23 MR. BLANK: Exhibit Number 8.
24 BY MR. BLANK:
25 Q. Do you see that table?
211
1 A. Yes.
2 Q. Can you tell me what accounts for the
3 missing values on Table 9?
4 A. Well, you can't see it without counting the
5 number of missing values in a row, but let's see. In
6 the first column the -- there's one missing value, so
7 that mean of 13.5 listed for the first column
8 station, geometric mean 13.5 would have been
9 calculated by taking the product of all 14 or the sum
10 of the logs of all 14 and in the case of some of the
11 logs, then you would divide by 13 instead of 14
12 because there was one missing value.
13 Q. All right. But do you know why that value
14 was missing?
15 A. I don't know.
16 Q. Did you make any efforts to inquire of the
17 District as to why that value was missing?
18 A. Not that specific value, no.
19 Q. Any of the values on Table 9?
20 A. It was my understanding that there were
21 sampling dates when no samples were taken from
22 certain stations.
23 Q. I'm sorry. Could you repeat your answer?
24 A. It was my understanding that there was
25 sampling dates in which no sample was taken from some
212
1 stations.
2 Q. Do you know why no sample was taken --
3 A. No, I don't.
4 Q. -- from those stations?
5 A. No.
6 Q. Do you know if field notes or log books
7 were kept by the individuals that collected the
8 samples?
9 A. I don't know.
10 Q. Did you ever ask to see those?
11 A. No, I didn't.
12 Q. Can you tell me why a ten percent rejection
13 limit was selected for the Refuge?
14 A. It was my understanding that this was
15 agreed upon by the parties concerned.
16 Q. Who did you receive that understanding
17 from?
18 A. Well, that understanding was based on the
19 fact that I heard no objections to the use of the ten
20 percent level.
21 Q. Who suggested using the ten percent level?
22 A. I'm not sure.
23 Q. When did you first become aware that a ten
24 percent level was being suggested be utilized?
25 A. Again, I'm not sure. Probably in May or
213
1 June.
2 Q. Did you conduct any analysis to determine
3 whether the ten percent rejection level was
4 appropriate to utilize for the Refuge?
5 A. I wouldn't know how to conduct such
6 analysis.
7 Q. So your answer is no?
8 A. Yes.
9 Q. On the same exhibit, Doctor, would you turn
10 to page E-16. In the second sentence of the first
11 paragraph, it states, "Sampling at these permanent
12 stations was initially bimonthly but became sparser
13 later in this POR as seen in the data Table 6."
14 Do you see that sentence?
15 A. Yes.
16 Q. Can you tell me why sampling at the 16
17 stations was initially bimonthly and then reduced?
18 A. I don't know why.
19 Q. Did you inquire?
20 A. I don't recall inquiring.
21 Q. In the regression model used for the
22 Loxahatchee limits, is it correct that the only
23 predictor variables are stage and an indicator
24 variable distinguishing the base period from later
25 dates?
214
1 A. You used the term, "predictor variable"?
2 Q. Yes, sir.
3 A. Those are the only dependent variables that
4 were used, yes.
5 Q. Does the model therefore assume there is no
6 year to year variability that is not accounted for by
7 stage?
8 A. No.
9 Q. Would you explain your answer, please.
10 A. We have the base year variable. That
11 accounts for differences between base year and
12 subsequent years.
13 Q. What about year to year variability after
14 the base period?
15 A. Any variability after the base period
16 that's not explained by stage goes into the residual
17 variance.
18 Q. What do you mean by residual variance?
19 A. It means squared area. That's measure of
20 the magnitude from the departure from the fitted
21 model observed by this fitted square and averaged.
22 Q. Going back to the permanence of the 16
23 interior marsh stations do you understand that the
24 model for the Refuge does not provide an estimate of
25 the average phosphorus concentration in the entire
215
1 refuge?
2 A. Yes.
3 Q. Could these stations be considered random
4 effects if you wanted to use the Refuge data to
5 estimate average phosphorus concentration in the
6 entire Refuge?
7 A. I doubt it. I don't know how the 16
8 stations were originally selected.
9 Q. Do you know who selected them?
10 A. No.
11 Q. What additional data do you feel would be
12 required if you wanted to estimate the average
13 phosphorus concentration in the entire Refuge?
14 A. The probability sample of locations in
15 which to take samples.
16 Q. How many additional samples would you feel
17 would be necessary?
18 A. I don't have a feeling. The more you take
19 the more precise you get.
20 Q. Do you know of any statistical method or
21 test that could be utilized to determine the
22 appropriate sample size in order to adequately model
23 the entire Refuge?
24 A. I don't know what you mean by model. I
25 don't know what model you're referring to.
216
1 Q. If you were to prepare or create a model in
2 order to estimate or establish average phosphorus
3 concentration throughout the Refuge, is there a
4 statistical method or test that you would utilize to
5 determine the adequacy of the sample size for that
6 purpose?
7 A. Adequacy is an undefined term. I'd have to
8 say no.
9 Q. Okay. I believe you answered this question
10 before, but I'd like to ask it again, just to make
11 certain that I understand your answer.
12 Did you consider deriving separate models
13 for each station in the Refuge?
14 A. I analyzed the model in which I included
15 station effects.
16 Q. But did you try fitting station specific
17 models?
18 A. I'm sorry. Again, I'm not sure what you
19 mean by -- I fit a model with certain specific
20 effects in the model.
21 Q. But you did not develop a model for each
22 station as I understand it.
23 A. That's correct. I did not.
24 Q. On page E-17 the sentence at the top of the
25 page, "Reasons for missing values in the POR are
217
1 unknown at this time but may have resulted from a
2 lack of water to sample at a station due to low stage
3 in the marsh."
4 Do you see that sentence?
5 A. Yes.
6 Q. Did you write that sentence?
7 A. I did.
8 Q. Are the reasons now known?
9 A. I don't know.
10 Q. Does that mean they are now -- they still
11 remain unknown?
12 A. They are unknown to me.
13 Q. Okay. That's what I was asking you.
14 Do you have any knowledge that station
15 CA1-1 and CA2 are more difficult to sample for some
16 reason at low stages than the other 14 stations?
17 A. I'm not sure.
18 Q. Are you saying you don't know?
19 A. Yes.
20 Q. Can you explain why there's a missing
21 observation at station CA-1 in March of '79 when
22 stage was at 15.21 of a value of stage that you
23 consider to be too low to be able to include in
24 observations from this date in your model?
25 A. That was not my understanding of the reason
218
1 for establishing that threshold of 15.42.
2 Q. Isn't that what the first sentence of the
3 paragraph on E-17 says?
4 A. Well, no, it doesn't say that 15.42 is the
5 stage at which --
6 Q. Where are you getting 15.42?
7 A. Toward the bottom of that paragraph.
8 Q. Well, I was asking about 15.21.
9 A. But your question I thought pertained to
10 the 15.42 -- 15.21 being less.
11 Q. That's correct. Less than your 15.42
12 threshold.
13 A. Right. But my understanding of that 15.42
14 threshold was not -- that stages below that were too
15 low to take a sample.
16 Q. I'm not sure that still answers my
17 question, but what I'm trying to understand is why we
18 had a value at station 2 on a date when we had a
19 stage of 15.21 whereas we had missing values on other
20 dates when the stage was higher.
21 A. I don't know why.
22 Q. Okay. Did you inquire?
23 A. No.
24 Q. The next sentence on page E-17, "The POR
25 data include stage measurement of three marsh gauges
219
1 CA1-7, 9 and 8C on each sampling date and the average
2 of these three readings is taken as an index of stage
3 for the marsh on that date."
4 Do you see that sentence?
5 A. Yes.
6 Q. Did you write that sentence?
7 A. Yes.
8 Q. Do you have any knowledge of how the stage
9 index you're referring to there relates to the actual
10 stage at all 16 or 14 marsh stations?
11 A. I can't know because I've never seen the
12 stages at these other stations.
13 Q. Do you have any data at all that would
14 indicate the amount of variability in stage at each
15 station on any given sampling date?
16 A. No, I do not.
17 Q. Do you know if any such data exists?
18 A. I don't know.
19 Q. Did you inquire as to whether any such data
20 existed?
21 A. No.
22 Q. Do you have any idea of a possible bias of
23 index of stage when used to estimate the average
24 stage in the whole marsh area?
25 MR. MCGRATH: Object. Asking for
220
1 speculation to the extent you're looking for
2 this witness to testify about possibilities.
3 MR. FITZGERALD: Could I get the question
4 read back?
5 (Thereupon, a portion of the record
6 was read by the reporter.)
7 THE WITNESS: Well, I'm not sure what you
8 mean by bias. An index is intended to be
9 proportional to, not equal to.
10 BY MR. BLANK:
11 Q. You don't understand the question?
12 A. No, I don't.
13 Q. What I'm asking you is do you think there
14 is any bias that results from using the index of
15 stage the way it's been used in the model for the
16 Loxahatchee?
17 A. I still have a problem with the term,
18 "bias." The average for these three gauges is not
19 intended to represent the average for the entire
20 marsh. As an index it's suppose to be proportional
21 to and the proportional factor is unknown.
22 Q. All right. How about when it's used to
23 estimate just the average stage of each of the 16 or
24 14 stations?
25 A. Same. Same argument applies.
221
1 Q. It's not used for that purpose?
2 A. It's used for that purpose as an index.
3 It's intended to be proportional to, but not equal to
4 the average for those 16 stations.
5 Q. Okay. Would you look at the next sentence
6 on E-17. "Excessively low stage introduces potential
7 biases in the field process of collecting a water
8 sample from a helicopter while also impairing
9 precision through missing values."
10 Did you write that sentence?
11 A. Yes.
12 Q. How does a low stage produce a missing
13 value?
14 A. Well, it was my understanding that at a
15 sufficiently low stage it was not possible to collect
16 a water sample from a helicopter.
17 Q. From which stations?
18 A. From the stations where the stage is
19 lowest.
20 Q. Do you know which ones they are?
21 A. No.
22 Q. Going back to the stage index, since the
23 three stations are an index, is it important to use
24 exactly or precisely the same stations to measure the
25 stage index when determining compliance?
222
1 A. Yes.
2 Q. Would the result be biased if these
3 stations were changed?
4 MR. FITZGERALD: Counsel, you are asking
5 about the locations of CA1-7, 9, 8C, those
6 stations?
7 MR. BLANK: Correct. The stage.
8 MR. FITZGERALD: Thank you.
9 THE WITNESS: Yes. That would produce a
10 bias.
11 BY MR. BLANK:
12 Q. The next sentence, "This exclusion is
13 implemented hereby establishing a threshold stage of
14 15.42 feet MSL per acceptance of five sampling dates
15 in the POR and in the future compliance monitoring
16 record."
17 Do you see that sentence?
18 A. Yes.
19 Q. Did you write that sentence?
20 A. Yes.
21 Q. How was that threshold stage of 15.42
22 determined?
23 MR. MCGRATH: Again, I believe this was
24 asked and answered but the witness can answer
25 it.
223
1 THE WITNESS: I don't know.
2 BY MR. BLANK:
3 Q. Would you turn to page E-50?
4 A. Five-oh?
5 Q. Yes, E-five-oh.
6 And read the first paragraph on that page.
7 A. Okay.
8 Q. Did you read that paragraph?
9 A. Yes.
10 Q. Can you now tell me where the 15.42 came
11 from?
12 A. No.
13 Q. Thank you. Just wanted to make sure I
14 wasn't missing something.
15 On page E-18, Doctor, Table 6, would you
16 look at observations for station 7 on February of
17 '80?
18 A. I'm sorry. Which station was that?
19 Q. 7, CA1-7. Do you see the value there, 96?
20 A. Yes, I do.
21 Q. Does that correspond to February of '80?
22 A. Right.
23 Q. And would you also look at the value for
24 CA1-9, station 9 on June of '80?
25 MR. FITZGERALD: Counselor, can you tell me
224
1 where you are finding a June '80 sampling event?
2 MR. MCGRATH: Yeah. I can't find it.
3 MR. FITZGERALD: I can't find a June '80.
4 MR. BLANK: You can find it in a couple of
5 places but I think the important thing is
6 whether the witness can find it.
7 THE WITNESS: No, I can't.
8 MR. FITZGERALD: June?
9 BY MR. BLANK:
10 Q. Did I say June?
11 A. Yes.
12 Q. I meant July.
13 Thank you.
14 MR. FITZGERALD: There was no trick to it.
15 I couldn't find it.
16 BY MR. BLANK:
17 Q. July of '80.
18 A. Yes.
19 Q. Do you see the value 169?
20 A. Yes.
21 Q. And look at station 12 on 9-80.
22 A. Yes.
23 Q. Do you see the value 494?
24 A. Yes.
25 Q. And station 15 on 9-80?
225
1 A. 121.
2 Q. Yes.
3 Do you have any explanation as to why these
4 numbers appear so large both relative to the same
5 station on different sampling dates and relative to
6 other stations on the same sampling date?
7 A. I don't know why.
8 Q. Did you inquire as to why --
9 A. No, I did not.
10 Q. -- that was?
11 Turning to page E-20, their regression
12 models. Was a weighted regression used to account
13 for the fact that the response variable is an average
14 based on a sample size that was not constant over all
15 sampling dates?
16 A. No.
17 Q. Why was a step change chosen to model an
18 increase in phosphorus levels?
19 A. This was incorporated simply as a contrast
20 between base year and subsequent years.
21 Q. And, again as I understand it, that wasn't
22 your decision to do it in that fashion, is that
23 correct?
24 A. That's correct.
25 Q. Do you know of any event in the Refuge
226
1 system that would have caused a step change in
2 phosphorus levels in the marsh between the base
3 period and later dates?
4 A. No.
5 Q. Did you consider other models that used a
6 linear or quadratic term to model increasing
7 phosphorus concentrations?
8 A. Could you read the question again?
9 (Thereupon, a portion of the record
10 was read by the reporter.)
11 THE WITNESS: You mean linear quadratic in
12 time?
13 MR. BLANK: Yes.
14 THE WITNESS: I didn't, no.
15 BY MR. BLANK:
16 Q. Do you know if anyone did?
17 A. It's my understanding that Walker did.
18 Q. Did you see the results of his analysis?
19 A. I believe I saw them on the screen of his
20 lap top computer.
21 Q. Did you request a copy so that you could
22 analyze it?
23 A. No.
24 Q. Why not?
25 A. Well, he was running the same regression
227
1 program that I was running, so I would get the same
2 results that he got.
3 Q. Is it true that the model that you used
4 with regard to the Refuge assumed the slope of the
5 line relating phosphorus concentration to stage is
6 the same in both time periods?
7 A. That would be one way of putting it.
8 Q. Is your answer yes?
9 A. Pardon me?
10 Q. Is your answer yes?
11 A. Yes.
12 Q. Did you look at other models that included
13 an interaction term between time periods and stage?
14 A. No, I did not.
15 Q. Why not?
16 A. It -- I'm not sure what you mean by time
17 period.
18 Q. I'm talking about the difference between
19 the base year and subsequent years.
20 A. It didn't occur to me to do so.
21 Q. Would you turn to page E-3, same exhibit
22 and look at the last sentence on that page, the
23 sentence that reads, "This rule allows a ten percent
24 risk of false rejection in contrast to the usual five
25 percent or one percent risk allowed in scientific
228
1 practice but the larger the risk of false rejection
2 the smaller the risk of erroneously accepting the
3 baseline hypothesis."
4 Do you see that sentence?
5 A. Yes.
6 Q. Does that sentence apply equally to the
7 model used for the Refuge?
8 A. Yes.
9 Q. Do you know, sir, how often sampling will
10 occur to determine whether the average phosphorus
11 concentration in the marsh is above what it would
12 have been in the base period?
13 A. I don't remember.
14 Q. Did you consider any methods from the field
15 of statistical procession control to set a limit for
16 phosphorus and the action that should be taken if
17 limit is exceeded?
18 A. No.
19 Q. Why not?
20 A. If the limit is exceeded, then that
21 triggers action of a committee of experts.
22 Q. Is it your understanding, sir, that if the
23 limit is exceeded the action that is taken will not
24 be enforcement?
25 A. We're getting outside my scope now.
229
1 Q. Well, you can simply say you don't know if
2 that's your answer.
3 A. Okay. I don't know.
4 Q. Did you perform any calculations to compute
5 the costs of making a Type I or Type II error?
6 A. No.
7 Q. Why not?
8 A. I wouldn't know the costs involved.
9 Q. Would you assume just for the moment that
10 there is no serial correlation between sampling dates
11 and assuming that the model for the Refuge is
12 adequate to describe the system and assuming that the
13 phosphorus concentrations behave as if they were in
14 the base period, can you tell me the probability or
15 do you know what the probability is of at least one
16 geometric mean being above the interim limit over a
17 period of one year?
18 A. This would depend on the frequency of
19 sampling during the -- one of the conditions under
20 the conditions you specified that's calculable using
21 binomial theorem.
22 Q. Do you know what the formula would be for
23 calculating it?
24 A. 1 minus .9 of the power end.
25 Q. Did you say 29?
230
1 A. .9.
2 Q. .9. okay. Thank you.
3 Turn to page -- I think we're already there
4 on E-20, are we not?
5 MR. FITZGERALD: We're on E-3.
6 MR. BLANK: We're back on E-20 now.
7 BY MR. BLANK:
8 Q. I'm looking at the last sentence of the
9 last full paragraph, the sentence which reads, "These
10 "clean 3" were adopted as representing the cleanest
11 TP condition during the POR."
12 Do you see that sentence?
13 A. Yes.
14 Q. Did you write that sentence?
15 A. Yes.
16 Q. Assuming that the marsh is impacted by
17 inputs from stations 5A and 6, the District
18 structures we referred to before, do you know of any
19 geographical and/or hydrological reasons why stations
20 5, 6 and 16 would be the "cleanest" stations out of
21 the 14 considered?
22 A. No, I don't.
23 Q. How do you know that the difference in
24 phosphorus concentrations between the three "clean"
25 stations and the other 11 stations isn't due to some
231
1 factor other than inputs from 5A and 6?
2 A. I don't know.
3 Q. Assume for a moment, no, absolutely no
4 phosphorus inputs from 5A and 6. Is it possible that
5 there is a high probability that the 14 station
6 average will exceed the long term limit?
7 MR. MCGRATH: Let me just object in calling
8 for speculation, asking the witness to testify
9 as to what's possible.
10 MR. BLANK: You can scratch the possible.
11 Is there a high probability that the 14
12 station average will exceed the long term limit?
13 MR. FITZGERALD: That's assuming no inflow?
14 MR. BLANK: No inflows from 5 and 6.
15 THE WITNESS: Are you going to specify a
16 period of time during what we're talking about?
17 BY MR. BLANK:
18 Q. You can pick whatever period of time you
19 like, Doctor.
20 A. If you pick a long enough period then the
21 probability will be high.
22 Q. How long would be long enough?
23 A. Well, you would apply the 1 minus .9 at the
24 power end.
25 Q. Okay. Inasmuch as you would be comparing a
232
1 14 station average to a 3 station average, wouldn't
2 it be even greater than 1 minus .9 of the power end?
3 A. You haven't included an assumption about
4 how these 14 stations -- how these other 11 behave
5 relative to the 3 and you gave me a lot of
6 assumptions to begin with.
7 Q. Do you want the question repeated?
8 A. No. I want to specify the assumption that --
9 I can't answer the question without your having
10 specified an assumption concerning how these other 11
11 behave relative to the 3.
12 Q. What type of assumption would you need?
13 A. An assumption of homogeneity.
14 Q. All right. Why don't you make that
15 assumption?
16 A. I can't see any reason why it would be
17 higher.
18 Q. Would you turn to page E-25, the last
19 sentence of the first full paragraph, "Extreme
20 reluctance to delete data on purely statistical
21 grounds without external evidence of an anomaly was
22 built into the present criterion by requiring the
23 P-value for rejection of a datum be .0001 instead of
24 the customary .01."
25 Do you see that sentence?
233
1 A. Yes.
2 Q. Did you write that sentence?
3 A. Yes.
4 Q. How would you characterize the grounds for
5 deleting the data at station C1 and C2, stations 1
6 and 2, from consideration in fitting the model?
7 A. That deletion was based on subject matter
8 considerations rather than statistical outlier
9 analysis.
10 Q. And how would you characterize the grounds
11 for deleting the data for the dates 3-79 and 6-79
12 from consideration in fitting the model?
13 MR. FITZGERALD: Counsel, you are talking
14 about the ENP, not the model?
15 MR. BLANK: No, sir, I'm not.
16 MR. FITZGERALD: You were on page E-25,
17 right?
18 MR. BLANK: Yes.
19 MR. FITZGERALD: I may be seriously
20 mistaken, but I always thought the 12 structures
21 were in Everglades National Park, so I would
22 object to your question then as vague, if you're
23 implying not.
24 (Discussion held off the record.)
25 BY MR. BLANK:
234
1 Q. Are you confused, sir?
2 A. No.
3 MR. FITZGERALD: I'll read it in the
4 record.
5 THE WITNESS: Well, those dates were
6 excluded on the basis of this threshold stage
7 15.42. Again, that is -- that's a subject
8 matter decision.
9 BY MR. BLANK:
10 Q. Okay. How was the level of P-value .0001
11 arrived at?
12 A. That's a reflection of my own extreme
13 reluctance to delete data on the basis of statistical
14 grounds.
15 Q. But you did -- did you attempt an analysis
16 with different P levels that is different than .0001?
17 A. No.
18 MR. FITZGERALD: Before your next question
19 counsel, could I ask the court reporter to flag
20 about the last two or three pages and could I
21 have those three on an expedited basis?
22 BY MR. BLANK:
23 Q. If you used a less stringent criteria of
24 .01 or .001 do you know which observations would
25 appear to be outliers?
235
1 A. Now are we referring to the Park or --
2 Q. How about the Refuge first?
3 A. I don't know.
4 Q. Would you know for the Park?
5 A. I could look through and identify some.
6 Q. How long would it take you sir, to do that?
7 A. Not long.
8 Q. I tell you what, Doctor, I'm not really as
9 interested as I am in the Refuge.
10 Could you tell me if any of the
11 observations that we flagged on Table 6, that is for
12 stations 7, 9, 12 and 15 -- do you remember those
13 values of 69, 169, 494 and 121?
14 MR. MCGRATH: Page E-18, Doctor.
15 BY MR. BLANK:
16 Q. E-18.
17 A. Yes.
18 Q. Could you tell me if any of those
19 observations would appear as outliers using a
20 criteria of .01 or .001?
21 A. I at the moment don't know how to relate
22 this sample index to the data in the table.
23 Q. You could make that calculation, though,
24 given time, could you not?
25 A. Yes, I could.
236
1 Q. Okay. Are you familiar with using outlier
2 analysis as a screening tool to indicate which
3 observations require further investigation and not
4 necessarily whether they should be deleted?
5 A. I understand that is used, yes.
6 Q. Did you use it in this case in that
7 fashion?
8 A. No.
9 Q. And why not?
10 A. I had no access to information attaching to
11 any one of these numbers.
12 Q. When you say you had no access, what do you
13 mean?
14 A. In field notes.
15 Q. Did you request access to that data?
16 A. No, I did not.
17 MR. BLANK: Let me mark this as 23.
18 (The document was marked Exb. No. 23.)
19 BY MR. BLANK:
20 Q. Do you have Exhibit 23 in front of you,
21 Doctor?
22 A. Yes.
23 Q. Did you attend a settlement conference or
24 meeting on the 3rd of March?
25 A. You mean the 12th of March?
237
1 Q. I'm sorry. The 12th of March.
2 And I will represent to you, Doctor, that I
3 believe whoever took these notes, put the wrong year
4 there. I think it was probably '91 and not '90.
5 But I see your name appears on the top of
6 the page, is that correct?
7 A. That's correct.
8 Q. Do you know who took these notes?
9 A. No.
10 Q. Do you remember attending the meeting
11 that's referenced by these notes?
12 A. I don't remember the particular date.
13 Q. Do you remember the meeting after reviewing
14 these notes?
15 A. Okay. Yes.
16 Q. You do recall attending the meeting?
17 A. I still don't remember the particular
18 meeting, but I was there.
19 Q. Would you look at the last page, the one
20 that's bates stamped 0863594. Do you note there's a
21 different date on the top of that page?
22 A. Yes.
23 Q. Do you recall whether the meeting lasted
24 from 3-12 to 3-15?
25 A. No, I don't recall.
238
1 Q. Do you recall any of the discussions that
2 would be reflected on 3594?
3 A. Doesn't ring a bell.
4 Q. All right. On this same page do you see
5 the reference "marsh 12.4 PPB"?
6 A. Yes.
7 Q. And then an arrow with the word, "if" over
8 it?
9 A. Yes.
10 Q. Do you know what that refers to?
11 A. No, I don't.
12 Q. Do you see the word, "rainfall"?
13 A. Yes.
14 Q. And then next to it 34.1?
15 A. Yes.
16 Q. Do you know what that value reflects?
17 A. No, I don't.
18 Q. Well, those three numbers, the number for
19 S5A, S6 and rainfall appear to be added together to
20 get 97.2, is that correct?
21 A. They do add to 97.2.
22 Q. Do you recall any discussions that would
23 indicate that the amount of phosphorus per year
24 entering the Refuge from atmospheric deposition would
25 equal 34 or 34.1 tons?
239
1 A. No.
2 Q. In your analysis of the limits established
3 for the Refuge did you take into consideration in any
4 way the amount of phosphorus inputted into the Refuge
5 through atmospheric deposition?
6 A. No.
7 Q. The phosphorus limits established in
8 Appendix E are based on total phosphorus, is that
9 correct?
10 A. Yes.
11 Q. Were you aware or are you aware of any
12 discussions relating to regulating based upon
13 orthophosphate rather than total phosphorous?
14 A. I seem to recall hearing some discussion,
15 not participating in it because I'm ignorant in this
16 area.
17 Q. You are ignorant in the difference between
18 ortho phosphorous and total phosphorus?
19 A. Yes, I am.
20 MR. BLANK: Can we take a short break?
21 MR. MCGRATH: Yeah. Sure.
22 (Thereupon, a recess was taken.)
23 BY MR. BLANK:
24 Q. Dr. Robson, with regard to the stage gauge
25 readings for the Refuge and the three gauges at C1-7,
240
1 9 and 8 were you provided the individual data for
2 each of those gauges or simply provided averages?
3 A. I don't remember.
4 Q. Can you tell me whether any one of the
5 gauges had consistently higher readings than the
6 other two gauges?
7 A. No.
8 Q. If that fact were true, that is that one
9 gauge was consistently higher than the other two
10 gauges, what effect would that have had on the way
11 you utilized stage in your model for the Refuge?
12 A. I would have been pleased to see that, but
13 it wouldn't affect the way I used it.
14 Q. Why would you have been pleased to see it?
15 A. It would tend to support the use of these
16 gauges as providing an index.
17 Q. An index of what?
18 A. Of stage.
19 MR. BLANK: Okay. I have no further
20 questions.
21 We would reserve, just for the record, the
22 right to resume Dr. Robson's deposition in the
23 event the computer disk that we were unable to
24 read has some information on it that would be
25 relevant to further questions for Dr. Robson.
241
1 MR. MCGRATH: That's understood.
2 MR. FITZGERALD: I reserve our right to
3 object to it. I assume it was timely provided.
4 They had three weeks to call you and say: Where
5 is it? Not that we necessarily object, but we
6 reserve our right to do so.
7 MR. MCGRATH: I wanted to clarify while we
8 were on the record. There's the matter of the
9 disk which I will represent I will look into. I
10 was wondering, there was the -- you wanted an
11 updated version of the vita?
12 MR. PERKO: Three things or four things
13 that I remember; disk, the updated vita, any
14 reports that Dr. Robson reviewed of Dr. Walker
15 and the Hirsch paper that he cited. I'll send
16 you a letter confirming that.
17 MR. MCGRATH: Is that it?
18 MR. BLANK: Unless someone's got other
19 questions.
20 Thank you, Doctor.
21 I would be remiss if I didn't further thank
22 you for coming down here for us to take your
23 deposition.
24 (Witness excused.)
25
242
1 (Thereupon, at 2:50 p.m.,
2 the deposition was concluded.)
3
243
1 C E R T I F I C A T E
2
The State of Florida )
3 County of Palm Beach. )
4
I, April Y. Sapp, Court Reporter and Notary
5 Public, State of Florida at large, do hereby certify
that Douglas Robson, Ph.D. was by me first duly sworn
6 to testify the whole truth; that I was authorized to
and did report said deposition in stenotype; and that
7 the foregoing pages, numbered from 1 to 241,
inclusive, are a true and correct transcription of my
8 shorthand notes of said deposition.
9 I further certify that the said deposition
was taken at the time and place hereinabove set forth
10 and that the taking of said deposition was commenced
and completed as hereinabove set out.
11
I further certify that I am not attorney or
12 counsel of any of the parties, nor am I a relative or
employee of any attorney or counsel or party
13 connected with the action, nor am I financially
interested in the action.
14
The foregoing certification of this
15 transcript does not apply to any reproduction of the
same by any means unless under the direct control
16 and/or direction of the certifying reporter.
17 In witness whereof I have hereunto set my
hand and seal this ____ day of_____________ 1993.
18
19
20 _______________________________
April Y. Sapp,
21 Notary Public, State of Florida
at large. My commission expires
22 August 3, 1993.