1

 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 
   
18                        Volume I
           Deposition of Douglas Robson, Ph.D.
19 
               Taken before April Y. Sapp, Court Reporter
20   and Notary Public in and for the State of Florida at
     large, pursuant to notice of taking deposition filed
21   by the Petitioners in the above cause.
                             -  -  -
22 
     Tuesday March 30, 1993
23   319 Clematis Street
     West Palm Beach, Florida 33401
24   9:05 - 11:50 a.m.
     1:05 - 5:10 p.m.
25                           -  -  -
                                                 2

 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 
   
19                         -  -  -
                                                 3

 1                         -  -  -
 2                      I  N  D  E  X
 3                         -  -  -
 4 
 5  WITNESS:              DIRECT  CROSS  REDIRECT  RECROSS
 6  Douglas Robson, Ph.D.
 7 
    BY MR. PERKO            6
 8 
   
 9 
                                                 4

 1                         -  -  -
 2                     E X H I B I T S
 3                         -  -  -
 4 
 5  NUMBER                                  PAGE
 6  EXB. NO. 1                               27
 7  Notice of taking Deposition Duces Tecum
 8  EXB. NO. 2                               32
 9  Expert Witness & Consultant Designation
10  EXB. NO. 3                               52
11  Draft on Derivation of Phosphorous 9-19-91
12  EXB. NO. 4                               59
13  Phosphorous Limits for Lox. Wildlife Refuge
14  EXB. NO. 5                               70
15  Handwritten paper by Robson
16  EXB. NO. 6                               78
17  Letter to Federico
18  EXB. NO. 7                               84
19  Analysis of Refuge phosphorous concentrations
20  EXB. NO. 8                               93
21  Appendix E SWIM Plan
22  EXB. NO. 9                               118
23  Comments on settlement agreement by Walker
24  EXB. NO. 10                              122
25  Print out and graph
                                                 5

 1                         -  -  -
 2                     E X H I B I T S
 3                         -  -  -
 4  NUMBER                                  PAGE
 5  EXB. NO. 11                              123
 6  Handwritten note by Robson
 7  EXB. NO. 12                              124
 8  Calculations of Seasonal Kendall Test
 9  EXB. NO. 13                              125
10  Calculations of Seasonal Kendall Test
11  EXB. NO. 14                              126
12  Vita of Robson
13 
14 
                                                 6

 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 first duly
 6   sworn, was examined and testified as follows:
 7              THE WITNESS:  I do.
 8             DIRECT (Douglas Robson, Ph.D.)
 9   BY MR. PERKO:
10       Q.    Please state your name and address for the
11   record.
12       A.    Douglas Robson.  150 MacLaren Street,
13   Penthouse 6, Ottawa, Ontario.
14       Q.    Dr. Robson, my name is Gary Perko.  I'm an
15   attorney for the Sugar Cane Growers Cooperative, Roth
16   Farms and Wedgeworth Farms in the pending
17   administrative challenge of the Everglades SWIM Plan.
18             Have you ever been deposed before, sir?
19       A.    Not in this case.
20       Q.    Have you ever been deposed before?
21       A.    Yes.
22       Q.    How many times?
23       A.    Just in one case, several depositions in
24   one case.
25       Q.    What did that case involve?
                                                 7

 1       A.    It was a lawsuit by Alpo versus Ralston
 2   Purina concerning misleading advertising.
 3       Q.    So it did not involve water quality?
 4       A.    No.
 5       Q.    So you are generally familiar with the
 6   procedure here.  I'll be asking you questions about
 7   your expected testimony in this proceeding as well as
 8   other knowledge you have of relevant facts.  And I
 9   have to caution you.  I've only had six hours of
10   statistics, so if you don't understand my question,
11   please tell me and I'll try to rephrase it and if you
12   ever need a break, just let me know.  We'll take a
13   few minutes, get some fresh air or drink of water or
14   whatever.
15             Dr. Robson, you've been listed as a
16   potential expert witness in this proceeding, is that
17   correct?
18       A.    Yes.
19       Q.    What is your understanding of your expected
20   testimony?
21       A.    To review the statistical analyses that
22   were done by Walker.
23       Q.    Done by William Walker?
24       A.    Yes.
25       Q.    What statistical analyses are you referring
                                                 8

 1   to?
 2       A.    Analyses of the water quality and inflow to
 3   the Park and water quality in the Refuge.
 4       Q.    Are you referring to the derivation of
 5   phosphorous concentration levels for the Refuge and
 6   phosphorus concentration limits for the Park that are
 7   explained in Appendix E of the SWIM Plan?
 8       A.    Yes.  I don't remember the Appendix E.
 9       Q.    But you are referring to the levels and
10   limits?
11       A.    Yes.
12       Q.    Have you reached final opinions that you
13   anticipate providing at final hearing in this matter?
14       A.    Yes.
15       Q.    And what are those opinions?
16             MR. MCGRATH:  Gary, I have to just object
17        to the breadth of that question.  That's
18        terribly over broad.
19             I'd also state on the record that, as far
20        as final opinions, it's my understanding that
21        there has yet to be anyone in this case whose
22        reached any final conclusions.  We would then
23        state as of this moment it is not anticipated
24        that Dr. Robson would need to perform any
25        additional work.  We would not want to preclude
                                                 9

 1        the possibility, should the circumstances arise,
 2        based on the incomplete work of all the other
 3        experts to, you know, have Dr. Robson do some
 4        additional analysis, so to the extent that he's --
 5        Dr. Robson is saying his opinions are final, we
 6        don't anticipate additional work at this time.
 7        We don't want to preclude the need for
 8        additional work should facts arise and the
 9        analyses be done.
10   BY MR. PERKO:
11       Q.    Dr. Robson, I'm just asking you what
12   opinions you've reached at this time.  If you could
13   identify those for me that's what I'm interested in.
14       A.    These are opinions concerning standards set
15   for the Park and for the Refuge.
16       Q.    And what is your opinion of those
17   standards?
18             MR. MCGRATH:  Again, let me object to the
19        over breadth.  I don't know if, in this
20        particular instance, if that's a question that
21        can be answered just on its face.  I mean if you
22        want to go through and find out what work he's
23        done, you know, having established that
24        foundation, based on the opinions, that might be
25        a better way to approach it.
                                                 10

 1             Doctor, if you can answer the question as
 2        stated please feel free.  If you can't, then let
 3        us know.
 4   BY MR. PERKO:
 5       Q.    Can you answer the question, Dr. Robson?
 6       A.    I'm not sure what the -- what you are
 7   referring to.
 8       Q.    You have reviewed the statistical methods
 9   utilized in developing the limits for inflows to the
10   Park and Refuge, is that correct?
11             MR. MCGRATH:  I would object to the form of
12        the question.  It mischaracterizes his
13        testimony.  I don't believe there's been any
14        testimony regarding how you phrased the
15        question, analysis of statistical methods.
16             MR. PERKO:  I'm simply quoting what was
17        listed on the expert designation.
18             MR. MCGRATH:  Ask him what's listed on the
19        expert witness designation.
20   BY MR. PERKO:
21       Q.    Dr. Robson, have you reviewed the
22   statistical methods utilized to develop the
23   phosphorous concentration limits for inflows to the
24   Park and the Refuge?
25       A.    I've reviewed the analyses.
                                                 11

 1       Q.    You have not -- are you saying that you
 2   have not reviewed the statistical methods used for --
 3   derived for those methods?
 4       A.    Statistical methods are standard methods.
 5   I've reviewed the analyses.
 6       Q.    Is there a distinction between analyses and
 7   statistical methods?
 8       A.    There's a distinction, yes.  In reviewing
 9   methods, I would say one would consider alternative
10   methods.
11       Q.    Have you reviewed alternative methods for
12   setting phosphorus concentration limits for inflows
13   to the Park and Refuge?
14       A.    No.
15       Q.    Do you know if Dr. Walker did?
16       A.    I know that he did many Seasonal Kendall
17   tests for trend.
18       Q.    Any other statistical methods other than
19   those identified in the SWIM Plan?
20             MR. MCGRATH:  I would just object to the
21        form of the question.  I -- I don't particularly
22        understand what you are referring to.
23   BY MR. PERKO:
24       Q.    Do you understand what I'm referring to
25   when I say the statistical methods identified in the
                                                 12

 1   SWIM Plan for setting the phosphorous limits for the
 2   Park and Refuge?
 3       A.    Vaguely.
 4       Q.    When you say you reviewed the analyses
 5   utilized to develop the phosphorus concentration for
 6   the inflows to the Park and Refuge what are you
 7   referring to?
 8       A.    Referring to the regression analyses that
 9   Walker performed.
10       Q.    What is your opinion of those regression
11   analyses?
12             MR. MCGRATH:  Opinion about what?  Again,
13        I'm objecting.
14   BY MR. PERKO:
15       Q.    Is it your opinion those regression
16   analyses are appropriate?
17       A.    That was not really my role to decide
18   whether they were appropriate.
19       Q.    What was your role?
20       A.    To verify the numerical correctness of the
21   results.
22       Q.    Is it fair to say that your role was
23   limited to determining whether to reduce the
24   calculations?
25       A.    Essentially, yes.
                                                 13

 1       Q.    Do you have an opinion as to whether the
 2   analyses performed by Dr. Walker are appropriate for
 3   setting limits on the phosphorous inflows to the Park
 4   and Refuge?
 5       A.    I am comfortable with his analyses, yes.
 6       Q.    When you say you are comfortable with the
 7   analyses, what do you mean?
 8       A.    It indicates approval.
 9       Q.    Dr. Robson, are you currently under
10   contract with the District?  By District I mean the
11   South Florida Water Management District.
12       A.    Yes.
13       Q.    When were you first contacted by the
14   District in this proceeding -- in connection with
15   this proceeding?
16       A.    You mean the current contract or --
17       Q.    The first time you were contacted.
18             MR. MCGRATH:  I would need to object as far
19        as the vagueness of the question.  I mean with
20        this proceeding are you referring to the
21        administrative proceeding presently ongoing?
22   BY MR. PERKO:
23       Q.    When were you first contacted by the
24   District regarding statistical methods for analyzing
25   environmental data associated with the Everglades?
                                                 14

 1             MR. MCGRATH:  I just need to object to the
 2        mischaracterization of Dr. Robson's role.
 3   BY MR. PERKO:
 4       Q.    Dr. Robson --
 5             MR. MCGRATH:  He can answer the question.
 6             I know what you are saying, but how you
 7        said it, I needed to place an objection on the
 8        record.
 9   BY MR. PERKO:
10       Q.    Dr. Robson, do you understand the question?
11       A.    Not entirely because I was not first
12   contacted by the District.  I was contacted by
13   Scadden, Arps.
14       Q.    Who at Scadden, Arps first contacted you?
15       A.    I can't remember the names.
16       Q.    When were you first contacted?
17       A.    It was in the fall of '89.
18       Q.    What were you first asked to do by Scadden,
19   Arps?
20       A.    I don't remember any specific instructions.
21       Q.    Do you know why they contacted you?
22       A.    I know they needed a statistician.
23       Q.    Do you know why they needed a statistician?
24             MR. MCGRATH:  Let me just object to the
25        form as a formality to the extent it encroaches
                                                 15

 1        on work product of my client's former counsel,
 2        as far as -- I mean Dr. Robson's free to testify
 3        as to his understanding of his role at the time
 4        he was initially contacted.
 5             MR. PERKO:  I'm only asking for
 6        Dr. Robson's knowledge.
 7             THE WITNESS:  My understanding is that
 8        because the other side had a statistician expert
 9        on their team, that Scadden, Arps needed one
10        also.
11   BY MR. PERKO:
12       Q.    Was that statistician expert Dr. Walker?
13       A.    That was my understanding.
14       Q.    By the other side, are you referring to the
15   United States?
16       A.    Yes.
17       Q.    Were you asked by Scadden, Arps to review
18   any work of Dr. Walker?
19       A.    Not initially.
20       Q.    When were you first asked to do something
21   by Scadden, Arps?  What was the first task that you
22   were given by Scadden, Arps?
23       A.    First task was to study some of the
24   background documents produced by the District.
25       Q.    What background documents?
                                                 16

 1       A.    One I remember was authored by Swift.
 2             MR. BLANK:  Excuse me.  I didn't hear the
 3        answer.
 4             MR. PERKO:  Swift.
 5             MR. MCGRATH:  One of the documents was
 6        authored by a Swift.
 7   BY MR. PERKO:
 8       Q.    Is that regarding periphyton?
 9       A.    Right.
10       Q.    Do you remember any other documents you
11   were asked to study?
12       A.    I remember a document.  I'm not sure that
13   it was presented to me at this time.  It was a
14   document in which Tony Federico was one of the
15   authors.
16       Q.    Do you remember what that document was
17   about?
18       A.    No, I don't.
19       Q.    Do you recall what the -- any other
20   documents that you may have reviewed at that time?
21       A.    No.
22       Q.    Did you review any of the work by
23   Dr. William Walker?
24             MR. MCGRATH:  At what period of time?
25   BY MR. PERKO:
                                                 17

 1       Q.    How -- when did you first review
 2   Dr. Walker's work?
 3       A.    Early 1990.
 4       Q.    What did you review at that time?
 5       A.    The one I remember is the Seasonal Kendall
 6   Test.
 7       Q.    Was that an analysis of trends in inflows
 8   to Everglades National Park?
 9       A.    Yes.
10       Q.    Did you provide the District or Scadden,
11   Arps with any written comments regarding that work?
12       A.    Not that I can recall.
13       Q.    Would you have retained copies of those
14   written comments had you provided them?
15       A.    Yes.
16       Q.    Would you have produced them in connection
17   with this deposition?
18       A.    Yes.
19       Q.    What was your opinion of Dr. Walker's
20   Seasonal Kendall Test analysis?
21       A.    The analysis was arithmetically correct.
22       Q.    Did you have any opinion as to the analyses
23   that Dr. Walker chose to perform, the methodology?
24       A.    The Seasonal Kendall Test was new to me at
25   that point in time.
                                                 18

 1       Q.    Does that mean you did not have an opinion
 2   on it?
 3       A.    I developed an opinion.
 4       Q.    What was that opinion?
 5       A.    That there were problems with the Seasonal
 6   Kendall Test.
 7       Q.    What problems?
 8       A.    There was a problem with serial correlation
 9   which invalidates the size of the test.
10       Q.    I'm sorry.  I didn't hear the last part.
11       A.    The size of the test.
12             MR. BLANK:  Did you say invalidates the
13        size of the test?
14             THE WITNESS:  Yes.
15    BY MR. PERKO:
16       Q.    What do you mean by serial correlation?
17       A.    Temporal correlation.
18       Q.    How did that invalidate the size of the
19   test?
20       A.    There is a published paper which I believe
21   I included in the material I provided you -- perhaps
22   not -- by the same authors that originally devised
23   the test and they demonstrated the impact of auto
24   correlation showing that it inflates the size of the
25   test.
                                                 19

 1       Q.    Do you recall the names of the authors of
 2   that paper?
 3       A.    I remember one of the names was Hirsh.
 4       Q.    Do you recall what publication that
 5   appeared in?
 6       A.    I'm not sure of the name of the journal.
 7       Q.    Do you remember the name of the paper?
 8       A.    No.
 9       Q.    Did you say no?
10       A.    No.
11       Q.    Did you have any other problems with
12   Dr. Walker's analysis of the inflows to Everglades
13   National Park?
14       A.    Nothing serious.
15       Q.    I take it that you did have other problems.
16       A.    No two statisticians can agree on anything.
17             MR. FITZGERALD:  We've noticed that.
18   BY MR. PERKO:
19       Q.    Did you perform any analysis to correct the
20   problems with Walker's analysis?
21       A.    Are you referring to any particular one?
22       Q.    Yes.  The problem with serial correlation
23   that you identified previously.
24       A.    I did some simulations.
25       Q.    What type of simulations did you do?
                                                 20

 1       A.    Generating synthetic data that incorporate
 2   auto correlation and model misspecification.
 3       Q.    What were the results of your simulations?
 4       A.    Well, that these violations do, in effect,
 5   destroy the validity of the test.
 6       Q.    Did you reanalyze the data that Dr. Walker
 7   analyzed?
 8             MR. MCGRATH:  I just have to say I believe
 9        he testified that he used synthetic data.
10             MR. PERKO:  He may have testified to that.
11        I'm asking him a different question.
12             MR. MCGRATH:  Okay.  Excuse me.
13             THE WITNESS:  Could you tell me which data
14        you are referring to?
15   BY MR. PERKO:
16       Q.    The inflow data that Dr. Walker analyzed.
17             MR. MCGRATH:  Let me just ask with respect --
18        are you still talking about the Seasonal Kendall
19        analysis?
20             MR. PERKO:  Yes.
21             MR. MCGRATH:  Okay.
22             THE WITNESS:  My recollection is I checked
23        his calculations.
24   BY MR. PERKO:
25       Q.    Did you reproduce his results?
                                                 21

 1       A.    That's my recollection, yes.
 2       Q.    Did you do any additional analyses of
 3   Dr. Walker's work on the Seasonal Kendall Test?
 4       A.    I did some sort of back of the envelope
 5   calculations at one point.
 6       Q.    Did you make any conclusions based on those
 7   calculations?
 8       A.    I concluded that the Spearman correlation
 9   was producing results very similar to the Seasonal
10   Kendall Test.
11       Q.    The Spearman correlation?
12       A.    Right.
13       Q.    Is this a different analyses than the
14   Seasonal Kendall Test?  The Spearman, I mean.
15       A.    Slightly different, yes.
16       Q.    To your knowledge, did anyone including
17   Dr. Walker, ever correct the problems with serial
18   correlation with this Kendall -- Seasonal Kendall
19   Test analysis?
20       A.    No.
21       Q.    Is it your -- is it still your opinion that
22   the serial correlation invalidates the size of the
23   test?
24       A.    Yes.
25       Q.    Do you have an opinion as to whether the
                                                 22

 1   Spearman's correlation analysis is more appropriate
 2   than the Seasonal Kendall Test?
 3       A.    I am undecided.
 4       Q.    What would you need to do to decide?
 5       A.    Examine the power of the two tests.
 6       Q.    Does the Spearman's correlation have any
 7   serial correlation problem?
 8       A.    Yes.
 9       Q.    Same type of problem as with the Seasonal
10   Kendall Test?
11       A.    Yes.
12       Q.    Are you aware of any other type of analyses
13   that could be used to examine the trends and inflows
14   to Everglades National Park?
15       A.    There are a number of non-parametric
16   analyses that are standard.
17       Q.    Have you ever performed any of those
18   analyses?
19       A.    Not on these data.
20       Q.    Prior to February of 1991 did you review
21   any other work of Dr. Walker?
22       A.    I don't recall the exact dates, but it was
23   around that time I did review documents he had
24   prepared for the Park.
25       Q.    Let me back up for one second.
                                                 23

 1             You stated that there are a number of
 2   non-parametric analyses that could be used to examine
 3   trans -- ENP inflows.  Do any of these non-parametric
 4   analyses not have a serial correlation problem?
 5       A.    Offhand, I can't think of any that would
 6   circumvent that.
 7       Q.    Are you aware of any other methodology,
 8   non-parametric or otherwise, that would not have a
 9   serial correlation problem?
10       A.    There are time series methods which attempt
11   to take into account serial correlation.
12       Q.    Did you ever perform any of these time
13   series analysis?
14       A.    No, I did not.
15       Q.    Do you know if Dr. Walker did?
16       A.    I don't know.
17       Q.    You previously testified that sometime
18   around February 1991 you reviewed some additional
19   documents prepared by Dr. Walker for the Park, is
20   that correct?
21       A.    Yes.
22       Q.    What additional documents were you
23   referring to?
24       A.    I recall one that dealt with exceedances.
25       Q.    Exceedances of what?
                                                 24

 1       A.    As I recall, the document did not deal with
 2   any specific data.  It was a document concerning the
 3   methodology.
 4       Q.    Methodology for setting limits?
 5       A.    Limits on exceedance, yes.
 6       Q.    I'm a little bit confused by your use of
 7   the word, "exceedance."  Exceedance of what?
 8       A.    In this case, it would be the frequency of
 9   samples in which the concentration exceeded any
10   specified limit.
11       Q.    That's the phosphorus concentration in
12   inflows to Everglades National Park?
13       A.    That would be one application, yes.
14       Q.    Could it also be applied to inflows to the
15   Loxahatchee National Wildlife Refuge?
16       A.    I don't know the nature of that data.
17       Q.    Do you know what specific concentrations
18   Dr. Walker performed or developed his exceedance
19   methodology for?
20       A.    Initially, it was for a number of different
21   levels.
22       Q.    Do you remember the levels?
23       A.    No.
24       Q.    Was 10 parts per billion one of the levels?
25       A.    I don't know.
                                                 25

 1       Q.    Do you recall if 50 parts per billion was?
 2       A.    I don't recall.
 3       Q.    What methodology did Walker use for this
 4   exceedance analysis?
 5       A.    Are you referring to this document that I
 6   was --
 7       Q.    Yes.
 8       A.    The document, as I recall, utilized model,
 9   a log normal model.
10       Q.    Did you develop an opinion as to the
11   appropriateness of Dr. Walker's model?
12             MR. MCGRATH:  With respect to that paper?
13             MR. PERKO:  The same paper we've been
14        talking about for ten minutes, counsel.
15             THE WITNESS:  I wouldn't call it
16        Dr. Walker's model.  It's log normal
17        distributions.  Nobody's in particular model.
18   BY MR. PERKO:
19       Q.    Do you have -- did you develop an opinion
20   as to the appropriateness of that log normal model
21   for examining exceedances in ENP inflows?
22       A.    The document I'm referring to didn't
23   address ENP data directly.
24       Q.    Did this document have a title?
25       A.    I'm sure it did.
                                                 26

 1       Q.    You don't recall what it was?
 2       A.    No.
 3             MR. PERKO:  Did you want to take a break,
 4        Doctor?
 5             THE WITNESS:  Yes, please.
 6             (Thereupon, a recess was taken.)
 7             MR. PERKO:  Back on the record.
 8   BY MR. PERKO:
 9       Q.    Dr. Robson, before we took a little break,
10   we were discussing a paper that you reviewed by
11   Dr. Walker dealing with methodology for determining
12   exceedances.  Did you provide any written comments on
13   that report?
14       A.    No, I did not.
15       Q.    You mentioned a couple of -- or additional
16   documents prepared by Dr. Walker that you reviewed
17   about that time period.  Do you recall what those
18   documents were?
19       A.    No, I don't recall.
20       Q.    Do you recall what the subject matter of
21   those documents were?
22       A.    The subject matter was setting standards.
23       Q.    Approximately how many additional documents
24   were there?
25       A.    One or two.
                                                 27

 1       Q.    Did you retain copies of the documents?
 2       A.    Yes.
 3       Q.    Did you produce them in connection with
 4   this deposition?
 5       A.    I'm not sure that I did, since they were
 6   not produced by me.
 7       Q.    Did you review the notice duces tecum for
 8   your deposition?
 9       A.    Will you repeat the question?
10       Q.    Did you review the notice of taking
11   deposition duces tecum for your deposition?
12       A.    I think so, yes.
13             MR. PERKO:  I'd like to mark this as an
14        exhibit.
15             (The document was marked Exb. No. 1.)
16   BY MR. PERKO:
17       Q.    Doctor, I'm showing you what's been marked
18   as Exhibit Number 1.  Do you recognize that document?
19       A.    I'm not certain.
20       Q.    Did you review a document request in
21   connection with this deposition?
22       A.    Yes.
23       Q.    Did you review it with counsel for the
24   District?
25       A.    No.
                                                 28

 1       Q.    Did you have any discussions with counsel
 2   for the District regarding the documents that you
 3   were to produce in connection with this deposition?
 4             MR. MCGRATH:  Dr. Robson, as far as
 5        discussions with the District and counsel for
 6        the District, if you had discussions, please --
 7             MR. PERKO:  I'm just asking if he had
 8        conversations.  I'm not asking for any work
 9        product or anything.
10             MR. MCGRATH:  I'm instructing the witness
11        in case his answer goes further.
12             Right now he's just asking you if you had
13        discussions with counsel for the District
14        regarding the request to produce for your
15        deposition.
16             THE WITNESS:  I had a telephone call.
17   BY MR. PERKO:
18       Q.    How did you go about compiling responsive
19   documents?
20       A.    I went through my, quote, files, unquote,
21   or, quote, piles, unquote.
22       Q.    And how did you determine which documents
23   to produce?
24       A.    I tried to produce all documents that
25   pertained to the -- this investigation.  I should
                                                 29

 1   point out that these piles included documents from
 2   other projects of mine.  My filing system is chaotic.
 3       Q.    Did you produce the reports that you
 4   reviewed that were authored by Dr. Walker?
 5             MR. MCGRATH:  I think a more appropriate
 6        question first would be:  Did he have the
 7        reports?
 8             MR. PERKO:  He's already testified to that,
 9        counsel.  He said he retained copies of the
10        documents that he reviewed.
11             THE WITNESS:  I don't recall whether I
12        found them.
13   BY MR. PERKO:
14       Q.    Did you specifically look for them?
15       A.    Yes.
16       Q.    Did you make any notes on the Walker
17   reports that you did review?
18       A.    Not that I recall.
19       Q.    Dr. Robson, you previously testified that
20   you were contacted by Scadden, Arps in the fall of
21   1989, is that correct?  And we've talked about some
22   reports of Dr. Walker that you reviewed, including
23   his work on the Seasonal Kendall Test and a number of
24   documents that you have reviewed, approximately
25   February of 1991.
                                                 30

 1             What else did you do for Scadden, Arps
 2   prior to February 1991?
 3       A.    In December I attended a orientation
 4   meeting here in West Palm.
 5       Q.    December of 1989?
 6       A.    Oh.  Maybe I've got the year wrong, huh?
 7   December '90.  Let me think a minute.  I must have
 8   the year wrong.
 9       Q.    So is it -- are you correcting your
10   testimony that you were contacted in the fall of
11   1990?
12             Let me see if we can get through it another
13   way.  How many contracts have you had with the
14   District or Scadden, Arps?
15       A.    Two.
16       Q.    Were those annual contracts?
17       A.    I don't recall.
18       Q.    Do you recall the terms of the first
19   contract?
20       A.    You mean the time limits?
21       Q.    Right.
22       A.    No, I don't recall.
23       Q.    Was that contract with Scadden, Arps?
24       A.    No.
25       Q.    Was it with the District?
                                                 31

 1       A.    Yes.
 2       Q.    Do you recall when you entered the second
 3   contract?
 4       A.    It was mid 1991.  I don't recall the exact
 5   date.
 6       Q.    Had the first contract expired?
 7       A.    Well, I didn't -- I don't recall how long
 8   that contract was for.
 9       Q.    Do you know why you entered a second
10   contract?
11       A.    No, I don't.
12       Q.    What did you contract -- with regard to the
13   first contract what did you contract to provide to
14   the District?
15       A.    I was simply to serve as an expert witness.
16       Q.    And this was in the federal lawsuit --
17       A.    Right.
18       Q.    -- brought by the United States against the
19   South Florida Water Management District --
20       A.    Yes.
21       Q.    -- and DER and others?
22             And what expert testimony did you
23   anticipate providing in that case?
24             MR. MCGRATH:  Let me clarify.  I think the
25        more appropriate question first would be to ask
                                                 32

 1        him if he reached any opinions in this case and
 2        that would seem to be the foundation for any
 3        expert testimony that would have been expected.
 4             MR. PERKO:  Counsel, he testified that he
 5        contracted to provide -- to serve as an expert
 6        witness for the District.
 7             MR. MCGRATH:  You are right.
 8             MR. PERKO:  I assume if he was going to be
 9        an expert witness he anticipated providing
10        testimony.  I'm asking him what his
11        understanding of his anticipated testimony was.
12             THE WITNESS:  At this point I can't recall
13        the terms of the contract in any detail.
14             MR. PERKO:  I apologize.  I didn't mean to --
15        first couple of exhibits I didn't provide
16        copies.
17             I'd like to mark this as Exhibit Number 2.
18             (The document was marked Exb. No. 2.)
19   BY MR. PERKO:
20       Q.    Dr. Robson, I show you Exhibit Number 2 and
21   ask if you recognize that document?
22             MR. FITZGERALD:  Counsel, since you don't
23        have copies of that, I would ask you to identify
24        it just for the rest of us.
25             MR. PERKO:  Sure.  It's the expert witness
                                                 33

 1        designation of the South Florida Water
 2        Management District in the federal suit.
 3             MR. MCGRATH:  Exhibit Number 2, let me just
 4        read.  It's case captioned United States of
 5        America versus South Florida Water Management
 6        District, case number 88-1886.  This particular
 7        pleading is titled, Designation by the
 8        Defendants South Florida Water Management
 9        District and Wodraska of Expert Witnesses and
10        Consultants.
11   BY MR. PERKO:
12       Q.    Do you recognize that document?
13       A.    Yes.
14       Q.    Dr. Robson, I'll show you page 4, item #13,
15   which is the expert witness designation for Douglas
16   S. Robson.  Does that refresh your recollection
17   regarding the anticipated expert testimony -- your
18   anticipated expert testimony in the federal suit?
19       A.    Yes.
20       Q.    What was that anticipated expert testimony?
21       A.    It says, "Statistical analysis of
22   environmental data."
23       Q.    Did you actually perform statistical
24   analysis of environmental data in preparation for
25   your testimony in the federal suit?
                                                 34

 1       A.    I don't recall whether I did under this
 2   contract.
 3       Q.    Other than reviewing the Walker reports
 4   that you've already testified about do you recall any
 5   other work that you performed on behalf of the
 6   District prior to February 1991?
 7       A.    No.
 8       Q.    Did you provide any written reports to the
 9   District prior to February 1991?
10       A.    No.
11       Q.    Did you formulate any opinions regarding
12   statistical analysis of environmental data prior to
13   February 1991?
14             MR. MCGRATH:  You are asking him if he
15        formulated opinions based on work he can't
16        remember doing?
17             MR. PERKO:  No, counsel.
18             The question stands.
19             THE WITNESS:  You are referring to
20        environmental data in general?
21   BY MR. PERKO:
22       Q.    Environmental data in connection with the
23   federal lawsuit.
24       A.    I had not formed firm opinions at that
25   stage.
                                                 35

 1       Q.    Had you formed any preliminary opinions?
 2       A.    I don't recall.
 3       Q.    Dr. Robson, you previously testified that
 4   you reviewed -- early in 1990 that you reviewed the
 5   work of Dr. Walker on the Seasonal Kendall Test for
 6   ENP inflows and you were of the opinion that there
 7   were problems with the Seasonal Kendall Test,
 8   specifically with serial correlation.
 9             MR. MCGRATH:  Let me just object to the
10        form, because I don't think it's clear, the time
11        periods, so to the extent the question
12        mischaracterizes testimony, the record will
13        reflect what he said.
14             MR. PERKO:  That's exactly right, counsel.
15   BY MR. PERKO:
16       Q.    Did you agree with the conclusions of
17   Dr. Walker's report on the Seasonal Kendall Test?
18       A.    What I agreed on was the correctness of the
19   calculations.
20       Q.    Did you agree with the conclusions of the
21   report, that there was an increasing trend in
22   phosphorus inflows?
23       A.    I had not formed a firm opinion at this
24   point.
25       Q.    Have you since formed a firm opinion?
                                                 36

 1       A.    One that I already expressed that the
 2   Seasonal Kendall Test has problems.
 3       Q.    Does that mean that you disagree with the
 4   conclusions of the report?
 5       A.    No.
 6       Q.    Do you have any opinion as to the
 7   conclusions, the correctness of the conclusions in
 8   Dr. Walker's report?
 9       A.    I conclude that you cannot reach an
10   opinion -- reach a conclusion based upon the Seasonal
11   Kendall Test.
12       Q.    And that is your opinion as you sit here
13   today?
14       A.    Yes.
15       Q.    Do you have an opinion as to whether
16   there's a more appropriate methodology for examining
17   trends in phosphorous inflows?
18             MR. MCGRATH:  Asked and answered.
19   BY MR. PERKO:
20       Q.    You can answer the question, Doctor.
21       A.    You are referring to non-parametric
22   methods?
23       Q.    Any methods.
24       A.    Would you repeat the question again?
25       Q.    Do you have an opinion as to whether there
                                                 37

 1   are -- there is more appropriate methods than the
 2   Seasonal Kendall Test methodology for examining
 3   whether there are trends in phosphorus inflows to
 4   Everglades National Park?
 5             MR. FITZGERALD:  Objection.  Asked and
 6        answered.
 7             THE WITNESS:  Yes.
 8   BY MR. PERKO:
 9       Q.    What is that opinion?
10       A.    That there are better methods.
11       Q.    What methods are you referring to?
12       A.    The method Walker ultimately used, the
13   regression method.
14             MR. FITZGERALD:  I'm sorry.  Doctor, I
15        couldn't hear the very last part of your answer.
16             THE WITNESS:  The regression method.
17   BY MR. PERKO:
18       Q.    Is that reflected in a report, the
19   regression method of Dr. Walker?
20       A.    Yes.
21       Q.    Do you recall the approximate date on that
22   report?
23       A.    Oh.  I thought you were referring to
24   Appendix E when you said report.
25             What report were you referring to?
                                                 38

 1       Q.    I asked if the regression method that
 2   Dr. Walker ultimately used is reflected in that
 3   report?
 4       A.    It's reflected in the Appendix E.
 5       Q.    Appendix E of the SWIM Plan?
 6       A.    If I have the letter correct, yes.
 7       Q.    Okay.  Does the regression method that
 8   Dr. Walker ultimately used have a serial correlation
 9   problem?
10       A.    Since he used annual averages, serial
11   correlation problem is a minor problem.
12       Q.    Did you have any discussions with
13   Dr. Walker regarding appropriate method -- methods to
14   use, be they regression or Seasonal Kendall Test?
15       A.    Not directly.
16       Q.    When did you first learn that Dr. Walker
17   was using a regression method?
18       A.    I'm not sure when I first learned it.
19       Q.    Early 1991, perhaps?
20       A.    Perhaps.
21       Q.    Do you recall if that was -- let me first
22   ask:  Are you aware that a stay was granted in the
23   federal suit in February of 1991?
24       A.    No.
25       Q.    Are you aware that settlement negotiations
                                                 39

 1   started on or about February 1991 in the federal
 2   suit?
 3       A.    No.
 4       Q.    Are you aware that settlement negotiations
 5   occurred in early 1991?
 6       A.    What do you mean by early?
 7       Q.    Before June.
 8       A.    Yes.
 9       Q.    Did you have any involvement in those
10   settlement negotiations?
11       A.    I attended some of the meetings.
12       Q.    How many meetings did you attend?
13       A.    I'm not sure.
14       Q.    More than five?
15       A.    I doubt it.
16       Q.    Less than five?
17       A.    Yes.
18             MR. FITZGERALD:  To be technically correct
19        with a statistician you have to say less than or
20        equal to.
21   BY MR. PERKO:
22       Q.    Do you remember when the first meeting you
23   attended was?
24       A.    I have to look at my records.
25       Q.    Do you have a calendar that indicates what
                                                 40

 1   meetings you attended or any other records for that
 2   matter?
 3       A.    I rely upon invoices that I sent in.
 4       Q.    How often do you send invoices to the
 5   District?
 6       A.    At that time it was monthly.
 7       Q.    What was discussed at the first meeting
 8   that you attended?
 9       A.    I can't remember what was discussed at the
10   individual meetings.
11       Q.    What was discussed generally at the
12   meetings?  Did you discuss phosphorus limits or
13   levels for the Park and Refuge?
14       A.    Yes.
15       Q.    Did you discuss different methodologies for
16   determining those limits or levels?
17       A.    Briefly.
18       Q.    Who did you discuss those methodologies
19   with?
20             MR. MCGRATH:  I would just object to the
21        form of the question as stated.  It would
22        indicate that the witness had those discussions.
23   BY MR. PERKO:
24       Q.    Did you personally participate in any
25   discussions regarding the methodologies to set limits
                                                 41

 1   or levels for the Park and Refuge?
 2       A.    That's what I meant when I said briefly.
 3       Q.    You participated in those discussions?
 4       A.    Yes.
 5       Q.    Who else participated in those discussions?
 6       A.    Bill Walker was present.
 7       Q.    Anyone else?
 8       A.    Federico, MacVicar and I don't remember the
 9   names of these other people.
10       Q.    Do you remember if Mark Maffei was one of
11   the participants?
12       A.    What is his role?
13       Q.    He's with the Loxahatchee National Wildlife
14   Refuge.
15       A.    There were one or two people there from the
16   Refuge.
17       Q.    But you don't remember their names?
18       A.    No.
19       Q.    Was anyone there on behalf of Everglades
20   National Park?
21       A.    Yes.
22       Q.    Do you remember who that was?
23       A.    I don't remember the name.
24       Q.    How many people were there on behalf of
25   Everglades National Park?
                                                 42

 1       A.    I believe there were two.
 2       Q.    I believe you testified that Mr. Federico
 3   and Mr. MacVicar were there on behalf of the
 4   District, is that correct?
 5       A.    Correct.
 6       Q.    Was there anyone else there on behalf of
 7   District?
 8       A.    There were other people there, but I'm not
 9   sure whether they represented the District or not.
10       Q.    Do you recall who those other people were?
11       A.    No.
12       Q.    What methodologies did you briefly discuss
13   at that time?
14       A.    The use of tolerance limits as a potential
15   alternative to confidence limits.
16       Q.    Any other methodologies?
17       A.    No.
18       Q.    What methodology was ultimately chosen?
19       A.    Confidence limits.
20       Q.    Do you know why tolerance limits were
21   rejected?
22       A.    I would have to guess.  I'm reluctant to
23   guess.
24       Q.    What is your understanding of why the
25   tolerance limits were rejected?
                                                 43

 1             MR. MCGRATH:  Doug, let me instruct you.
 2        If you have an understanding that's based on
 3        more than just guesswork or sheer speculation.
 4             THE WITNESS:  My understanding is that
 5        there was a lack of familiarity and
 6        understanding of the conceptive tolerance
 7        limits.
 8   BY MR. PERKO:
 9       Q.    Who ultimately calculated the levels and
10   limits for the Park and Refuge?
11       A.    I'm having a problem with ultimately.  They
12   were calculated a number of times along the way.
13       Q.    Okay.  Who first calculated the levels and
14   limits?
15       A.    Walker.
16       Q.    Do you know if anyone else calculated the
17   limits and levels?
18       A.    I did.
19       Q.    Did you use the same methodologies as
20   Dr. Walker?
21       A.    Yes.
22       Q.    Did you use any other methodologies other
23   than confidence limits to calculate levels or limits?
24       A.    Not to calculate the limits, no.
25       Q.    I take it you did use other methodologies
                                                 44

 1   in some way.
 2       A.    Yes.
 3       Q.    What were the results of those analyses?
 4       A.    These were outlier analyses.
 5       Q.    Did you ever calculate limits or levels
 6   using tolerance limits?
 7       A.    Not using real data, no.
 8       Q.    Using synthetic data?
 9       A.    Yes.
10       Q.    Do you know if Dr. Walker ever calculated
11   limits or levels using tolerance limits?
12       A.    I don't know.
13       Q.    Do you know if Dr. Walker used any
14   methodologies other than the confidence limits to
15   calculate limits or levels?
16       A.    I don't know.
17       Q.    Approximately how many times did Dr. Walker
18   calculate limits or levels?
19       A.    I can't say.  I didn't see all of his
20   calculations.
21       Q.    How about you -- how many times did you
22   calculate limits and levels?
23       A.    Two or three times for the Park, I would
24   guess.
25       Q.    Over what period of time?
                                                 45

 1       A.    May through August.
 2       Q.    Did you ever calculate limits for the Park?
 3   Prior to May 1991 that is.
 4       A.    I may have.
 5       Q.    Did you ever calculate levels for the
 6   Loxahatchee National Wildlife Refuge?
 7       A.    Yes.
 8       Q.    How many times?
 9       A.    I'm not sure.
10       Q.    In the range of two to three or four?
11       A.    Four I will say.
12       Q.    Over the same period of time?
13       A.    No.  It would be the latter part of that
14   period.
15       Q.    The latter part of May to August?
16       A.    No.  I guess it would be in late June,
17   July.
18       Q.    Do you know when Dr. Walker first
19   calculated levels for the Loxahatchee?
20       A.    I don't recall.
21       Q.    Was it prior to June or July?
22       A.    Yes.
23       Q.    Did you -- during this time period,
24   February through July of 1991 did you review any
25   written reports or documents authored by Dr. Walker?
                                                 46

 1       A.    Is that period starting February 1 or end
 2   of February?
 3       Q.    After the documents you previously talked
 4   about.
 5       A.    In February I did review some documents
 6   which were presented as exhibits at his deposition.
 7       Q.    Did you review any other documents authored
 8   by Dr. Walker during that time period?
 9       A.    I don't recall any.
10       Q.    Did you review Dr. Walker's calculations of
11   the limits and levels for the Park and Refuge?
12             MR. MCGRATH:  In what period of time, that
13        February period?
14   BY MR. PERKO:
15       Q.    I'm speaking now from February to July of
16   1991.
17       A.    Yes.
18       Q.    Do you recall approximately how many -- how
19   many series of calculations you reviewed of
20   Dr. Walker's?
21       A.    I would have to guess again.
22       Q.    Approximately how many?
23       A.    Two or three.
24       Q.    Okay.  Did you review any other documents
25   authored by Dr. Walker during that time period?
                                                 47

 1       A.    I don't recall.
 2       Q.    Did anyone else besides you and Dr. Walker
 3   calculate levels for the Loxahatchee or limits for
 4   Everglades National Park?
 5       A.    I recall that somebody from the Refuge
 6   checked calculations.
 7       Q.    Do you recall who that was?
 8       A.    No.
 9       Q.    Did they check your calculations?
10       A.    I don't recall if it was mine or Walker's.
11       Q.    They simply checked the calculations, they
12   didn't do any other -- is that correct?
13       A.    That's correct.
14       Q.    Did the actual limits and levels calculated
15   change during that time period?
16             MR. MCGRATH:  Gary, do you mind if I have
17        that question read back?  I don't know if I
18        caught it all.
19             MR. PERKO:  That's fine.
20             Did the actual limits calculate change
21        during that time period?
22             MR. MCGRATH:  Okay.
23             THE WITNESS:  The calculated limits depend
24        upon which pieces of data are excluded.
25   BY MR. PERKO:
                                                 48

 1       Q.    But the equations themselves stayed the
 2   same?
 3       A.    The models stayed the same.
 4       Q.    Were different data bases used to calculate
 5   levels at different times?
 6       A.    To my knowledge the only difference was in
 7   the matter of exclusion of data.
 8       Q.    Were different data excluded at different
 9   periods of time?
10       A.    Yes.
11       Q.    Ultimately, 2 of the 16 stations were
12   excluded, is that correct?
13       A.    Correct.
14       Q.    Do you know if the levels or limits -- I'm
15   speaking of the Loxahatchee now -- do you know if the
16   Loxahatchee levels were ever calculated using all 16
17   stations?
18       A.    I don't recall.
19       Q.    Were the levels calculated different --
20   using different combinations of the 16 stations?
21       A.    You mean subsets?
22       Q.    Right, exactly.
23       A.    Not to my knowledge.
24       Q.    Just the 14?
25       A.    Just the 14.
                                                 49

 1       Q.    Do you know why the two stations were
 2   excluded?
 3       A.    That decision was not made by me, so I'm
 4   not sure.
 5       Q.    Do you know who made the decision?
 6       A.    I don't know if it was one individual.
 7       Q.    Would Dr. Walker have had input into that
 8   decision?
 9             MR. FITZGERALD:  Objection.  Calls for
10        speculation.
11             THE WITNESS:  Yes.
12   BY MR. PERKO:
13       Q.    Anyone else to your knowledge?
14       A.    No.  I don't know who all had input into
15   it.
16       Q.    Do you know why the two stations were
17   excluded?
18             MR. MCGRATH:  That was asked, plus
19        answered.
20             MR. PERKO:  I'm sorry.  Strike that.
21             MR. BLANK:  Counsel, I'm not sure it was
22        answered.
23             MR. MCGRATH:  The answer was he didn't
24        know.
25   BY MR. PERKO:
                                                 50

 1       Q.    I'll ask it again just to make sure.
 2             Do you know why the two stations were
 3   excluded?
 4       A.    I don't know for certain.
 5       Q.    What is your understanding?
 6       A.    Which may be incorrect.
 7       Q.    I understand.
 8       A.    That the stations were infrequently
 9   sampled.
10       Q.    Anything else?
11       A.    I would have to guess.
12       Q.    In addition to the two stations that were
13   excluded there were two sampling dates that were
14   excluded, is that correct?
15       A.    Correct.
16       Q.    Do you know why those sampling dates were
17   excluded?
18       A.    Again, this decision was not made by me.
19       Q.    Would it have been the same people that
20   decided on excluding the two stations?
21       A.    I would have to guess.
22       Q.    What is your understanding of why the two
23   sampling dates were excluded?
24       A.    Well, I misspoke before.  It was the
25   frequency of sampling in these -- no.  No.  Sorry.
                                                 51

 1   Sampling dates.  Let me think a minute.  No.  Oh.  I
 2   recall.  The stage was below the 15.12 feet.
 3             MR. BLANK:  I'm sorry.  I didn't hear the
 4        response.  15. what?
 5             THE WITNESS:  I said 15.12.  I'm not sure
 6        if I'm correct in that.
 7   BY MR. PERKO:
 8       Q.    How was the minimum stage determined, do
 9   you know that?
10       A.    I don't know the details.
11       Q.    From a statistical standpoint, do you agree
12   with the exclusion of the two stations and dates?
13       A.    I have to rely on the judgment of the
14   subject matter experts in this issue.
15       Q.    In what subject matter would you say that
16   decision was based on?  Was it the result of a
17   statistical analysis or was it --
18       A.    I recall that there was an analysis.  I
19   wouldn't call it a statistical analysis, using a
20   model.
21       Q.    Do you know who developed the model?
22       A.    Pardon me?
23       Q.    Do you know who developed the model?
24       A.    I didn't hear you.
25       Q.    Do you know who developed the model?
                                                 52

 1       A.    No, I don't.
 2       Q.    When you said that you would have to rely
 3   on the experts in the subject matters, did you have
 4   someone specific in mind?
 5       A.    Not some single individual, no.
 6       Q.    Do you have anyone in mind?
 7       A.    Yes.  The representatives of the Park or
 8   the Refuge and Bill Walker.
 9             MR. MCGRATH:  Gary, when you find a good
10        break point, do you mind if we take a short
11        break?
12             MR. PERKO:  We can do it right now.
13             (Thereupon, a recess was taken.)
14             Mark this, please.
15             (The document was marked Exb. No. 3.)
16   BY MR. PERKO:
17       Q.    Dr. Robson, I'm showing you what's been
18   marked as Exhibit Number 3.  Do you recognize this
19   document?
20       A.    Yes.
21       Q.    And what is it?
22       A.    Derivation of Phosphorus Limits for
23   Everglades National Park and Phosphorous Levels for
24   Loxahatchee National Wildlife Refuge dated August 19,
25   1991.
                                                 53

 1       Q.    That's a draft, correct?
 2       A.    Draft.
 3       Q.    Do you know who drafted this document?
 4       A.    I did.  I did a first draft.
 5       Q.    Prior to this August 19th version?
 6       A.    Yes.
 7       Q.    Did you produce that first draft in
 8   connection with this deposition?
 9       A.    Yes.
10       Q.    Are there any substantive changes between
11   the first draft and this August 19th draft?
12       A.    I would have to see that earlier draft to
13   answer that.
14       Q.    Did anyone review your earlier draft?
15       A.    I believe Walker did.
16       Q.    Did he provide you with written comments?
17       A.    I recall one page of written comments.
18   These were not provided to me by Walker.  They were
19   provided to me by Federico.
20       Q.    Were those written comments dated, do you
21   recall?
22       A.    I don't recall.
23       Q.    Do you remember the approximate time you
24   received those comments?
25       A.    I couldn't --
                                                 54

 1       Q.    Would it have been in July, perhaps?
 2       A.    Perhaps.
 3       Q.    But you are not sure?
 4       A.    No.
 5       Q.    Did you produce those written comments in
 6   connection with this deposition?
 7       A.    Yes.
 8       Q.    Did you actually develop the calculations
 9   that are reflected in this draft document?
10       A.    Yes.
11       Q.    Did the -- to your knowledge, did the
12   calculations in the earlier draft differ from the
13   calculations included in this August 19th draft?
14       A.    I don't believe so.
15       Q.    These calculations for the Loxahatchee
16   phosphorous concentration levels exclude the two
17   stations and two sampling dates, is that correct?
18       A.    Correct.
19       Q.    Were you instructed to exclude these two
20   stations and dates?
21       A.    Yes.
22       Q.    By whom?
23       A.    Federico.
24       Q.    Let me ask you this.  What were
25   Mr. Federico's instructions to you?
                                                 55

 1       A.    It would have been to drop those two dates
 2   and stations.
 3       Q.    Okay.  Who instructed you to draft this
 4   document, Exhibit Number 3?
 5       A.    Federico.
 6       Q.    Did he give you any other instructions
 7   other than telling you to draft it and exclude the
 8   two stations and dates?
 9       A.    This draft was prepared after I had done
10   the analysis.  His instructions to exclude were prior
11   to his instructions about the draft.
12       Q.    Do you know when he instructed you to
13   exclude the two stations and dates?
14       A.    July 5th.
15       Q.    Did you ever question that instruction?
16       A.    Not directly to him, no.
17       Q.    Did you have any problems with the
18   instruction?  Did you feel comfortable with it?
19       A.    Yes.
20       Q.    Why?
21       A.    Because I had analyzed the data and
22   satisfied myself that these merited exclusion.
23       Q.    Why did they merit exclusion?
24       A.    The stations merited exclusion because of
25   the low sampling rate.
                                                 56

 1       Q.    What about -- I'm sorry.  Go ahead.
 2       A.    The dates merited exclusion on the basis I
 3   mentioned earlier, experts in subject matter
 4   judgment.
 5       Q.    Did you do any statistical analysis to
 6   determine that it was reasonable to exclude these
 7   dates and stations or was that based on a qualitative
 8   judgment of someone else?
 9       A.    The latter.
10       Q.    I believe you previously testified that you
11   would rely on the judgment of representatives from
12   the Refuge as well as Dr. Walker, is that correct?
13       A.    Correct.
14       Q.    Did you receive any written comments from
15   anyone other than Dr. Walker on the draft of Appendix
16   E, Exhibit Number 3 or the earlier draft?
17       A.    None that I recall.
18       Q.    Are you familiar with the settlement
19   agreement that was entered into in the federal
20   lawsuit?
21       A.    I think so, yes.
22       Q.    Are the calculations used to derive the
23   limits and levels in the settlement agreement the
24   same as in Exhibit Number 3?
25       A.    The calculations are the same.  I believe
                                                 57

 1   there was a minor change in the data base.
 2       Q.    What was the change in the data base?
 3       A.    At one point there was a change in one of
 4   the phosphorus concentrations.
 5       Q.    Do you recall or could you tell from
 6   Exhibit Number 3 and/or the settlement agreement what
 7   change was made?
 8       A.    You couldn't tell.
 9       Q.    Do you know?
10       A.    I just mentioned it.
11       Q.    Okay.  But you don't specifically know what
12   change was made?
13       A.    Not offhand.
14       Q.    Could you tell from looking at the
15   documents?
16       A.    No.
17       Q.    Why was that change made?
18       A.    It was considered an erroneous datum.
19       Q.    On what basis?
20       A.    I'm not qualified to say.
21       Q.    Did you do any work on the limits or levels
22   for the Park and Refuge after drafting Exhibit Number
23   3?
24       A.    No.
25       Q.    Have you performed any other work for the
                                                 58

 1   District after drafting Exhibit Number 3 after August
 2   1991?
 3       A.    All I can recall is attending the McClave
 4   deposition.
 5       Q.    That was in December of 1992?
 6       A.    Yes.  And phone conversation with
 7   Fitzpatrick concerning another deposition.
 8       Q.    Do you mean Fitzgerald?
 9       A.    Fitzgerald.  Sorry.
10       Q.    What deposition was that concerning?
11       A.    Millard.
12       Q.    Did you attend the Millard deposition?
13       A.    No.
14       Q.    Did you provide Mr. Fitzgerald with any
15   areas of inquiry for the Millard deposition?
16       A.    Yes.
17       Q.    Do you recall specifically what those
18   inquiries were?
19       A.    I don't recall the inquiries.  I just
20   recall a table I prepared for him.
21       Q.    What was that table concerning?
22       A.    Concerned the frequency distribution of the
23   number of base year stations generated by the Millard
24   simulations.
25       Q.    Did you reach any conclusions regarding the
                                                 59

 1   Millard simulations?
 2       A.    Well, the conclusions were represented in
 3   that frequency distribution that I presented.
 4       Q.    What were they?
 5       A.    That a substantial fraction of his
 6   simulations included less than five base year dates.
 7       Q.    What's the significance of that conclusion?
 8       A.    That the simulations are not representative
 9   of the data that were used.
10       Q.    Have you done any additional work since the
11   McClave deposition?
12       A.    Only in reviewing some of the documents
13   that I provided you.
14       Q.    Did you review those documents in
15   preparation for this deposition?
16       A.    Some of them, yes.
17       Q.    Do you recall which documents you reviewed?
18       A.    The one I have in hand, Exhibit 3.
19       Q.    Any others?
20       A.    The earlier draft.
21             MR. PERKO:  Mark this, please.
22             (The document was marked Exb. No. 4.)
23   BY MR. PERKO:
24       Q.    Dr. Robson, I show you what's been marked
25   as Exhibit 4.  Do you recognize this document?
                                                 60

 1       A.    Yes.
 2       Q.    Is that the draft -- earlier draft that you
 3   have been speaking of?
 4       A.    It's part of it.
 5       Q.    Part of it.
 6             The part pertaining to phosphorus limits
 7   for the Loxahatchee National Wildlife Refuge,
 8   correct?
 9       A.    It is part of that part, yes.
10       Q.    Okay.  Let's see.  Appears to be the only
11   part I received.
12             Dr. Robson, do you anticipate doing any
13   additional statistical analyses in preparation for
14   the final hearing in this matter?
15       A.    No, I do not anticipate any.
16             MR. MCGRATH:  Just for the record, that
17        goes with the caveat in the event that
18        circumstances arise, that will not preclude us
19        from doing so and, of course, given proper
20        notice if any additional work is done, but not
21        as of this date and time.
22             MR. BLANK:  Can we take a short break?
23             MR. PERKO:  Sure.
24             (Thereupon, a recess was taken.)
25   BY MR. PERKO:
                                                 61

 1       Q.    Dr. Robson, I'd like to refer you again to
 2   Exhibit Number 3, which is the August 19, 1991 draft
 3   of Appendix E.  Whose calculations were ultimately
 4   included in this draft, yours or William Walker's?
 5       A.    Mine.  There is one of William Walker's
 6   included.
 7       Q.    And which one is that?
 8       A.    It's on page 11.
 9       Q.    Could you read it into the record, please?
10       A.    There are two formulas there for the
11   exceedance limit.  One of them is based on my
12   calculations and the other on Walker's.
13       Q.    Which one was based on your calculations?
14       A.    I was just trying to recall.  I would have
15   to check the settlement agreement to know which one
16   was mine and which one was Walker's.
17       Q.    Whose calculations went into the settlement
18   agreement?
19       A.    His with respect to this particular
20   formula.
21       Q.    Okay.  What about the other formulas?
22       A.    For the Park, they're all Walker's.  For
23   the Refuge, they're mine.
24       Q.    Just so I understand your testimony, in the
25   August 19th draft of Appendix E, all the calculations
                                                 62

 1   are yours except for one of the two relating to
 2   exceedance formulas?
 3       A.    Correct.
 4       Q.    And in the settlement agreement all of the
 5   calculations relating to the Park limits are
 6   Dr. Walker's and all of the calculations relating to
 7   Refuge levels are yours?
 8       A.    Correct.
 9       Q.    Are the calculations related to the Refuge
10   the same in the draft Appendix E as they are in the
11   settlement agreement?
12       A.    Yes.
13       Q.    What about the calculations related to the
14   Park?  I assume there's differences.
15       A.    There are small differences.
16       Q.    Do you know what those differences are?
17       A.    All of the equations are slightly
18   different.  The coefficients are slightly different.
19       Q.    Why are the coefficients different in the
20   draft Appendix E?  Why didn't you use the same
21   coefficients?
22       A.    My recollection is that one data point was
23   changed.
24       Q.    Okay.  It is fair to say that the
25   methodology is the same in Appendix E as it is in the
                                                 63

 1   settlement agreement, the only difference is related
 2   to that single datum change?
 3       A.    Yes.
 4       Q.    Dr. Robson, did you draft any portion of
 5   the settlement agreement?
 6       A.    No.
 7       Q.    Do you know who drafted the provisions of
 8   the settlement agreement relating to the ENP limits?
 9       A.    I would have to guess.
10       Q.    Do you know who drafted the Refuge -- the
11   portions relating to the Refuge limits -- levels
12   rather?
13       A.    Again, I'd have to guess.
14       Q.    Who do you have in mind?
15             MR. FITZGERALD:  Objection.
16             MR. MCGRATH:  The question calls for
17        speculation.  Either he knows or he doesn't
18        know.
19    BY MR. PERKO:
20       Q.    You can answer the question, Doctor.
21             MR. FITZGERALD:  Counsel, are you going to
22        persist in that question?
23             MR. PERKO:  If he knows, if he has an
24        answer to the question I'd like to hear it.
25             MR. FITZGERALD:  You are expressly going to
                                                 64

 1        seek speculative answers from the witness?
 2             MR. PERKO:  I'll withdraw the question.
 3   BY MR. PERKO:
 4       Q.    Dr. Robson, have you reviewed any of the
 5   work of Dr. Walker relating to phosphorous uptake or
 6   settling rates?
 7       A.    No.
 8       Q.    Do you plan to?
 9       A.    I have no plans to, no.
10       Q.    Okay.  Dr. Robson, what was the purpose of
11   setting limits -- or I'm sorry -- phosphorus
12   concentration levels for the Refuge?
13       A.    I'm not qualified to say.
14       Q.    When you set out to calculate levels, what
15   were you trying to do?
16       A.    To calculate limits based on base year
17   conditions.
18       Q.    And what was the base year?
19       A.    It was the first year of record.
20       Q.    1979 -- '78-'79, rather?
21       A.    Yes.
22       Q.    Are you familiar with the 50 PPB discharge
23   concentration target included in the settlement
24   agreement?
25       A.    I know there is such a limit.  That's the
                                                 65

 1   extent of my knowledge.
 2       Q.    Do you know how that was derived?
 3       A.    No, I don't.
 4       Q.    Were you present at any meetings during the
 5   settlement negotiations in which the 50 PPB target
 6   was discussed?
 7       A.    I don't recall any specific discussion.
 8       Q.    During the settlement negotiations did you
 9   participate in any discussions on topics other than
10   the phosphorus concentration limits for the Park and
11   levels for the Refuge?
12       A.    No.
13       Q.    Is it fair to say that your input was
14   limited to that particular topic?
15       A.    Yes.
16       Q.    Did you ever correspond directly with
17   Dr. Walker during settlement negotiations?
18       A.    No.  Let me add, not to my knowledge, not
19   to my recollection.
20             MR. PERKO:  I think we're at a good
21        breaking point for lunch.
22             MR. FITZGERALD:  Fine with me.
23             (Thereupon, a recess was taken.)
24   BY MR. PERKO:
25       Q.    Dr. Robson, just before we broke for lunch
                                                 66

 1   I asked you about the purpose of setting the limits
 2   for the Park and the levels for Refuge.  I believe
 3   you said the purpose was to set levels or limits
 4   reflecting the conditions on the base line period --
 5       A.    Right.
 6       Q.    -- of 1978 through '79.
 7       A.    Right.
 8       Q.    Do you know why the '78 through '79 base
 9   line period was selected?
10       A.    I don't know.  I thought it was dictated by --
11   again, I'm not sure -- OFW.  I'm speaking out of
12   ignorance.
13       Q.    So I take it you hadn't had any experience
14   applying the OFW rule before?
15       A.    No.
16       Q.    Now, in setting the levels for the Refuge,
17   the calculations used data outside the '78 through
18   '79 base period, is that correct?
19       A.    Yes.
20       Q.    If the goal is to set limits at the
21   conditions during the base period, why would you use
22   data outside the base period?
23       A.    In order to improve the precision.
24       Q.    What do you mean by the precision?
25       A.    Magnitude of the standard area used in
                                                 67

 1   calculating elements.
 2       Q.    Was there a concern that if only 1978
 3   through '79 data were utilized that the magnitude of
 4   the standard error would be too large?
 5       A.    Yes.
 6       Q.    At what threshold would you consider the
 7   magnitude of standard error to be too large?
 8       A.    I wouldn't have a threshold.
 9       Q.    Do you know who determined or who decided
10   that data outside the '78 through '79 base period
11   should be utilized?
12       A.    Again, it's my understanding that OFW rules
13   require us to utilize all available data.
14       Q.    Did you utilize all available data?
15       A.    I used all that was available to me.
16       Q.    Okay.  Who did you obtain the data from?
17       A.    Kevin Rodberg.
18       Q.    Is Rodberg, is he with the District?
19       A.    Yes.
20       Q.    Did you request specific data or did he
21   send you what was available?
22       A.    He sent me.  I didn't request anything
23   specific.
24       Q.    Did you ever request additional data?
25       A.    I don't recall that I did, no.
                                                 68

 1       Q.    Were you ever concerned that there was not
 2   sufficient data to calculate these?
 3       A.    My understanding was that we were obliged
 4   to use whatever data are available.
 5       Q.    Is it your opinion that the available data
 6   was sufficient to calculate enforceable limits?
 7       A.    I'm not qualified to say on that.
 8       Q.    Why aren't you qualified?
 9       A.    For one thing, you said enforceable.  I
10   don't know what's involved in the enforcement.
11       Q.    Well, if I deleted the word enforceable,
12   would you have an opinion?
13       A.    Would the question be -- read it again,
14   then.
15       Q.    Do you have an opinion as to whether the
16   available data is sufficient to establish limits or
17   levels for the Refuge?  And by that I mean
18   statistically reliable limits or levels.
19       A.    I don't see it as a question of
20   reliability.
21       Q.    Do you have questions about the sufficiency
22   of the data?
23       A.    No.
24       Q.    Did you ever express a concern to the
25   District or anyone else that there was insufficient
                                                 69

 1   data to establish limits for the Refuge?
 2       A.    No, I did not.
 3       Q.    Once again, Dr. Robson, before we broke for
 4   lunch, you testified regarding Dr. Walker's Seasonal
 5   Kendall methodology for examining whether there's a
 6   trend in phosphorous inflows to Everglades National
 7   Park and I believe you testified that you cannot
 8   reach a conclusion regarding trends based upon the
 9   Seasonal Kendall Test, is that correct?
10       A.    Yes.
11       Q.    I believe you said that the regression
12   analysis would be a more appropriate methodology, is
13   that correct?
14       A.    Yes.
15       Q.    When -- let me ask you this.  Does the
16   regression methodology indicate that there is a trend
17   in phosphorus inflows to Everglades National Park?
18       A.    Yes.
19       Q.    How would you describe that trend?
20       A.    There's a significantly positive slope to
21   the regression.
22       Q.    Do you know what that slope is?
23       A.    No.
24       Q.    Did you actually perform a regression
25   analysis to determine whether there was a trend?
                                                 70

 1       A.    Yes.
 2       Q.    Is that reflected in any of the documents
 3   that you produced in connection with this deposition?
 4       A.    It's reflected in the standard of the
 5   slopes.
 6       Q.    In Appendix E?
 7       A.    Yes.
 8       Q.    Who first proposed to use the regression
 9   methodology?
10       A.    I don't know.
11       Q.    Referring you to Exhibit Number 3, the
12   draft Appendix E, could you identify for me the slope
13   to which you were referring to just a minute ago or
14   slopes?
15       A.    On page 6 equation that says, "fitted TP as
16   11.8309 plus 5932."  5932 is the slope I'm referring
17   to.
18       Q.    And that indicates a trend --
19       A.    Yes.
20       Q.    -- in phosphorus inflows?
21             MR. PERKO:  Mark this, please.
22             (The document was marked Exb. No. 5.)
23   BY MR. PERKO:
24       Q.    Dr. Robson, I'd like you to look at what's
25   been marked as Exhibit Number 5 to this deposition.
                                                 71

 1   Could you tell me if you recognize that document?
 2       A.    I do.
 3       Q.    And what is this document?
 4       A.    This is a note which I sent to Federico
 5   attempting to explain the concept of tolerance
 6   limits.
 7       Q.    When did you send this note to Mr. Federico --
 8   Dr. Federico?
 9       A.    March 26, '91.
10       Q.    Did you send this letter before or after
11   you learned of the proposed use of regression
12   analysis?
13       A.    After.
14       Q.    Do you know if any limits or levels have
15   been calculated using the regression analysis before
16   you prepared this note?
17       A.    By whom?
18       Q.    Anyone.
19       A.    Are you asking whether tolerance limits can
20   be applied in a regression context?
21       Q.    No.  No.  No.  I'm asking whether any
22   limits or levels had been calculated using the
23   regression approach before you drafted the note in
24   Exhibit 5?
25       A.    Yes.
                                                 72

 1       Q.    By whom?
 2       A.    Walker.
 3       Q.    Had you critiqued those limits before you
 4   wrote this note?
 5       A.    You mean prepare a written critique?
 6       Q.    Had you reviewed them at all?
 7       A.    I knew they existed and I knew what they
 8   were.  I had seen them.
 9       Q.    Did you prepare a written critique?
10       A.    No.
11       Q.    Did you provide any oral comments to
12   Dr. Walker or anyone else regarding those limits?
13       A.    Other than --
14             MR. MCGRATH:  At what point?  Are we
15        talking before the drafting of the Exhibit 5?
16             MR. PERKO:  Yes.
17             THE WITNESS:  Other than my suggestion that
18        tolerance limits be considered as an
19        alternative.
20   BY MR. PERKO:
21       Q.    None other than that, is that what you're
22   saying?
23       A.    Right.
24       Q.    Dr. Robson, I'd like to refer you to the
25   second full paragraph on page 1 of Exhibit Number 5.
                                                 73

 1       A.    Yes.
 2       Q.    About the middle of that paragraph it
 3   states that, "We begin with the simulation of upper
 4   tolerance levels since (to my mind) they are more
 5   relevant to standard setting than are confidence
 6   limits.  This relevance derives from the fact that in
 7   the standard setting mode of operation we have one
 8   POR for calculating the upper limit followed by a
 9   sequence of years of compliance checking."
10             Was this a true statement when you wrote it
11   in March of 1991?
12       A.    Yes.
13       Q.    Do you agree with it today?
14       A.    Yes.
15       Q.    What is the basis for that statement?
16       A.    I think it's explained here in subsequent
17   pages.
18       Q.    So you would simply refer to the document?
19       A.    Yes.
20       Q.    When you speak of confidence limits is that
21   essentially the regression approach?
22       A.    Confidence limits is a very general
23   concept.  It applies not only in regression but in a
24   variety of circumstances.
25       Q.    Let me ask you this.  Why did you write
                                                 74

 1   this note?
 2             MR. MCGRATH:  Asked and answered.
 3    BY MR. PERKO:
 4       Q.    You can answer the question.
 5       A.    Because at a meeting shortly before this I
 6   had offered the suggestion that we consider tolerance
 7   limits and the audience I was addressing didn't seem
 8   to understand what tolerance limits constituted.
 9       Q.    Who was the audience that you were
10   addressing?
11       A.    The ones I mentioned earlier, Federico,
12   MacVicar and --
13       Q.    The ones you mentioned in connection with
14   the settlement agreement meeting?
15       A.    Yes.
16       Q.    Or settlement negotiation meeting?
17       A.    Yes.
18       Q.    Did you have concerns about the confidence
19   limit approach?
20       A.    Well, concerns enough to mention the
21   consideration of this alternative.
22       Q.    Dr. Robson, I refer you to page 3 at the
23   very bottom.  States, "Since the 1978-'90 POR is now
24   history the SFWMD/ENP is locked into a row of either
25   table.  The question therefore becomes:  Knowing you
                                                 75

 1   had to be locked into a randomly chosen row which
 2   table (i.e., which upper limit procedure), would be
 3   the more appropriate choice?"
 4             Did you ever receive an answer to this
 5   question?
 6       A.    Indirectly, no.
 7       Q.    Indirectly?
 8       A.    This was followed by some discussion with
 9   Federico.
10       Q.    And what did Mr. Federico say in those
11   discussions?
12       A.    He asked me whether the confidence limit
13   could be interpreted as a tolerance limit --
14       Q.    And what did you tell him?
15       A.    -- at a certain level of confidence.
16       Q.    What did you tell him?
17       A.    I said yes.
18       Q.    Did you have any other discussions with
19   Mr. Federico regarding tolerance limits?
20       A.    No.
21       Q.    Did you discuss anything else at the time
22   you discussed whether tolerance limits could be
23   interpreted in such a manner?
24       A.    Not with Federico, no.
25       Q.    Did you discuss it with anyone else?
                                                 76

 1       A.    I discussed it with George Shih.
 2       Q.    What did Mr. Shih say?
 3       A.    The concept was also new to him and he
 4   expressed interest in it.
 5       Q.    At what level is the tolerance limit
 6   equivalent to a confidence limit?
 7       A.    It's impossible to say in general.
 8       Q.    Dr. Robson, who ultimately decided on the
 9   methodology to be used in calculating the levels for
10   the Refuge?
11       A.    My understanding it was a group decision.
12       Q.    Who was among the group?
13       A.    The same people I mentioned earlier.
14       Q.    Dr. Walker?
15       A.    Yes.
16       Q.    Mr. MacVicar?
17       A.    Yes.
18       Q.    Mr. Federico?
19       A.    Yes.
20       Q.    And others from the Park and Refuge?
21       A.    Right.
22       Q.    Were you among the group?
23       A.    No.
24       Q.    Is it your opinion, as you sit here today,
25   that the tolerance limit approach is more appropriate
                                                 77

 1   than the confidence limit approach in establishing
 2   limits?
 3       A.    Yes.
 4       Q.    What is the basis for that opinion?
 5       A.    Exhibit 5.
 6       Q.    Was the confidence limit approach
 7   ultimately utilized to develop the limits for the
 8   Refuge?
 9       A.    Yes.
10       Q.    Did you ever -- I may have asked you this
11   question, but I want to make sure the record is
12   clear.  Did you ever utilize the tolerance limit
13   approach to calculate levels for the Refuge or limits
14   for the Park?
15       A.    I did some calculations but I think it was
16   with synthetic data rather than Park data.
17       Q.    Is there any reason why you cannot use the
18   tolerance limit approach to calculate levels for the
19   Refuge or limits for the Park?
20       A.    No.
21       Q.    Do you know why the group that you referred
22   to previously rejected tolerance limits in favor of
23   confidence limits in establishing the limits of the
24   Park and Refuge?
25             MR. MCGRATH:  That's been asked and it's
                                                 78

 1        been answered.
 2             THE WITNESS:  I don't know why, no.
 3             (The document was marked Exb. No. 6.)
 4   BY MR. PERKO:
 5       Q.    Dr. Robson, if you could, direct your
 6   attention to Exhibit Number 6 and I ask you if you
 7   recognize this document?
 8       A.    Yes.
 9       Q.    What is that document?
10       A.    It's a letter to Federico.
11       Q.    Does this reflect your use of the tolerance
12   limit methodology?
13       A.    I'm sorry.  I was studying this.  Would you
14   repeat?
15       Q.    You previously mentioned that you used the
16   tolerance limit approach to calculate levels,
17   utilizing synthetic data, I believe you said.
18       A.    That's what I said, yes.
19       Q.    Does this document reflect that analysis
20   that you performed?
21       A.    This document appears to use real data
22   rather than synthetic data.
23       Q.    Did you prepare this document?
24       A.    Yes.
25       Q.    Did you -- does it utilize the tolerance
                                                 79

 1   limit approach?
 2       A.    Yes, it does.
 3       Q.    Did you actually calculate levels for the
 4   Refuge utilizing the tolerance limit approach?
 5       A.    Yes.
 6       Q.    And where does it reflect the results of
 7   that analysis or does the document reflect the
 8   results?
 9       A.    For the Refuge it appears at 614.
10       Q.    That's bates page 0850614?
11       A.    Yes.
12       Q.    How do the limits reflected in this
13   document compare to those ultimately included in
14   Appendix E?
15       A.    This analysis appears to have made no
16   attempt to incorporate the stage effect in the model
17   so there's a single tolerance limit.
18       Q.    And what is the single tolerance limit?
19       A.    Listed as 15.04 or 1411 depending on the
20   level of confidence.
21       Q.    Could you use the tolerance limit approach
22   to establish limits and still incorporate stage into
23   the model?
24       A.    Yes.
25       Q.    Did you do that?
                                                 80

 1       A.    No.
 2       Q.    Dr. Robson, refer your attention to bates
 3   pages 0850612 through 14.  There's a discussion in
 4   here regarding station effects, in particular, on
 5   page 613.
 6             MR. MCGRATH:  Is there a question?
 7             MR. PERKO:  I'm sorry.
 8   BY MR. PERKO:
 9       Q.    What does the discussion of station effects
10   on page 613 concern?  What were you doing here?
11       A.    Well, I was utilizing an additive model,
12   additive in a logarithmic scale --
13       Q.    For what purpose?  I'm sorry.  Go ahead.
14       A.    -- in which I incorporated three factors,
15   year effect, bimonth effect and station effect,
16   treating stations as fixed effects.
17       Q.    And what did you conclude for station
18   effects?
19       A.    They are listed, the estimates are listed
20   on page 613.
21       Q.    Were the station effects significant?
22       A.    A few were.
23       Q.    Does Appendix E take station effects into
24   account for the Loxahatchee?
25       A.    It does in the outlier analysis.
                                                 81

 1       Q.    How does it take station effects into
 2   account in the outlier analysis?
 3       A.    I can't be sure looking at these tables.
 4       Q.    How are station -- how, if at all, are
 5   station effects taken into account in the actual
 6   setting of the limits in Appendix E?
 7       A.    The data are averaged across stations.
 8       Q.    Why were the data averaged across the
 9   station?
10       A.    In order to produce an index which is
11   non-specific to station.
12       Q.    Why did you want to produce an index that
13   was non-specific to station?
14       A.    It wasn't me who wanted to, but it would
15   otherwise require setting limits individually for
16   each station.
17       Q.    Do you lose any information regarding
18   variability from station to station when you average
19   data across the stations?
20       A.    You would lose an interaction effect.
21       Q.    What effect would that have on the limits
22   themselves?  Would it raise them or lower them?
23       A.    The limit on the average across stations
24   would be considerably less than the limit on an
25   individual station.
                                                 82

 1       Q.    Given the fact that in Exhibit Number 6 you
 2   determined that at least some -- that their station
 3   effects were significant for at least some of the
 4   stations, did you have concerns about averaging
 5   across the stations?
 6       A.    No.
 7       Q.    Why not?
 8       A.    For the reason I just cited.
 9       Q.    Because you were seeking to produce an
10   index that was non-specific to station?
11       A.    Yes.
12       Q.    Could a model with station effects
13   incorporated have been used to set a limit for the
14   average station?
15             MR. MCGRATH:  Just want to object to the
16        extent it calls for speculation.
17             THE WITNESS:  Could be done, yes.
18   BY MR. PERKO:
19       Q.    Did you ever, in fact, utilize such a
20   model?
21       A.    Not for setting confidence limits, no.
22       Q.    For setting tolerance limits?
23       A.    I would have to look.
24       Q.    Direct your attention to page 614 to help
25   refresh your recollection.
                                                 83

 1       A.    If I remember the question, the answer is
 2   yes.
 3       Q.    You did --
 4       A.    I did utilize stations in this.
 5       Q.    In the tolerance limits reflected in
 6   Exhibit Number 6?
 7       A.    Yes.
 8       Q.    In your opinion, as a statistician, is it
 9   more appropriate to incorporate station effects into
10   the model before averaging or to average before you
11   set the limits?
12       A.    It depends.
13       Q.    What does it depend on?
14       A.    There's a problem here of spatial auto
15   correlation.
16       Q.    What is the problem with spatial
17   correlation?
18       A.    In this tolerance limit analysis I ignored
19   the issue of say neighboring stations being
20   positively correlated.
21       Q.    Is it your opinion that averaging solves
22   the spatial correlation problem?
23       A.    Yes.
24       Q.    How so?
25       A.    The effect of the spatial correlation is
                                                 84

 1   automatically incorporated into the estimate for
 2   standard error.
 3             MR. PERKO:  Mark this as Exhibit 7, please.
 4             (The document was marked Exb. No. 7.)
 5   BY MR. PERKO:
 6       Q.    Dr. Robson, if you would, please direct
 7   your attention to Exhibit Number 17 and tell me if
 8   you recognize this document.
 9             MR. BLANK:  Exhibit 7.
10             MR. FITZGERALD:  Exhibit 7.
11   BY MR. PERKO:
12       Q.    I'm sorry.  7.
13       A.    Yes.
14       Q.    What is that document?
15       A.    This is an analysis of the Refuge
16   phosphorus concentrations --
17       Q.    What does this analysis -- go ahead.
18       A.    -- using a linear model which incorporated
19   a dummy model for base year, a dummy variable for
20   stations and stage.
21       Q.    What do you mean by the term, "dummy
22   variable"?
23       A.    Variable that takes on values 1 or 0 where
24   the 1 indicates the date falls in a base year and the
25   0 indicates that it's not in the base year.
                                                 85

 1       Q.    Did you actually utilize the model
 2   reflected in Exhibit Number 7 to develop limits
 3   for -- or levels for the Loxahatchee?
 4       A.    No.
 5       Q.    Why not?
 6       A.    Maybe I misspoke.  This does calculate
 7   limits right here.
 8       Q.    Okay.  You're referring to the table
 9   underneath the heading, Example of 90 Percent
10   Confidence Limits?
11       A.    Yes.  Yes.
12       Q.    Does this exhibit provide an example of a
13   model that determines confidence limits taking into
14   account station effects?
15       A.    Yes.
16       Q.    Were you still considering the use of
17   tolerance levels when you drafted this document in
18   June 1991?
19       A.    I don't believe so.
20       Q.    Why had you foregone the use of tolerance
21   levels by that point -- tolerance limits?
22       A.    I agreed to go along with the idea that
23   confidence limits do represent tolerance limits at
24   some unknown level of confidence.
25       Q.    Why had you agreed to go along with that
                                                 86

 1   assumption?
 2       A.    Because the people that I was dealing with
 3   were not familiar with the concept.
 4       Q.    Were you instructed to adopt that
 5   assumption by the District?
 6       A.    I wouldn't say I was instructed to.  I
 7   agreed to.
 8       Q.    Okay.  Dr. Robson, in your opinion, as a
 9   statistician, do you -- is it your opinion that the
10   model reflected in Exhibit Number 7 is more
11   appropriate than the models in Appendix E?
12             MR. MCGRATH:  Let me just raise an
13        objection for the record to the series of "more
14        appropriate" questions that have been asked,
15        just so -- as much as they're misleading and
16        ambiguous and may incorrectly or improperly
17        suggest that, while something in Dr. Robson's
18        opinion may be more appropriate, that models or
19        methods actually used are not appropriate.  So
20        just to the extent that when we are having these
21        "more appropriate" questions, I object to the
22        question.
23   BY MR. PERKO:
24       Q.    Do you understand the question?
25       A.    Could you repeat it?
                                                 87

 1       Q.    From a statistical standpoint is it your
 2   opinion that the model reflected in Exhibit Number 7
 3   is more appropriate than the model used to develop
 4   the levels for the Refuge in Appendix E?
 5       A.    No.  I would still be concerned about the
 6   effect of this spatial correlation.
 7       Q.    Did you ever calculate the spatial
 8   correlation?
 9       A.    No.
10       Q.    Do you know if anyone else did?
11       A.    I don't know.
12       Q.    Dr. Robson, you have indicated that in
13   various models you looked at station effects, stage
14   and base years.  Did you consider any other
15   parameters in developing models for the Loxahatchee
16   levels?
17       A.    One of these documents include a bimonth
18   effect.
19       Q.    Any other parameters?
20       A.    No.
21       Q.    Did you ever try to determine whether there
22   was a correlation between phosphorous inflows to the
23   Loxahatchee and interior marsh concentrations?
24       A.    No.
25       Q.    Do you know if anyone else has tried to
                                                 88

 1   determine that, if there's such a correlation?
 2       A.    Yes.
 3       Q.    Who?
 4       A.    McClave.
 5       Q.    Anyone else?
 6       A.    I don't know.
 7       Q.    Dr. Robson, is it your understanding that
 8   the interim and long term levels for the Refuge are
 9   ultimately going to be used to regulate inflows to
10   the Loxahatchee?
11       A.    That's my understanding today.
12       Q.    Given that understanding, do you think it's
13   important to know if there's a correlation between
14   inflows and interior marsh concentrations?
15       A.    Yes, but.
16             MR. BLANK:  I'm sorry.  I didn't hear the
17        answer.
18             THE WITNESS:  Yes, but.
19   BY MR. PERKO:
20       Q.    Why are you qualifying your answer?
21       A.    Because I wouldn't know what antecedent
22   period to use.
23       Q.    Could you -- I'm not sure I understand why
24   would that be a concern of yours, an antecedent
25   period.
                                                 89

 1       A.    Well, the concentration in the marsh today
 2   is related to inflow of some previous period, not
 3   today's inflow.
 4       Q.    Okay.  Do you have any intention of
 5   determining whether there's a correlation between
 6   inflows into the Refuge and interior marsh
 7   concentrations?
 8       A.    No.
 9       Q.    Do you know if anyone else plans to look at
10   that?
11       A.    I don't know.
12       Q.    Dr. Robson, how is stage incorporated into
13   the models used to set the limits for the
14   Loxahatchee?
15       A.    As a independent variable in linear
16   regression.
17       Q.    Is it fair to say that the model predicts a
18   relationship between stage and concentration?
19       A.    Yes.
20       Q.    What stage data are utilized in the model?
21       A.    There were three gauges.  Offhand, I don't
22   know how they are identified.
23       Q.    I'm sorry?
24       A.    Offhand, I don't know how they are
25   identified, but there are three gauges involved.
                                                 90

 1       Q.    They are identified in Appendix E?
 2       A.    Yes.  CA1-7, 9.
 3       Q.    What page --
 4       A.    CA1-7, 9 and 8C.  Page 18.
 5       Q.    Is it your opinion, Dr. Robson, that these
 6   three gauge stations are representative of stage
 7   throughout the Loxahatchee National Wildlife Refuge?
 8       A.    I wouldn't know.
 9       Q.    Do you think it's important to determine
10   whether these stations are representative?
11       A.    It would be desirable to know.
12       Q.    And why is that?
13       A.    Well, these are being used as an index, the
14   water level in the Refuge.  It would be desirable to
15   have them interspersed among the stations.
16       Q.    Dr. Robson, on page 18 of Exhibit 3 which
17   is the draft Appendix E, states that, "The POR data
18   includes stage measurement at the three marsh gauges
19   CA1-7, 9 and 8C on each sampling date and average of
20   these three readings is taken as index of stage for
21   the marsh on that date."
22             Why did you average across these three
23   gauge stations?
24       A.    To produce a single number which is an
25   index.
                                                 91

 1       Q.    Could you develop a model that uses
 2   individual stage values rather than the average?
 3       A.    You mean a model to determine limits?
 4       Q.    Yes, sir.
 5       A.    Are you suggesting a limit for each stage,
 6   each gauge?
 7       Q.    Well, first of all, could you use such a
 8   model to develop a -- the type of limits that are
 9   explained in Appendix E?
10       A.    Well, ultimately I would somehow have to
11   average over the station -- over the gauges,
12   otherwise you have gauge specific limit.
13       Q.    Do you lose any information by averaging
14   across the gauge stations?
15       A.    Any information you lose I'm not sure how I
16   could exploit it to begin with.
17       Q.    Let me ask you this.  What effect would --
18   does averaging across the three stations have on the
19   limits?
20       A.    As compared to what?
21       Q.    Would it -- do you know if the limits would
22   be higher or lower if you did not average across the
23   stations?
24       A.    No, I don't know.
25       Q.    Dr. Robson, Appendix E establishes a
                                                 92

 1   threshold stage of 15.42 feet, is that correct?
 2       A.    Yes.
 3       Q.    And below -- any reading below that value
 4   is -- compliance isn't determined whenever there's a
 5   reading below that value, is that right?
 6       A.    I think so.
 7       Q.    I believe you previously testified -- I
 8   think you spoke about a 15.12 threshold value.  Were
 9   you referring to this 15.42?
10       A.    Yes, I was.
11       Q.    I believe you stated that was based on
12   qualitative judgment of others --
13       A.    Right.
14       Q.    -- is that correct?
15             Was that judgment -- let me refer you to
16   Appendix III of Exhibit Number 3 page -- I believe it
17   is 18.
18       A.    48.
19       Q.    48.  I'm sorry.
20             It's entitled, Least Squares Estimation of
21   a Threshold Level for Stage.
22       A.    Yes.
23       Q.    What does this analysis reflect?
24       A.    I was attempting to use a model and let the
25   data dictate threshold level of stage.
                                                 93

 1       Q.    Did you ultimately derive threshold for
 2   stage utilizing this model?
 3       A.    Utilizing the -- all 16 stations and 16
 4   dates I did.
 5       Q.    And what was the threshold level that you
 6   derived?
 7       A.    14.41.
 8       Q.    Did you perform this analysis after the
 9   qualitative judgment had been made to establish 14.  --
10   15.42 as the threshold stage value?
11       A.    I don't recall the order in which it was
12   done.
13             (Thereupon, a recess was taken.)
14   BY MR. PERKO:
15       Q.    Dr. Robson, before we broke we were talking
16   about the averaging of stage across the gauge
17   stations in the Loxahatchee.  Could you tell me whose
18   idea it was to average across the gauge stations?
19       A.    No, I can't tell you.  I don't know.
20             (The document was marked Exb. No. 8.)
21   BY MR. PERKO:
22       Q.    I'd like to refer you to what's been marked
23   as Exhibit Number 8.  Take a look at that.  Do you
24   recognize that document?
25       A.    Yes.
                                                 94

 1       Q.    And what is that document?
 2       A.    It's Appendix E of the Everglades SWIM
 3   Plan.
 4       Q.    Dr. Robson, I should have introduced this
 5   in the record previously rather than referring to the
 6   draft.  From now on when I refer to Appendix E I'll
 7   be referring to this Exhibit Number 8 and if you
 8   happen to notice any changes that would change your
 9   previous answers, please let me know.  I don't
10   believe that there were any substantive changes.
11             Dr. Robson, I'd like to refer you to page
12   E-16 of Exhibit 8 labeled Phosphorous Levels for
13   Loxahatchee National Wildlife Refuge.  Refer your
14   attention to the second paragraph, the statement
15   that, "Permanence of the 16 interior marsh stations
16   implies that in the statistical sense they have fixed
17   effects on phosphorus concentration, each effect
18   being peculiar to the particular permanent local
19   environmental features at that station.  Effects are
20   relative and on the logarithmic scale, approximately
21   additive; this implies that in a comparison between
22   two particular stations their total phosphorus values
23   will tend to stand in the same ratio over time."
24             What is the basis for this statement?
25       A.    Which aspect of the statement are you
                                                 95

 1   referring to?
 2       Q.    Let's take them individually.  The
 3   implication that the, "Permanence of the 16 interior
 4   marsh stations implies that in a statistical sense
 5   they have fixed effects on phosphorus concentration."
 6   What's the basis for that implication?
 7       A.    Well, the phrase "fixed effects" is really
 8   the statistical concept.  It refers to the average at
 9   a given station over an infinite number of years.
10       Q.    What is a basis for determining that such
11   effects are fixed?
12       A.    The permanence of the stations.
13       Q.    What do you mean by the statement that
14   "effects are relative"?
15       A.    That's explained in the last half of that
16   sentence.
17       Q.    Okay.  And you're referring to the
18   statement that, "This implies that in a comparison
19   between two particular stations their TP values will
20   tend to stand in the same ratio over time."
21       A.    Yes.
22       Q.    What's the basis for that statement?
23       A.    It's based on the assumption of additive
24   model on the logarithmic scale.
25       Q.    Could you tell me what you mean in laymen's
                                                 96

 1   terms?  Is that possible?
 2       A.    Well, actually the second half of the
 3   sentence says the same thing, that the ratio over
 4   time tends to be constant.
 5       Q.    Why does the ratio tend to be constant over
 6   time?
 7       A.    This is based on a comparison of the two
 8   figures in the last -- page E-60, the last page of
 9   the appendix.
10       Q.    Okay.  Figure 3?
11       A.    Yes.
12       Q.    Suppose the additive assumption is
13   incorrect and the stations -- the TP value at
14   individual stations would not -- did not stand in the
15   same ratio over time.  What effect would that have on
16   the limits?
17             MR. FITZGERALD:  I think I have object to
18        the question because you've mischaracterized
19        what the witness said earlier.  Correct me if
20        I'm wrong.
21             THE WITNESS:  I may have misunderstood.
22        But you just said to assume that the additive
23        assumption was incorrect and there is no
24        testimony that the additivity is based on an
25        assumption.  I think the exhibit says just the
                                                 97

 1        opposite.
 2   BY MR. PERKO:
 3       Q.    Well, assume the stations' TP
 4   concentrations at the particular stations did not
 5   vary, they were the same ratio -- in the same ratio
 6   over time.  What effect would that have on the
 7   limits?
 8       A.    The greater the amount of variability in
 9   such a ratio, the larger would be the standard area
10   used in calculating limits, so the limits would go
11   up.
12       Q.    Your discussion of additive effects in this
13   paragraph, do you consider that to be an assumption
14   of the model?
15       A.    The approximate additivity is the
16   assumption, yes.
17       Q.    Did you do any independent analysis to
18   verify that assumption?
19       A.    I did not do a formal test in additivity.
20       Q.    Could you explain to me how Figure 3 on
21   page E-60 supports the additive assumption?
22       A.    It's really the comparison between the top
23   panel and the lower panel.
24       Q.    What -- I'm sorry.  Go ahead.
25       A.    The lower panel indicates a more uniform or
                                                 98

 1   homogeneous error distribution.
 2       Q.    What is the difference between the upper
 3   panel and the lower panel?  What is being --
 4       A.    Every panel represents residuals based upon
 5   analyzing the phosphorus concentration untransformed
 6   whereas the lower one is transformed on the log
 7   scale.
 8       Q.    And this indicates to you that the effects
 9   are additive?
10       A.    It indicates that the additive model
11   produces relatively homogeneous error variance.
12       Q.    Does anything else in Appendix E support
13   the additivity assumption?
14       A.    On page E-57 this supports the same
15   conclusion I reached comparing the two panels.
16       Q.    How does it support that conclusion?  You
17   are referring to the, Outlier Analysis for the 14
18   Stations X 14 Dates Refuge TP Data?
19       A.    Comparing the extreme P values on the
20   untransformed scale versus on the transformed scale.
21   Transformation achieves greater symmetry.
22       Q.    How did you transform the data?
23       A.    The log transform.
24       Q.    Did you ever use a different method of
25   transforming the data?
                                                 99

 1       A.    I didn't, no.
 2       Q.    Do you know if anyone else did?
 3       A.    Yes.
 4       Q.    Who?
 5       A.    Walker.
 6       Q.    What did he -- what other method did he
 7   use?
 8       A.    As I recall, it was the square root of the
 9   reciprocal.
10       Q.    Do you know why Dr. Walker used the square
11   root of the reciprocal?
12       A.    No.  I can't be sure.
13       Q.    Do you know whether a -- I'll refer you
14   back to Figure 3 on page E-60.
15             Do you know whether a non-additive model
16   would produce an even better error distribution than
17   the bottom panel of Figure 3, a non-additive model
18   that still uses the log transform?
19       A.    I couldn't rule out that possibility.
20       Q.    Have you ever attempted to use such a
21   model, a non-additive model?
22       A.    By non-additive do you mean non-linear on
23   the log scale?
24             MR. MCGRATH:  Just what do you mean,
25        counsel?
                                                 100

 1             MR. PERKO:  Yes.
 2             MR. FITZGERALD:  Counsel, could you explain
 3        that in detail?
 4             THE WITNESS:  You said yes?
 5   BY MR. PERKO:
 6       Q.    Yes.  I did say yes to your question.  I
 7   ignored Mr. Fitzgerald.
 8       A.    I did not attempt to use a nonlinear model
 9   on the log scale.
10       Q.    Do you know if Dr. Walker did?
11       A.    I'll have to take that back.  I did use a
12   nonlinear model on this when I was trying to estimate
13   a threshold that involved a nonlinear parameter.
14       Q.    Are you referring to the Appendix E in
15   which you developed a threshold on stage?
16       A.    Yes.
17       Q.    Is additivity an assumption independent of
18   whether log transformation is used?
19       A.    No.
20       Q.    No?
21       A.    No.
22       Q.    Is the threshold model that you just
23   referred to in Appendix III of Appendix E, is it
24   non-additive with respect to stations?
25       A.    I have forgotten what page that's on.
                                                 101

 1       Q.    That is on page E-49.
 2       A.    Here I have averaged over stations.
 3       Q.    So it is --
 4       A.    I don't have the station effect in the
 5   model.
 6       Q.    So, in other words, it is additive with
 7   respect to stations, is that correct?
 8       A.    Any non-additivity with respect to stations
 9   would be incorporated into the error of variance.
10       Q.    Dr. Robson, referring your attention again
11   to page E-16 in the second paragraph where you speak
12   of "permanence of the 16 interior marsh stations."
13   When you use the term "permanence," do you mean
14   permanence in their location over time?
15       A.    Yes.
16       Q.    When you say in the next sentence that,
17   "effects are relative," would it be more appropriate
18   to say that effects are assumed to be relative?
19       A.    Yes.
20       Q.    Did you perform any analysis to determine
21   whether, in deed, the effects are relative?
22       A.    I've already answered that question.
23       Q.    Referring to Figure 3?
24       A.    No.  You asked that specific question
25   before, I think.
                                                 102

 1       Q.    Okay.  Did you -- could you refresh my
 2   recollection on what you said?
 3       A.    My reply was I did not perform a formal
 4   test of additivity.
 5       Q.    Dr. Robson, I'd like to refer your
 6   attention to figure number 6 on page -- Table 6 on
 7   page E-18.  To your knowledge is the data reflected
 8   in this table, is the data accurately reflected in
 9   this table?
10       A.    To my knowledge, yes.
11       Q.    Who actually inputted the data?  Do you
12   know?
13       A.    I'm trying to recall whether I was provided
14   with a raw data file to begin with and extracted this
15   from the data file or whether I was provided with
16   these data as such and I don't remember.
17       Q.    Did you actually type Appendix E into your
18   computer?
19       A.    Yes.
20       Q.    Okay.
21             MR. MCGRATH:  Do you mean Appendix E or the
22        table that we're discussing right now?
23             MR. PERKO:  The table we're discussing
24        right now.
25             THE WITNESS:  Yes.
                                                 103

 1   BY MR. PERKO:
 2       Q.    Dr. Robson, how is time accounted for in
 3   the models used to develop the limits for the Park
 4   and levels for the Refuge in Appendix E?
 5       A.    The Park itself incorporated is a linear
 6   regression term.
 7       Q.    What about the Refuge?
 8       A.    It's incorporated as a dummy variable.
 9       Q.    For the Refuge why was a dummy variable
10   used?
11       A.    My recollection is that with stage in the
12   model the -- an attempt to incorporate time produced
13   a non-sufficient regression for time.
14       Q.    And who performed that attempt to integrate
15   time into the model?
16       A.    Walker.
17       Q.    Whose decision was it to use a dummy
18   variable in the equations?
19       A.    That followed from the conclusion that the
20   regression on time was insignificant with stage in
21   the model.
22       Q.    Who ultimately decided to use the dummy
23   variable?
24       A.    I don't know.
25       Q.    What is the effect of using the dummy
                                                 104

 1   variable?
 2       A.    It provides a means of utilizing all the
 3   data to project back to the base year.
 4       Q.    Does the use of the dummy variable assume a
 5   step function increase in phosphorus concentrations
 6   between May and August 1979?
 7       A.    That assumption would be compatible with
 8   the use of the dummy variable.
 9       Q.    What is a step function increase?
10       A.    It's a jump from one level to another.
11       Q.    Do you know what that jump would be, based
12   on the data used to develop the Loxahatchee levels?
13       A.    I don't know offhand, no.
14       Q.    Can you determine that from Appendix E?
15       A.    On page E-21 the fitted TP equals 2.59
16   minus .4856, etc.  Minus .4856 exponential would be
17   the jump you referred to.
18       Q.    Is it reasonable to assume that such a jump
19   occurred?
20       A.    No.  I didn't make that assumption myself.
21       Q.    Okay.  Who made that assumption?
22       A.    Nobody to my knowledge.
23       Q.    In your opinion is that assumption
24   reasonable?
25       A.    I can see no reason to accept that kind of
                                                 105

 1   assumption.
 2       Q.    You can see no reason to accept it?
 3       A.    Right.
 4       Q.    Did Dr. Walker ever perform a regression
 5   analysis to determine a trend in ENP inflows?
 6       A.    A trend in discharge you mean?
 7       Q.    Phosphorus inflows to the ENP.
 8       A.    Yes.
 9       Q.    And what did that analysis -- what were the
10   results of that analysis?
11       A.    Well, we've already talked about the
12   Seasonal Kendall Test.  The other analysis was the
13   regression model.
14       Q.    Right.  And the regression model was used
15   in Appendix E, is that correct?
16       A.    Yes.
17       Q.    And what did that -- what were the results
18   of that model?  Did it show a trend?
19       A.    Yes.
20       Q.    What trend did it show?
21       A.    We have been over this earlier.
22       Q.    We have?
23       A.    Yes.  On page E-6 fitted TP equals are
24   there.
25       Q.    5.91 exponentiation --
                                                 106

 1       A.    No.
 2       Q.    I'm sorry.  Go ahead.
 3       A.    That's it.  That's the per year change in
 4   the TP.
 5       Q.    I follow you.
 6             In your calculation of the limits for the
 7   Refuge did you ever use any other ways of expressing
 8   time other than the dummy variable?
 9       A.    I didn't, no.
10       Q.    Did Dr. Walker?
11       A.    I testified a moment ago that my
12   recollection is that he did use time and found it
13   nonsignificant.
14       Q.    You're right.
15             Dr. Robson, what data was utilized to
16   derive the long term levels for the Refuge?
17       A.    Data from the clean 3 stations.
18       Q.    Who decided to use the clean 3 stations?
19       A.    I don't know who all was involved in that
20   decision.
21       Q.    Would it have been the same people that you
22   mentioned previously?
23       A.    Yes.
24       Q.    Federico, MacVicar, Walker, representatives
25   from the Refuge and representatives from the Park?
                                                 107

 1       A.    I'm not sure that all of these people were
 2   involved in that decision.
 3       Q.    Would Dr. Walker have been involved in the
 4   decision?
 5       A.    I would guess that he was, yes.
 6       Q.    What about Mr. Federico?
 7       A.    I don't know.
 8       Q.    How were the clean 3 stations selected or
 9   why were they selected as the basis for the long term
10   levels?
11       A.    It's explained in the appendix somewhere.
12   Page E-20.
13       Q.    Are you referring to the third paragraph --
14       A.    Yes.
15       Q.    -- where it states that, "Long term levels
16   are intended to achieve water quality comparable to
17   those existing at the least impacted stations during
18   the 1978-'79 water year"?
19       A.    Yes.
20       Q.    "Development of such levels utilizing data
21   from the entire POR rather than just in 1978-'79 data
22   was based on selection of 3 clean stations."  And
23   later on in the paragraph it states that, "These
24   clean 3 were adopted as representing the cleanest TP
25   condition during the POR and the geometric means over
                                                 108

 1   these 3 stations over 14 sampling days of the POR in
 2   calculating long term phosphorous levels for the
 3   Refuge."
 4             Is that your understanding of the basis for
 5   selecting the 3 clean stations?
 6       A.    Yes.
 7       Q.    I'm sorry.  I skipped a sentence there.
 8             Also says that, "Table 9 reveals that the 3
 9   clean stations CA1 through 5, 6 and 16 were
10   distinctive in having the lowest phosphorus
11   concentration action POR with geometric means
12   approximating the classical bound of 10 PPB for
13   oligotrophic waters, as well as having the lowest
14   geometric means for the 1978-89 base year."
15             Why was that criteria utilized in selecting
16   the 3 clean stations?
17             MR. MCGRATH:  I would just object to the
18        form, seeing as how Dr. Robson previously
19        testified he wasn't involved in the decision
20        making process.
21             MR. PERKO:  Doesn't necessarily mean he
22        doesn't know why.
23             THE WITNESS:  Well, my understanding is
24        limited to what is said here.
25   BY MR. PERKO:
                                                 109

 1       Q.    Do you know where the classical bound of 10
 2   PPB came from?
 3       A.    No.
 4       Q.    Do you know who originally proposed it?
 5       A.    I first heard it from Walker.
 6       Q.    What is your understanding of the purpose
 7   of the long term levels?
 8       A.    The first sentence, "Long term levels are
 9   intended to achieve water quality comparable."
10       Q.    Comparable to those existing at the least
11   impacted stations during the 1978-'89 water year?
12       A.    Yes.
13       Q.    Least impacted -- do you know what the
14   term, "least impacted" refers to?
15       A.    I don't know for sure.
16       Q.    What is your understanding of the meaning
17   of that term?
18       A.    Having the lowest phosphorus concentration
19   for the POR.
20       Q.    Do you know why the long term levels were
21   intended to achieve the water quality comparable to
22   those existing at the least impacted stations?
23       A.    Again, I don't know why.  I can only
24   surmise.  It wasn't my decision.
25       Q.    Do you know if it was based on the
                                                 110

 1   regulatory requirement?
 2       A.    No, I don't know.
 3       Q.    Do you know if it was based on a perceived
 4   water quality requirement?  By that I mean some
 5   perceived environmental need.
 6       A.    I have to reiterate.  My understanding is
 7   reading what we have in front of us here.
 8       Q.    Did you ever consider more than 3 clean
 9   stations in calculating the long term limits?
10       A.    I didn't.
11       Q.    Do you know if Dr. Walker did?
12       A.    I don't know.
13       Q.    When the decision was made to use the 3
14   clean stations was there ever any consideration of
15   going back to individual station values rather than
16   averaged values?
17       A.    Not to my knowledge.
18       Q.    Could a model be developed to determine
19   long term levels using individual as opposed to
20   averaged values?
21       A.    The question would then arise whether you
22   want specific -- station specific limit for each of
23   the 3 or other clean stations.
24       Q.    If you were to use individual values rather
25   than average values would you necessarily have to
                                                 111

 1   determine station specific limits?
 2       A.    If not, you'd be back to averaging.
 3       Q.    How did the model used to develop the ENP
 4   interim limits differ from the model used to develop
 5   the Refuge interim levels?
 6       A.    ENP did incorporate the independent
 7   variable time.
 8       Q.    It -- did you say did not?
 9       A.    Did.
10       Q.    It did incorporate.
11       A.    And it did not use the log transformation.
12       Q.    How did it incorporate the independent
13   variable time?
14       A.    Linearly.
15       Q.    Why were different models used to calculate
16   the ENP interim limits?
17       A.    Well, for one thing, we didn't have
18   discharge to work with and we had fewer data, shorter
19   time span.
20       Q.    You are referring to the Loxahatchee,
21   correct?
22       A.    Loxahatchee versus ENP.
23       Q.    You didn't have a discharge to deal with in
24   the Loxahatchee and you had data over a shorter
25   period of time for the Loxahatchee?
                                                 112

 1       A.    Yes.
 2       Q.    Just so I understand you you say you didn't
 3   have discharge data for the Loxahatchee?
 4       A.    I didn't have it.
 5       Q.    Okay.  But you did have it for the ENP?
 6       A.    Yes.
 7       Q.    To your knowledge did anyone else have
 8   discharge data for the Loxahatchee?
 9       A.    I don't know.
10       Q.    You also flow-weighted for the ENP limits,
11   is that correct?
12       A.    Correct.
13       Q.    Why didn't you flow-weight for the
14   Loxahatchee levels?
15       A.    We don't have flow at the sampling
16   stations.
17       Q.    Why did you time average for the ENP
18   limits?
19       A.    You mean within years?
20       Q.    Uh huh.
21       A.    Again, I can only guess.
22       Q.    Whose decision was it to time average?
23       A.    I don't know.
24       Q.    What effect did time averaging have on the
25   limits?
                                                 113

 1       A.    It has the effect of, in large, overcoming
 2   the auto correlation between sampling dates.
 3       Q.    Would it result in higher or lower levels
 4   then if you had not time averaged?
 5       A.    It would depend on what time series
 6   analysis achieved.  I can't say.
 7       Q.    What effect would flow-weighting have on
 8   the limits?
 9       A.    As compared to what?
10       Q.    Non flow-weighted calculations.
11       A.    I don't know.
12       Q.    How does the Park model treat 0 or negative
13   flow values?
14       A.    You are referring to the model we're
15   looking at here?
16       Q.    Yes.
17       A.    As I recall they, in effect, were excluded
18   in computing the flow-weighted average because the
19   flow as 0 would automatically exclude itself.
20   Negative I understand was excluded.
21       Q.    Dr. Robson, let me refer you to page E-10.
22   It's labeled, Exceedance Frequency Guideline for the
23   Estimated Percent of time that Park Inflow Exceeds 10
24   Parts Per Billion.
25       A.    Yes.
                                                 114

 1       Q.    What was the purpose of this analysis.
 2       A.    This was to develop supplemental guidelines
 3   based on the exceedance concept.
 4       Q.    These guidelines -- are these guidelines to
 5   be used in determining compliance with the limits?
 6       A.    My understanding is they are not.
 7       Q.    Was a frequency analysis performed for the
 8   Loxahatchee?
 9       A.    Not to my knowledge.
10             MR. PERKO:  Let's take a break.
11             (Thereupon, a recess was taken.)
12   BY MR. PERKO:
13       Q.    Dr. Robson, I believe you previously
14   testified that you did not attempt to determine
15   spatial correlation in establishing these limits in
16   the models --
17       A.    Right.
18       Q.    -- is that correct?
19             If you were to assume that spatial
20   correlation were not there would a model with station
21   effects be more appropriate than using average of the
22   station concentrations?
23       A.    Again, you're assuming they're going to
24   produce a single rather or a station specific --
25       Q.    Yes.
                                                 115

 1       A.    -- limit?
 2             I can't answer that offhand.  I don't know.
 3       Q.    What would you need to do to answer that
 4   question?
 5       A.    Well, I'd have to see exactly how we're
 6   going to calculate this single limit.
 7       Q.    Dr. Robson, I refer your attention to
 8   tables -- Table 1 on page E-6 and Table 2 on E-8.
 9   Table 1 is the Calculation of Interim Limits and
10   Table 2 is the Calculation of Long Term Limits.
11             I notice here on Table 1 that you're
12   comparing the S12 plus S333 flow-weighted mean versus
13   the S12 S333 composite annual flow, is that correct?
14       A.    Yes.
15       Q.    And on Table 2 for the long term limits
16   you're comparing S12 flow-weighted mean TP versus S12
17   plus S333 composite annual flow, is that correct?
18       A.    Yes.
19       Q.    Why does Table 1 use S12 plus S333
20   flow-weighted mean whereas Table 2 uses S12
21   flow-weighted mean by itself?
22       A.    I thought that that point was addressed
23   somewhere in the document.
24       Q.    Are you speaking of page E-5 at the bottom?
25       A.    Yes.
                                                 116

 1       Q.    It states that, "With calculations now
 2   based strictly on POR data for the S12 structures,
 3   excluding the S333 POR water quality data on the
 4   grounds that in the long term they will be minimal
 5   inflow to Shark River Slough through this structure."
 6       A.    Yes.
 7       Q.    That's the basis for distinction?
 8       A.    Yes.
 9       Q.    Is that a result of the modified water
10   deliveries GDM that there will be minimal inflow to
11   Shark River Slough through the S333 structure?
12       A.    I don't know.
13       Q.    Dr. Robson, does Appendix E utilize the
14   same outlier analysis for the ENP as it does for the
15   Refuge, same model?  Specifically refer you to
16   Appendix 1 on page E-25 which is the Outlier Analysis
17   for S12 and S333 and Appendix 4 on page E-55 the
18   Outlier Analysis for the Refuge.
19       A.    Well, it's similar in several instances.
20   Uses the log transform, incorporates month effects
21   since we do have data almost twice a month and it
22   incorporates year effect.
23       Q.    Is this a different model than was used to
24   actually calculate the limits in the levels in the
25   model?
                                                 117

 1       A.    Yes.
 2       Q.    Why was a different model used to calculate
 3   the limits and levels, then, to perform the outlier?
 4       A.    Well, for one thing the outlier analysis
 5   does suffer from the effect of serial correlation
 6   between biweekly samples versus ultimate model serial
 7   correlation between years.
 8       Q.    Why would serial correlation be important
 9   in determining the model for developing the limits
10   and not important for developing the -- doing the
11   outlier analysis?
12       A.    This is an admitted weakness of the outlier
13   analysis and does not incorporate serial correlation.
14       Q.    Let me ask you, Dr. Robson.  Was the
15   outlier -- were the outlier analyses performed for
16   the Loxahatchee after exclusion of the two sampling
17   dates and stations?
18       A.    Yes.
19       Q.    Is it normal statistical practice to
20   exclude data prior to outlier analysis?
21       A.    If you have subject matter reasons for
22   excluding data, yes.
23       Q.    Absent subject matter reasons, there would
24   be no reason?
25       A.    Right.
                                                 118

 1       Q.    Were the outliers identified through the
 2   outlier analysis actually excluded from the
 3   calculation of the limits?
 4       A.    Yes.
 5       Q.    For both the Park and the Refuge?
 6       A.    My recollection is that there were none for
 7   the Refuge.
 8       Q.    Okay.  And there was one for the Park?
 9       A.    Right.
10             MR. PERKO:  Would you mark this, please.
11             (The document was marked Exb. No. 9.)
12   BY MR. PERKO:
13       Q.    Dr. Robson, I direct your attention to
14   what's been marked as Exhibit Number 9.  Do you
15   recognize that document?
16       A.    Yes.
17       Q.    And what is this document?
18       A.    This is Walker's comments on the settlement
19   agreement of July 11, 1991.
20       Q.    Refer you to the paragraph on page 1
21   labeled Appendix A-5.  Second sentence states, "It
22   was my impression that Doug Robson's results were to
23   be used in the final settlement agreement.  I don't
24   know whether he was ever able to reproduce my results
25   (still trying to resolve differences.)"
                                                 119

 1             Who actually calculated the ENP limits
 2   first?  Was it Dr. Walker or you?
 3       A.    Walker.
 4       Q.    Were you ever able to reproduce
 5   Dr. Walker's results?
 6       A.    I don't recall whether I actually did the
 7   calculation.
 8       Q.    Later on in that paragraph it states that,
 9   "To be consistent, it seems preferable to use
10   Robson's results in the settlement agreement."
11             Is it your understanding that your results
12   were incorporated into the settlement agreement?
13       A.    No.
14       Q.    So, in other words, Walker's results were
15   incorporated in the settlement agreement?
16       A.    That's my understanding.
17       Q.    Did you ever resolve the discrepancies
18   described in the data bases?
19       A.    Yes.
20       Q.    How were they resolved?
21       A.    That's explained on page E-10 of the
22   appendix.
23       Q.    Where on page E-10?
24       A.    The sentence just above the first equation,
25   "rejection limit as calculated."
                                                 120

 1       Q.    Okay.  If I understand you correctly, it's
 2   your understanding that the data utilized by
 3   Dr. Walker is what is actually reflected in the
 4   calculation of the limits, is that correct?
 5       A.    That's my understanding, but I'm not sure
 6   of that.
 7       Q.    I'm going to refer you to the second page
 8   of Exhibit Number 9.  There's a discussion of degrees
 9   of freedom and T values.
10       A.    Yes.
11       Q.    Could you explain to me what Dr. Walker was
12   talking about in this paragraph?
13       A.    This issue is also addressed somewhere in
14   the document.  Page E-50.  That entire paragraph is a
15   discussion of the fact that I attempted to get a
16   threshold stage value and when I was using the 14
17   stations and 14 dates, data set attempting to, that
18   model dictated that instead of the four parameter
19   model I should be using this three parameter model.
20   With the threshold stage incorporated in the model,
21   there are four degrees of freedom accounted for the
22   by the model.  But the model I actually ended up
23   using is only three unknown parameters to estimate so
24   there are three degrees of freedom associated with
25   that.  But the fact that I was led to this three
                                                 121

 1   parameter from a four parameter model indicates that
 2   there's a penalty for that degree of freedom, penalty
 3   for that search and, as I remark here, another
 4   statistician might have started with this three
 5   parameter model to begin with, so it raises a
 6   philosophical question of whether I should pay the
 7   penalty for that one degree of freedom.
 8       Q.    You ultimately used the three degrees of
 9   freedom assumption in your calculations, is that
10   correct?
11       A.    I used the three degree of freedom
12   assumption in calculating the standard error and I
13   have used the four degree freedom of assumption in
14   looking up the T value in the T tables.  This was my
15   own compromise.
16       Q.    What effect would that have on the limit?
17       A.    It would have a small effect because the
18   difference between the two T values is -- it's a
19   difference only in the second decimal place, as I
20   recall.
21       Q.    Was it your decision to make that
22   assumption?
23       A.    Yes, it was.
24             MR. PERKO:  Mark this, please.
                                                 122

 1             (The document was marked Exb. No. 10.)
 2   BY MR. PERKO:
 3       Q.    Dr. Robson, I refer you to Exhibit Number
 4   10.  Do you recognize this document?
 5       A.    Yes.
 6       Q.    Did you author this document?
 7       A.    Yes.
 8       Q.    What does it reflect?
 9       A.    The first two pages are a print out of the
10   start of a data file by Kevin Rodberg.
11       Q.    By Kevin Rodberg?
12       A.    Yes.
13       Q.    What's on the third page?
14       A.    Third page is a graph comparing confidence
15   limit and tolerance limit.
16       Q.    Was this graph prepared during the analysis
17   we discussed previously which you used synthetic data
18   in calculating limits using the tolerance approach?
19       A.    It looks like it used real data on the
20   second page.
21       Q.    What do the X and Y axes on the table on
22   the second page reflect?  I'm sorry.  On the third
23   page.  This one, third page of the document.
24       A.    The X axis would represent S which I
25   presume is stage and the why X axis would represent
                                                 123

 1   concentration, phosphorus concentration.
 2       Q.    The fourth page shows a similar table
 3   reflecting a line for -- says tolerance limit and a
 4   line for confidence limit.  What model did you use in
 5   developing this table?
 6       A.    This would have been the model that Walker
 7   used in analyzing the reciprocal of the square root
 8   again.
 9       Q.    And that was for the Loxahatchee?
10       A.    Yes.
11       Q.    Was the same model utilized to develop the
12   tolerance limit as the confidence limit?
13       A.    Yes.
14       Q.    And does, once again, the X axis reflect
15   stage and the Y axis reflect concentration?
16       A.    Yes.  This is the same figure.
17       Q.    Same figure.
18             (The document was marked Exb. No. 11.)
19   BY MR. PERKO:
20       Q.    Dr. Robson, if you would, take a look at
21   Exhibit Number 11.  Do you recognize this document?
22       A.    Yes.
23       Q.    Did you draft this document?
24       A.    Yes.
25       Q.    What is this document?
                                                 124

 1       A.    Just a note to myself after having seen
 2   examples of Kruskal Wallis Test applied by somebody
 3   from the Park.
 4       Q.    What was the purpose of Park employees' use
 5   of the Kruskal Wallis Test?
 6       A.    Well, the object was to -- apparently to
 7   test the homogeneity of years 1978 through '88.
 8       Q.    Did you ever pursue the Kruskal Wallis Test
 9   further?
10       A.    No.
11       Q.    Do you intend to?
12       A.    No.
13       Q.    Do you recall who the Park employee was?
14       A.    No, I don't.
15       Q.    Is the Kruskal Wallis Test appropriate in
16   your opinion?
17       A.    No.
18       Q.    Why not?
19       A.    The reasons are indicated here.
20             (The document was marked Exb. No. 12.)
21   BY MR. PERKO:
22       Q.    Dr. Robson, if you would, take a look at
23   Exhibit Number 12 and tell me if you recognize this
24   document.
25       A.    Yes.
                                                 125

 1       Q.    Did you draft this document?
 2       A.    Yes.
 3       Q.    What does it reflect?
 4       A.    The calculations I did when I was exploring
 5   the properties of the Seasonal Kendall Test.
 6       Q.    Dr. Walker's analysis we discussed
 7   previously?
 8       A.    These were synthetic data.
 9       Q.    Is the analysis reflected in this document
10   consistent with your conclusion that the Seasonal
11   Kendall Test was not appropriate in determining
12   whether there's a trend in phosphorus inflows?
13       A.    Yes.
14             MR. PERKO:  Let's mark this as Composite
15        Exhibit 13.
16             (The document was marked Exb. No. 13.)
17   BY MR. PERKO:
18       Q.    Dr. Robson, I direct your attention to
19   what's been marked as Composite Exhibit Number 13,
20   two documents.  Do you recognize these documents?
21       A.    I produced them, yes.
22       Q.    Did you also produce them in connection
23   with your review of Dr. Walker's Seasonal Kendall
24   Test analysis?
25       A.    Yes.
                                                 126

 1       Q.    Are both of these documents consistent with
 2   your conclusion that the Seasonal Kendall Test was
 3   not appropriate for determining whether there's a
 4   trend in phosphorus inflow to Everglades National
 5   Park?
 6       A.    Yes.
 7             (The document was marked Exb. No. 14.)
 8   BY MR. PERKO:
 9       Q.    Dr. Robson, I refer you to what's been
10   marked as Exhibit Number 14.  Do you recognize this
11   document?
12       A.    Yes.
13       Q.    And what is it?
14       A.    It's my vita.
15       Q.    Is that an accurate and complete copy of
16   your vita?
17       A.    Pardon me?
18       Q.    It is an accurate and complete copy of your
19   vita?
20       A.    It's not quite up to date.
21       Q.    Have you since prepared an updated version?
22       A.    Yes.
23       Q.    When did you prepare that?
24       A.    I believe it was 1992.  It was for a
25   different project.
                                                 127

 1       Q.    How would it differ from this version?
 2       A.    It would include, on page 2, some more
 3   recent professional activities and on the last page
 4   would include further publications.
 5       Q.    Could you provide us a copy of the more
 6   recent one?
 7       A.    Yes.  Sure.
 8       Q.    What additional recent professional
 9   activities would you include on page 2?
10       A.    Consultant for the U.S. Fish and Wildlife
11   Service on designing a monitoring program for the
12   spotted owl project.
13       Q.    Is that the project you went to Alaska on?
14       A.    No.
15       Q.    That was polar bear?
16       A.    That was polar bear so that would be
17   another one.
18       Q.    Was that also for the U.S. Fish and
19   Wildlife Service?
20       A.    Yes.
21       Q.    Also a monitoring program --
22       A.    Yes.
23       Q.    -- for the polar bears in Alaska, is that
24   correct?
25       A.    Right.  Alaska and Russia.
                                                 128

 1       Q.    Any other recent experience that's not
 2   indicated on this copy of the resume?
 3       A.    Yes.  I was a consultant to a bioeconomics
 4   consulting firm to design angler surveys on 26 rivers
 5   in western Montana.
 6       Q.    You said angler surveys?
 7       A.    Right.
 8       Q.    Fishing?
 9       A.    Yes.
10       Q.    In western Montana?
11       A.    Yes.
12       Q.    Any other recent experience that's not
13   indicated on this exhibit?
14       A.    There's another U.S. Fish and Wildlife
15   project on monitoring the population of Pacific
16   Grant, g-r-a-n-t.
17       Q.    What is a Pacific Grant?
18       A.    Referring to the Pacific population of
19   grants.  Grant is a small goose.  Looks like a Canada
20   goose only it's smaller.
21       Q.    Any other recent professional activities?
22       A.    Yes.  There's one more I know.
23       Q.    What was that project?  Sorry.  I didn't
24   mean to rush you.
25       A.    Critiquing a document which is attempting
                                                 129

 1   to develop alternative experimental designs for
 2   testing the impact of pesticide on non-target
 3   species.
 4       Q.    Would that be vegetative species or --
 5       A.    No.  Primarily birds.
 6       Q.    Okay.  Is there any other recent
 7   professional activities that are not indicated in
 8   Exhibit 14?
 9       A.    I can't think of any.
10       Q.    Okay.  I believe you mentioned that you
11   would also add a few new publications, is that
12   correct?
13       A.    Yes.
14       Q.    What publications would you add?
15       A.    There's a book which I'm a second author
16   of.
17       Q.    Do you recall the title?
18       A.    No.  I'm sorry.
19       Q.    What was the subject matter?
20       A.    It's --
21             MR. MCGRATH:  Guess.
22             THE WITNESS:  That I do know.  Subject
23        matter Mark Recapture Analysis in Control
24        Experiments.
25   BY MR. PERKO:
                                                 130

 1       Q.    What is meant by the term, "Mark Recapture
 2   Analysis"?
 3       A.    Marking animals and later recovering them.
 4       Q.    That's what I thought.
 5             Any other publications that are not
 6   indicated on this version of your resume?
 7       A.    I'm not sure.
 8             MR. PERKO:  We'll send you a letter
 9        requesting that information.
10             MR. MCGRATH:  Sure.
11   BY MR. PERKO:
12       Q.    Referring your attention back to page 2,
13   Dr. Robson, you indicate on here that you have recent
14   professional experience as a statistical consultant
15   to W.S. Overton Oregon State University, EPA on the
16   design and analysis of the National Surface Water
17   Survey, 1985 to '88.
18       A.    Yes.
19       Q.    What did that project involve?
20       A.    It was selecting water quality samples from
21   samples of lakes and streams.
22       Q.    What was the purpose of the project?  Was
23   it an inventory?
24       A.    The EPA is planning to undertake a
25   continuing monitoring program of all natural
                                                 131

 1   resources in the U.S.
 2       Q.    Did you assist in the design of this
 3   survey?
 4       A.    Yes.
 5       Q.    What did you -- what particular aspects did
 6   you assist on?
 7       A.    On developing estimation formulas.
 8       Q.    Estimation formulas for what?
 9       A.    Well, for example, one of the summary
10   statistics is supposed to represent the frequency
11   distribution of streams within a given region of the
12   U.S. which have -- well, frequency distribution of
13   water quality parameters for streams in the U.S.
14       Q.    What was the purpose of developing the
15   frequency distribution plan?
16       A.    To allow administrators to determine the
17   number of streams that might be below some desired
18   level of water quality.
19       Q.    What specific water quality parameters did
20   you analyze in connection with this project?
21       A.    It was all water chemistry.
22       Q.    What specific parameters?
23       A.    Understand, I didn't analyze data myself.
24       Q.    Okay.
25       A.    I was concerned with the development of
                                                 132

 1   formulas.
 2       Q.    How many formulas did you actually develop
 3   approximately?
 4       A.    I don't count formulas.
 5       Q.    I'll withdraw the question.
 6       A.    Okay.
 7       Q.    Did you utilize -- in connection with this
 8   project did you utilize any of the statistical
 9   analyses used in Appendix E?
10       A.    No.
11       Q.    Was your work in any way similar to
12   Appendix E?
13       A.    I would have to say no.
14       Q.    Do you anticipate relying on any experience
15   gained in this project in support of your testimony
16   in this proceeding?
17       A.    No.
18       Q.    What was the EPA Episodic Response Project
19   Research Plan?
20       A.    I have forgotten.
21       Q.    Let me ask you this.  Of the professional
22   activities listed on page 2 did any involve
23   statistical analysis of water quality data?
24       A.    I did some analyses of water quality data
25   collected in the Maryland Stream Survey.
                                                 133

 1       Q.    And what did that project involve?
 2       A.    My part was assisting them in the design of
 3   the survey and the selection of sample streams and
 4   suggesting methods of estimation.  Again, one of the
 5   targets of estimation was the frequency distribution
 6   of water quality parameters across streams in
 7   Maryland.
 8       Q.    Was it -- would it be fair to characterize
 9   this Maryland Synoptic Stream Survey as similar to
10   the National Surface Water Survey, but more limited
11   in scope --
12       A.    Right.
13       Q.    -- geographically?
14       A.    Right.
15       Q.    Were either the Maryland Synoptic Stream
16   Survey or National Surface Water Survey intended to
17   provide information to be utilized in setting water
18   quality standards or effluent limitations?
19       A.    No.
20       Q.    Do any of the professional activities
21   listed on page 2 of your vita involve regression
22   analyses?
23       A.    Yes.
24       Q.    Which ones?
25       A.    Toward the middle, consultant to Leighton
                                                 134

 1   and Regnery.
 2       Q.    Leighton and Regnery is a law firm?
 3       A.    Yes.
 4       Q.    Is that the proceeding that you previously
 5   discussed in which you were deposed?
 6       A.    Yes.
 7       Q.    What type of regression analyses did you
 8   perform in connection with that work?
 9       A.    This involved multiple regression comparing
10   the effects of two different nutritional diets.
11       Q.    How large were the data sets that you were
12   analyzing?
13       A.    Only in the hundreds of data points.
14       Q.    I may have asked you this question already,
15   but did this case actually proceed to trial?
16       A.    Yes.
17       Q.    It did?
18       A.    Yes.
19       Q.    Did you testify at that trial?
20       A.    Yes.
21       Q.    Were you qualified as an expert witness?
22       A.    Yes.
23       Q.    And in what field were you qualified?
24       A.    Statistics.
25       Q.    When was that trial?
                                                 135

 1       A.    1988.
 2       Q.    Have you ever -- other than this case have
 3   you ever testified as an expert witness?
 4       A.    In court?
 5       Q.    Yes.
 6       A.    No.
 7       Q.    In any administrative proceeding?
 8       A.    No oral testimony.
 9       Q.    Any written testimony?
10       A.    Yes.
11       Q.    And in what proceeding was that testimony
12   given?
13       A.    This goes back before the -- this period.
14       Q.    Okay.
15       A.    This concerned the impact of power plants
16   on striped bass in the Hudson River.
17       Q.    What impacts did your testimony concern?
18   Was it water quality?
19       A.    No.
20       Q.    What impacts was it?
21       A.    Concerned the analysis of historical data
22   on the aged composition of the striped bass stock in
23   the Hudson River.
24       Q.    Have you previously had any experience, in
25   your professional activities, any experience with
                                                 136

 1   dummy variables?
 2       A.    Oh, yes.
 3       Q.    Approximately how many projects have you
 4   worked on?
 5       A.    I couldn't count them.
 6       Q.    Have you ever had any experience in --
 7   involving the setting of water quality standards or
 8   effluent limits?
 9       A.    No.
10       Q.    No?
11       A.    No.
12       Q.    Any environmental standards of any kind?
13       A.    I mentioned the pesticide project.  When
14   EPA originally formulated guidelines for the field
15   experiments for chemical companies to perform in
16   preparation for applying for registration, I was on
17   the committee that critiqued these standards.
18       Q.    That's the scientific advisory panel
19   subpanel?
20       A.    Subpanel.
21       Q.    EPA Office of Pesticides and Toxic
22   Substances?
23       A.    Yeah.
24       Q.    Did you perform any statistical analyses in
25   connection with that project?
                                                 137

 1       A.    No.  This, again, was formulas.
 2       Q.    Okay.  What prior professional activities,
 3   in your mind, are most similar to your work in the
 4   instant proceeding?
 5       A.    I would have to say my consulting
 6   experience in general.
 7       Q.    How much of your time -- how long have you
 8   served as a private consultant?
 9       A.    As a private consultant only since 1987.
10       Q.    Is that full time --
11       A.    No.
12       Q.    -- for all the projects you've worked on
13   listed in the -- in a private capacity, listed on
14   page 2 under recent professional activities?
15       A.    No.
16       Q.    Do any other projects come to mind since
17   you entered the private sector?
18       A.    I had a contract with Norwich
19   Pharmaceutical to analyze collection of field trials
20   on a drug for osteoporosis.
21       Q.    Did you perform any statistical analyses in
22   connection with that project?
23       A.    Yes.
24       Q.    What type of analyses?
25       A.    These were non-parametric analyses.
                                                 138

 1       Q.    Any other private projects that are not
 2   listed on page 2 that we haven't previously
 3   discussed?
 4       A.    I don't recall any.
 5       Q.    What, if any, of your publications do you
 6   intend to rely upon in any part of your testimony in
 7   this proceeding?
 8       A.    I hadn't planned to rely on any particular
 9   one.
10       Q.    Do any of your publications involve
11   statistical analysis of environmental data?
12       A.    One might qualify in page 6, about fifth
13   from the bottom, Environmental Monitoring to Assess
14   Impact of Drilling Discharges.
15       Q.    What statistical analyses did you perform
16   in connection with this paper?
17       A.    Regression analyses.
18       Q.    Any other papers involve statistical
19   analysis of environmental data?
20       A.    I guess it depends on what you mean by
21   environmental data.
22       Q.    Let's limit it to water quality data.
23       A.    I think none of them.
24       Q.    Dr. Robson, what, if any, documents do you
25   intend to rely upon in support of your testimony in
                                                 139

 1   this proceeding?
 2       A.    Other than Appendix E I would rely upon the
 3   paper by Hirsh and somebody else I mentioned who had
 4   examined the effect of serial correlation on the
 5   Seasonal Kendall Test.
 6       Q.    Do you currently possess a copy of that
 7   paper?
 8       A.    Yes.  Somewhere.
 9       Q.    Is that something that you can supply to
10   us?
11       A.    Yes.
12       Q.    Any other documents that you might rely
13   upon in support of your testimony?
14             MR. MCGRATH:  Just object to the form to
15        the extent that that question might also
16        encroach on attorney/work product aspects of
17        strategy preparing for the final hearing and
18        things of that nature.
19             MR. PERKO:  In that regard, counsel, has a
20        privilege list been developed for Dr. Robson?
21             MR. MCGRATH:  No.  There is, I don't
22        believe, any documents that are being claimed to
23        be privileged.  I think everything has been
24        produced.  That is my understanding.
25   BY MR. PERKO:
                                                 140

 1       Q.    Dr. Robson, are there any other documents
 2   you can think of?
 3       A.    I can't think of any more.
 4       Q.    Have you ever performed analyses of time
 5   series data incorporating serial correlation?
 6       A.    No.
 7       Q.    Dr. Robson, we've previously discussed some
 8   reports by William Walker that you reviewed over your
 9   involvement in this project.  Is that something that
10   you can identify in your files?
11             MR. MCGRATH:  Are you asking him if he can
12        identify it or find it?
13             MR. PERKO:  If he can find it.
14             THE WITNESS:  If I can find it, yes.  I
15        know the document.  I'll recognize it when I see
16        it.
17             MR. PERKO:  We would like to get copies of
18        that if possible.
19             MR. MCGRATH:  Obviously, there would be no
20        objection if Dr. Robson can find it.
21             MR. PERKO:  Did you want to break for
22        today?  I'm about ready to get into a different
23        subject matter.
24             MR. MCGRATH:  Sure.
25             (Discussion held off the record.)
                                                 141

 1             (Whereupon, a recess was taken.)
 2 
 3 
 4 
 5 
 6 
 7 
 8 
 9 
10 
11 
12 
13 
14 
15 
16 
17 
18 
19 
20 
21 
22 
23 
24