1 1 STATE OF FLORIDA DIVISION OF ADMINISTRATIVE HEARINGS 2 Case Nos. 92-3038, 92-3039, 92-3040 3 SUGAR CANE GROWERS COOPERATIVE ) 4 OF FLORIDA, a Florida Agricultural ) Cooperative Marketing Association, ) 5 ROTH FARMS, INC., and ) WEDGWORTH FARMS, INC., ) 6 ) and ) 7 ) FLORIDA SUGAR CANE LEAGUE, INC. and) 8 UNITED STATES SUGAR CORPORATION; ) ) 9 and ) ) 10 FLORIDA FRUIT AND VEGETABLE ) ASSOCIATION, LEWIS POPE FARMS, ) 11 W.E. SCHLECHTER & SONS, INC., ) AND HUNDLEY FARMS, INC., ) 12 ) Petitioners, ) 13 ) vs. ) 14 ) SOUTH FLORIDA WATER MANAGEMENT ) 15 DISTRICT, an Agency of the State ) of Florida. ) 16 ) and ) 17 ) MICCOSUKEE TRIBE OF INDIANS OF ) 18 FLORIDA, the UNITED STATES OF ) AMERICA, and FLORIDA DEPARTMENT OF ) 19 ENVIRONMENTAL REGULATION, and ) the FLORIDA WILDLIFE FEDERATION, ) 20 ) Intervenors. ) 21 ) 22 99 N.E. 4th Street, 3rd Floor Miami, Florida 23 February 15, 1994, 9:00 a.m. 24 25 Deposition of Dr. Steven P. Millard JACK BESONER & ASSOCIATES, INC. 2 1 2 Taken before Lora Lee Knorr, Certified 3 Shorthand Reporter in and for the State of Florida 4 at Large, pursuant to Notice of Taking Deposition 5 filed in the above cause. 6 - - - - - - - 7 8 APPEARANCES: 9 10 ON BEHALF OF THE PETITIONERS: 11 PEEPLES, EARL & BLANK, P.A. 12 One Biscayne Tower, Suite 3636 Two South Biscayne Boulevard 13 Miami, Florida 33131 BY: Robert H. Blank, Esq. 14 ON BEHALF OF THE RESPONDENT/INTERVENORS: 15 UNITED STATES ATTORNEYS OFFICE 16 99 N.E. 4th Street, 3rd Floor Miami, Florida 33132 17 BY: Thomas A. Fitzgerald, Assistant United States Attorney 18 ON BEHALF OF SOUTH FLORIDA WATER MANAGEMENT 19 POPHAM HAIK 20 SCHNOBRICH & KAUFMAN, LTD. 100 S.E. Second Street 21 P.O. Box 019101 Miami, Florida 33131 22 BY: Jose A. Loredo, Esq. 23 24 25 - - - - - - - JACK BESONER & ASSOCIATES, INC. 3 1 I N D E X Witness Direct Cross Redirect Recross 2 Dr. Millard 4 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 JACK BESONER & ASSOCIATES, INC. 4 1 Thereupon: 2 DR. STEVEN P. MILLARD, 3 was called as a witness by the Intervenor, and 4 after being first duly sworn, was examined and 5 testified under oath as follows: 6 DIRECT EXAMINATION 7 BY MR. FITZGERALD: 8 Q. Good morning, Doctor. On the record, 9 Tom Fitzgerald on behalf of the 10 Respondent/Intervenors in this matter. You were 11 called for your deposition back in -- it was March 12 of 1993, that the fundamentals, as I recall, you 13 indicated you hadn't testified about. 14 If you want to take a break at any 15 time, just tell me. If you don't understand a 16 question, please advise me, and I'll try and put 17 it into some kind of format that makes sense for 18 you. And if, in the course of the questions, in 19 order to answer a question you need to assume 20 something or need additional parameters in order 21 for to you be able to formulate a reasonable 22 response, please just advise me, and we will try 23 and sort that out as we go along. 24 Doctor, in February of this year, 1994, 25 the firm with whom you maintain a consulting JACK BESONER & ASSOCIATES, INC. 5 1 relationship, Peeples, Earl and Blank, filed a 2 document in this matter captioned "Disclosure of 3 Expert Fact Witnesses" on behalf of Florida Sugar 4 Cane League, Incorporated and United States Sugar 5 Corporation. In paragraph 14 they designated you 6 as a potential witness in this matter and 7 identified a subject matter of expected testimony. 8 Since July of 1993, have you had 9 occasion to speak with anyone concerning what the 10 subject matter of your expected testimony will be 11 in this case? 12 A. Yes. 13 Q. And with whom did you speak? 14 A. Mr. Blank. 15 Q. Okay. And did you discuss the areas 16 that you would be designated an expert in or 17 simply look at the areas of work that you were 18 performing on behalf of the firm? 19 A. I talked about both. 20 Q. And what was your understanding as a 21 result of those conversations of the subject 22 matter of expected areas of testimony for you? 23 A. Expected areas of testimony would 24 include my opinion on the limits that are set in 25 Appendix "E", the validity of those limits, how JACK BESONER & ASSOCIATES, INC. 6 1 the entry and access data was gathered in the 2 period of '92 on to '93 validates or does not 3 validate or changes the model used in Appendix "E" 4 to set limits for the -- I'm talking about 5 Loxahatchee. 6 Also, with regard to Loxahatchee, I 7 expect I'll testify about whether I've been able 8 to find any sort of correlation between the 9 phosphorus load coming into Loxahatchee and the 10 ambient water phosphorus at the fourteen sampling 11 stations that are used to determine the limit. 12 Let's see, possibly I'll testify on 13 work that I'm expected to do that has to do with 14 looking at phosphorus loads coming out of the EAA 15 and phosphorus loads coming into the Everglades 16 National Park. 17 Q. Did I understand you just now said the 18 work you may do in those areas -- those last two 19 on phosphorus loads out of EAA and phosphorus 20 loads into Everglades National Park? 21 A. That's correct. 22 Q. So you have not commenced that work 23 yet? 24 A. No, I'm still waiting on some data for 25 that work. JACK BESONER & ASSOCIATES, INC. 7 1 Q. What data are you awaiting? 2 A. I'm waiting to receive phosphorus -- 3 let's see, I'm waiting to receive data from a 4 number of structures that border the EAA and also 5 data that flows into the Everglades, data from 6 structures where the water flows into Everglades 7 National Park. I'm waiting for data that will be 8 in the form of, I believe, monthly loads, because 9 my current understanding -- and I believe I 10 received data from some of those structures a 11 while back. 12 But as I recall, the form of the data 13 was phosphorus loads, say, every other week, 14 something like that, and the kind of sample was 15 not always the same. There were about ten or 16 eleven different samples codes, several of which I 17 think indicated a composite and several of which 18 indicated a grab. And, so, I'm waiting for data 19 from Environmental Services and Permitting where 20 they would be computing what the monthly load is. 21 Based on my understanding -- my 22 understanding is an algorithm -- that is what the 23 district uses. 24 Q. When have you been advised you will 25 receive that data? JACK BESONER & ASSOCIATES, INC. 8 1 A. I wasn't given a specific date. I 2 expect it will probably be within a couple of 3 weeks. 4 Q. How long will it take you or how long 5 do you anticipate it will take you to conduct your 6 analysis of phosphorus loads out of the EAA and 7 into ENP once you receive the data, assuming the 8 data is in usable format? 9 A. I would anticipate anywhere from two 10 weeks to a month. 11 Q. So you expect it within, say, two 12 weeks, and then two weeks to a month -- so we're 13 looking at about six weeks. And as soon as you 14 complete that work, are your final opinions 15 manifest then in that subject area upon completion 16 of the work or do you have to sit there and 17 eliminate the scatter plots and curves and the 18 graphs that evidently you experts in this case 19 seem to enjoy producing to the baffling of the 20 attorneys? 21 MR. BLANK: I feel compelled to object 22 to the form of the question. 23 MR. FITZGERALD: Noted, but I think the 24 hearing officer will see the inherent veracity of 25 that. JACK BESONER & ASSOCIATES, INC. 9 1 THE WITNESS: No, when I say my estimate 2 of two weeks to a month, that includes coming to 3 an opinion -- final opinion. 4 BY MR. FITZGERALD: 5 Q. So a minimum of six weeks from today -- 6 somewhere in the range of four to six weeks from 7 today you may have final opinions on that subject 8 area? 9 A. Uh-huh. 10 Q. What phosphorus trend models have you 11 examined since our March '93 initial deposition? 12 MR. BLANK: If you don't understand the 13 question, Doctor, ask him for clarification. 14 THE WITNESS: Okay. Yes, I guess when 15 you say "phosphorus trend", I guess I need you to 16 be more specific. 17 BY MR. FITZGERALD: 18 Q. Paragraph 14 of the document I've 19 previously described, subparagraph B, says your 20 subject of expected testimony -- the first three 21 words are "phosphorus trend models". So what does 22 that mean to you in the context of the work you've 23 been doing, your consultation with other attorneys 24 -- experts and the attorneys with regard to what 25 your testimony is going to be about in this case? JACK BESONER & ASSOCIATES, INC. 10 1 MR. BLANK: You're not compelled to 2 speculate, Doctor. If you don't understand, say 3 so. 4 THE WITNESS: Right. Okay. You could 5 classify the work that I'm doing for Loxahatchee, 6 analyzing the Loxahatchee model and the entry and 7 access data -- analyzing that. You could classify 8 that under phosphorus trend analysis because there 9 is a separation in time between the data that was 10 gathered between '78 and '83 and the data that's 11 -- also entry and access data of '92 to '93. 12 And there -- you could look at the 13 question of whether the phosphorus concentration 14 and the entry and access period is really 15 different from the phosphorus concentration in 16 either the base line period of '78 to '79 or the 17 post base line period of '79 to '83. Whether I'd 18 really call that looking for a trend or -- I 19 probably wouldn't call it looking for a trend. 20 I'd call it looking for a difference just between 21 those two time periods. 22 BY MR. FITZGERALD: 23 Q. Have you completed your work on that 24 particular aspect? 25 A. No, I'm still working on it. JACK BESONER & ASSOCIATES, INC. 11 1 Q. When do you on anticipate you'll have 2 final opinions on whether there are differences in 3 phosphorus concentrations from the base line 4 period -- the post base line period and the entry 5 and access period? 6 A. Within a couple of weeks. 7 Q. Well, you're going to testify about 8 statistics I see, and I'm going to skip over that 9 one just as I did in March. That's just a little 10 too generic. 11 You indicate, back in March, that you 12 were not engaged at that time in any STA, storm 13 water treatment area in modelling. Have you 14 engaged in any analysis or modelling the storm 15 water treatment areas since then? 16 A. Would the work that I did for the 17 mediation plan fall under that? I wouldn't think 18 so. 19 Q. No, I don't think so. 20 A. No. 21 Q. Okay. You bring up the mediation plan. 22 When you say that, do you mean the work that's 23 reflected in some of your documents provided by 24 counsel that occurred roughly in the July through 25 October, November time frame of 1993 relating to JACK BESONER & ASSOCIATES, INC. 12 1 settlement negotiations in the case? 2 A. Yes. 3 Q. Okay. Let me make one thing clear, 4 Doctor, based on a ruling of the hearing officer, 5 as we understand it, I will not be asking you 6 questions intended to elicit anything involving 7 your work related to the settlement negotiations. 8 However, if in order to answer another question 9 you have to address that, please advise me so we 10 can sort that out before you go into what you did 11 there. 12 The only, I guess, exception to that is 13 your ultimate factual opinions for the areas 14 you're not designated in are, in part, based on 15 some work did you during that period which is just 16 occidental, then you need to tell me that, too, so 17 we can sort that out. 18 I did recognize that in the documents 19 provided there was a fair amount of material 20 related to things like payment schedules and risk 21 assessments and some models which you did there, 22 and I don't intend to ask you about that. So when 23 I say STA models, I mean STA models in the sense 24 that it appears in the SWIM Plan adopted by the 25 board in 1992. JACK BESONER & ASSOCIATES, INC. 13 1 A. Okay. 2 Q. So I take it then your answer, 3 essentially, is no. 4 A. That's correct. 5 Q. All right. So when it says STA models, 6 we can sort of ignore that, except incidently the 7 phosphorus concentration limits, I think, you've 8 probably already covered to some degree. 9 Have you done any work related to 10 phosphorus concentration limits other than what 11 you've just described initially about the earlier 12 testimony, in other words, Loxahatchee work, entry 13 and access, that sort of thing? 14 A. I haven't done any work addressing 15 phosphorus concentration limits, no, as stated in 16 the SWIM Plan. 17 Q. Except as previously outlined in your 18 initial remarks; is that correct? 19 A. That's right. That's right, except as 20 I had previously explained. 21 Q. Have you done any analysis or are you 22 in the process of doing anything regarding water 23 quantity trends in the Everglades protection area 24 or the Everglades agricultural area? 25 A. No. JACK BESONER & ASSOCIATES, INC. 14 1 Q. Is it your understanding you would be 2 doing any water quality trend work? 3 A. No, I don't believe so. 4 Q. Okay. One area you've indicated of 5 subject matter is the QAQC data base work. What 6 QAQC data work have you done in connection with 7 this case? 8 A. With regard to the entry and access 9 data in Loxahatchee, there were field replicates 10 taken. So that gave me an opportunity to look at 11 how much variability is constructed by just small 12 scale field variability, how much of that -- how 13 much of the overall variability in the data is 14 contributed just by field replicate data. That's 15 just using the ES&P data and the Department of 16 Justice data. I understand field replicates were 17 taken. We were only given averages of those field 18 replicates. 19 Q. Did you compare the ESP data to the 20 federal average data for samples at the same 21 sites, on the same dates, etc., etc.? 22 A. Yes, I have -- 23 Q. And what opinions, if any, did you 24 derive from that comparison? 25 A. They don't agree very well. JACK BESONER & ASSOCIATES, INC. 15 1 Q. How much variance or what analysis did 2 you subject them to? 3 A. I've done some very simple plots. So 4 far I have looked at -- on a station by station 5 basis. So for each station I've looked at a plot 6 of the ES&P data for each date. 7 So the picture that I have in my head 8 is that I'm -- on the access you have the date 9 12/92, 1/93, 2/93, etc., etc. And you can plot 10 the three ES&P field replicates or make a box 11 based on those three, but it's a joint -- 12 basically it's just you look at the three 13 obvious. So you can look at those where on the 14 one access I'm thinking of the logarithm of the 15 total phosphorus. 16 So you can look at those observations, 17 and then you can also plot on the same plot, the 18 Department of Justice observation, which is 19 apparently the average of the three field 20 replicates. And, so, you can just look at that by 21 eye to see where the Department of Justice data 22 follows relative to the ES&P data. So you can 23 make a plot like that for each station. 24 And then another simple plot that I've 25 made is simply looking at the fourteen station JACK BESONER & ASSOCIATES, INC. 16 1 average for the ES&P data for each date and 2 comparing that to the fourteen station average 3 with the Department of Justice data. 4 Q. And what were the results of that 5 comparison? 6 A. Ideally, if there was fairly good 7 agreement between the two data sets, you would see 8 something that should tend to cluster along a line 9 that ideally has slope of one and an intercept of 10 zero. But if you do that -- if you make a plot, 11 it looks -- the data are scattered all over the 12 place. It doesn't look like very good fit. 13 There's not very good agreement. 14 Q. Back in March you talked about four 15 different factors that would affect your review of 16 any data, things like, you know, whether the 17 samples were taken in the same fashion, processed 18 the same way, etc. Are those the types of factors 19 that you would look at to account for the 20 variability or variance in the data set -- the two 21 data sets, the ESP and the DOJ data set? 22 A. In order to try and investigate why 23 you're not getting the results that you expect, 24 you need to go back and look at the whole sampling 25 process. JACK BESONER & ASSOCIATES, INC. 17 1 Q. Did you encounter any problems with the 2 data set produced by ES&P? 3 A. What do you mean by problems? 4 Q. Well, did you employ all the data in 5 your analysis provided by ESP for the three 6 replicates, on every occasion, from every site? 7 A. When I'm talking about the plot, the 8 plots that I was just describing, the first set of 9 plots where I looked at each station, I compared 10 the Department of Justice data with the three 11 field replicates, what I looked at -- and then 12 when I talked about the plot where I compared the 13 fourteen station average, initially I took the 14 average of the field replicates. This is all 15 after log transforming the data. 16 After taking the log of the data, I 17 took the average of the field replicates at each 18 station on each date, and then I took the fourteen 19 station average. And, so, I compared that against 20 the fourteen station average based on the 21 Department of Justice data. 22 Q. I'm not sure I heard the answer to the 23 question I was trying to ask. Maybe I didn't put 24 it too artfully. 25 Did you or anyone screen the data JACK BESONER & ASSOCIATES, INC. 18 1 before those two plot analyses to account for 2 outliers or data that was suspect to -- based on 3 the best professional judgment of anybody involved 4 in the process? 5 A. I have done some -- what's called 6 outlier analysis on the ES&P data. I also have 7 gotten information from John Davis, where he has 8 gone back and looked in log books, the sampling 9 log books, and he has identified approximately 10 five particular observations where it was noted in 11 the log book that it took several passes -- at 12 least two passes through the filter before they 13 felt that the -- whatever was suspended in the 14 water they basically had gotten rid of in order to 15 get a "clean sample". 16 Whenever you see -- so I take it from 17 John Davis, in his best professional judgment, he 18 said that it would possibly be legitimate to leave 19 those particular observations out of an analysis. 20 Q. Okay. On these five -- or whatever 21 number that Dr. Davis decided in his best 22 professional judgment or suspect or whatever 23 reasons, did you then exclude those from the two 24 plots you were discussing? 25 A. I haven't done that yet. I haven't JACK BESONER & ASSOCIATES, INC. 19 1 done that yet. 2 Q. Am I correct in my understanding that 3 if the samples were tainted with excess dissolve 4 particularly or they were clogged filters, they 5 had little chunks of fish or titan in that, that 6 would result after the acid digest process used 7 ESP and elevator process for phosphorus -- 8 A. That's my understanding. 9 Q. And from whom do you derive that 10 understanding? 11 A. From conversations with Dennis 12 Lettenmaier and John Davis. 13 Q. In your experience doing environmental 14 modelling -- your experience was detailed in the 15 earlier depos in March. Is it unusual to do this 16 type of outlier analysis and exclude some field 17 testing data because of -- it's called obvious 18 problems in the collection or processing process. 19 A. No, it's not unusual to have problems 20 with certain samples in environmental monitoring 21 data or industrial data in any kind of process. 22 Q. When that arises, is it your practice 23 then to use the best available data that the -- 24 after your outlier analysis and consultation with 25 people who would have an understanding of the JACK BESONER & ASSOCIATES, INC. 20 1 processes involving the parameters you're 2 examining? 3 A. My usual procedure is to use all the 4 data first and do both parametric and 5 nonparametric analysis, if possible. And then 6 based on the conversation with the people who are 7 experts in their field -- experts in the subject 8 matter, and their opinions on what can and cannot 9 be discarded, then I would go back and do the same 10 analysis with that revised data set as well. So I 11 would end up doing several analyses. 12 Q. Would it be possible to take two sets 13 of data that are in a related area, you know, like 14 phosphorus concentrations where the two sets were 15 gathered by different techniques and, therefore, 16 can't be treated as a single base and, through 17 some statistical method, pool that data and 18 examine the data through some pooling process that 19 would give you a feel for the variables that 20 you're interested in? 21 A. Your question is a little too vague for 22 me. 23 Q. All right. Is it possible to look at a 24 specific data set and pool it with a general data 25 set, not unique to the system you're examining, to JACK BESONER & ASSOCIATES, INC. 21 1 develop a trend model? 2 A. I don't understand what you mean by 3 specific and general -- a specific data set and 4 general data set. 5 Q. If you have a data base that covers 6 fifty marsh systems or wetlands systems, and 7 phosphorus data is available in that universal 8 data set -- constructive data set, ESP goes out to 9 Loxahatchee, for example, and develops the data 10 set that you've examined, is it possible to pool 11 that data and learn something or get a feel for 12 the functioning of the systems and their 13 phosphorus processes that are involved? 14 A. It's certainly possible to combine data 15 that have been gathered from several different 16 areas, using several different techniques, but you 17 have to be very careful about reaching any 18 conclusions. 19 Q. Why is that? 20 A. Well, because of all the factors that 21 are involved. If you have data gathered from 22 several different marsh systems, and you're trying 23 to pool them to try and figure out some general 24 law, it may be that "general law" operates 25 different in the different systems. And if you JACK BESONER & ASSOCIATES, INC. 22 1 pool all the data together and come up with some 2 supposedly general law based on that pool data, it 3 may be that the only reason you see that is 4 because the general law holds in maybe eight out 5 of ten cases, and there are the other two cases 6 where it maybe works the opposite way. Those are 7 overshadowed by the fact that you've got eight out 8 of ten which work one way so that the data masks 9 out the effects of the other two. 10 Q. If you wanted to data to see if -- say 11 the model you were developing for your specific 12 situation, in fact, bears a relationship with the 13 general data set, could you, as Dr. Davis did with 14 the Loxahatchee data, could you prescreen the 15 other data sets -- other eight or ten systems with 16 people knowledgeable in the appropriate 17 disciplines and the processing of the variables 18 you're trying to predict and maybe eliminate some 19 of those that would mask and give you an erroneous 20 result? 21 A. Certainly you can always try to do 22 that. 23 Q. In pooling data like that or in trying 24 to construct through a pooling mechanism, would a 25 bayesian analysis be a way to do that? JACK BESONER & ASSOCIATES, INC. 23 1 A. Certainly you could do it, use a 2 bayesian analysis, but I have to tell you that my 3 expertise on bayseian approach is very limited, if 4 nonexistent. 5 Q. Okay. All right. If you were going to 6 -- if you were tasked with that process, how would 7 you do it? 8 MR. BLANK: I'm going to enter a 9 standard objection as to the form and the line of 10 questioning. It seems we're getting into 11 hypothetical questions here, and I'm not certain 12 the witness is really in a position to really 13 discern what you're trying to accomplish here, but 14 go ahead, as best you can. 15 MR. FITZGERALD: Well, I'm not too 16 concerned with him trying to discern what I'm 17 trying to say, but he is designated as a 18 statistician, and I'm questioning him on the 19 statistical methods that he would employ or that 20 are available to do this sort of thing. 21 MR. BLANK: My only concern is I'm not 22 sure that he understands this sort of thing. 23 THE WITNESS: Yeah, I mean, I wasn't -- 24 I'm sort of getting lost now in the general line 25 of questions. Can we step back for a minute, and JACK BESONER & ASSOCIATES, INC. 24 1 you describe to me exactly what it is that you 2 want? 3 BY MR. FITZGERALD: 4 Q. If I want to take the ESP data and 5 compare it to a national data basis on wetlands 6 that has phosphorus data and pool that data, after 7 appropriate screening of the national data base, 8 how would you, as an expert statistician, have 9 experience in environmental modelling to pool that 10 type of data? 11 MR. BLANK: When you refer to ESP data 12 -- 13 MR. FITZGERALD: On Loxahatchee. 14 MR. BLANK: You're talking about the 15 entry and access data? 16 MR. FITZGERALD: Dr. Millard has been 17 using that term. 18 THE WITNESS: It would involve several 19 steps. As far as what I would do, first of all, I 20 will have to familiarize myself with the national 21 data base, understand where that data came from, 22 how it was collected, what laboratory procedures 23 were used to analyze it, etc., etc., and then I 24 would have to integrate that knowledge with the 25 knowledge I have about how the ES&P data were JACK BESONER & ASSOCIATES, INC. 25 1 collected and the laboratory protocol that was 2 used to analyze that data. 3 I would have to talk with people 4 knowledgeable in the field of wetlands ecology and 5 ask them about what, if any, particular 6 differences they might expect to see since you 7 have, in the national data base -- I assume that's 8 composed of data from several different areas, 9 several different wetlands. 10 MR. FITZGERALD: Let's just, for the 11 sake of argument, assume hundreds. 12 THE WITNESS: Okay. Hundreds of 13 wetlands. And then based on those conversations 14 with experts and integrating them, all that 15 knowledge, at that point I may decide that it may 16 or may not be feasible to pool the data for 17 whatever particular purpose it is that I've been 18 called to do. 19 BY MR. FITZGERALD: 20 Q. Have you been asked to conduct that 21 type of process in this case? 22 A. What I've been asked to do is take the 23 ES&P data -- entry and access data and compare 24 that data with the data that are given in Appendix 25 "E" that were used to set the limits for the JACK BESONER & ASSOCIATES, INC. 26 1 Loxahatchee National Wildlife Refuge. 2 Q. Back in March, when we were looking at 3 this, you stated that you had not examined the 4 district's phosphorus model for limit purposes in 5 Loxahatchee contained in Appendix "E", but rather 6 you had conducted your analysis tests based on Dr. 7 Walker's model which was different. Have you 8 since conducted an analysis of the Appendix "E" 9 district model? 10 A. Yes, I've done some analyses based on 11 that data. 12 Q. Okay. Now, using the ESP data you've 13 just described, have you conducted a comparison of 14 that against Dr. Walker's model contained in his 15 original model that you looked at in his '91 16 paper? 17 A. No, I've only been looking at comparing 18 the ES&P entry and access data with the modelling 19 data that's contained in Appendix "E". 20 Q. Okay. In analyzing Dr. Walker's '91 21 model, you produced a draft report that reflected 22 a bootstrapping analysis and some other 23 consideration and listed a number of areas of 24 possible further work and inquiry. Did you ever 25 finalize that report? JACK BESONER & ASSOCIATES, INC. 27 1 A. No. 2 Q. Why not? 3 A. I wasn't asked to. 4 Q. Did you ever conduct a bootstrap 5 analysis with Appendix "E" and phosphorus trend 6 for Loxahatchee? 7 A. No. 8 Q. Why not? 9 A. I wasn't asked to. 10 Q. You testified in March the 11 bootstrapping method was a valid statistical tool 12 for getting a feel for the marsh and variance of 13 -- and just sort of a subject of the validity of 14 the modelling effort. Why would that not be 15 equally valid with respect to the district 16 Appendix "E" model? 17 A. Well, the main point of the draft or 18 report that I produced that commented on Walker's 19 report was to get a feel for the amount of wiggle, 20 the amount of variability in the limits, because 21 the limits are -- the way the limits were set, 22 they were really a realization of a random 23 variable. And the point is that had I gone back 24 and done the same, gathered the same sort of data 25 that I gathered in the base line period '78 to JACK BESONER & ASSOCIATES, INC. 28 1 '83, had I gone back and possibly gathered on 2 different days, in slightly different locations, 3 those are two possible sources of variability. 4 Had I done that, what kind of limits would I come 5 up with then? Would they be comparable? How much 6 different would they be? 7 And, so, the point of that, of my 8 report on the bootstrap analysis -- the main point 9 was to try and get a feel for the record for that. 10 Q. Okay. In doing that, a bootstrap as I 11 understood, it, with the substation program, the 12 way you wrote it, it was limited observation 13 obviously to the existing data set. You were not 14 able to substitute data from other locations, on 15 other dates, per se, but were merely able to 16 provide an analysis on any given day by substation 17 of the fourteen by fourteen data you had. 18 A. That's correct. 19 Q. Okay. So when you say it gives you a 20 feel for what might have happened, if you sampled 21 at different places, different days, then those 22 apply to develop the model, and it's that's sort 23 of best professional judgment kind of wiggle 24 analysis, isn't it? I mean, you don't have the 25 data -- JACK BESONER & ASSOCIATES, INC. 29 1 A. You can't construct the data, right, 2 but that's the point of the bootstrap. 3 Q. Okay. But inherent in that, isn't the 4 weakness -- isn't the weakness of bootstrap why it 5 only gives you a feel or an approximation? 6 A. It's a limits mechanism. It kind of 7 puts a bound around the data set you do have and 8 gives you a feel for what the variance might be. 9 Yes, it probably gives you a low data on the 10 variance. 11 Q. Why does it give you a lower bound? 12 A. Because of what you just said. 13 Q. So why didn't you do that for the 14 district's model? It sounds like a pretty good 15 idea. 16 A. Well, since the ES&P data were 17 gathered, we've been using that to compare to the 18 district's model. Rather than doing bootstrap 19 analysis, we've been using the ES&P data to 20 compare the data with the district's data -- 21 compare as far as whether the relationship between 22 the fourteen station average of the log TP, the 23 relationship of that variable with the average 24 station still loads, whether the slope and the 25 intercept of the line are about the same, and JACK BESONER & ASSOCIATES, INC. 30 1 also, whether there's real any difference -- any 2 difference in the intercept of that line, in other 3 words, in layperson's term, whether the phosphorus 4 concentration, the entry and access period is 5 really any different than the phosphorus 6 concentration in the base line period. 7 Q. You indicated that in conducting your 8 analysis you log transformed the data first. Why 9 did you do that? 10 A. Well, the district model is based on 11 doing that. 12 Q. Okay. But you just told me you weren't 13 testing the district's model in the way you did 14 Dr. Walker's model, and you essentially had a 15 clean slate. You've got distribution period of 16 record data base. You've got the ESP data from 17 the entry and access Loxahatchee. 18 So it seems to me you probably have the 19 freedom to handle the data however you choose to 20 determine if, in fact, there is a quantifiable 21 correlation between the phosphorus concentration 22 and the station in Loxahatchee, yet you choose to 23 do a log transformation. Was there something that 24 had led you to do that independent -- 25 A. No, I think you're misrepresenting what JACK BESONER & ASSOCIATES, INC. 31 1 my directed task was. My directed task was to 2 take the district's model, and assume that that's 3 the model that you have to use, and then look at 4 the entry and access period data to see how that 5 compares with the district's model. 6 Q. So you had a preconceived goal when you 7 went into the analysis? 8 MR. BLANK: Object to the form of the 9 question. 10 BY MR. FITZGERALD: 11 Q. Did you have a preconceived goal when 12 you went into the analysis? 13 A. My task, as I understood it, was to 14 look at the ES&P data and entry and access data 15 and compare it with the district data in Appendix 16 "E". If you're talking about a goal, my goal was 17 to compare -- look at the entry and access data 18 and see how closely it matched the district's 19 model. 20 MR. FITZGERALD: I'd like to mark this 21 document as deposition exhibit 13. It is a four 22 page document, bate numbered 1210001 through 23 1210004. 24 (The document referred to 25 was thereupon marked as JACK BESONER & ASSOCIATES, INC. 32 1 Respondent's Exhibit Number 2 13 for Identification, 3 a copy of which is attached 4 hereto.) 5 BY MR. FITZGERALD: 6 Q. Is that your handwriting? 7 A. Yes, it is. 8 Q. Okay. So these are your notes from 9 some point in time. In fact, the numbering may 10 not match, and that's one thing I wanted to ask 11 you, whether the four pages are from the same 12 Orlando meeting on Thursday, 12/9/93, which is 13 what's at the top of the second page, and down at 14 the bottom it says 1210002. 15 A. Uh-huh. 16 Q. And then there's a page number 2 17 handwritten behind it, which is 0003, page number 18 3, which is 0004 at the top page, which is 0001. 19 Is that part of the same notes taken or is that -- 20 they came stapled together. 21 MR. BLANK: What's the question? 22 MR. FITZGERALD: I'm trying to find out 23 if the top page of notes belong as part of the set 24 of three pages thereafter, which are sequentially 25 numbered in the doctor's handwriting. JACK BESONER & ASSOCIATES, INC. 33 1 THE WITNESS: They were probably written 2 on the same pad of paper. Whether they were 3 separated by separate pages or not, I can't tell 4 you. 5 BY MR. FITZGERALD: 6 Q. All right. So you can't tell me if it 7 was part of the same meeting, the first page 8 0001. 9 A. These notes were written during this 10 meeting in Orlando. 11 Q. Okay. So you were sort of going side 12 by side with a list of things that occurred to you 13 to do as a result the discussion in Orlando. For 14 example, number 5 on that first page says: "Redo 15 the percent cattail analysis use DOJ data." 16 A. Uh-huh. 17 Q. Had you done a cattail analysis before 18 that? 19 A. Yes. Yes, I was given data from 20 Breedlove, Dennis and Associates on percent 21 cattail that was taken along approximately -- 22 let's see, A, B, C, D, E, F -- about six transects 23 in the Loxahatchee, and along the transect they 24 took measurements of percent cattail at a 25 particular place, what the ambient water total JACK BESONER & ASSOCIATES, INC. 34 1 phosphorus concentration was, also what the soil 2 phosphorus concentration was and what the water 3 depth was. 4 Q. Did you attempt to correlate the data 5 to determine if there was any predictor variable? 6 A. I looked at several plots of percent 7 cattail or some transformation of percent cattail 8 versus the other variables, water TP, soil TP, 9 water depth. 10 Q. And what opinion, if any, did you form 11 based on your analysis of those plots and 12 comparison of the cattail percentage against other 13 variables? 14 A. It turned out that the percent cattail 15 appeared to be correlated with water depth in -- 16 for water depth less than four. It was a positive 17 correlation. There were about four observations 18 where water depth was six feet or higher, and in 19 those instances there was no cattail. 20 So it could very well have been some 21 sort of parabolic relationship, but the data only 22 went up to a certain water depth. And then there 23 weren't -- I don't think -- I don't recall there 24 were any observations, say, between four feet and 25 six feet. JACK BESONER & ASSOCIATES, INC. 35 1 So, except for the observations that 2 were six feet or over, there were about four 3 observations where the water depth was six feet or 4 over. So there was correlation between percent 5 cattail and water depth. There was no correlation 6 between percent cattail and ambient water TP or 7 soil TP. 8 Q. You say that the correlation is 9 positive. How positive? By what statistical 10 measure was it positive? 11 A. If you're asking me to recall, say, 12 something like R-squared value, I can't -- I don't 13 recall that, but the picture clearly showed a 14 positive correlation. 15 MR. BLANK: I think you may be 16 misconstruing the way he's using the term 17 positive. 18 MR. FITZGERALD: Well, that's what I'm 19 trying to explore. If he can give us an R-2 value 20 or if he could give the "P" value, I would know 21 exactly what he meant, and that's how we went with 22 the -- because the analysis of Dr. Walker's work 23 was discussed at great length back in March. 24 BY MR. FITZGERALD: 25 Q. So when you say positive, what do you JACK BESONER & ASSOCIATES, INC. 36 1 mean? 2 A. I mean, if you look at the scatter 3 plot, the plot shows that as water depth 4 increased, percent cattail cover increases. 5 There's a general relationship there. 6 Q. Did you seek to model that 7 relationship? 8 A. Yes, I did. 9 Q. And what -- 10 A. It was statistically significant. I 11 don't remember the "P" value. I think, in fact, I 12 probably modelled it initially as a quadratic 13 effect so that in increase with water depth and 14 then at a certain point it started to decrease 15 with water depth. 16 Q. So you modelled it with a parabolic 17 function then. 18 A. That's correct. 19 Q. Did you make any effort to determine, 20 by discussions with field specialists, whether, in 21 fact, it was in parabolic but acidotic at the end? 22 A. I talked with some people at Breedlove, 23 Dennis and Associates, George Carlson, for one, 24 about the behavior of the cattail species and 25 whether it made sense that the cattail would JACK BESONER & ASSOCIATES, INC. 37 1 behave that way according to water depth. 2 Q. And what did he tell you? 3 A. He told me that, based on his 4 knowledge, for the species of cattail, that was 5 the predominant cattail in that area. If I recall 6 correctly, it was his understanding that its 7 favored water depth was somewhere between two and 8 four feet, something like that. 9 Q. Is the plot you're referring to 10 included in the materials that you provided to us? 11 A. I don't believe so. 12 Q. Why not? 13 A. It's part of a draft report that's 14 preliminary. 15 Q. All right. In March you provided us 16 your draft report on the bootstrapping effort; did 17 you not? 18 A. Yes. 19 MR. FITZGERALD: Counsel, I have a 20 little problem with that not being provided. The 21 witness is supposed to have his final opinions now 22 anyway. The hearing officer made that very 23 clear. And he's telling me that his final opinion 24 as to one of the most significant areas he's going 25 to testify in won't be provided for six weeks at JACK BESONER & ASSOCIATES, INC. 38 1 the soonest which, of course, is beyond the end of 2 discovery in this case. He's not providing a 3 draft report on which he has stated here he is 4 going to be relying. 5 BY MR. FITZGERALD: 6 Q. What remains to be done on the draft 7 report? 8 A. As far as what remains on the report, 9 one thing is this item number 5 because, as I 10 said, in the report I initially did the analysis 11 with a parabolic fit. Another way to do the 12 analysis would be to just use only water depths 13 that were four feet or less and then model that. 14 Q. Assuming you did that, it seems 15 intuitive to me that it gives you a higher 16 positive correlation; is that correct? 17 A. No, not necessarily. In fact, it might 18 make it a little bit lower because of the way the 19 data is spread out. As I said, we had water 20 depths between about zero and four feet, and then 21 it might have been a little higher, and then not 22 really anything between four and six, and then at 23 six feet or more there were about four 24 observations. 25 The problem with fitting a parabolic JACK BESONER & ASSOCIATES, INC. 39 1 curve like that is that these observations over -- 2 at six feet are more -- are basically the ones are 3 determining a lot about the parabolic curve. So 4 if you cut out those observations and just use the 5 depths at four feet or less and just stay with a 6 straight line, then you may not necessarily get as 7 high an R-squared. It's not necessarily going to 8 be thorough. You'll get a higher R-squared if 9 that's what you're saying. 10 Q. This effort or possible effort to 11 finalize your draft report sounds like, to me -- 12 and correct me if I'm wrong -- a matter of a 13 couple of minutes work with your software 14 package. It does not seem to be a particularly 15 involved operation; is that correct? 16 A. Well, it would take not too much time 17 probably to refit the model, but it would take a 18 while to then compare that model to the original 19 fit and then draw conclusions from that. 20 Q. Did you calculate R-square for the 21 original fit? 22 A. For the quadratic fit? 23 Q. Yes. 24 A. Yes. I don't recall what it is. 25 Q. And you don't have that report with JACK BESONER & ASSOCIATES, INC. 40 1 you? 2 A. No. 3 Q. And it's not in any of the materials 4 you provided to us? 5 A. No. 6 Q. Okay. Now, in addition -- 7 MR. BLANK: I'm not all together certain 8 of that. You might want to have him look at the 9 materials you have. 10 MR. FITZGERALD: We can certainly do 11 that, but I have gone through it a number of 12 times, and I see no draft or final reports that I 13 know of that discuss that, but I could be 14 mistaken. And it will certainly be gone through 15 in terms of giving Dr. Millard an opportunity to 16 find it if it's there, and we will reserve to move 17 our right to strike the witness if it's not 18 there. 19 BY MR. FITZGERALD: 20 Q. Number 7 on the first page of exhibit 21 13 says: "Redo percent cattail with DOJ data." 22 Did you do that as part of this draft report? 23 A. At this point I'm not sure what I'm 24 referring to in point number 7. 25 Q. How about point number 8? Did you look JACK BESONER & ASSOCIATES, INC. 41 1 at soil phosphorus, water depth, percent cattail 2 in WCA-1 and WCA-2 using ready soil phosphorus, 3 water phosphorus interpolated -- 4 A. I have started to analyze some data, 5 soil phosphorus, water depth and percent cattail 6 using ready soil for WCA-1 and WCA-2 and also 7 using percent cattail that was derived by taking a 8 map that was produced by Breedlove, Dennis and 9 Associates for WCA-2, a map of percent cattail 10 cover in that area. Someone at Environmental 11 Services got the ready station coordinates from 12 the Breedlove, Dennis and Associates cattail map 13 and attempted to determine what the percent 14 cattail cover was at that station. 15 Now, in some instances the station 16 appeared to be on the border line of two different 17 levels of percent cattail cover. So I've just 18 started to do some analyses looking at percent 19 cattail cover and the soil phosphorus at the ready 20 stations and then some data on a monthly average 21 water depth that I also obtained from 22 Environmental Services & Permitting. I haven't 23 done very much with that right now. 24 Q. When do you anticipate that will be 25 complete? JACK BESONER & ASSOCIATES, INC. 42 1 A. Well, actually what we're going to do 2 is, as far as using the data that I just described 3 to you, where I'm only looking at values, either 4 actual values of the station or values 5 interpolated the percent cattail and the water at 6 the station, that analysis should be done within a 7 couple of weeks. 8 We're also going to do an analysis 9 where we take soil phosphorus and WCA-2 based on 10 creeding procedure and percent cattail based on 11 some sort of grid interpolation procedure and 12 water depth also based on a grid interpolation 13 procedure and do an analysis using all that data. 14 Q. Who is doing that? 15 A. Environmental Services & Permitting. 16 Q. Okay. That's a company -- in my 17 experience, companies rarely do anything people 18 tend to do. Who are those people you keep 19 referring to at Breedlove and ESP with whom you're 20 dealing with? 21 A. Well, Breedlove, Dennis and Associates 22 are the people that produced the percent cattail 23 map, but I'm not -- for this analysis in WCA-2, 24 I'm not talking with them directly. I'm getting 25 all my data from Environmental Services. JACK BESONER & ASSOCIATES, INC. 43 1 Q. Do you have that percent cattail map? 2 A. No. 3 Q. Then how are you going to employ the 4 data depicted on it or the data used to develop 5 it? 6 A. Environmental Services & Permitting is 7 doing that for me. They will give to me the data 8 in the form of what they have given -- what they 9 have given me so far is they've given me estimated 10 percent cattail as to each of the ready stations 11 and also average monthly water depth for certain a 12 time period -- I don't recall exactly what it is 13 right now -- at each of the ready stations. 14 Q. Okay. And what's the purpose of your 15 analysis utilizing that data? 16 A. The purpose was to try and determine if 17 there's a relationship between the percent cattail 18 cover and the level of the soil phosphorus and/or 19 the average monthly water depth. 20 Q. So you're not examining water 21 phosphorus or relationship between water to the 22 phosphorus as a variable. 23 A. No. 24 Q. You're aware from your review of the 25 SWIM Plan and Appendix "E" that the district JACK BESONER & ASSOCIATES, INC. 44 1 believes there's correlation between water 2 phosphorus and cattail; are you not? 3 A. Yes, I was aware of that. 4 Q. Why then are you choosing to not 5 consider water phosphorus or total phosphorus 6 concentration as a potential predictor variable 7 when that is exactly what you did with regard to 8 the percent cattail analysis you did for 9 Loxahatchee? 10 A. Well, the data were available for the 11 Loxahatchee because the measurements were taken on 12 those transects. 13 Q. Dr. Reddy also produced soil or -- I'm 14 sorry -- water phosphorus data, did he not, for 15 his tests results? 16 A. I don't know for his analysis. 17 Q. What did you mean in Exhibit 13 when 18 you said that the goal of the program is to show 19 that historically the center of Loxahatchee is not 20 impacted? 21 MR. BLANK: What page are you looking 22 at? 23 MR. FITZGERALD: The top of the page, 24 number 2, which is 1210003, the second entry 25 down. JACK BESONER & ASSOCIATES, INC. 45 1 THE WITNESS: Oh, goal of -- okay. 2 Goal of program show that historically the center 3 of Loxahatchee is not impacted. I would have to 4 speculate what that refers to at this point. 5 BY MR. FITZGERALD: 6 Q. Well, from whom did you derive the 7 notion that that was a preset goal of your 8 analysis program? 9 A. It's not -- I don't believe it's 10 referring to any analysis that I'm doing. I 11 believe that's referring to -- it's my guess that 12 I'm referring to -- 13 MR. BLANK: No, Doctor, don't guess. 14 THE WITNESS: Okay. I don't know. 15 BY MR. FITZGERALD: 16 Q. Is it customary to, in conducting an 17 analysis or doing the type of work you have in 18 this matter, to start out with a preconceived goal 19 as opposed to testing a hypothesis? 20 A. In my line of work as a consultant, I'm 21 obviously aware that my clients will have goals, 22 but as a statistician, it's my duty, if you want 23 to call it, or it's my purpose to analyze the data 24 and find what I find. 25 Q. The draft report that you've been JACK BESONER & ASSOCIATES, INC. 46 1 referring to on Loxahatchee and the analysis 2 you've conducted, to whom have you shown that? 3 A. Which draft report are you referring 4 to? 5 Q. Do you have more than one that we 6 haven't seen? 7 A. What draft report are you talking 8 about? 9 Q. Your correlation with the positive 10 value regarding the percent cattail. With whom 11 have you shared that, the draft report? 12 A. Let's see, I sent a copy to Mr. Blank. 13 I sent a copy to John Davis. I sent a copy to 14 Mike Dennis at Breedlove, Dennis and Associates. 15 I don't recall whether I gave a copy of that to 16 Dennis Lettenmaier or not. 17 Q. When did you produce that draft? 18 A. Approximately August or September of 19 1993. 20 Q. Have you received comments back from 21 Dr. Davis or Dr. Dennis on your draft report? 22 A. We discussed it a little bit at the 23 meeting in Orlando, but these -- at which time 24 these notes were taken. 25 Q. Did they give you any written comments JACK BESONER & ASSOCIATES, INC. 47 1 as to your draft report? 2 A. No. 3 Q. Back in June of '93, you sent a draft 4 report to Dennis Lettenmaier before sending 5 anything to Bob, and in that context you were 6 speaking about Mr. Blank. Was that the draft 7 report that you were referring to? Do you recall 8 that? 9 A. Yes. I sent a draft of that report to 10 Dennis, and then I was waiting for his comments on 11 the report before I sent it to Bob, and then I 12 went ahead and sent it to Bob and John and Mike 13 before I got any comments back from Dr. 14 Lettenmaier. In fact, I never got any comments 15 back. 16 Q. But it's your recollection now that, in 17 fact, you sent the original draft report to Dr. 18 Lettenmaier? 19 A. Yes, I did send a draft to him. 20 Q. Okay. And if your computer mail was 21 dated sometime in June of '93, then the report 22 must have existed at least since then. 23 A. That's correct, some form of it. 24 Q. Well, that -- and you anticipated the 25 next question. Dr. Lettenmaier gave you no JACK BESONER & ASSOCIATES, INC. 48 1 comments on the report? 2 A. That's correct. 3 Q. Dr. Davis and Dr. Dennis discussed it 4 with you in Orlando in December of 1993? 5 A. That's correct. 6 Q. As a result of your conversations with 7 them, did you alter the report? 8 A. No. 9 Q. Have you altered the report since 17 10 June, 1993? 11 A. I think I may have made some changes, a 12 few changes. 13 Q. What caused you to make changes in the 14 report? 15 A. I think I just made some minor changes. 16 I think there's probably minor differences between 17 whatever draft I gave Dennis Lettenmaier and the 18 draft that I sent to Mr. Blank, Dr. Davis and Dr. 19 Dennis. 20 Q. Have you sent that out for a peer 21 review of any sort? 22 A. No. 23 Q. Do you know Jim Loftis? 24 A. I know of him. 25 Q. Have you ever had a conversation with JACK BESONER & ASSOCIATES, INC. 49 1 him regarding this case? 2 A. No. 3 Q. Have you ever met him? 4 A. No. 5 Q. In the materials you provided there are 6 notes that purport to reflect statements by Mr. 7 Loftis regarding Dr. Walker's work. Do you recall 8 where you came you up with that information? 9 A. It would probably -- it would have to 10 have been either in a conversation with Mr. Blank 11 or Dr. Lettenmaier. 12 Q. Has Dr. Lettenmaier had conversation 13 Mr. Loftis to your knowledge? 14 MR. BLANK: Dr. Loftis you mean. 15 MR. FITZGERALD: Dr. Loftis, correct. 16 THE WITNESS: I think they've talked, 17 yes. 18 BY MR. FITZGERALD: 19 Q. Do you know for certain? 20 A. No, I can't tell you for certain. It's 21 my belief that they have. 22 Q. When do you anticipate receiving from 23 ESP the data derived from the Breedlove, Dennis 24 Associates cattail map 2-A? 25 A. Either this week or next week. JACK BESONER & ASSOCIATES, INC. 50 1 Q. How long will that analysis effort 2 take? 3 A. Once I have the data it should probably 4 take a good -- about -- oh, anywhere between three 5 and four days. 6 Q. So that will not affect or be impacted 7 by the work on "P" loads out of EAA and into 8 Everglades National Park that you're going to do 9 during that same period. 10 A. No. 11 Q. Okay. And I guess I'm still a little 12 unclear, and maybe you can put it in a nutshell 13 for me. What remains to be done on your draft 14 report to finalize it? 15 A. Go back and redo the analysis by 16 looking only at water depths between zero and four 17 feet, fit the model and incorporate any comments 18 that I get from Dr. Lettenmaier. 19 Q. Okay. You had said you sent that to 20 him in June, June 17th, and that you've gotten no 21 comments back. Do you anticipate comments now? 22 A. Well, if you recall, in June -- I'm not 23 sure exactly when the mediation plans started -- 24 when the mediation process started, but I know it 25 was around that time. And, so, once that started, JACK BESONER & ASSOCIATES, INC. 51 1 everything that we had been working on up to that 2 point basically got dropped, and we concentrated 3 on tasks that were asked of us during the 4 mediation at one point. 5 Q. What makes you believe that Dr. 6 Lettenmaier has comments? Has he told you he has 7 comments on the draft report? 8 A. He had requested me to send him a draft 9 of the report so that he could give me his 10 comments. 11 Q. There is no additional data for 12 Loxahatchee that you're awaiting, is there? 13 A. No. 14 Q. So all that data was from the initial 15 transecting work and vegetative work done in 16 Loxahatchee by ESP? 17 MR. BLANK: Counsel, you're referring to 18 the data related to this draft report? 19 MR. FITZGERALD: Yes. 20 THE WITNESS: I think the data -- 21 MR. BLANK: Can we go off the record for 22 just a minute? 23 MR. FITZGERALD: Sure. 24 (Discussion off the record.) 25 THE WITNESS: Could you repeat the JACK BESONER & ASSOCIATES, INC. 52 1 question? 2 (Thereupon the referred to question was 3 read back by the reporter as above 4 recorded.) 5 THE WITNESS: Okay. And I was about to 6 respond it was Breedlove, Dennis and Associates 7 who took that data. 8 BY MR. FITZGERALD: 9 Q. Okay. The vegetative analysis was done 10 by Breedlove, Dennis and Associates -- the 11 transecting in Loxahatchee? That's BDA, for the 12 record. 13 A. Yes. 14 Q. Yes? 15 A. Yes. 16 Q. And ESP did the analysis and sampling 17 and -- 18 A. Actually, I'm not sure whether BDA did 19 that or ES&P did that, the transecting. 20 Q. Didn't you tell me that to do this kind 21 of analysis you needed to talk to the people who 22 know about the sampling mechanisms and lab 23 techniques and everything else to assess the data 24 set? 25 MR. BLANK: I object to the form of the JACK BESONER & ASSOCIATES, INC. 53 1 question. I think you were referring to, counsel, 2 at that time with your concept with regard to 3 pooling the data base. 4 BY MR. FITZGERALD: 5 Q. Okay. Well, back in March we talked 6 about what a statistician needs to know about this 7 data set in order to evaluate, assist and draw 8 conclusions. Is that true with respect to the 9 data from WCA-1 that you employed in your draft 10 report? 11 A. The general things as far as talking 12 with people, how they collect it in the field and 13 how they did the analysis in the lab, yes. 14 Q. Okay. Did you do that with respect to 15 the vegetative assessments conducted by Breedlove 16 -- did you say it was BDA? 17 A. Breedlove, Dennis and Associates. 18 Q. Did you talk to the people who did 19 that? 20 A. Yes, I talked with Mike Dennis and 21 George Carlson who was the project manager. 22 Q. Okay. Did he go out and do the work 23 himself or did others do it? 24 A. I believe a number of -- I'm not sure 25 how many people were involved in that. JACK BESONER & ASSOCIATES, INC. 54 1 Q. Did you ever speak to Dr. Joe Birch? 2 A. Yes, I've had one telephone 3 conversation with him. 4 Q. Was it regarding the vegetative 5 community data from Loxahatchee? 6 A. No, it was regarding the percent 7 cattail map and WCA-2. 8 Q. Have you ever spoken with Courtney 9 Hackney? 10 A. No. 11 Q. Have you ever spoken with Miles M. Bud 12 Smart? 13 A. No. 14 Q. How about W. Michael Dennis? 15 A. Yes. 16 Q. Have you spoken with anyone regarding 17 the vegetative community data in Loxahatchee 18 collected February 16th of '93? 19 A. Yes. 20 Q. Okay. When did that occur? 21 A. Sometime between when the data were 22 collected and when I sent my draft report to Dr. 23 Lettenmaier in June. 24 Q. Okay. On how many occasions did you 25 speak with Dr. Dennis regarding that subject? JACK BESONER & ASSOCIATES, INC. 55 1 A. I probably spoke with Dr. Dennis about 2 twice and with George Carlson another two or three 3 times at least. 4 Q. Did Mr. Carlson ever tell you that he 5 went out and did any of this work himself? 6 A. No, as far as I know he was the project 7 manager. 8 Q. Since the March deposition did you 9 supply your bootstrap data to anyone other than 10 Dr. Lettenmaier, the bootstrap that was done, Dr. 11 Walker's work? 12 A. You mean did I supply it in the form of 13 a report? 14 Q. Uh-huh. 15 A. Besides Mr. Blank and Dr. Lettenmaier, 16 as far as I know, no, I haven't sent it to anybody 17 else as far as I can recall. 18 Q. Have you reviewed any subsequent work 19 by Dr. Walker since the May '91 report that was 20 the subject of your bootstrap analysis? 21 MR. BLANK: With the exclusion of the 22 mediation materials? 23 MR. FITZGERALD: Yes. 24 THE WITNESS: Yes, I've read a couple of 25 different reports by him. I've read the report JACK BESONER & ASSOCIATES, INC. 56 1 that had to do with trends in Everglades National 2 Park. I don't remember the exact title. I've 3 also reviewed his report on the model that he used 4 to size the STA's. 5 I have several of his reports in my 6 possession, but I believe those are the only two 7 that I really looked at. 8 BY MR. FITZGERALD: 9 Q. Okay. With regard to the trend in 10 Everglades National Park, the first report you 11 see, have you formed any opinions based on your 12 review of that report? 13 A. I don't recall. I recall having a 14 problem with one of the things he did, as far as 15 the variables that he looked at. As I recall, his 16 response variable and his predictor variable were 17 inherently correlated because the response 18 variable were, in fact, function of the predictor 19 variable, which produces some that you call 20 spurious correlation. 21 That was -- I believe I started to 22 review that report, and then I don't think I 23 actually finished reviewing the whole report, and 24 then the mediation process came up, and so I never 25 went back to that report. JACK BESONER & ASSOCIATES, INC. 57 1 Q. Do you intend to go back and reanalyze 2 that report? 3 A. No, I don't believe so. 4 Q. Do you expect to offer any expert 5 testimony or opinions at hearing in this matter 6 regarding Dr. Walker's analysis of trends at 7 Everglades National Park? 8 A. The only way that it might possibly 9 come up is, as I stated earlier, I have been asked 10 to look at loads coming out of the EAA and loads 11 coming into the Everglades National Park and 12 trying to see if there's a relationship between 13 those two sets of data. Whether that would touch 14 on Dr. Walker's analysis or not -- I don't think 15 it would, but that's the only thing. 16 Q. With regard to the model developed by 17 Dr. Walker, besides STA's, what have you done with 18 respect to that report? 19 A. I reviewed it in preparation for a 20 meeting that I had with Dennis Lettenmaier and 21 Carlos Marin. 22 Q. What was the purpose of that meeting? 23 A. The purpose of that meeting was to 24 discuss how to evaluate Walker's model and how to 25 perform sensitivity analyses on the model. JACK BESONER & ASSOCIATES, INC. 58 1 Q. Did you conduct a sensitivity analysis? 2 A. No, I believe Dr. Marin is performing 3 those analyses. I was sitting in on the meeting 4 mainly just for the possibility of giving my input 5 to whatever would be done. 6 Q. I had understood from our earlier 7 discussion of your designation of testimony that 8 you don't intent to offer any expert opinions or 9 testimony at the hearing in this matter regarding 10 STA models. Is that still true or did I 11 understand your earlier testimony correctly? 12 A. Yes, I don't anticipate discussing that 13 topic. 14 Q. You haven't been involved in Dr. 15 Marin's sensitivity analysis? 16 A. No. 17 Q. You haven't reviewed it? 18 A. No. 19 Q. You haven't seen it? 20 A. I think I saw some -- I know I saw some 21 preliminary results at the meeting in Orlando. 22 Q. And what were those results? 23 A. I don't recall the details. I believe 24 the general result that I recall was that the 25 STA's would have to be sized considerably larger JACK BESONER & ASSOCIATES, INC. 59 1 -- would possibly have to be sized considerably 2 larger if you change certain assumptions about the 3 model. 4 Q. Do you recall that in conducting that 5 analysis what settling rate or settling velocity 6 was used by Dr. Marin in his analysis? 7 A. I believe he used a wide range. 8 Q. So he sort of did a mini/max solution 9 on the sizing? 10 A. Yes, if I recall correctly. 11 Q. Okay. And you indicate you have 12 several other of Walker reports in your 13 possession, and you don't think you've read them. 14 May I deduce from that with keen insight that you 15 do not anticipate offering expert testimony in 16 this matter at hearing regarding those records and 17 the subject included therein? 18 A. To the best of my knowledge, yes. 19 Q. Approaching the subject from the other 20 end, in case I wasn't smart enough to ask the 21 question the right way, what expert opinions, if 22 any, do you intend to offer based on your review 23 of Dr. Walker's work at the hearing in this case? 24 A. Does Dr. Walker's work include Appendix 25 "E" in the SWIM Plan? JACK BESONER & ASSOCIATES, INC. 60 1 Q. All right. Now, you understood, back 2 in March, Dr. Walker's '91 model was different 3 from Appendix "E"; did you not? 4 A. That's correct. 5 Q. Okay. And you understood that Appendix 6 "E" is the SWIM Plan. 7 A. That's correct. 8 Q. Okay. And what your clients have 9 challenged is the SWIM Plan. 10 A. Okay. Then I don't believe I'll be 11 testifying on anything that Dr. Walker has done. 12 Q. May I take it from your hesitation and 13 in cross examination of me, that you do plan to 14 offer some expert opinions about the model in 15 Appendix "E"? 16 A. Yes. 17 Q. Okay. What opinions, if any, have you 18 formed regarding Appendix "E", and are these final 19 opinions? 20 A. Preliminary. 21 Q. When do you plan to have a final 22 opinion on that? 23 A. Within the next couple of weeks. 24 Q. What is going to lead you from your 25 current preliminary opinion status to final JACK BESONER & ASSOCIATES, INC. 61 1 opinions in a couple of weeks? 2 A. Mostly my analysis of the Department of 3 Justice Data. 4 Q. Which DOJ data? 5 A. The Loxahatchee data during the entry 6 and access period from '92 to '93. 7 Q. What analysis of those data are you 8 conducting that will finalize your opinions with 9 regard to Appendix "E"? 10 A. How they compare with the '78 to '83 in 11 Appendix "E". 12 Q. What statistical mechanism are you 13 using to conduct that comparison or evaluate the 14 manner in that they may or may not compare? 15 A. Well, the model in Appendix "E" is 16 based on a regression model, and assuming that 17 there is a negative correlation between the 18 fourteen station average of log TP and the stage, 19 and the limits are then derived based on this 20 regression model. 21 So one question you can ask is whether 22 the data gathered in the entry and access periods 23 still exhibit that same relationship between the 24 fourteen station average log TP and average 25 station because those limits are based on the JACK BESONER & ASSOCIATES, INC. 62 1 assumptions that that relationship holds -- 2 Q. No. With the correlation, I 3 understand. Have you analyzed the ESP data to see 4 if there exists the same correlation? 5 A. Yes, I've looked at ESP data, and I've 6 started to look at DOJ data. 7 Q. Okay. Now, extracting from the ESP 8 data, does the correlation appear to exist or hold 9 true with the approach taken in Appendix "E" to 10 that data? 11 A. Yes. That's a good question. 12 Q. Thank you. We sort of understood 13 that. I mean, that's sort of getting to the hub 14 of the thing. 15 A. There are some interesting features 16 about the ESP data, and there are also very 17 interesting features about the DOJ data, features 18 which are different. It's a different issue. In 19 the ES&P data, the measurements for April of '93 20 are extremely low. In fact, I think they are 21 almost all at the detection limit. 22 Q. What was the detection limit for ESP? 23 A. The detection limit was between two and 24 four parts per billion is my understanding. 25 Q. Okay. Did any of the subject matter JACK BESONER & ASSOCIATES, INC. 63 1 experts from ESP or BDA or any of the other 2 consultant groups suggest to you an explanation 3 for the extremely low values in April of '93 in 4 their data? 5 A. I asked John Davis about that, and he 6 was not aware of any particular mechanism as far 7 as the sampling procedure or the laboratory 8 procedure that was any different from any of the 9 other dates. 10 Q. Was anyone else able to offer an 11 explanation? 12 THE WITNESS: Well, we talked about that 13 yesterday. 14 MR. BLANK: Please don't offer me. 15 MR. FITZGERALD: I mean, I'm willing to 16 listen to you testify, but -- 17 THE WITNESS: No, Nobody that I've 18 talked to, who is an expert on wetlands -- the 19 wetlands systems or no one from ESP or BDA has 20 offered a sane explanation. 21 BY MR. FITZGERALD: 22 Q. Did you see the same dip in the DOJ 23 data from April of '93? 24 A. No, the DOJ data -- well, when you say 25 "dip" -- JACK BESONER & ASSOCIATES, INC. 64 1 Q. The same ask extremely low values in 2 comparison to -- 3 A. Right. No, the DOJ values were about, 4 if I recall -- let's see, the ES&P values were 5 about around three and four. I think they were 6 all three, except there was four -- well, sorry. 7 I'm averaging the field replicates again. 8 All the data, they were around three or 9 four. I think maybe there was one that was seven 10 parts per billion. The DOJ data was about two or 11 three times higher. They had values ranging from 12 about five to eleven, I think, something like 13 that. So if you talk just about the ES&P data and 14 look at a -- 15 MR. BLANK: Is there a question 16 pending? 17 MR. FITZGERALD: Yes, there is. He's 18 actually answering it, too. 19 THE WITNESS: As I recall, the question 20 was whether the relationship between the fourteen 21 station average log TP and average station holds 22 for the ES&P data. 23 MR. BLANK: Okay. 24 THE WITNESS: If you include that April 25 '93 observation, then there is not a significant JACK BESONER & ASSOCIATES, INC. 65 1 relationship between the fourteen station average 2 log TP and the stage. 3 BY MR. FITZGERALD: 4 Q. Have you done an outlier analysis on 5 that data? 6 A. Yes, if you do a statistical outlier 7 analysis -- well, I have -- there's a couple ways 8 that you can do the analysis. One way to do the 9 analysis is the way that an outlier analysis was 10 done, the Appendix "E" in the '78 to '83 data, and 11 in that case they allowed for a different affect 12 for each date and each station. 13 Okay. So, if do you that analysis on 14 the field replicates, for '93 -- '92 outlier -- if 15 you do an analysis where -- well, you have to 16 figure out what you're going to do with the field 17 replicates. One thing to do is just average the 18 field replicates and then take the fourteen 19 station average and then plot the fourteen station 20 average with the log TP versus station. 21 And if you do that and fit a 22 regression, you do an outlier analysis based on 23 that model or you where you assume that model is 24 correct, then 4/93 would be an outlier at some 25 small alpha level. JACK BESONER & ASSOCIATES, INC. 66 1 Q. What is your opinion on the proper way 2 to handle the April '93 data from ESP? 3 A. Well, the proper way to handle it is to 4 go back, first of all, examine thoroughly the log 5 books, laboratory procedure, etc., etc., to make 6 sure that nothing unusual shows up there. 7 Q. And have you done that? 8 A. I haven't done it. 9 Q. Okay. But you testified you sort of 10 did that indirectly by asking John Davis to do 11 that. 12 A. Right. 13 Q. And he has not communicated to you that 14 he could find anything to explain this data set? 15 A. That's correct. 16 Q. Okay. 17 A. So then at that point -- 18 Q. All right. 19 A. So then at that point the question is 20 whether there's really a problem with that data or 21 if that data is actually showing you that your 22 assumptions about the model are not correct. 23 Q. Okay. Now, assuming there is something 24 inherently suspect about the April '93 data, from 25 what you said few moments ago, if I understood JACK BESONER & ASSOCIATES, INC. 67 1 that, you, in conclusion of that data, and 2 treating it as valid data, would suggest that the 3 relationship identified in Appendix "E" is not 4 correct. If you exclude that data -- and the way 5 you worded it sounded to me like you probably 6 tried that, too. 7 A. Uh-huh. 8 Q. What is the affect of your analysis on 9 the mechanism and model identified in Appendix 10 "E"? 11 A. If you include that observation, then 12 there is a significant negative slope in the 13 model. There is a negative -- a significant 14 negative correlation between average station, the 15 predictor variable and fourteen station average 16 log TP. 17 Q. All right. Did you fit a regression to 18 that? 19 A. Yes. 20 Q. And what was your R-square in that 21 regression? 22 A. I don't recall the R-squared on that 23 one. I actually fit several regressions. I fit a 24 regression just using the ES&P data. I also did 25 some regressions where I went back and used the JACK BESONER & ASSOCIATES, INC. 68 1 '78 to '83 data and combined that with the ES&P 2 data. And in that case I allowed for, just as 3 Appendix "E", they allowed for a difference 4 between '78 and '89 versus '79 and '83 and then 5 also allowed for a -- basically they assumed that 6 the slope was the same for those two periods, but 7 they allowed the '79 to move up to '83. 8 Q. Do you understand why they did that? 9 A. Because it's my understanding that the 10 way the particular log law is written, the water 11 year '78 is the basis for determining the limits. 12 Q. So you understood it to be an external 13 constraint over which there was no statistical 14 mechanism. It was legally imposed requirements. 15 A. That was my understanding. 16 Q. Okay. When you did that and you 17 employed the same mechanism allowing for the 18 differences, what result did you achieve? 19 A. Well, it depends. There are a couple 20 of issues involved. One is how do you handle the 21 field replicates because the base line data, as 22 far as I know, after repeating the examination, 23 John Davis -- and as far as he knows, the data in 24 Appendix "E", from '79 to '83, only involved one 25 field replicate, where the ES&P data involved JACK BESONER & ASSOCIATES, INC. 69 1 three field replicates, and maybe there was an 2 occasion miseval here and there. 3 Q. So the question is if you want to 4 combine these two data sets and fit a model, how 5 do I handle the field replicates? And the problem 6 arises because, in your model, your average is 7 over the fourteen stations. 8 Okay. So it's not quite clear exactly 9 how you handle those field replicates. You could 10 just average them in the ES&P data, but then you 11 really have a slightly different quantity in your 12 fourteen station average where you've averaged 13 field replicates, which really doesn't -- is not 14 going to really behave in the same way, and 15 variables, in essence, will be reduced compared to 16 the data that you used to conduct the model in 17 Appendix "E" since that was only based on one 18 field replicate. 19 So you could -- one way is to just 20 average the field replicates for the ES&P data, 21 combine those with the '78 to '83, and do a test 22 to see whether the slope of that -- I'm sorry -- 23 the intercept of that line is different. 24 A. Well, first you want to check to make 25 sure the slopes are the same for all three lines, JACK BESONER & ASSOCIATES, INC. 70 1 and you don't really -- that's called a test for 2 interaction. You don't really get a significant 3 effect, nor if you average all the field 4 replicates that I'm talking about, leaving out 5 April of '93. 6 So then if you do that, the regression 7 line for the ES&P data is marginally -- I would 8 say is marginally higher than the base line 9 periods '78 to '79. There's probably a slight 10 statistical significant difference. 11 Q. If you -- 12 A. Now, there is the problem with certain 13 field replicates being identified as possible 14 outliers. And as I mentioned earlier, there were 15 about four or five observations when Dr. Davis 16 went back and looked at the log books. Those are 17 identified as being problems. If you leave those 18 out, the "P" value is going to go up slightly. 19 So whether you want to say there's a 20 significant difference between the average 21 concentrations for the entry and access percent 22 and the average concentration for the base line 23 period depends on what alpha level you want to 24 use. 25 Q. Okay. Would you explain what you mean JACK BESONER & ASSOCIATES, INC. 71 1 by alpha level? 2 A. We talked about that at my last 3 deposition. 4 MR. FITZGERALD: Well, you weren't here. 5 MR. LOREDO: I wasn't here. 6 THE WITNESS: Whenever -- 7 MR. FITZGERALD: I don't think we did 8 talk about the alpha levels the last time. 9 THE WITNESS: Oh, sure we did. There's 10 an exhibit. If you brought that with you, in 11 fact, that exhibit would help me explain what it 12 is. That's it right there. 13 MR. FITZGERALD: Now I remember. We did 14 cover it. We are referring to Exhibit 12. We can 15 pick up where we were talking about the possible 16 correlation of the evaluation of Appendix "E" and 17 what alpha level you would accept or you would 18 use, and you were about to explain what you meant 19 by alpha level. 20 THE WITNESS: Right. In this particular 21 case, what you're testing graphically is on the 22 "Y" axis. You've got fourteen station average log 23 TP. On the "X" axis you've got average station, 24 and you've got a fitted line for the period '78 to 25 '79, the base line period. And you've also got a JACK BESONER & ASSOCIATES, INC. 72 1 fitted line for the entry and access period, and 2 -- 3 BY MR. FITZGERALD: 4 Q. That's your fitted line that you've 5 done? 6 A. Right. It's the -- well, it's the 7 fitted line where I've actually combined all the 8 data together. And I've allowed for -- first of 9 all, I've tested to see whether the line -- it has 10 the same slope for all three periods. 11 Q. Does it? 12 A. Yes. Well, there's not -- 13 Q. Any significant difference? 14 A. It's not significantly different. 15 Q. What does that imply or tell you about 16 the three lines? 17 A. It tells you that they have the same 18 slope. 19 Q. Right. 20 MR. BLANK: Subject to the April date? 21 THE WITNESS: Yes, this is all -- 22 BY MR. FITZGERALD: 23 Q. Excluding April '93 EDP data. 24 A. Right, it tells you that there's not 25 enough evidence to say that they don't have the JACK BESONER & ASSOCIATES, INC. 73 1 same slope. That's what a statistician talks 2 about to prove the null hypothesis. 3 Q. Okay. 4 A. So then assuming that the slopes of the 5 line are the same, the question is whether the 6 intercept is different between the fitted line for 7 '78 to '79 versus the fitted line for the entry 8 and access period. 9 Q. Okay. That's the "Y" intercept? 10 A. The "Y" intercept, yes. And, so, 11 that's what you're testing, whether the line for 12 the ES&P data -- whether the line intercept is 13 different for that line compared to the "Y" 14 intercept for the base line period of '78 to '79. 15 So any time you do a test hypothesis, 16 based on data that is inherently random, you have 17 the possibility of making two kinds of mistakes. 18 You can -- in this case, the null hypothesis would 19 be that the "Y" intercept would be the same "Y" 20 null hypothesis, which would basically be to the 21 argue phosphorus concentration. The entry and 22 access period are no different from the phosphorus 23 concentration in the base line period after 24 adjusting for the stage. 25 Q. So that's your null hypothesis and then JACK BESONER & ASSOCIATES, INC. 74 1 your alternative hypothesis -- you actually have a 2 choice, and the alternative hypothesis -- one 3 alternate hypothesis would be that it's 4 different. 5 A. But it might be higher or it might be 6 lower, and the other alternative hypothesis would 7 be simply that it's higher for the entry and 8 access period. A third one, which you may want to 9 test, is whether it's lower for the entry and 10 access period. 11 So let's assume, for the moment, that 12 the null hypothesis that -- well, the null 13 hypothesis is that there's no difference in the 14 intercepts and the alternative is, assume for the 15 moment, the intercept higher for the entry and 16 access data. Okay. There you can make two kinds 17 of mistakes. 18 First of all, you can say, the "Y" 19 intercept is higher for the entry and access data, 20 when, in fact, it's not. It's just that for your 21 particular sample of data that you got, it 22 happened to look a little higher than -- so the 23 test says that it's higher, and that's called a 24 type 1 error because you're rejecting the null 25 hypothesis. You're saying -- JACK BESONER & ASSOCIATES, INC. 75 1 Q. When you shouldn't. 2 A. Right. You're rejecting the null 3 hypothesis when, in fact, it's true. That's 4 called a type 1 error. Then the other mistake you 5 can make is you can say I don't have sufficient 6 evidence to reject a null hypothesis. So I'm 7 going to have to say that there is no difference 8 between the two intercepts when, in fact, there 9 is, And that's called type 2 error. And as far 10 as this chart -- 11 Q. During the break counsel and I looked 12 at that, and we said that's a wonderful chart, and 13 we understand it. I don't know if we need to go 14 too further. 15 All that is, by way of clearly 16 addressing the issue, is when you raised the point 17 of what alpha value you want to use. 18 A. That's right because when you do a test 19 hypothesis, then you get out of that test 20 something called a "P" value. And in layman's 21 terms, the "P" value is the probability of seeing 22 a result as extreme or more extreme than the one 23 that you saw, assuming the null hypothesis is 24 true. 25 So the general procedure is, if you JACK BESONER & ASSOCIATES, INC. 76 1 have a small "P" value, if you see something that 2 the "P" value indicates would not happen very 3 often, if, in fact, the null hypothesis is true, 4 that tends to make you believe the null hypothesis 5 is not true. So then the question is: Well, how 6 small does the "P" value have to be? 7 Are you going to reject a null 8 hypothesis if you see something that happens only 9 ten percent of the time or only five percent of 10 the time or only one percent of the time or point 11 one percent? There's no answer from statisticians 12 to tell you that. That's a policy decision. 13 Q. Okay. When you conducted your testing 14 of the "Y" intercepts, comparing the base line and 15 the ESP, what did you utilize as an alpha value? 16 A. I only reported "P" values. See, 17 that's what statisticians out of that. 18 Q. And what "P" values did you report? 19 A. Well, I generated a table of "P" 20 values, and I also generated plots of "P" values. 21 And let me explain that in a minute. The table of 22 "P" values that you generated were based on 23 whether you include April '93 or not, and the ES&P 24 data, when you include April '93 or not, whether 25 you include all the data, or whether you get rid JACK BESONER & ASSOCIATES, INC. 77 1 of the four or five observations that were 2 indicated to be problem observations based on the 3 log books. 4 And then there's actually one more 5 possibility, and that also even after you get rid 6 of the four or five observations that were 7 considered to be problems, based on the log books, 8 if you run an outlier analysis on the remaining 9 data, similar to the outlier analysis that was 10 done in Appendix "E", there is one station and 11 date that turns up as an outlier, which I believe 12 is station 9 in June of '93. So you could leave 13 that out of the analysis, too, as I recall it 14 did. Whether you include that or leave that out 15 didn't make a lot of difference. 16 So I produced a table of "P" values. 17 Based on those "P" values, based on those -- and 18 those -- that comes up to six pockets. Then with 19 or without April '93, whether you get rid of the 20 four or five problem field replicates, and then 21 whether -- you also get rid of station 9 in June 22 of '93. 23 Now, I did that where you average the 24 station replicates. So that's one way to do it. 25 Those "P" values, in a sense, are not really JACK BESONER & ASSOCIATES, INC. 78 1 technically correct because, as I said before, if 2 you average the field replicates and then take the 3 fourteen station average, that quantity is not 4 really the same kind of random variable as 5 averaging the fourteen stations, based on the base 6 line data because the base line data didn't have 7 field replicates. 8 So I produced a "P" value table -- "P" 9 values that way. I also did it 10 non-parametrically, testing to see if there's a