1
1 Division of Administrative Hearings
2 Department of Administration, State of Florida
3 SUGAR CANE GROWERS COOPERATIVE)
OF FLORIDA; ROTH FARMS, INC.; )
4 and 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._____)
) VOLUME I
8 FLORIDA SUGAR CANE LEAGUE, )
INC.; UNITED STATES SUGAR )
9 CORPORATION; and NEW HOPE )
SOUTH, INC., )
10 Petitioners, )
V ) DOAH
11 SOUTH FLORIDA WATER MANAGEMENT) Case 92-3039
DISTRICT, an agency of the )
12 State of Florida, et al., )
_____________Respondents._____)
13 )
FLORIDA FRUIT AND VEGETABLE )
14 ASSOCIATION; LEWIS POPE FARMS;)
W.E. SCHLECHTER & SONS, INC., )
15 and HUNDLEY FARMS, INC., )
Petitioners, )
16 V ) DOAH
SOUTH FLORIDA WATER MANAGEMENT) Case 92-3040
17 DISTRICT, an agency of the )
State of Florida, et al., )
18 _____________Respondents._____)
19
Deposition of Kenneth Rutchey
20
Taken before Robin L. Merker, Court
21 Reporter and Notary Public in and for the State of
Florida at large, pursuant to notice of taking
22 deposition filed by the Petitioners in the above cause.
23 - - -
Monday, February 7, 1992
24 319 Clematis Street
West Palm Beach, Florida 33401
25 9:10 - 12:15 p.m.
2
1 APPEARANCES:
2 On behalf of the Petitioners Florida Sugar
Cane League, Inc., United States Sugar Corp,
3 and New Hope, Inc.:
4 Earl, Blank, Kavanaugh & Stotts
One Biscayne Tower
5 Suite 3636
Two South Biscayne Boulevard
6 Miami, Florida 33131
By: MARK KOBELINSKI, ESQUIRE
7
On behalf of the Petitioners Sugar Cane Growers
8 Cooperative, Roth Farms, Inc., and WEDGEWORTH
Farms, Inc.:
9
Hopping, Boyd, Green & Sams
10 123 South Calhoun Street
Tallahassee, Florida 32314
11 BY: WILLIAM H. GREEN, ESQUIRE and
CAROLYN S. RAEPPLE
12
On behalf of the Respondent SFWMD:
13
Popham, Haik, Schnobrich & Kaufman, Ltd.
14 100 Southeast 2nd Street
Miami, Florida 33131
15 BY: GREGORY M. CESARANO, ESQUIRE
16 On behalf of the Intervenor, United States of
America:
17
THOMAS A.W. FITZGERALD, ESQUIRE
18 Assistant United States Attorney
155 South Miami Avenue
19 Suite 600
Miami, Florida 33130-1693
20
ALSO PRESENT:
21
JOE B. BIRCH, Ph.D.
22 EDWARD DOWNING
MICHAEL SOUKUP
23 MICHAEL STORY
24
25
3
1 - - -
2 I N D E X
3 - - -
4 WITNESS: DIRECT CROSS REDIRECT RECROSS
5 Kenneth Rutchey
6 BY MR. KOBELINSKI: 5
BY MS. RAEPPLE: 226
7
- - -
8
E X H I B I T S
9
- - -
10 Rutchey Exb. No. 1 8
Development of an Everglades Vegetation Map
11 Using a SPOT Image and the Global Positioning System
12 Rutchey Exb. No. 2 33
WCA-2A Field Data Collection Sites for 30
13 classes 10-2-91 to 1-13-92
14 Rutchey Exb. No. 3 56
WCA-2A Tree Islands
15
Rutchey Exb. No. 4 67
16 WCA-2A Accuracy Assessment 3-23 to 4-24-92.
17 Rutchey Exb. No. 5 80
Laboratory Notebook
18
Rutchey Exb. No. 6 105
19 WCA 3S for SPOT 5-11-92
20 Rutchey Exb. No. 7 105
WCA 3N for SPOT 5-11-92
21
Rutchey Exb. No. 8 108
22 Inland Wetland Change Detection in the
Everglades Water Conservation Area 2A Using a Time
23 Series of Normalized Remotely Sensed Data
24 Rutchey Exb. No. 9 196
Holeyland file, Bates Nos. 1220441 through
25 1220572
4
1
Rutchey Exb. No. 10 228
2 Color Map of WCA-2A, using SPOT image 4-4-87
3 Rutchey Exb. No. 11 273
Satellite imagery as of 1-25-93
4
Rutchey Exb. No. 12 278
5 Resume of Ken Rutchey
6 Rutchey Exb. No. 13 298
Memorandum dated 11-5-92 from Ken Rutchey
7
Rutchey Exb. No. 14 316
8 Memorandum dated 3-30-1990 from Michael Maceina
9 Rutchey Exb. No. 15 354
Latitudes and longitudes, coordinates
10
Rutchey Exb. No. 16 355
11 Review of Remote Sensing activities by John
Jensen.
12
Rutchey Exb. No. 17 357
13 Phosphorus data
14 Rutchey Exb. No. 18 364
Memorandum from Ken Rutchey and Les Vilchek
15 dated 10-22-92
16 Rutchey Exb. No. 19 370
Memorandum from Ken Rutchey dated 7-29-91
17
Rutchey Exb. No. 20 370
18 Memorandum from Ken Rutchey dated 6-18-90
19 Rutchey Exb. No. 21 373
Memorandum from Ken Rutchey dated 10-30-89.
20
Rutchey Exb. No. 22 382
21 WCA-2A fire records
22 Rutchey Exb. No. 23 384
Draft page of Everglades SWIM Plan with
23 handwritten note on back.
24
25
5
1 P R O C E E D I N G S
2
3 - - -
4 Thereupon,
5 Kenneth Rutchey
6 being by the undersigned Notary Public first duly
7 sworn, was examined and testified as follows:
8 THE WITNESS: I do.
9 DIRECT (Kenneth Rutchey)
10 BY MR. KOBELINSKI:
11 Q. Good morning, Mr. Rutchey. My name is
12 Mark Kobelinski. Have you been deposed before?
13 A. Yes.
14 Q. You have, all right. Good. I think I
15 knew that.
16 And I think you're aware that a deposition
17 is a means by which people involved in litigation can
18 ask questions of individuals who may or may not have
19 facts dealing with the issues in the case under oath.
20 I'll be asking you a number of questions. I believe
21 one or another of the attorneys here will be asking you
22 some questions. I'd like you to answer to the best of
23 your ability as truthfully as possible. If you don't
24 understand a question, please let me know and I'll
25 attempt to rephrase it in a manner in which you can
6
1 understand. If you don't know the answer to a question
2 or don't remember, I don't know or I don't remember are
3 the best responses you can give.
4 If you do assume something, please let us
5 know. Otherwise we're going to go on our own
6 assumption that you are not assuming. If you feel
7 compelled to assume, please tell us.
8 Under those ground rules, you were served,
9 sir, I believe with a subpoena duces tecum?
10 A. Right.
11 Q. All right.
12 You were given a notice and, as I
13 understand it, you've gathered documents in response to
14 that?
15 A. Yes.
16 Q. You produced those all to us?
17 A. No.
18 Q. All right.
19 A. This morning I brought in, after reading
20 it further, a report. You might have this, you might
21 not. I don't think that was -- the C-51 study, you
22 said all publications that I've been involved in.
23 Q. Okay.
24 A. And this is another one that was by Dewey
25 Worth in the early '80's, work that he did in
7
1 documenting the vegetation, and I have basically all
2 the data for that report since Dewey has now left the
3 District.
4 Q. Great, all right.
5 Well, we'll get some copies of that later.
6 A. And the other item I was asked to provide
7 was the digital data also.
8 Q. Um-hum.
9 A. The techs and I haven't put that all
10 together yet.
11 Q. All right.
12 A. I'm working on that.
13 Q. Now I note that you have a document there
14 in front of you, titled Development of an Everglades
15 Vegetation Map Using a SPOT Image and the Global
16 Positioning System. Is that the final or most recent
17 copy of that particular paper?
18 A. Yes.
19 Q. All right.
20 And I know you did produce that; is that
21 correct?
22 A. Yes.
23 Q. All right.
24 Well, why don't you give me few moments to
25 go through it. We weren't quite sure which was the
8
1 final version.
2 (Thereupon, a discussion was held off the
3 record.)
4 MR. KOBELINSKI: All right. We all set?
5 Why don't you mark that as 1?
6 (Thereupon, the document was marked
7 Rutchey Exb. No. 1 for Identification.)
8 BY MR. KOBELINSKI:
9 Q. Mr. Rutchey, I've handed you what we've
10 marked as Petitioner's Exhibit 1 to your deposition,
11 which is a document entitled Development of an
12 Everglades Vegetation Map Using a SPOT Image and the
13 Global Positioning System, dated September of 1992,
14 Bates No. 1220147 through 1220440. Would you look
15 through that and identify that for me, please, if you
16 can?
17 A. This is the work that Les Vilchek and I
18 did, looking at vegetation in the Everglades and using
19 multispectral satellite data to delineate the
20 vegetation.
21 Q. Okay.
22 And what portion of Everglades was that?
23 A. Water Conservation Area 2A.
24 Q. Okay.
25 When was this work done? When did you
9
1 start on this project?
2 A. I'd say initially right after the data was
3 acquired on 8-10-91.
4 Q. Okay.
5 And when did you finish this project?
6 A. I'd say approximately eight months after
7 the actual -- all the data was collected and then maybe
8 another four months the final report was done.
9 Q. So that was just about September '92 is
10 when you finished up with this?
11 A. Yeah, I'd say that was the last, you know,
12 the revised. You know, you get comments back and...
13 Q. What we've marked Exhibit 1 is the final
14 product from that study, or are there any additional or
15 subsequent drafts?
16 A. There was drafts prior to this, yes.
17 Q. Were there any drafts subsequent to, after
18 this?
19 A. No.
20 Q. Okay.
21 Who is Les Vilchek?
22 A. He's a co-worker, same level as I am,
23 working for South Florida Water Management.
24 Q. What does he do?
25 A. He's -- he does pretty much what I do. He
10
1 has a strong background in aerial photography
2 delineation, dealing with aerial photography. He
3 worked with the National Wetland Inventory before he
4 came to the District and basically learned the image
5 processing from me, the digital image processing and we
6 used his background in aerial interpretation pretty
7 extensively in this project.
8 Q. Thank you.
9 Prior to August of '91, did you have any
10 experience in using satellite imagery, showing
11 vegetation through satellite imagery?
12 A. Yes.
13 Q. Could you just briefly take me through
14 when you first started doing that actually through
15 today, what your experience has been?
16 A. Okay.
17 I'd say I started, I would say it was in
18 '87 when I first became interested in this method. I
19 basically had done a lot of research in the Everglades
20 prior to that using standard ground reconnaissance, you
21 know, going out in the field as a biologist and looking
22 at vegetation, put a transect out. Basically that's --
23 the results of that are in contained in this report
24 that Dewey Worth did. But we saw there was so much
25 more out there that we weren't covering, so we wanted
11
1 to move on to maybe looking at the area in its
2 entirety. And then initially we started just taking
3 photographs from helicopters, putting poles out in the
4 fields so we can basically tell distances, so that when
5 we delineate the vegetation we can come up with an
6 acreage of what things were that we were delineating
7 were or so -- were on the ground.
8 And so we then basically moved to
9 satellite data and Dewey, you have to give Dewey the
10 credit. He's the one that really headed us in that
11 direction. I just sort of, I guess, tagged along. But
12 it piqued my interest and I became very interested in
13 it. In fact, it became my niche at the District. And
14 at that time we had the I2S system, which is an image
15 processing system and I learned it and I did some
16 initial work. And basically at this point I was
17 getting a real education at the District learning the
18 software and different techniques and...
19 Q. Put an approximate period of time around
20 this.
21 A. I'd say '87 to '89.
22 Q. Okay.
23 A. I did a preliminary map of Water
24 Conservation Area 2A using a 1987 SPOT scene. And then
25 I started working on this -- this project, completed
12
1 this work. At this point I'm moving more now towards
2 using satellite imagery and aerial photography. I
3 consider all remotely sensed data or image processing
4 of data. That's it pretty well.
5 Q. Okay.
6 A. Sums it up.
7 Q. You say a 1987 SPOT vegetative map, was
8 that -- what area was that?
9 A. Water Conservation Area 2A.
10 Q. With regard to the photo surveys using
11 helicopters for use in vegetative mapping, what period
12 of time was that?
13 A. (No response.)
14 Q. I believe you said you used, did some
15 photo surveys using helicopters where you put posts in
16 or poles in to mark --
17 A. That was before any actual digital image
18 processing. That was real early like, I'd say, in the
19 '81 through '85.
20 Q. Okay.
21 Do you recall what areas you did using
22 that method?
23 A. Basically it was sawgrass areas, tree
24 islands and slough areas.
25 Q. Okay.
13
1 Was that in any of the Conservation Areas?
2 A. Water Conservation Area 2A.
3 Q. Did you cover the entire Conservation
4 Area?
5 A. No.
6 Q. Do you still have the results of those
7 surveillances or not?
8 A. Some of it.
9 Q. Did you publish a report or put out any
10 type of report based on that?
11 A. Basically all that work is probably in
12 here.
13 Q. Okay.
14 A. We collected a lot of data and we put this
15 report together. We didn't use all the data. We used,
16 you know, we collected a lot of data, so it was...
17 Q. Okay.
18 Going then towards this report which we
19 marked as Exhibit 1, if you could walk me through the
20 process.
21 And I believe, just for the record, the
22 copy that has been marked as Exhibit 1, I do not
23 believe has color photos on the back of it. I know
24 that you have another copy of that here -- is that a
25 fact, Mr. Rutchey --
14
1 A. Correct.
2 Q. -- of this report which does have color
3 photos?
4 A. (Shakes head up and down.)
5 Q. Feel free if, during the deposition you
6 need to refer to color photos as opposed to those black
7 and white photocopies to do so. We will then attach a
8 corrected one with the color photos to the deposition
9 itself.
10 If you could literally, from the point
11 where you became involved, walk me through the steps it
12 took to get to this final report, and if I need a
13 little more detail I'll ask you. Otherwise --
14 A. Okay.
15 Q. When did you -- I assume the first part of
16 that was selecting a -- well, you tell me. Step Number
17 1, when did you get the assignment?
18 A. Well, we got the imagery that was
19 collected on August 10, 1991 from SPOT data
20 multispectral imagery.
21 Initially the first thing we do is went
22 out and got ground truth information to rectify the
23 imagery.
24 Q. Before you get to that point, how was the
25 August SPOT imagery selected?
15
1 A. Basically the overriding consideration was
2 availability. In the Everglades we don't have a lot of
3 clear days where we have a satellite where it's totally
4 cloud free.
5 Q. Um-hum.
6 A. So that was why that one was selected.
7 Q. Okay.
8 Who actually made the selection of the
9 August 10, '91 SPOT imagery?
10 A. I did.
11 Q. Okay.
12 And was this -- was SPOT imagery the only
13 satellite imagery that you considered at this point?
14 A. No, no. We looked at LandSat too. Again
15 the overriding consideration though was a scene that
16 was totally cloud free for that area.
17 Q. Okay.
18 A. And there was some other work done using
19 LandSat in the Everglades, and they compared it to
20 SPOT, and they showed that SPOT was the better way of
21 looking at the vegetation and delineating.
22 Q. Okay.
23 Who had done this LandSat work and
24 compared it to SPOT?
25 A. Terry Gilbert of Game and Fish.
16
1 Q. Did you consult with Gilbert with regard
2 to selecting the SPOT imagery?
3 A. Yes.
4 Q. Had you worked with the LandSat before?
5 A. Yes.
6 Q. And did you concur with Mr. Gilbert's
7 opinion that SPOT would be a better imagery to use as
8 opposed to LandSat?
9 A. Yes.
10 Q. Okay.
11 What was the specific purpose or goal at
12 this point in time of the assignment? What were you
13 attempting to do?
14 A. Use satellite imagery to delineate the
15 vegetation in Water Conservation Area 2A.
16 Q. Any particular vegetation types or
17 literally all the vegetation you could?
18 A. All the vegetation we possibly could.
19 There was no emphasis on any specific species.
20 Q. Okay.
21 And did you select 2A as the area of study
22 or was that just, again, part of the assignment?
23 A. Selected it. Because that's where I had
24 worked since my beginning of employment with South
25 Florida Water Management District, and that area also
17
1 has had the most work done in it in previous.
2 Q. Okay.
3 Was the assignment literally just to map
4 any area within the Water Conservation Areas, open area
5 selection? How was the assignment developed?
6 A. Basically just we mapped Water
7 Conservation Area 2A.
8 Q. I'm just curious whether or not the
9 assignment was we need to start doing a vegetative
10 mapping, and you said fine, I'd like to do 2A because
11 that's what I'm most familiar with. How was that
12 assignment developed?
13 A. I would say that the direction at that
14 point came from Dewey Worth to start off doing Water
15 Conservation Area 2A.
16 Q. Okay. All right.
17 Did you review various SPOT imagery for 2A
18 and then select August 10 from a number of options or
19 was August 10 essentially the one that was proposed to
20 you?
21 A. Again, there wasn't a whole lot available
22 due to cloud cover. And it's -- I didn't want to go
23 back to a scene that was collected earlier because then
24 it makes it hard to ground truth. So I had to wait for
25 a scene to become available and then start working on
18
1 it.
2 Q. Okay. All right.
3 In what format did you receive the August
4 10 SPOT imagery data?
5 A. In digital format.
6 Q. Did you still have those digital files?
7 A. Yes.
8 Q. Unaltered, unmodified?
9 A. Yes.
10 Q. Okay.
11 Once you received the digital files for
12 the '80 -- excuse me -- 8-10-91 SPOT imagery, what did
13 you do next?
14 A. I went out, collected ground truth data
15 for rectification of the imagery. We collected 20
16 ground control points. That was the first thing we
17 did. Then we cut the actual area out that we were
18 interested in, which was Water Conservation Area 2A,
19 out of the imagery.
20 Q. Okay.
21 So let me -- first, after getting the
22 data, I assume you load it on to a computer, then went
23 up to selected 20 sites to ground truth --
24 A. To ground truth for rectification
25 purposes.
19
1 Q. And what does ground truthing for
2 rectification purposes mean?
3 A. When you get the data it's not to any real
4 world map projection. So we wanted to get it into a
5 real world map projection using real world ground truth
6 information.
7 Q. Now in other words, when you get it, it
8 doesn't have the boundary of 2A drawn on it; is that
9 right? Is that --
10 A. No. The way you get a boundary is to cut
11 the imagery from the satellite, original satellite
12 data. This rectification is so that when you look at
13 an area and -- you know where the location is at.
14 Q. Okay.
15 A. You don't know that when you first get it.
16 You have to establish ground controls and rectify it.
17 Q. So what are the types of sites you use for
18 that purpose?
19 A. Some of the fish camps within the area and
20 then points all the way around the outside perimeter.
21 Q. That including the structures?
22 A. I believe so.
23 Q. And did you do the 20 sites, inspection of
24 the 20 sites for ground truthing?
25 A. Les Vilchek and myself.
20
1 Q. Approximately how long did that take?
2 A. I would say a day to two days to collect
3 the information.
4 Q. Okay.
5 What's the next step?
6 A. Okay.
7 Initially we took the raw data and did an
8 unsupervised classification and we chose 30 clusters.
9 We did the unsupervised classification within the 30
10 clusters. We tried to find areas that we can visit in
11 the field. And this was all based on the accuracy of
12 the GPS. So basically the imagery was initially
13 rectified to .4 RMS, root means square, it means that
14 pixel within it is accurate to approximately eight
15 meters.
16 Q. With regard to this process, how many
17 classes did you -- did the computer or did you collect?
18 A. 30 clusters.
19 Q. 30 clusters, all right.
20 Let me ask you this, perhaps as we go
21 through, for the sake of -- we received a fair number
22 of documents, a lot of which are, as you know, lists of
23 data. With regard to Step 1, which was selecting the
24 August 10, 91 imagery, did you produce any documents
25 specifically related to that fact?
21
1 I'm just trying to organize the documents
2 that you produced.
3 A. Could you repeat that question?
4 Q. You've produced to us about a foot and a
5 half of documents in the last few days.
6 A. Okay.
7 Q. We're trying to order them to understand
8 exactly what they are and identify them. Perhaps it
9 might be easiest to see which of those documents fit in
10 a particular project you were doing and in what steps.
11 I'm asking you, did you produce any documents with
12 regard to your selection of the August 10, 91 SPOT
13 imagery?
14 A. Yes.
15 Q. Okay.
16 What documents would have you produced in
17 regard to that?
18 A. If I remember, all -- I think you have all
19 the GPS location points that were used in the
20 rectification process, in the ground truthing and in
21 the your analysis of the mapping project.
22 Q. Okay.
23 I understand we probably have all the
24 documents. Our problem is trying to figure out which
25 documents are for what task. Are there specific
22
1 documents with relation to your selection of the August
2 10, '91 SPOT imagery?
3 A. I don't think so.
4 Q. With regard to your selecting 20 sites for
5 ground truthing and for rectification purposes, were
6 there any documents specifically related to that task
7 that you've produced to us?
8 A. I believe so.
9 Q. What would those look like?
10 A. Well, I identify them based on the folder
11 that I have them in.
12 Q. Well, we received those things clipped or
13 rubber banded together. There's the first batch,
14 perhaps if you went through that and can identify the
15 20 sites or the documents related to the 20 sites for
16 ground truthing.
17 While you're doing that, let me ask a
18 quick question. Are you aware that we received your
19 documents in two batches?
20 A. No.
21 Q. Okay.
22 So that wouldn't help, telling you I
23 received it in Batch 1 or Batch 2.
24 (Thereupon, a discussion was held off the
25 record.)
23
1 BY MR. KOBELINSKI:
2 Q. It's not there. Okay. This is another
3 set of documents we received.
4 A. I don't see them.
5 Q. Okay. At the break we can take a closer
6 look at that.
7 Moving on to the next step that you
8 described to me as taking the raw data and looking at
9 the computer, doing an unsupervised classification,
10 would there be any specific documents generated from
11 that? Would the computer generate a particular file
12 or --
13 A. I believe all it's all digital. I don't
14 think there's any hard copy of that.
15 Q. Okay.
16 Now, do I understand correctly that the
17 computer collected more than 30 classifications, but
18 ultimately narrowed it down to 30?
19 A. No.
20 Q. How does that work?
21 A. I chose 30.
22 Q. Oh, you essentially told the computer I
23 want 30 classifications, go through and find them?
24 A. I want 30 clusters to be my final number
25 of classes.
24
1 Q. Okay. Well, you have to excuse me, I'm --
2 the difficulty is you have someone who is not expert in
3 your field deposing an expert in your field.
4 When it's an unsupervised classification
5 by the computer, does that mean the computer just goes
6 through and selects what it considers to be vegetation
7 of the same class or cluster?
8 A. Basically it looks at the band information
9 of the satellite data and clusters it out based on
10 means, basically it's a standard deviation, into 30
11 separate clusters.
12 Q. But you're the one that first plugs in to
13 the computer and tells it I want 30 to start with?
14 A. Right.
15 Q. So you could have done 40 and it would
16 have given you 40?
17 A. That's correct.
18 Q. Or you could have given it one and it
19 could have -- would have given you the entire map I
20 guess; is that essentially correct?
21 A. Yes, I think that's probably what would
22 happen.
23 Q. Did you write any memos or anything
24 related to that particular document?
25 A. Not to my knowledge.
25
1 Q. Okay.
2 After you received then the computer
3 classification of 30 clusters -- would that be the
4 appropriate term?
5 A. Yes.
6 Q. Okay. 30 clusters.
7 -- what was the next step you did?
8 A. Okay.
9 Basically, we tried to go out in the field
10 and identify what those clusters were.
11 Q. Okay.
12 How did you go about doing that?
13 A. We used GPS, Global Positioning System,
14 and tried to pick a minimum of five points in each one
15 of these clusters, geographically locate it throughout
16 the area. And that's what we used for trying to
17 determine those 30 clusters.
18 Q. Okay.
19 Just so I understand how you went about
20 determining the 30 clusters, did you print out a map?
21 Was there a map printed out of it showing the 30
22 clusters, where they were?
23 A. That's correct.
24 Q. All right.
25 Do you still have a copy of that map?
26
1 A. It's digital.
2 Q. All right.
3 Did you allow the computer to select the
4 sites where you would go out and ground truth or
5 determine what clusters represented or did you do it
6 yourself?
7 A. Did it myself.
8 Q. Okay.
9 Do you have any type of map or -- showing
10 where those sites were that you visited for the
11 determination of classes?
12 A. Yes. I -- there might be a hard copy, I
13 can't remember. But there's definitely a digital copy.
14 Q. Okay.
15 And in the digital copy, would it have the
16 coordinates of where you were going?
17 A. That's correct.
18 Q. Okay.
19 How did you go about selecting sites you'd
20 visit?
21 A. It was based on whether we were able to
22 accurately navigate to them use GPS technology. There
23 was some inaccuracy in the actual GPS and some
24 inaccuracy in the rectification process of the
25 satellite imagery. So basically we determined that we
27
1 had to have a minimum of 3 by 3 pixel cluster in order
2 to accurately navigate to. In most cases we prefer to
3 have bigger than that.
4 Q. Now with a 3 by 3, you're talking about,
5 what is that, 60 meters squared?
6 A. 60 meters on side. 60 by 60 by 60 by 60.
7 Q. All right.
8 And how many sites did you go to identify
9 the clusters?
10 A. 129.
11 Q. All right.
12 And that was yourself who picked the 129
13 sites?
14 A. I would say I picked them along with Les
15 Vilchek.
16 Q. Okay. All right.
17 What was the next step then, visiting the
18 sites?
19 A. We visited the sites and we wrote down
20 what we saw there and came back and wanted to see if
21 the cluster was accurate and it -- it had broken it out
22 into 30 clusters, whether they were all unique. And at
23 this point we determined they weren't unique, that
24 there was real heterogeneity with a number of clusters.
25 Q. Now prior to the visitation to the site,
28
1 how did you get to the site?
2 A. Either by air boat or helicopter.
3 Q. Okay.
4 And you did that using GPS, I believe you
5 mentioned?
6 A. That's correct.
7 Q. Okay.
8 Once you got to the site, I believe you
9 stated you identified the vegetation?
10 A. That's correct.
11 Q. What would you do, check a site in the
12 center of the pixel, in the center of the 9 square or 9
13 pixels?
14 A. Approximately.
15 Q. Then how would you go about identifying
16 what was there?
17 Who would do that? Let's start with that.
18 A. We stuck with one person so that we
19 wouldn't -- it's just better if you do it with one
20 person than have a number of people working on that
21 aspect of it, because he stays consistent throughout
22 the process.
23 Q. Okay.
24 A. And Les Vilchek was that person.
25 Q. Okay. All right.
29
1 Once you arrived at a site, where would he
2 look to make the identification?
3 A. If we were in a helicopter he would look
4 down at the site, because you want to try to determine
5 what the satellite was seeing. If you were on an air
6 boat, we have cages on the air boat, we got on top of
7 the cage and looked basically down so we can see was it
8 sparse or moderate or dense or what the prevailing
9 vegetation signature was for that area.
10 Q. Okay.
11 With regard to the helicopter, were you
12 hovering, trying to hover at any particular level or...
13 A. Yeah, I'd say it was approximately 50
14 feet.
15 Q. Okay.
16 A. 25 feet.
17 Q. All right.
18 You now -- as I understand it, you are in
19 the approximate, what you've identified as being in the
20 approximate center of a 9 pixel square, each pixel
21 being 20 meters by 20 meters. Was Les attempting to
22 identify the community in all 9 pixels or was he trying
23 to look at what he considered to be the boundary of
24 that center pixel?
25 A. I would say he was looking at a 20 meter
30
1 by 20 meter area.
2 Q. Was that something you discussed with him
3 so you know that or is that just your --
4 A. No, we discussed that.
5 Q. Okay.
6 Now at -- during this trip would you also
7 observe and take any notes if, for instance, if there's
8 a change in the type of vegetation surrounding -- from
9 what you've identified as the center pixel?
10 A. I believe he did that, yes.
11 Q. Now were there written documents? Were
12 there written observations from this trip?
13 A. I believe there is, yeah. I think he --
14 they might be included here somewhere.
15 Q. But he made field notes of every, all 129
16 visits?
17 A. I believe so, yeah.
18 Q. Okay.
19 Was he keeping the field notes or were
20 you?
21 A. He kept them pretty much.
22 Q. Okay.
23 Did you make any separate observations or
24 field notes of any sort?
25 A. No. I was basically the GPS man. I
31
1 also -- I put it in in navigating to points and making
2 sure we got to the right place.
3 Q. Okay.
4 Now, is it your belief that the field
5 notes from these site visits were included in the
6 documents that were produced to us?
7 A. I think so, but I'm not positive.
8 Q. Okay.
9 Can we go through and again identify those
10 so we can start segmenting out what these documents
11 are? If it's easier, as I said, take all the rubber
12 bands off.
13 (Thereupon, a discussion was held off the
14 record.)
15 THE WITNESS: This is in digital form. I
16 don't see the --
17 BY MR. KOBELINSKI:
18 Q. Let's set that aside for the moment.
19 Everything that's -- everything you have that relates
20 to the -- is it just that one document you have in your
21 hand or is there anything underneath that also?
22 A. Well, this actually gives you the
23 coordinate and what we saw there. I'll keep looking
24 through.
25 Q. Well, I'd like to see if we do have field
32
1 notes. Let's -- why don't we set that aside. I'll
2 identify that for the record in a moment. But if
3 there's anything else related to that visit to the 129
4 sites in the field and the identification of
5 vegetation.
6 A. I think this pertains to that too. I
7 think these are it. We actually made sheets up, field
8 inspection sheets. I think this is it.
9 Q. All right. Let me try find these here.
10 MR. CESERANO: You want the number?
11 MR. KOBELINSKI: Yes.
12 MR. CESARANO: 1202944.
13 THE WITNESS: No, this, I think, was for
14 ground truthing.
15 BY MR. KOBELINSKI:
16 Q. Why don't we put -- this is yours, and I
17 think you turned some stuff over here. I just want to
18 make sure.
19 MR. CESARANO: Pardon me, let me look at
20 this for a moment if I could.
21 MR. KOBELINSKI: All right. What does
22 that start with?
23 MR. CESARANO: 1203.
24 MR. KOBELINSKI: All right. And I think
25 we'll just mark this as a large composite
33
1 exhibit.
2 (Thereupon, the document was marked
3 Rutchey Exb. No. 2 for Identification.)
4 BY MR. KOBELINSKI:
5 Q. Mr. Rutchey, did you -- were there any
6 photos taken during the three site visits, the 129
7 sites?
8 A. I think we did do it for this one, I'm not
9 sure.
10 Q. Who would have those photos?
11 A. I'd have to look in my slides for that,
12 because we have gone there on and off on different
13 things.
14 Q. Would they be in just slide form or would
15 you also have photos, do you know?
16 A. They'd be slides.
17 Q. Okay.
18 Who would have taken the -- snapped the
19 photo or slide or whatever, done the camera work?
20 A. Yeah, I -- if I recall, I don't think we
21 did it for this particular project, we didn't take
22 pictures. I'm not sure.
23 Q. Okay.
24 A. I'd have to look.
25 Q. All right.
34
1 Mr. Rutchey, I'm showing you a composite
2 exhibit that we've marked as No. 2. There's actually a
3 number of different portions to it, but the first page
4 appears to be a photocopy of a possibly a file folder
5 which reads WCA-2A field data collection sites for 30
6 classes, 10-2-91 through 1-13-92. And all these
7 documents bear consecutive Bates numbers in the lower
8 right-hand corner, 1202881 all the way through 1203072.
9 Some of these are stapled together.
10 I'd ask you first of all, is this a -- one
11 entire file that you keep, this particular file?
12 A. It's not an entire file, it's a number of
13 different files.
14 Q. Perhaps -- because it was produced to us
15 in this consecutive order, I'd like you to go through
16 and perhaps see if you can identify each set, the first
17 portion of which is stapled together at 1202882. Do
18 you know what that particular document relates to?
19 A. This is the actual data that I sent to
20 John Jensen of the 30 ground truth sites and their
21 coordinate location.
22 Q. And it's actually referring to 129 sites?
23 A. Yes.
24 Q. Okay.
25 Now -- so this does relate to this initial
35
1 site identification of the 129 sites and the classes
2 that are listed there in this document. Where did you
3 get that information?
4 A. These are the classes that we determined
5 from looking during our aerial reconnaissance is the
6 site locations.
7 Q. Okay.
8 Just going through this top page here,
9 1202882, the first column there are a number of
10 numbers. What would that indicate?
11 A. Just basically it was a way of telling the
12 different ones apart, 1 through 129.
13 Q. So that's just essentially the numbers
14 that, for instance that first page we're on, 61, 63,
15 80, these are just the various sites; is that correct?
16 A. Um-hum.
17 Q. All right.
18 The next one is labeled class, I believe
19 you just stated that classification, you identified it
20 based upon your observations; is that correct?
21 A. Yes.
22 Q. I'm still on Page 1202882.
23 A. Yes.
24 Q. The next has UTM.
25 A. That's correct.
36
1 Q. What is that UTM, coordinate?
2 A. Universal Transverse Mercator
3 Q. That's your GPS coordinates there?
4 A. Yes.
5 Q. All right.
6 Next document that, at least it's stapled
7 together, bears Bates No. 2202885 through 887. And
8 could you -- do you know what this document reflects?
9 A. I'm going to tell you these don't look
10 like they're in order to me. And not to have my
11 folders here, my original stuff, I'm guessing basically
12 what I'm seeing here.
13 Q. All right.
14 Well, to the best of your ability do you
15 do you recognize what this could be?
16 A. It's basically -- it's classes grouped
17 together and their location, such as Class 1 is -- you
18 see all together and Class 2 and so on, and so on to
19 30.
20 Q. This is based upon the original 129 sites?
21 A. I believe so.
22 Q. Okay.
23 And the first column there says point old,
24 what does that mean?
25 A. I'm going to have to go back. We had two
37
1 sites, point old and -- well, not point old. And we
2 had new and old -- not new. It was basically two
3 different trips. We combined them together in the
4 final documents or data set.
5 Q. When you say two different trips, is this
6 all part of identification of the 129 sites where the
7 vegetation was?
8 A. Yes. Two different trips that basically
9 it might have been more than two trips, but at one
10 point in time we had one data set and then we had
11 another data set, then we combined them, the old data
12 set, old and new date set.
13 Q. Good classifications. All right.
14 Well, let me back up a moment. How long
15 did it take you to visit all 129 sites?
16 A. I think we did it over a period of two or
17 three months.
18 Q. Okay.
19 And what months would those be?
20 A. I think it says right here on the
21 beginning of Exhibit 2, 10-2-91 to 1 -- it look likes
22 13-92.
23 Q. Okay.
24 And did you go to all these sites?
25 A. Yes.
†⁏歡礮ഊഊ††††㈴†††††††††䅮搠摩搠祯甠杯⁴漠慬氠瑨敳攠獩瑥猿ഊഊ††††㈵††††††䄮††奥献ഊ††††††††††††††††††††††††††††††††††㌸ഊഊ††††‱††††††儮††䤠扥汩敶
38
1 Q. I believe you said Les Vilchek did all the
2 identification. He also obviously went too.
3 A. Right.
4 Q. Did anybody else participate in the site
5 identification trips?
6 A. No.
7 Q. Approximately how many trips did it take?
8 A. Without having my folders, I don't recall.
9 Q. Okay.
10 And did you visit certain sites more than
11 once?
12 A. I don't think so.
13 Q. Okay.
14 I was just trying to understand why there
15 were two different data sets that you then merged.
16 A. I'm going to have to go back and look. I
17 remember that we collected one set and then we
18 collected another set, not repeating, two different
19 sets. We called one old and one new, and then we
20 combined it. That's the best of my knowledge, without
21 going back to my notes.
22 Q. All right.
23 Are those notes that you've produced to
24 us?
25 A. No, they're in -- it's basically in my --
39
1 yeah, they're probably here. But it's all -- not like
2 in my folders.
3 Q. Okay.
4 So you had field notes then from this
5 aspect of the study?
6 A. Well, those were already part of that.
7 Q. We're coming up on those.
8 A. Right.
9 Q. Okay.
10 Let's move on to the next document which
11 is Bates No. 120288 through 120291. And this likewise
12 has point new. My question would be that in the
13 approximately fourth column there, a number of data
14 points have 5 by 5 or 5 X 5. What does that refer to?
15 A. I believe that was if it had 5 by 5 pixels
16 or more, instead of a 3 by 3 pixel.
17 Q. Okay. All right.
18 Moving on then to next stapled portion of
19 this, which bears Bates No. 1202892 through 1202898,
20 and what would this relate to?
21 A. These are all out of the software called
22 PFINDER, and it's our software that we use to
23 differentially correct our field GPS data. This is the
24 output.
25 Q. Okay.
40
1 This is output from the unit you had out
2 in the field?
3 A. I'm not sure without having my folders in
4 front of me. I mean it's either output from the unit
5 in the field or it's the output after it's been
6 differentially corrected.
7 Q. All right. Let me just explore that for a
8 moment.
9 You have a GPS unit and it takes you to a
10 spot; is that correct?
11 A. That's correct.
12 Q. All right.
13 What are you -- how do you get to a spot
14 with a GPS unit?
15 A. You can enter way points into the GPS
16 unit, basically navigation points. You can navigate to
17 the point. The military's been jamming GPS signals
18 since the Gulf incident, so about the worse accuracy is
19 about a hundred meters. But there's -- we've found
20 that you can take that data set back to your office,
21 you have a community base station set up collecting GPS
22 data out of a known location that's been surveyed to
23 less than one centimeter. You can differentially
24 correct the recovery unit data to get an accurate
25 location of where you really were.
41
1 Q. Okay.
2 The unit itself, the recovery unit as you
3 called it, records then time and place, it was --
4 A. It records the time and the position.
5 Q. Okay.
6 Is that where you just press a button and
7 it does that for you, or how does it do it?
8 A. Pretty much press a button.
9 Q. So every time you got a spot that Les
10 would do a site identification you would press a
11 button, it would record the time and coordinates?
12 A. Right. We took a number of coordinates at
13 each place that we stopped, because that gives you
14 better accuracy the more you can average the data. And
15 basically we went to known benchmarks during our field
16 trips that we found, identified and checked the unit to
17 see how accurate we really were getting. And accuracy
18 ranged from three to seven meters.
19 Q. Okay.
20 That was during all your visits to the 129
21 sites?
22 A. Yes.
23 Q. Okay.
24 (Thereupon, a discussion was held off the
25 record.)
42
1 BY MR. KOBELINSKI:
2 Q. The next document is Bates No. 1202899
3 through 1202904.
4 Can you identify what that document is?
5 A. Again, this is output from the PFINDER
6 software for GPS, PFINDER software.
7 It looks like we can take, for instance,
8 the first page, Site No. 61, these are all the points
9 that were collected at that site. And then this is
10 like until -- before we actually average it and get
11 then a final number.
12 Q. Okay.
13 Are these -- looking at 61, you have
14 approximately 10 readings there. I didn't count it, is
15 that about right?
16 A. Yeah.
17 Q. 1, 2, 3, 4, 5. Yeah I think it's about
18 10. Is that all from the same holding spot or did you
19 actually move around to get those?
20 A. No, same spot.
21 Q. Okay. All right.
22 And then going to the next document
23 1202905 through 909, what would this be?
24 A. Again, this is output from the GPS PFINDER
25 software.
43
1 Q. Okay.
2 Just drawing your attention to the
3 leftmost column where it says text, and following
4 through, is that an identification of some sort?
5 A. In this case, not on that page, it
6 represents the site number.
7 Q. And the following pages?
8 A. Bug in the program. We were able to
9 overcome that bug by looking at the time in the second
10 column to determine where we were.
11 Q. All right.
12 In next document, 1202910 through 913,
13 what is that?
14 A. GPS PFINDER output.
15 Q. Okay.
16 Would that hold true likewise for the next
17 document, 1202914 through 2920?
18 A. That's correct.
19 Q. Okay.
20 And next documents 1202921 through 927?
21 A. Yes.
22 Q. Okay.
23 And then drawing your attention to the
24 next four documents bearing Bates numbers 1202928
25 through 936, what are these documents?
44
1 A. PFINDER GPS data output.
2 Q. Okay.
3 And looking at 1202937 there's handwriting
4 on there, it says this starts quote the new data set
5 quote. Now does that indicate that the new data set
6 essentially starts on December 3, '91 and went only
7 after that period? Is that how you'd read this?
8 A. Yeah, I'd say that's a fair statement.
9 Q. Does this help you recall as to why there
10 were two data sets?
11 A. No.
12 Q. On this sheet, this 1202937, there's
13 handwriting on here in a couple of places, it says
14 miss. What does that refer to?
15 A. I don't recall.
16 Q. Okay.
17 Whose handwriting is that; do you
18 recognize it?
19 A. It looks like mine.
20 Q. Drawing your attention to 1202939, the
21 document goes through 940, the leftmost column in
22 handwriting, what are these, the handwriting identified
23 as 96, 104, 54 as I'm going through?
24 A. I think this means -- dealing with the
25 accuracy of GPS being a hundred meters, that when we
45
1 actually went to a site that we thought was that
2 cluster, so the unit was off a hundred meters, it threw
3 us outside the cluster when we went back and checked
4 it. So we missed a cluster. So we couldn't use the
5 ground truth statement, because we weren't at that
6 location. That's what miss means, we missed.
7 Q. Okay. All right. Fair enough.
8 Well, just look at 1202939, just so I
9 understand your first eight readings, are those all the
10 same site or are those different sites?
11 A. I would say they're different sites.
12 Q. Each of these represents a different site?
13 A. Um-hum.
14 Q. All right.
15 And then quickly finishing up the
16 following two documents, 1202941 through 943, these are
17 likewise GPS documents; is that right?
18 A. Yes.
19 Q. Okay.
20 What would do you when you missed a site,
21 would you go back on a different day?
22 A. We would try to attempt to go back to it.
23 Q. Did you ultimately locate all 129 sites to
24 your satisfaction?
25 A. Initially there were more sites, there
46
1 were -- we just never could get to them. They -- some
2 of the clusters broke out to just 3 by 3 pixels, and if
3 they were jamming for a hundred meters we never found
4 that site. We couldn't include it.
5 Q. Of the 129 original sites that you set out
6 to do, how many did you ultimately get to?
7 A. No, there was more than 129. Initially
8 there was 150.
9 Q. Oh, I see.
10 And do I understand correctly then that
11 you were able to get to a hundred -- get to 129 of the
12 150?
13 A. Right.
14 Q. Okay. All right.
15 Then looking at Bates Nos. 1202944 through
16 1203016, the first page of which has -- states SFWMD
17 Environmental Sciences Digital Mapping Field Inspection
18 Sheet, what are these documents, sir?
19 A. These are the ground truth data, what we
20 saw there at the site depicting the vegetation.
21 Q. Okay.
22 And who kept these, who prepared these
23 documents?
24 A. Les Vilchek.
25 Q. And he prepared these out in the field?
47
1 A. No.
2 Q. Okay.
3 How would he record what he was seeing out
4 in the field?
5 A. On a piece of notebook paper or whatever,
6 then he came back in, he transcribed them to here.
7 Q. Okay.
8 Does he still retain his field book or his
9 field notes themselves?
10 A. I'm not sure. I sort of doubt it.
11 Those things, especially in your air boat,
12 they get all smudged up and get wet, so that's why we
13 did this.
14 Q. Now under a vegetation type, for instance
15 on the first page you have 1, 2, 3, 4, 5, 6, 7, 8, 9,
16 10, 11 categories, how are the 11 categories selected?
17 A. Those are some of the major things that we
18 thought were out there initially, so to make it easier
19 so we don't have write them down every time we went to
20 a site, we made a list.
21 Q. Okay. All right.
22 And I notice -- I draw your attention to
23 documents in the next rubber band set, which is 1203017
24 through 1273072. What are these documents?
25 A. The same. Ground truth information.
48
1 Q. All right.
2 And that is all -- this ground truthing
3 relates to the original visitation of the approximate
4 129 sites; is that correct?
5 A. I believe so.
6 Q. All right.
7 Are there any other documents then that
8 would relate to this portion or this task in the
9 project?
10 A. Just the digital data.
11 Q. Okay.
12 Once you received these field notes or the
13 field inspection sheets, what would do you with them?
14 A. Basically we went to each of the clusters
15 to see if they had clustered out uniquely, and we found
16 that wasn't the case, that a lot of them exhibited
17 heterogeneity within a single cluster within -- based
18 on our field reconnaissance.
19 Q. Okay.
20 Who did that work?
21 A. I did that work.
22 Q. What was the type of recovery unit you
23 used for the GPS?
24 A. A PathFinder Basic.
25 Q. All right.
49
1 You've done all the field work, you have
2 all these field notes, you come back, what's the next
3 step in the process?
4 A. Well, once we determined that everything
5 just didn't cluster out the way we had hoped, we used
6 that data set to do a supervised classification.
7 Q. All right.
8 How did you go about doing that?
9 A. It's basically the computer does -- they
10 do -- you feed it the ground truth information and run
11 a supervised classification and you look at the
12 results, see if they mean anything, see if you have
13 unique spectral signatures for the final data. And we
14 determined that, yeah, we came up with initially 19
15 unique signatures based on that kind of ground truth
16 information. And we used those, initially we saw some
17 problem areas and we attacked those and came up with
18 one additional class.
19 Q. What was the additional class?
20 A. Periphyton.
21 Q. All right.
22 Now with supervised classification, a
23 little more detail there. Did you select 19 classes
24 based upon your review of field notes and make a
25 determination there were actually 19 identifiable
50
1 classifications out there, or did the computer then
2 select 19?
3 A. No, it's based on supervised, based on
4 ground truth data.
5 Q. Okay.
6 So you took the data from your visits to
7 the 129 sites?
8 A. Um-hum.
9 Q. You yourself came up with 19 classes from
10 that?
11 A. Right.
12 Q. Okay.
13 Then you tell -- identify them to the
14 computer, the computer reruns the digital data to see
15 where it finds those classes?
16 A. Based on the spectral signature of those
17 data sites, basically the band information within those
18 areas. It's used in the supervised classification to
19 delineate based on these statistics.
20 Q. Okay.
21 Is this something you were doing or Les
22 was doing?
23 A. I was doing.
24 Q. All right.
25 Are there any documents related to that
51
1 particular activity?
2 A. Hard copy, no. Digitally, yes.
3 Q. What would have you digitally related to
4 this?
5 A. Well, we have all the seed locations for
6 the ground truth information. And basically all the
7 steps from there to the final supervised
8 classification, the output maps.
9 Q. Okay.
10 You didn't make hard copies of the maps
11 though, is that you're saying?
12 A. No. Well, I did make hard copies of --
13 not hard copies, no, no, I didn't. Just digital.
14 Q. Okay.
15 What was the next step?
16 A. Well the two problem areas, as I said, was
17 one was sawgrass/cattail sparse class, we knew from
18 aerial photography that we had a lot of, what we were
19 seeing in the southern end was periphyton that was
20 coming up as sawgrass/cattail sparse, so we wanted to
21 see if we can break that out. So we actually took that
22 class out of the supervised classification and ran an
23 unsupervised classification on that and broke it out
24 into, I believe it was five new spectral classes. We
25 look at each individually, looked at each of those
52
1 class individually and determined that we could break
2 out the periphyton from the sawgrass/cattail sparse
3 class. And then we came up with one additional class,
4 periphyton in the southern region.
5 Q. Okay.
6 I believe you said you knew from aerial
7 photography that the sawgrass/cattail part showing up
8 in the southern end was actually periphyton.
9 A. Um-hum.
10 Q. That aerial photography you were reviewing
11 as part of this particular project?
12 A. Yes.
13 Q. Okay.
14 What aerial photography were you using?
15 A. We had an area flown approximately at the
16 same time by Southern Resource Mapping at 1:24,000
17 scale, color infrared.
18 Q. That was by who?
19 A. Southern Resource Mapping.
20 Q. Now you're talking about the area, you're
21 talking about all of WCA-2A?
22 A. That's correct.
23 Q. Do you recall approximately when they did
24 that?
25 A. I think it was -- I'd be guessing.
53
1 Q. Well --
2 A. It was within a year of this, the actual
3 data that we acquired, this satellite imagery.
4 Q. Within a year of August 10, '91?
5 A. Right.
6 Q. Okay.
7 Now did you retain the infrared
8 photography from that?
9 A. Yes.
10 Q. Okay.
11 A. Yes.
12 Q. Who is doing the vegetative identification
13 using infrared photography?
14 A. Both Les Vilchek and myself.
15 Q. Did you produce any of those infrared
16 photos?
17 A. No.
18 Q. Okay.
19 Do you still retain those?
20 A. Um-hum, yes.
21 Q. With regard to the southern area which the
22 computer was identifying as sawgrass/cattail sparse,
23 and that you had determined from aerial photography was
24 periphyton, were there any of the 129 sites located
25 within that area?
54
1 A. I don't recall. The sites were pretty
2 well distributed through out the whole area, I would
3 guess yes.
4 (Thereupon, Mr. Downing entered the room.)
5 MR. KOBELINSKI: Okay. Let's take a five
6 minute break as long as Ed just came in.
7 (Thereupon, a brief recess was taken.)
8 BY MR. KOBELINSKI:
9 Q. Let's see the step No. 6 that we were
10 discussing, or at least I had written down, is
11 supervised classification, which was done after you had
12 visited 129 sites for the unsupervised. And were there
13 any documents generated for this particular step of the
14 project?
15 A. Just what's available digitally, to the
16 best of my knowledge.
17 Q. Is that -- did you have hard copies that
18 you produced to us?
19 A. I don't believe so.
20 Q. Okay.
21 So the documents that you produced to us,
22 there are none related to this step?
23 A. The supervised?
24 Q. Yes.
25 A. Just the, you know, the ground truth
55
1 information and locations and what we saw there.
2 Q. Yeah, that's fine. We've already gone
3 through that?
4 A. Right.
5 Q. Okay. I appreciate that. I'm trying to
6 make some sense of what we have so we can sort of
7 follow along with what you did out there.
8 Now, you ended up with 19 and then you
9 added a 20th class of periphyton; is that correct?
10 A. That's correct.
11 Q. What was the next step?
12 A. There was another problem area, tree
13 islands.
14 Q. Um-hum.
15 A. They were being classified to something
16 other than tree islands. And basically we went and
17 digitized the tree islands on the supervised
18 classification, ran another unsupervised classification
19 just on tree islands, ten spectral classes, and then
20 collapsed that down into four already existing classes,
21 and mosaicked them back into the output of the
22 supervised classification
23 Q. Now, is there any documentation related to
24 that step?
25 A. That's probably -- I did see something,
56
1 ground truthing of the tree islands, the visits that we
2 made to those, stated tree islands.
3 Q. Why don't we quickly identify that?
4 (Thereupon, a discussion was held off the
5 record.)
6 BY MR. KOBELINSKI:
7 Q. It might be the bottom thing in that
8 entire stack. Take a look at the -- I have it as the
9 very bottom. Is that -- yeah, look at that one. I
10 don't know if that's it, but...
11 A. Yes, I believe this is it.
12 MR. KOBELINSKI: Why don't you mark that
13 as 3.
14 (Thereupon, the document was marked
15 Rutchey Exb. No. 3 for Identification.)
16 BY MR. RUTCHEY:
17 Q. All right.
18 Mr. Rutchey, drawing your attention to
19 what's been marked as Exhibit 3, and which record is a
20 document, bears Bates Nos. 1203073 through 1203105, and
21 the first page of which appears to be a cover, a file
22 cover, bearing the name WCA-2A tree islands and bears a
23 date of 2-20-92.
24 By the way, would February of 1992 be
25 about the time that you were doing this
57
1 reclassification of tree islands?
2 A. Yeah, this is the ground truthing of these
3 tree islands.
4 Q. Now, did this result in separate ground
5 truthing? In other words, did you go back out in the
6 field?
7 A. Yes.
8 Q. The third time you've gone out in the
9 field if I've counted correctly? Well, the third task.
10 First one was for, you referred to it as rectification,
11 that was your first field visit; is that right?
12 A. Field visits.
13 Q. Field visits, plural.
14 The second one was visits to the 129
15 sites, and this would be third set of visits you took
16 to the field?
17 A. That's correct.
18 Q. All right.
19 And there was -- Les Vilchek did this?
20 A. Both of us. Again, I'm the GPS and he's
21 the --
22 Q. All right. How many --
23 A. -- vegetation.
24 Q. How many sites did you visit?
25 A. I could count them, how many sheets are
58
1 here. Probably that's how many sites.
2 Q. All right.
3 A. It looks like 31.
4 Q. Okay.
5 Let me draw your attention back to
6 something -- or before I do that, looking at what is
7 the Bates page of this document, 1203074, what would
8 this be?
9 A. That's the location of the tree islands'
10 approximately center points.
11 Q. Okay.
12 Did you follow the same GPS technique of
13 taking a number of readings and then correcting it for
14 any distortion caused by the GPS system?
15 A. No, the tree islands are all fairly large
16 that we did digitized. And once you're navigating to
17 them, to your navigation points, you can see the trees,
18 so it wasn't necessary.
19 Q. Then drawing your attention then to the
20 field inspection sheets, which commences at 123075, in
21 the upper right-hand corner it says photo number R1-1.
22 Does that indicate that those photos were taken during
23 these visits?
24 A. I would say yes. Well, that's a good
25 point.
59
1 Q. I know, that's what I just noticed myself.
2 Going back then to Exhibit 2, I believe
3 start of the field sheets, it's at 1202944 --
4 A. Right.
5 Q. -- these likewise have photo numbers in
6 them. Does that refresh your recollection as to
7 whether or not photos were taken?
8 A. Obviously, yes, they were.
9 Q. All right.
10 Do you know whether or not those photos
11 are still retained?
12 A. I haven't looked at them if they are. I
13 haven't looked at them. I don't even recall them. But
14 if they are, I might have them.
15 Q. Okay.
16 Were they used for any purpose in
17 refreshing your recollection when making
18 identifications?
19 A. I don't think so. I think we used our
20 visual view of the area as we saw it in real time with
21 our eyes.
22 Q. Okay.
23 Did you actually help Les with these or,
24 again, was he -- these are all just his notes as to
25 what he identified at the tree islands? I'm referring
60
1 back to Composite Exhibit 3.
2 A. He took the notes, but we would sit there
3 and he would tell me what he thought he saw there. And
4 we would, you know, between the both of us, I would
5 say, we came up with it. But he was the primary
6 person. But he would say this is what I see. I would
7 say well, you know, maybe sawgrass was 60 percent, he'd
8 say 70 percent. But we were pretty close. But we
9 still kept it to one person's view. I just wanted to
10 make sure he wasn't totally out of line. It's good to
11 have a second opinion there when you -- he's doing
12 that.
13 Q. Now with regard to the tree islands, did
14 you attempt to identify one entire tree island or were
15 you likewise attempting to identify a 20 by 20 pixel or
16 area?
17 A. Basically we were looking at the entire
18 tree island.
19 Q. Okay.
20 So of the 31 sites, are these essentially
21 then 31 different tree islands?
22 A. Yes.
23 Q. Okay.
24 Any notation as to the approximate area
25 that was covered by the tree islands?
61
1 A. Repeat that again.
2 Q. Would you be able -- is there any notation
3 on these sheets which would tell me how large the tree
4 island was?
5 A. No.
6 Q. Okay. All right.
7 Based upon these field trips, what did you
8 then do, how did you use this information?
9 A. We used this ground truth information, as
10 I said before, we cut out the tree islands from the
11 supervised classifications, 31 of them, and did
12 unsupervised classifications on them and made ten
13 clusters. Then we determined from this ground truth
14 information what those 10 clusters represented. And
15 some of them were duplication, so we collapsed it down
16 to four, which were already four classes we had within
17 the original 19 classes of supervised classification,
18 and then collapsed it -- well, collapsed to those four,
19 and then mosaicked it back into the supervised
20 classification.
21 Q. All right.
22 Just so I understand what you were doing
23 before, let me see if I understand what you're doing
24 here. A tree island in this case could actually be
25 comprised of approximately how many in this case, a
62
1 pixel being a 20 by 20 meter square?
2 A. Hundreds, if not thousands.
3 Q. Okay.
4 When you went out there to a particular
5 tree island, were you attempting to identify all of the
6 various pixel types that were out there or were you
7 doing a general vegetation type for the entire tree
8 island?
9 A. Basically, no, we were trying to look at
10 from the point of -- I have to look here.
11 Q. You want to use your original, it has the
12 nice color photo?
13 A. Yeah.
14 Basically we were looking at tree islands
15 from the point of a sawgrass/brush mixture 1, which is
16 what we categorize it as, sawgrass/brush mixture 2, and
17 these are all delineated within a table by -- what we
18 mean by these categorizations, tree island and other
19 keys. It might have been just been sawgrass, in some
20 cases cattail.
21 Q. Okay. All right.
22 What was the next step?
23 A. Okay.
24 Well, then we started mapping our accuracy
25 assessment of the overall map product.
63
1 Q. Once you did the tree islands, did you
2 then have what you believed to be a final or a pretty
3 much a final vegetation map?
4 A. That's correct.
5 Q. Okay.
6 And is there -- is that the same one
7 that's attached to Exhibit 1?
8 A. That's correct.
9 Q. Okay.
10 And that would be Bates -- would that be
11 Bates No. 1220435?
12 A. That's correct.
13 Q. Okay. All right.
14 Go ahead, what was the next step then?
15 A. Map accuracy assessment. We generated 241
16 random points from the final supervised classification.
17 We used a stratified random sampling technique based on
18 class. And the number of points selected was based on
19 the required minimum of 204 for an 85 percent map
20 accuracy with an error of plus or minus five percent.
21 And this was based on the binomial formulas.
22 Q. Now 241, would that -- were you
23 essentially trying to get X number of samples for each
24 of the particular 20 classs?
25 A. Well, the stratified random sampling by
64
1 class implies that you're going to take samples within
2 each of the classes.
3 Q. Is that -- does it by necessity result in
4 a minimum number taken within each particular class?
5 A. No.
6 Q. Okay.
7 And is there data showing approximately
8 how many were taken for each class?
9 A. Yes.
10 Q. Okay.
11 Where would that be?
12 A. You might have it in hard copy. It's
13 definitely in digital form.
14 Q. Okay. All right.
15 Once you determines these 241 points,
16 actually I believe then the computer then generates
17 these 241 points; is that correct?
18 A. That's correct.
19 Q. Is there a map showing the 241 points?
20 A. Yes, a map was generated digitally.
21 Q. You get the points out of that digital
22 map?
23 A. Yes.
24 Q. Did you produce that one?
25 A. It might be here, I'm not sure.
65
1 Q. Okay.
2 A. I'm not sure if it's in my files.
3 Q. All right.
4 You had 241 points. What did you do at
5 that point?
6 A. Basically you visit all these points out
7 in the field.
8 Q. Okay.
9 A. Ground truth them and see if they are what
10 you said they were in the map.
11 Q. All right.
12 Were you able to visit all 241 points?
13 A. I believe so.
14 Q. Okay.
15 A. If I remember correctly --
16 Q. I'm sorry?
17 A. In acquiring 241 points we were again
18 constrained when the computer said this is where it
19 will be, that it had to be within a minimum of a 3 by 3
20 pixel area. I mean if it fell on one pixel and it was
21 surrounded by other things, well then, we weren't
22 obviously going to find that area, so...
23 Q. With regard to -- is that an instruction
24 that you gave the computer to pick 241 sites, was to
25 make sure it was picking a 3 by 3 or a 9 pixel area?
66
1 A. (No response.)
2 Q. Do you understand my question?
3 A. Right. I understand.
4 I'm not sure how we actually went about
5 doing that. I can't remember if there was a way we
6 could do that on the computer or whether we just picked
7 the closest point that was -- met the minimum 3 by 3
8 requirement.
9 Q. Ultimately you ended up with 241 3 by 3
10 sites?
11 A. Or larger.
12 Q. Or larger.
13 A. Right.
14 Q. You visited these?
15 A. Right.
16 Q. Okay.
17 Do you have documents related to those
18 visits?
19 A. I believe so.
20 Q. Similar to what we've already seen?
21 A. I think so.
22 Q. Okay.
23 Why don't we flip this one over so we
24 don't get confused. That's the tree island one. And
25 let's see if we can -- you can just keep them there,
67
1 that's your set.
2 A. Okay.
3 Q. That too. You don't need to go through
4 that, that's the other. It will be probably be here or
5 in here.
6 A. I believe this is it.
7 Q. All right.
8 And you're -- could you read off the Bates
9 numbers? These are consecutive. Just take, if you
10 just look at the top page there.
11 A. 1202231.
12 Q. Through? The final page --
13 A. Yeah.
14 Q. -- of that entire.
15 A. Well, I'm make making sure all this is.
16 Q. Okay.
17 Well, I'll take you through each and every
18 one, trust me.
19 A. 1202541.
20 Q. Okay. 2541. Okay good.
21 MR. KOBELINSKI: Mark this as Composite 4.
22 (Thereupon, the document was marked
23 Rutchey Exb. No. 4 for Identification.)
24 BY MR. KOBELINSKI:
25 Q. Showing you then, Mr. Rutchey what's been
68
1 marked as Composite Exhibit 4, which bears the Bates
2 number you just read off into the record, the first
3 page of which appears to be a file cover reading WCA-2A
4 Accuracy Assessment, 3-23 through 4-24-92, is that the
5 period of time that you did the site visits for the
6 accuracy assessment?
7 A. Yes.
8 Q. Okay.
9 Now I want to go through this, rather
10 quickly identify what we have inside here. The first
11 document in there which bears Bates No. 120232 through
12 36, is this a summary of the site visits to all 241
13 sites?
14 A. Yes.
15 Q. Okay.
16 And so if I understand, isn't that, being
17 similar to a document previously identified in regard
18 to the 129 site visits, this is a summary based upon
19 the field notes and the visitations for the entire
20 field study?
21 A. Yes.
22 Q. Okay.
23 And what would be 1202237 through 40?
24 A. Well, the first part is the class number
25 that was assigned to each of the 20 classes.
69
1 Q. Okay.
2 Now with regard to points, would that be
3 the 240 some-odd sites?
4 A. Well, it's 210 of them.
5 Q. Okay.
6 Were there only 210 sites then that were
7 visited?
8 A. No, there was 241, but it was collected,
9 again, two different times. There was a second set
10 that was added on, 31 sites.
11 Q. Okay.
12 So those cover 210 sites. I can see that,
13 going along the columns here on this first page,
14 1202237, you have the points, you have the GPS
15 coordinates; is that correct?
16 A. (Shakes head up and down.)
17 Q. What would the, I guess it would be the
18 fourth column, which starts at the top of 10, 1, 2, 4,
19 1, 1 and et cetera? Is that class identification?
20 A. Yeah, but from looking at this I can't
21 tell whether that's what we saw in the field or what
22 was actually in the map, the final map that we
23 produced.
24 Q. Okay.
25 What is the final column when it says
70
1 negative 1 everywhere?
2 A. It looks like a bug.
3 Q. A bug, all right.
4 When you went out to the field to make
5 these classifications of 241 sites, were you aware
6 already what the computer was identifying each site as?
7 A. Les wasn't. I wasn't either, no.
8 Q. Okay.
9 The next document is 1202241 through 45.
10 It appears similar to what we just looked at, however
11 there's a fair amount of handwriting here. And drawing
12 your attention to that final, to the top of that where
13 the first page says final-map-CAT --
14 A. Um-hum.
15 Q. -- what are the numbers in that left-hand
16 first column, first one? It looks like possibly 19.41.
17 It's difficult to read. The next one is clearer,
18 18.34, 7.76 and down the line. Do you know what those
19 are?
20 A. I'd be guessing, but I think it's
21 percentage.
22 Q. Percentage of the entire area?
23 A. Percentage of each of these classes.
24 Q. All right. I'm afraid you lost me on
25 percentage.
71
1 Are you talking about what percentage of
2 WCA-2A is made up of those classes or what
3 percentage -- why don't you answer that.
4 A. These numbers in the first column --
5 Q. Um-hum.
6 A. -- okay, I believe they represent the
7 percentage of that class within Area 2. But again I'm
8 not sure. This is something I wrote on this paper two
9 years ago.
10 Q. All right.
11 Going along you have then the next column
12 or the, just class numbers 1, 2 through 20, and the
13 following column is a column of handwritten numbers,
14 41, 39, 9 16. Do you recall what those numbers are?
15 A. No.
16 Q. All right. Let's move along.
17 Drawing your attention then to a large
18 piece of map, which bears Bates No. 1202246, do you
19 recall what this is?
20 A. I believe this is our path that we
21 followed to look at the map accuracy assessment points.
22 Q. So this essentially has all 241 points on
23 here?
24 A. I think it's got the first 210.
25 Q. Okay.
72
1 Did you -- were 31 points added on after
2 the initial 210?
3 A. Yes.
4 Q. Why was that?
5 A. One of the things we wanted to know with
6 confidence is the cattail, and the stratified random
7 sampling didn't come up with enough points on that
8 cattail area. So we wanted to get more points to
9 verify whether the map was accurate in respect to
10 cattail. So we did a stratified -- another stratified
11 random sampling, checking 31 points within the cattail
12 zone. So an additional 31 sites were visited.
13 Q. All right.
14 When you say within the cattail zone,
15 which of the 20 classes were used for that stratified
16 selection?
17 A. (No response.)
18 Q. If you go back one, you can see the 20
19 classes, if that helps you.
20 A. Without actually seeing the digital data,
21 it's hard to be definite, but I would say it was
22 cattail dense, moderate, sparse cattail/brush mixture,
23 and I believe those are the only ones.
24 Q. Okay.
25 Do you recall whether sawgrass/cattail
73
1 mixture was reviewed, or part of that 31?
2 A. Oh, excuse me, yes, I would say those two.
3 I'm not sure.
4 Q. Would there be some documents in here that
5 we could look at to confirm that one way or the other?
6 A. It's all digital as far as I know.
7 Q. Okay.
8 The next set within this composite exhibit
9 1202247 through 253, what are these? What are these
10 documents?
11 A. This is ground truth data and the first
12 page is GPS output, PFINDER output.
13 Q. Okay.
14 And this is with regard to just these
15 seven sites?
16 A. That's correct.
17 Q. Okay.
18 Did you have any problems with the GPS
19 system during the inspection at 241 sites?
20 A. Other than the one I previously told you
21 about, as far as the navigation aspect.
22 Q. Okay.
23 And is there a particular name that's
24 referred to as, I think there is, where the data is
25 that it's not telling you the accurate location, what
74
1 is that?
2 A. Well, you either have raw GPS data or you
3 have differentially corrected GPS data. The raw can be
4 off by as much as a hundred meters. Differentially
5 corrected, we found we could get it down to three to
6 seven meters.
7 Q. And did you have that type problem when
8 you were doing the site visitation at the 241 sites?
9 A. As far as the inaccuracy?
10 Q. Right.
11 A. Yes.
12 Q. Okay.
13 Now then, going on to the next document
14 which appears as Bates No. 1202254 through 2275, what
15 are these documents?
16 A. Again, it's GPS PFINDER output data along
17 with ground truth.
18 Q. Are these grouped into any particular
19 manner? For instance, are these grouped into classes?
20 A. It's probably grouped based on that day's
21 work that we went out, went out in the field.
22 Q. Oh, all right. Okay.
23 And if you would just then go through the
24 remainder of this exhibit and confirm that what we have
25 here then are similar GPS data sheets along with field
75
1 inspection reports, and that would be up through Bates
2 No. 1202500.
3 A. (Indicating.)
4 Q. Would -- through that Page 500, is that
5 the initial 210 sites?
6 A. Yes.
7 Q. Okay.
8 And drawing your attention to Bates No.
9 1202501 through 541, what would these be?
10 A. This has the additional -- let me see. It
11 looks like 32 sites.
12 Q. Okay.
13 A. That we visited.
14 Q. All right.
15 Now, looking at these documents here in
16 regard to the additional sites for the cattail accuracy
17 assessment --
18 A. Um-hum.
19 Q. -- does this help you determine what
20 classes were used or included in this?
21 A. No.
22 Q. Okay.
23 Nothing in here would identify that?
24 A. You're asking -- well, could you repeat
25 the question?
76
1 Q. My question is, going back to -- on the
2 additional 31 or 32 sites for cattail accuracy, I'm
3 trying to see which of the 20 classifications were used
4 for that --
5 A. All right.
6 Q. -- random selection of 31 or 32 sites.
7 A. Again, I think it's -- without looking at
8 the digital data, I believe it was cattail dense, not
9 moderate, sparse and cattail brush mixture.
10 Q. Okay.
11 A. And you added sawgrass/cattail mix, so I'm
12 not sure about that.
13 Q. I didn't mean to add it. I was just
14 asking whether or not it was part of it. To the best
15 of your recollection it was Classes 10, 11, 12 and 13?
16 A. Right.
17 Q. What about 15, which is brush/cattail
18 mixture?
19 A. I don't think so so. It started with 10.
20 Q. And would your digital file show which
21 classes were identified or used?
22 A. Yes.
23 Q. Was that 31 or 32 additional sites that
24 were added on?
25 A. It looks like 32.
77
1 Q. Would it make a total of -- the total
2 would be 242 points then?
3 A. No, it's 241. If we went back and
4 counted, I think it was a total of 241 altogether.
5 Q. All right. All right.
6 What was the next step then?
7 A. Okay.
8 So we did the map accuracy assessment.
9 Q. Um-hum.
10 A. And we came up with the results from that.
11 And we -- at that point we wanted to see if we can
12 better those results by collapsing the -- within
13 species density characteristics that we were trying to
14 break out using the 20 class map. For instance
15 sawgrass, we had a dense, moderate to sparse. Well,
16 maybe the satellite imagery was having a hard time
17 breaking out with a species with that fine line between
18 graduations of density. So we combined all the
19 sawgrass and all the cattail into one class and came up
20 with a 12 class map.
21 Q. Okay.
22 A. Final map.
23 Q. Would there be a particular document
24 related to that step?
25 A. Yes. It's part of this paper. It's all
78
1 explained.
2 Q. Other than -- well, I understand the
3 explanation of the paper. Are any specific documents
4 going through that work step that you've produced?
5 A. Not to my knowledge. Digitally there is.
6 Q. All right.
7 And any additional steps after that?
8 A. We then did a map accuracy assessment on
9 that final 12 class map using the same data set that we
10 used on the 20 class map and came up with the overall
11 map accuracy.
12 Q. Okay.
13 Now as I understand it, you had -- then
14 you had the computer rerun the map using 12
15 classifications; is that correct?
16 A. We collapsed the 20 class map to a 12
17 class map, yeah.
18 Q. The computer would generate a new figure
19 showing just the 12 classes?
20 A. That's correct.
21 Q. Now, again, if I understand your testimony
22 that you did not make any additional field visits, then
23 you just took the data that was from the 241 sites
24 visited and used that for accuracy of the final map
25 with the 12 classifications?
79
1 A. That's correct.
2 Q. All right.
3 Any additional steps?
4 A. That about sums it up.
5 (Thereupon, a discussion was held off the
6 record.)
7 BY MR. KOBELINSKI:
8 Q. Mr. Rutchey, I'm going to hand you a
9 number of documents, see if you could identify what the
10 documents relate to.
11 MR. FITZGERALD: Give him a hint, these
12 are his documents?
13 MR. KOBELINSKI: We'll use that
14 definition.
15 BY MR. KOBELINSKI:
16 Q. Yeah, these are all from documents that
17 you've been looking at today in response to a duces
18 tecum that you've produced to us today in response to
19 the duces tecum; first of which is a laboratory
20 notebook, 1202643 through 688.
21 Do you know what that lab book relates to?
22 A. Yeah. This is the actual at the site that
23 we visited, it has in the left-hand margin the site
24 visited and what actually the pixels that were within
25 the imagery, 3 by 3 matrix of every one of them.
80
1 Q. This relates then to the paper we were
2 just discussing?
3 A. Yes.
4 Q. Okay.
5 MR. KOBELINSKI: Why don't you go ahead
6 and mark that.
7 (Thereupon, the document was marked
8 Rutchey Exb. No. 5 for Identification.)
9 MR. KOBELINSKI: I will identify this
10 better for the record. What we've marked as
11 Exhibit 5 -- is that correct? -- is a
12 laboratory notebook. It bears Bates Nos.
13 1202643 through 688.
14 BY MR. KOBELINSKI
15 Q. Is that your notebook, sir?
16 A. Yes.
17 Q. All right.
18 Why don't we use just the first -- second
19 page of it which is 1202644 and it says Class 1. What
20 exactly does this reflect?
21 A. It almost jiggles my memory about old and
22 new, but I'll try to guess.
23 Um, basically this is the -- what was,
24 let's say, Site No. 4, for the ground truthing, what
25 the pixel data was, such as was it all -- in this case
81
1 it looks like on No. 4 that it was all one, which would
2 have been, it looks like sawgrass dense. So we had a
3 whole 3 by 3 dense homogeneous patch of dense sawgrass.
4 Q. Well, let me break this apart a bit, if I
5 can. On Page 1202644 in the left-hand margin outside
6 the grid there are two numbers written there, 4 and 5.
7 Do you see that?
8 A. Yeah.
9 Q. Is that site Site No. 4, is that what that
10 reflects?
11 A. Yes.
12 Q. All right.
13 Now, following across this grid of
14 approximately 30 or 40 squares across, this is a grid?
15 A. Yes.
16 Q. And within that, just following the row
17 next to 4, interspersed in there are, in six places,
18 are the number 1 space, space, 1, space, space, 1, et
19 cetera. What exactly does that mean? Is there a
20 reason you have 6 columns there? How would I -- how do
21 you interpret this?
22 A. Okay.
23 I separated them out. If you look -- I'm
24 going further on so you can understand better.
25 Q. Okay.
82
1 A. If that's --
2 Q. Okay.
3 Just give a Bates page and I'll turn
4 there.
5 A. 1202679.
6 Q. Done.
7 A. Okay.
8 Let's go -- let's go like 131 there in
9 upper right-hand site. Hold on one moment.
10 (Thereupon, a discussion was held off the
11 record.)
12 BY MR. KOBELINSKI:
13 Q. Okay.
14 131 appears to be on the upper right-hand
15 corner.
16 A. Okay.
17 That Site 131, the computer is telling us
18 within that location of Site 131, of the accuracy of
19 the site location, that every pixel showed to be 20,
20 which this case was periphyton.
21 Q. Okay.
22 A. If you look at the column over there we
23 say yes, it's 20, what we determined it to be. We went
24 out in the field and they came up to be 6, so this one
25 was wrong.
83
1 Q. Okay.
2 So this actually classification No. 6?
3 A. That's correct.
4 Q. This is out in the field, okay. So
5 reading next to 131 I have a box with nine 20's in it.
6 A. Right.
7 Q. On those 9 pixels in the computer I would
8 find them all as Class No. 20 --
9 A. Right.
10 Q. -- and that's why you have 20 written down
11 then in the third column. Over where 6 is is what
12 actually the field notes show?
13 A. That's correct.
14 Q. All right.
15 Now, just out curiosity, would that mean
16 to you that, according to your field observations, all
17 nine of those were Category 6 or is that the
18 predominant category within the nine?
19 A. What it means to me is that our computer
20 said that that 9 by 9 matrix was periphyton and our
21 field reconnaissance showed otherwise, in that it was
22 counted as wrong.
23 Q. Okay.
24 How would you go about determining whether
25 your field observations showed all 9 pixels to be
84
1 Category 6?
2 A. You go out in the field and you ground
3 truth it and you write, like on field sheets, the
4 notes. And if it's not periphyton it's something other
5 then.
6 Q. Perhaps I'm not explaining myself.
7 Is that final column, where you have
8 written 6, does that indicate that that -- that all
9 nine pixels would be Category 6, or would 6 be
10 predominant of these nine, for instance, five out of
11 the nine were 6, would 6 be written there or would 6
12 and 4 or whatever classification?
13 A. I believe it represents predominant
14 vegetation that was at that site.
15 Q. Okay. All right.
16 This is from the whole 141 site field
17 trips when you're doing the accuracy assessment?
18 A. Yes.
19 Q. Okay.
20 Is this something that was out in the
21 field?
22 A. No.
23 Q. Okay.
24 So this was developed after you took all
25 the field notes and did this back in the office?
85
1 A. Yeah. The only thing that pertained to
2 the field notes was, there are like four columns in
3 each one of these sets. The initial site number, the
4 pixel representation and in the computer what we
5 determined that the overriding majority was from that
6 pixel representation from that, up to that point that
7 was all computer-generated. The last column was the
8 only one that was derived in the field.
9 Q. I understand, okay. All right. This was
10 5.
11 MR. CESARANO: You want to put a rubber
12 band around Number 4, however many there were?
13 MR. KOBELINSKI: Oh, yeah, let's do that.
14 BY MR. KOBELINSKI:
15 Q. All right.
16 I show you a group of documents which bear
17 Bates No. 1202689 all the way through 1202772, and
18 actually that first 2689 appears to be a file folder
19 cover stating georectification points, areas 3N, 3S and
20 Holeyland. Could you just -- could we go through that
21 and tell me those sheets relate to?
22 A. This is a rectification that we did on
23 some satellite imagery of the area, what I call 3
24 North, everything north of Alligator Alley. Then 3
25 South, everything south, and Holeyland.
86
1 Q. Okay.
2 Now this has not been related to the 2A
3 study?
4 A. No.
5 Q. Okay.
6 When was this 3A study done?
7 A. I have it, it looks -- it appears this
8 initial data set was collected in May -- no, June,
9 1992.
10 Q. Okay.
11 Now, is this the rectification stage then
12 of the project, essentially using satellite imagery for
13 vegetative mapping of 3A?
14 A. Yes.
15 Q. All right.
16 When did you start that project?
17 A. Same dates, May to June, 1992.
18 Q. That using SPOT imagery?
19 A. This, yeah, for rectification, yes.
20 Q. Okay.
21 Are you still -- is an that ongoing
22 project right now?
23 A. Yes.
24 Q. Okay.
25 What stage are you at in that project?
87
1 A. Basically we have the satellite imagery
2 rectified and we have some aerial photography that was
3 acquired from National Park Service cooperative
4 agreement that we did with them in acquiring NASA
5 photography.
6 Q. What is the date of the aerial
7 photography?
8 A. November, December, 1992.
9 Q. And that's infrared?
10 A. That's correct.
11 Q. Are you following essentially the same, I
12 have them as eight steps, but essentially the same
13 steps that you used for 2A for this project?
14 A. No.
15 Q. How are they different?
16 A. Using -- I'm using straight
17 photointerpretation of aerial photography as opposed to
18 a digital process.
19 Q. Okay.
20 Did this start with a SPOT imagery digital
21 information?
22 A. Yes.
23 Q. Okay.
24 Did there come a time you abandoned using
25 that?
88
1 A. Yes.
2 Q. Are you still going to use it?
3 A. Yes.
4 Q. All right.
5 I misunderstood your response then. How
6 does this differ from the 2A study?
7 A. I'm going to be using more conventional
8 aerial photointerpretation methods as opposed to
9 digital.
10 Q. Okay.
11 With regard to the digital, did you
12 complete the rectification of the digital information?
13 A. Yes.
14 Q. Okay.
15 Did you do an unsupervised classification?
16 A. Preliminary, yes.
17 Q. With approximately how many classs?
18 A. I tried a number of different classes.
19 Q. Approximately how many?
20 A. 25 to 150.
21 Q. All right.
22 The data that you have there, does that
23 just deal with the rectification set that you have
24 there, the sheet that I put in front of you?
25 A. Yes.
†††′㈠††††††††⁔桥慴愠瑨慴⁹潵慶攠瑨敲攬潥猠瑨慴ഊഊ††††㈳††畳琠摥慬⁷楴栠瑨攠牥捴楦楣慴楯渠獥琠瑨慴⁹潵慶攍ਠ†††′㐠††瑨敲攬⁴桥桥整⁴桡琠䤠灵琠楮牯湴映祯甿ഊഊ††††㈵††††††䄮††奥献ഊ††††††††††††††††††††††††††††††††††㠹ഊഊ††††‱††††††儮††佫慹
89
1 Q. Okay.
2 How many rectification points did you use?
3 A. I don't recall. It's more than what was
4 used for Area 2A because it's a larger area.
5 Q. Okay.
6 Have you done any site visits to verify
7 the unsupervised the classifications?
8 A. Yes.
9 Q. Approximately how many?
10 A. I think maybe one or two field trips.
11 Q. Which have covered how many sites?
12 A. I don't recall. I'd be guessing. Maybe
13 50.
14 Q. Why don't we close that up?
15 And did you produce documents similar to
16 the field inspection sheets that we have seen with
17 regard to those, the 3A field inspection for the -- of
18 the unsupervised classes?
19 A. I don't believe any were generated.
20 Q. Okay.
21 Did you generate any field notes of what
22 you saw out there?
23 A. Yeah, they're probably around somewhere.
24 They might be in these piles or ...
25 Q. Why don't we find them. Why don't you
90
1 take a quick look?
2 A. I don't see it there.
3 Q. All right.
4 The other place it could possibly be is --
5 well that was -- you did go through that. That's
6 everything.
7 Do you believe you still have those?
8 A. They might have been in my folders or Les
9 Vilcheck might have them. It probably is still at the
10 raw data stage. I don't even think it's been
11 transcribed to these summary sheets that you --
12 Q. The raw data would be some sort of a field
13 notebook with notes on it?
14 A. Yeah, or just a pad similar to what you
15 have there.
16 Q. And Les Vilchek is assisting you with this
17 project?
18 A. Yes.
19 Q. Okay.
20 Were you taking photos at the field sites,
21 photos as you did in 2A?
22 A. I don't believe so.
23 Q. Would your field notes reflect that?
24 A. Yeah. I don't think we did though.
25 Q. Do you recall approximately how many sites
91
1 you visited?
2 I know I asked that, but I don't recall
3 your answer. It was two dates I believe you said?
4 A. It would be a guess, 50.
5 Q. And when did you do that?
6 A. Oh, probably not too shortly after we
7 acquired that -- that data set.
8 Q. So probably late '92 or early '93?
9 A. Probably '92.
10 Q. Okay.
11 After that field trip, did you then
12 perform a supervised classification?
13 A. No.
14 Q. All right. I'm getting a little too
15 disorganized, because I've jumped in the middle of it.
16 When did you start the 3A vegetative
17 mapping project?
18 A. Initially I think we made the first
19 efforts when we acquired the first SPOT satellite
20 imagery which was in May of '92.
21 Q. Okay.
22 And is anyone else working on this project
23 other than Les Vilcheck and yourself?
24 A. No.
25 Q. Okay.
92
1 After obtaining a SPOT imagery, I
2 understand you did a georectification of the map?
3 A That's correct.
4 Q. What was your next step?
5 A. We just basically analyzed the data, some
6 very preliminary analysis, in looking at it digitally
7 to see if we could get a feel for patterns in Area 3.
8 And we groundtruthed some of these pattern areas, some
9 very preliminary stuff.
10 Q. Is this different than the two day field
11 trip we were just discussing?
12 A. I think so.
13 Q. Okay. All right.
14 The field notes from that?
15 A. We didn't find them. I think they're
16 available somewhere. Les probably still has them
17 somewhere.
18 Q. You just looked through everything we had
19 here. Do you recall seeing them?
20 A. No.
21 Q. All right.
22 After that field trip, did that result in
23 any type of modification to your map? Or was it just,
24 as I understand it, just a beginning to understanding
25 the vegetative communities at this time that were out
93
1 there?
2 A. Which map?
3 Q. You have a digitized SPOT imagery map of
4 3A; is that correct?
5 A. Correct.
6 Q. You georectified it?
7 A. Okay.
8 Q. Then I understand you analyzed the data
9 and, as you put it, made some site visits to look at
10 vegetative communities?
11 A. Right.
12 Q. As a result of these site visits, did you
13 change the map at all or start changing
14 classifications?
15 A. No.
16 Q. It was just to familiarize yourself with
17 the area?
18 A. Yes.
19 Q. I gather from what you said, Area 2A
20 you've done a fair amount work in, is that right?
21 A. Yes.
22 Q. And 3A you've not?
23 A. That's correct.
24 Q. Okay.
25 After making this field trip with Les to
94
1 familiarize yourself with the communities, what was
2 your next step?
3 A. Basically we haven't done too much since
4 then.
5 Q. Okay.
6 I believe you did say you did an
7 unsupervised classification of several of them?
8 A. That was part of the preliminary --
9 Q. Oh, all right.
10 A. -- just looking at the data and just
11 generally looking at it and trying to figure out land
12 patterns.
13 Q. Okay.
14 You also said you made a second field
15 trip, the two day field trip to points selected?
16 A. Yeah, one, two days.
17 Q. Okay.
18 Is it -- just so I understand it, that
19 makes two field trips you made?
20 A. No, it's -- I mean this is going back a
21 while, but either one or two days we went out in the
22 field after we acquired the data and just have a
23 preliminary looking at data, trying to come up with
24 some generalized land patterns.
25 Q. Okay. I'm sure you're explaining it
95
1 perfectly. I just want to make sure I understand.
2 In my notes I had thought you had told me
3 you actually did go to a particular site to do some
4 site identification, that was approximately a two day
5 trip where you have field notes, but you have not
6 written up the inspection field inspection sheets for.
7 A. We did some very preliminary unsupervised
8 classifications. We had patterns and we visited them
9 and just tried to look at the general land patterns in
10 the area.
11 Q. Okay.
12 A. That's as far as we got.
13 Q. I understand that.
14 Was this preliminary unsupervised site
15 inspection that you were doing one or two days, is that
16 the second field trip with regard to 3A?
17 A. The first one was for the actual
18 rectification.
19 Q. All right. What was the second one?
20 A. That's basically those one or two days was
21 the second.
22 Q. Just two trips then?
23 A. Two trips.
24 Q. I misunderstood you. Thank you.
25 And you finished that in late '92?
96
1 A. Approximately '92 sometime.
2 Q. Okay.
3 Have you done any additional work on the
4 vegetative mapping of 3A?
5 A. No.
6 Q. Okay.
7 Is there a reason it --
8 A. Just working on a lot of different things
9 and other projects and it's taken me away from it at
10 this point.
11 Q. Is that an ongoing project, the mapping of
12 3A?
13 A. Yeah, I would say so.
14 Q. Are you going to attempt to get new SPOT
15 imagery or never mind?
16 A. No.
17 Q. Okay.
18 Do you intend to use the May '92 SPOT
19 imagery?
20 A. Yes.
21 Q. Okay.
22 Since approximately a year and a half has
23 passed since the collection of the May SPOT imagery,
24 how do you intend to ground truth it?
25 A. I don't. I plan to use that rectification
97
1 process to rectify a more current scene that's already
2 been acquired.
3 Q. What scene is that?
4 A. Basically it's late 1993.
5 Q. Is that SPOT imagery?
6 A. Yes.
7 Q. Do you recall what month of '93?
8 A. No, we -- no.
9 Q. All right.
10 Are you again then going to be using the
11 '93 SPOT imagery for the purpose of doing vegetative
12 mapping of 3A?
13 A. As part of the process, yes.
14 Q. Okay.
15 Will you also be using infrared as you did
16 with the aerial photography done with the National Park
17 Service?
18 A. That is a possibility, yes.
19 Q. Do you intend to use any infrared
20 photography?
21 A. Maybe.
22 Q. Do you already have it?
23 A. No.
24 Q. Okay. Are there plans to have it flown?
25 A. It's possible.
98
1 Q. Is it in the works?
2 A. No.
3 Q. Let me see what else I have here. Let me
4 see if we can identify it.
5 (Thereupon, a discussion was held off the
6 record.)
7 BY MR. KOBELINSKI:
8 Q. I'm showing you, sir, what has been marked
9 as Bates No. 1220375 through 1220400. Again this is a
10 document that you produced to us in response to your
11 notice of your deposition, the top of which appears to
12 be a file folder marked M. Koch.
13 Can you identify that for me, what that
14 relates to?
15 A. This is some work that Marguerite
16 initiated in a project that I was working on with her,
17 basically looking at transects in Area 2A. And she was
18 taking the lead on this. Basically we were looking at
19 the transition zone going from the impacted area to the
20 nonimpacted area of Water Conservation Area 2A.
21 Basically we were looking at hydrology, soils and
22 vegetation.
23 Q. Okay.
24 When was that project being conducted?
25 A. This was pretty much right before she left
99
1 the District, and I don't remember the exact data she
2 left, but this is what I had done up to that point when
3 she left. I haven't done much with it since.
4 Q. Okay.
5 Is that an ongoing project?
6 A. Not that I know of.
7 Q. Did you come to any conclusions from that
8 project?
9 A. No.
10 Q. Okay.
11 To your knowledge, is anyone still working
12 on that project?
13 A. Not this specific one, no.
14 Q. Okay.
15 Is anyone working on a similar project
16 that you're aware of?
17 A. Yes.
18 Q. Who is that?
19 A. I'll not sure of all the players. There's
20 a number of people in my division working on looking at
21 transects and gradient changes within the transects
22 within Water Conservation Area 2A.
23 Q. Do you know who is the head of that
24 project?
25 A. Currently I believe it's Fred Sklar.
100
1 Q. Fred, could you spell that name?
2 A. S K L A R.
3 Q. Okay.
4 And what do you understand the project is
5 doing?
6 A. Looking at -- well, collecting and
7 analyzing data along these transects within Water
8 Conservation Area 2A.
9 Q. Are these transects south of the 10
10 structures?
11 A. That's correct.
12 Q. And what type of data are they collecting?
13 A. All types of water chemical constituents,
14 soil data, algae, water depth, these areas are the
15 major components.
16 Q. Are they also doing vegetative
17 identification?
18 A. I believe at the points, the sites, yes.
19 Q. Do you know if they're using your
20 vegetative map that we've just been discussing, the 2A
21 vegetative map?
22 A. Yes.
23 Q. And do you know what the purpose is?
24 A. I think its final purpose is to determine
25 why the area is becoming impacted due to the inflow
101
1 through the S-10 structures.
2 Q. Well, is it attempting to analyze the
3 causes for vegetative change?
4 A. Causes for vegetative change is one
5 component.
6 Q. What are the other components?
7 A. Water quality, soil phosphorus or soil
8 chemical constituents, change that occurs with
9 vegetation, things like that.
10 Q. Is there anything else?
11 A. There's probably other things, but that's
12 the ones I know about.
13 Q. Who is working with Fred Sklar on that
14 that you're aware of?
15 A. Peter Rawlik. We just hired a new woman,
16 I'm not each sure of her name at this point.
17 Q. You know her first name?
18 A. Miao, I think that's how you say it. Jim
19 Grimshaw, Paul McCormick, Jim Laing. I think that's
20 the major people. There could be others that I'm not
21 sure what their role in it.
22 Q. Peter Rawlik, how do you spell that name?
23 A. R A W L I K.
24 Q. All right. Let's see what's left.
25 Let me show you again a number of
102
1 documents that were received. Now they were banded
2 together, I don't know whether they're related to each
3 other, Bates No. 1220096 through 1220189. First page
4 is all handwritten. I believe it says on the bottom
5 WCA-2 period DAT, N R O W, N C O L S. Could you
6 identify that?
7 A. This is the data from the vegetative map
8 that Les Vilchek and I produced that shows various
9 aggregations in the number and rows of columns that
10 resulted from those aggregations.
11 Q. Now is this the 2A study we were just
12 discussing?
13 A. No.
14 Q. All right.
15 What area is this study?
16 A. This was within Water Conservation Area
17 2A.
18 Q. Okay.
19 But this is a different project?
20 A. Yes.
21 Q. What project is this?
22 A. This is the paper I'm working on with
23 Jayantha Obeysekera, looking at scale issues as they
24 pertain to the Everglades.
25 Q. Could you just review what's in that and
103
1 see if that all relates to the same project?
2 A. No, it doesn't. It's not all pertaining
3 to this project.
4 Q. Okay. Hold on one moment.
5 Before I have you go through the documents
6 on this project with Jayantha Obeysekera, you're saying
7 it deals with scale issues with regards to the
8 Everglades. What exactly did that mean?
9 A. There's a number of these in existence or
10 being developed for the Everglades. And one of the
11 considerations of any model development is that you
12 consider the scale of what you're modeling or when
13 changes occur or what scale, and that's what this paper
14 is about. We're trying to figure out, at least for one
15 component or stated variable of a model, in this
16 particular instance it's vegetation, at what scale the
17 model should be looking at or addressing.
18 Q. All right.
19 And have you determined a proper scale to
20 look at for vegetation change?
21 A. It's very preliminary at this point.
22 We're in the data analysis.
23 Q. Okay. When did you start that project?
24 A. I'd say it was eight months, ten months
25 ago.
104
1 Q. Okay.
2 And how are you doing it, what's your
3 methodology?
4 A. Basically we have the original -- well,
5 basically we used the 12 class map output as opposed to
6 the 20 class map output. We aggregated it at a number
7 of different scales, aggregations going anywhere from
8 20 to a thousand and increasing by basically 40,
9 except -- with the exception of going from 20 to 40.
10 So each pixel is going to represent -- we have 20
11 meters, 80 meters, so on and so on, and we're looking
12 at those final map outputs using a program developed by
13 Monica Turner from Oakridge National Laboratories
14 called SPANS. And we're mainly interested in the
15 components, the fractal components of that program.
16 Fractals have been used to look at scale issues of all
17 types of physical phenomena, clouds, shorelines of
18 lakes, land cover. So we're -- this is resident in the
19 literature and we're using it to look at Everglades
20 vegetation.
21 Q. Okay.
22 Have you ruled out any of those
23 classifications or sizes from 20 to 1,000 as yet?
24 A. The results at this point are very
25 preliminary.
105
1 Q. What do the preliminary results show?
2 A. Somewhere around 120 meters.
3 Q. Okay.
4 Could you then identify by Bates numbers
5 which documents are related to that project starting
6 with initial page 1220096?
7 A. Everything up to 1220166.
8 Q. All right.
9 A. We've identified that.
10 Q. Let me show you two documents, or a
11 document, bears Bates No. 1202849 through 1202867, the
12 first page appears to be a file folder which indicates
13 WCA-3S for SPOT 5-11-92. What is that?
14 A. This is the sites that we visited in Water
15 Conservation Area 3, the satellite imagery we
16 previously spoke about, and those are Les's notes that
17 he had of all of his sites.
18 MR. KOBELINSKI: All right.
19 You can mark that as 6.
20 (Thereupon, the document was marked
21 Rutchey Exb. No. 6 for Identification.)
22 MR. KOBELINSKI: I'd like to mark this as
23 7.
24 (Thereupon, the document was marked
25 Rutchey Exb. No. 7 for Identification.)
106
1 MR. KOBELINSKI: For the record, I have
2 marked the document previously handed to Mr.
3 Rutchey as Exhibit No 6 that bears Bates No.
4 1202849 through 867 and has a first page which
5 appears to be a file folder labeled WCA 3S SPOT,
6 5-11-92.
7 BY MR. KOBELINSKI:
8 Q. I show you, Mr. Rutchey, what has been
9 marked as Exhibit No. 7, bears Bates No. 1202868
10 through 1202880 and appears to be a file folder on the
11 front which says WCA 3N for SPOT 5-11-92. I ask you if
12 you've ever seen that document before.
13 A. This is the same as the previous data set
14 but -- except for different days and south of Alligator
15 Alley.
16 Q. This again then relates to the 3A visit?
17 A. That's correct.
18 Q. Okay. Okay.
19 Next I'm showing you what's been marked as
20 Bates Nos. 12002773 through 1202848 and the front cover
21 which is Holeyland data for SPOT 5-11-92. Can you
22 explain what that is?
23 A. This is the same satellite imagery except
24 a different area. It's called the Holeyland area. And
25 this is a preliminary data analysis of that data and
107
1 some ground truth information along with some GPS
2 PFINDER output sets.
3 Q. That's a part of the 3A mapping then,
4 you're also mapping the Holeyland?
5 A. I would say Holeyland is a separate
6 project.
7 Q. Okay.
8 Is that Holeyland vegetative mapping still
9 an ongoing project?
10 A. I would say yes.
11 Q. Okay.
12 Do you have any additional data or
13 documents you've produced related to this Holeyland
14 mapping?
15 A. I believe there is some here, yes.
16 Q. If you could go through and identify
17 those.
18 A. This is all that's left of yours.
19 Q. Leave that here.
20 (Thereupon, a discussion was held off the
21 record.)
22 THE WITNESS: I believe all this is
23 similarly related, somewhat related.
24 BY MR. KOBELINSKI:
25 Q. These various documents are all Holeyland,
108
1 also part of that project?
2 A. I pulled this out of that.
3 Q. Okay.
4 MR. KOBELINSKI: And the witness has
5 identified documents bearing Bates No. 1220441
6 through 1220583, however he has taken out as
7 unrelated a document bearing Bates No. 1220491
8 through 1220494.
9 BY MR. KOBELINSKI:
10 Q. And these documents are, again, relate to
11 Holeyland mapping; is that correct?
12 A. Yes.
13 Q. Okay.
14 It appears to me we've finally identified
15 pretty much all the documents you've sent to us. We'll
16 just take a quick break for lunch and then we'll start
17 finding out what they mean. All right.
18 (Thereupon, a luncheon recess was taken
19 from 12:10 to 1:25 p.m..).
20 BY MR. KOBELINSKI:
21 Q. All right, Mr. Rutchey. I remind you that
22 you're still under oath.
23 MR. KOBELINSKI: One thing before we get
24 on, why don't we mark this 8?
25 (Thereupon, the document was marked
109
1 Rutchey Exb. No. 8 for Identification.)
2 BY MR. KOBELINSKI:
3 8. Mr. Rutchey, I'm showing you what's been
4 marked as Rutchey Exhibit No. 8, which is a paper
5 bearing Bates No. 1220001 through 1220032, which is
6 entitled Inland Wetland Change Detection in the
7 Everglades Water Conservation Area 2A Using a Time
8 Series of Normalized Remotely Sensed Data. And I
9 gather this is a paper that you have participated in?
10 A. Yes.
11 Q. Is this the final draft of it?
12 A. Yes.
13 Q. Is this -- will this be published in any
14 type of a --
15 A. Yes, it's been accepted by the Journal of
16 Photogrammetric Engineering and Remote sensing.
17 Q. Do you know when it will be published?
18 A. They -- right now, I called them just the
19 other day. There's about an 18 month turn around for
20 papers that they have a backlog right now.
21 Q. This was sent out for peer review?
22 A. Yes.
23 Q. Okay.
24 And have you received peer review comments
25 back?
110
1 A. Jensen has all that.
2 Q. Okay.
3 You didn't receive any peer review
4 comments?
5 A. Nope.
6 Q. Was the document revised after it was
7 peer-reviewed?
8 A. I'm not sure.
9 Q. Okay.
10 Are the documents that you've produced to
11 today -- which, if any, of the documents -- when I say
12 today, in response to this notice of deposition.
13 Which, if any, of these documents that are remaining
14 relate to this paper?
15 A. That are remaining? Probably none.
16 Q. Okay.
17 And of the documents previously marked,
18 which, if any, relate to this paper?
19 A. The original 129 ground truth sites and
20 their locations and what we found there.
21 Q. Okay.
22 What about the accuracy analysis for the
23 241 sites?
24 A. I believe that was sent too, but I don't
25 think that was used in this paper.
111
1 (Mr. Downing entered the room.)
2 BY MR. KOBELINSKI:
3 Q. Okay. All right.
4 What was your participation in this paper,
5 what did it encompass?
6 A. Basically just provide the original data
7 set.
8 Q. Okay.
9 And the data set you provided to -- was
10 that Dr. Jensen?
11 A. Right.
12 Q. Okay.
13 Is there a list of data sets you can just
14 tick off for me?
15 A. Yeah, the 129 ground truth sites is one
16 and then all the original multispectral satellite
17 imagery data sets that we used in this analysis along
18 with the final 20 class and 12 class vegetation maps
19 that I produced, Les Vilcheck and I produced.
20 Q. All right. I can't write as fast as she
21 can.
22 The original digital data set, right?
23 From the August of '91 or '92? August '91, yeah,
24 August 10, '91 digital data set; is that right?
25 A. Right.
††††㈵††††††䄮††剩杨琮ഊ†††††††††††††††††††††††††††††††††‱ㄲഊഊ††††‱††††††儮††ㄲ㤠杲潵
112
1 Q. 129 ground truth sites?
2 A. Right.
3 Q. And the final 20 class and 12 class
4 digital map data sets; right?
5 A. Correct.
6 Q. Was there anything else?
7 A. And all the original multispectral data
8 sets that were used in this paper.
9 Q. Now you're referring to, if I understand
10 correctly, data sets for April 4, 1987?
11 A. That's one.
12 Q. January 17, 1982?
13 A. Yeah.
14 Q. April 2, 1976 and and February 22, 1973?
15 A. That's correct.
16 Q. Now did you just send him the original
17 digital data set you had for those four dates?
18 A. That's correct.
19 Q. Okay.
20 You had not manipulated them to any
21 extent?
22 A. No.
23 Q. Other than providing these data sets to
24 Dr. Jensen, what other participation did you have in
25 this paper?
113
1 A. That's about it.
2 Q. Okay.
3 Did you draft any portions of it?
4 A. No.
5 Q. Okay.
6 Do you know what Marguerite Koch's
7 participation was in this paper?
8 A. She provided the soil or water phosphorus
9 data.
10 Q. Did she perform any analysis comparing
11 vegetation to the porewater data?
12 A. I'm not sure.
13 Q. What about Sunil Narumalani? Last name is
14 N A R U M A L A N I, his first name is S U N I L.
15 A. I'm not sure what his role was in this.
16 You'll have to talk to Dr. Jensen about that.
17 Q. Have you ever met Dr. Narumalani?
18 A. Yes.
19 Q. Okay.
20 What exactly is his specialty, if you
21 know?
22 A. I'm not sure. I just met him a conference
23 one time. I know what he looks like. He's an image
24 processing -- I think he's got a doctorate basically
25 in -- basically digital image processing. That might
114
1 not be a right term, but that's what his field of
2 expertise was.
3 Q. What was Dr. Jensen's role with regard to
4 this paper?
5 A. He was the primary author.
6 Q. Okay.
7 To your knowledge is he the one that
8 conducted most of the analysis?
9 A. Right.
10 Q. Okay.
11 Drawing your attention then, sir, back to
12 Rutchey Exhibit 1, the 2A vegetative analysis, and I
13 have a few questions on some of the steps you had taken
14 with regard to what I have written, Step 4, I'm talking
15 about after the computer had gone through and broken
16 out the information into the 30 classifications or
17 clusters, and at which point I believe you said 154
18 sample sites were selected by the computer; is that
19 correct?
20 A. No, we just chose 30 classes, within each
21 one of those clusters we attempted to find five ground
22 truth sites.
23 Q. Oh, okay.
24 So the computer didn't generate the five
25 sites, that was something that you mechanically did?
115
1 A. That's correct.
2 Q. Okay. Pardon me.
3 And as I understand the criteria, was that
4 finding a -- essentially a three by three pixel square,
5 in other words, nine pixels in the same classification;
6 is that correct?
7 A. That was the minimum, right, requirement.
8 Q. And you, of course, would take more than
9 that.
10 With regard to locating the the test sites
11 what -- what was the exact instrument that you used?
12 It might be in here somewhere.
13 A. The exact instrument?
14 Q. Yes.
15 A. It was called ERDAS. It's a software
16 image processing software package.
17 Q. But I'm talking about going out on the
18 site, what GPS equipment did you use to find the actual
19 sites, to find the sites with?
20 A. GPS basic unit.
21 Q. Is there one standard GPS basic unit or
22 are there different models of that?
23 A. There's different models.
24 Q. Do you know what model you had?
25 A. That's the name of it.
116
1 Q. Okay..
2 I believe also you mentioned that there
3 was a base station. Where was the base station
4 located?
5 A. At the South Florida Water Management
6 District, Gun Club Road.
7 Q. Are you responsible for that base station?
8 A. Yes.
9 Q. And how many -- as I understand base
10 stations, they keep track of the various satellites; is
11 that correct?
12 A. Yes.
13 Q. Do you know what particular type of base
14 station this was?
15 A. It's a Trimble Pathfinder community base
16 station.
17 Q. Does that -- how many satellites did that
18 keep track of at any given point in time?
19 A. Well, it varied from then till now. Right
20 now it can track up to nine.
21 Q. Then what about back in '91?
22 A. We didn't have a full ephemeris window
23 back then, window of opportunity. They didn't have all
24 the satellites in place.
25 Q. You lost me on that term.
117
1 A. Ephemeris.
2 Q. All right.
3 What were you operating with in 1991?
4 A. I don't remember the exact number, but our
5 window was limited in that actual span of time that we
6 can acquire an accurate GPS data. You can actually ask
7 the software for when the window is open that you can
8 acquire the data based on the geometry of satellites in
9 space, the available satellites in space.
10 Q. Okay.
11 Was this then one of the governing factors
12 as to when you would go on the site visits or field
13 visits?
14 A. Within a day, yes.
15 Q. Okay.
16 Perhaps if you would walk me through,
17 again, the GPS portion of it. How would you plan out
18 the site visit from the GPS aspect?
19 A. Well, I had my points I wanted to visit.
20 I put that coordinate into the GPS unit. I'd navigate
21 as best I could to that point.
22 Q. All right.
23 You were saying you had to -- perhaps I
24 misunderstood. I thought you told me you first had to
25 contact someone to find out whether or not the base
118
1 unit could read or contact certain satellites. Take me
2 from the initial step if you would.
3 A. I would look at the data, window of
4 availability, and figure out the span of time that I
5 needed to be out there in order to collect, you know,
6 accurate GPS data. So I go out during that time and
7 visit a point, log it over a period of two to three
8 minutes, that way you have multiple points, it makes
9 sure your final solution is more accurate. Take it
10 back, differentially correct each one of those points,
11 average the solution from that, and come up with a
12 point that you were really at when you went out in the
13 field.
14 Q. All right.
15 With regard to the window of availability,
16 who do you contact to find out what it is or is this
17 just a publication that you view through --
18 A. No, it's all in -- what you can do is any
19 almanac, that unit logs, whenever you're looking for
20 data, you can download that to the computer and it can
21 figure out from that almanac when the windows of
22 opportunity are open for collecting GPS data.
23 Q. All right.
24 Now this almanac is in the base unit or is
25 that in the unit you're taking out in the field?
119
1 A. It's in both.
2 Q. Okay.
3 A. It's in both.
4 Q. Is that something you load in right before
5 you go out in the field, this almanac?
6 A. No. The almanac is generated
7 automatically. You don't have to -- just take the unit
8 out, let it make contact to the satellites. It
9 automatically puts an almanac in there, you can
10 download that almanac into your PFINDER software.
11 Q. Okay. All right.
12 The first time you are going out there,
13 you have a window of opportunity, window of
14 availability. You're starting at, I don't know where
15 you would start, the 10 structures or on the --
16 A. For which?
17 Q. For the first trip you took out there.
18 A. For rectification?
19 Q. No, the 129.
20 A. It was random where I start, Site 1. Not
21 really known.
22 Q. Okay.
23 What do you do then? Do you punch in
24 coordinates for the first place you want to go?
25 A. Right.
†††′㌠††††††††⁗桡琠摯⁹潵漠瑨敮㼠⁄漠祯甠灵湣栠楮ഊഊ††††㈴††潯牤楮慴敳潲⁴桥楲獴⁰污捥⁹潵⁷慮琠瑯漿ഊഊ††††㈵††††††䄮††剩杨琮ഊ†††††††††††††††††††††††††††††††††‱㈰ഊഊ††††‱††††††儮††坨敲攠摯
120
1 Q. Where do you get those coordinates?
2 A. That came from -- that's been asked and
3 answered.
4 Q. All right.
5 So where do you get those coordinates?
6 A. Asked and answered.
7 Q. But you have to -- I don't remember the
8 answer, sir, can't you just say?
9 A. It's part of this report. You've already
10 asked that question.
11 Q. All right. I don't have it written down,
12 sir, you have to respond to my question. You can't
13 just say asked and answered.
14 THE WITNESS: I have to keep answering --
15 if he asks me the same question over and over do
16 I have to keep answering?
17 MR. CESARANO: Well, why don't you go
18 ahead and answer it once more. Maybe he'll
19 write it down this time.
20 THE WITNESS: Unsupervised, the
21 classification within each of 30 clusters we
22 tried to find five sites. That would have made
23 150 and we found 129. That's where the sites
24 came from.
25 BY MR. KOBELINSKI:
121
1 Q. Okay.
2 But where did you get coordinates
3 themselves? I still don't understand.
4 A. The coordinates are part of the digital
5 imagery, we rectified it.
6 Q. So it's part of the digital package you
7 received, the data?
8 A. Well, no. You use the -- the
9 rectification points to rectify the original imagery.
10 Then you have a real world map projection, then when
11 you point to an area, you know the location within the
12 software.
13 Q. Okay.
14 The computer tells you the --
15 A. Right.
16 Q. Okay.
17 Did you -- did you assign any error as to
18 the rectification as to whether or not there was any
19 error in that?
20 A. Yes, it was a root mean square of .4.
21 Q. By mean, what as far as meters?
22 A. Eight meters.
23 Q. Okay.
24 So you're about to start your trip, you
25 punch your coordinates for the first site; is that
122
1 correct?
2 A. Yes, it could be the first site, yeah,
3 probably start at one.
4 Q. Whichever one starts with, I'm just trying
5 to follow through and understand the procedure.
6 A. Okay.
7 Q. I guess it's directional, it just takes
8 you there?
9 I've never used a GPS unit, I'm just a
10 lawyer, okay.
11 How do you get out there, you punch it in
12 and there's a little arrow?
13 A. If we -- depending on the method, yeah,
14 it's basically that simple. It's -- you put way points
15 in and you navigate to them, and it basically tells you
16 your bearing and where you need to go and bearing and
17 direction. And it has actually a little compass thing
18 that says well, you should turn this way, a little bit
19 this way, and when you get there, you know, within a
20 certain distance and it, you know, like we could get
21 down almost to where it says you're on the spot.
22 Q. Okay. All right.
23 Now you talked about taking several
24 readings, if I understand it correctly?
25 A. Right.
†††′㌠††††††††⁎潷⁹潵⁴慬步搠慢潵琠瑡歩湧敶敲慬ഊഊ††††㈴††敡摩湧猬映䤠畮摥牳瑡湤琠捯牲散瑬礿ഊഊ††††㈵††††††䄮††剩杨琮ഊ†††††††††††††††††††††††††††††††††‱㈳ഊഊ††††‱††††††儮††佮捥⁹潵
123
1 Q. Once you get to the spot, what do you do?
2 A. I take a number of readings over a period
3 of two to three minutes, and that varied depending on
4 whether we used the helicopter or airboat.
5 Q. Okay.
6 With the helicopter how many readings
7 would you take?
8 A. Five. It's hard for a helicopter to hover
9 for two to three minutes in one exact spot.
10 Q. And with an airboat?
11 A. Probably over a period of two minutes,
12 three minutes, one every 15 seconds, approximately
13 maybe ten altogether, if I remember correctly,
14 that's...
15 Q. Okay.
16 How do you physically do a reading? What
17 do you do? Is that something you just program into the
18 machine itself?
19 A. Right. The machine actually logs the
20 data. You can actually set the machine to collect data
21 at intervals based on where you're at, you can say the
22 interval, say collect data every fifteen minutes and it
23 will log your position every fifteen minutes and store
24 it.
25 Q. Just trace your path then using that data?
124
1 A. Right.
2 Q. Okay.
3 A. You could use it for that purpose also.
4 Q. All right.
5 What did you have it set then for this
6 particular -- your field trips?
7 A. I don't understand the question.
8 Q. I thought you said you could set it
9 automatically to read?
10 A. Oh, it was -- I'm not -- I don't remember
11 exactly, but in the field I think it was every 15
12 seconds and in the plane, I think it was every --
13 helicopter -- every five seconds.
14 Q. Okay.
15 Would you still have the -- well, all
16 right.
17 Before I find out what you have, this is
18 presumably recording in some sort of a computer or some
19 data base --
20 A. Um-hum.
21 Q. -- out in the field?
22 A. Yes.
23 Q. What do you with it when you get --
24 A. We download it into the PFINDER GPS
25 software.
125
1 Q. That automatically takes in -- looks at
2 the input from the base station, determines where you
3 actually were, in case there was a type of
4 interference?
5 A. That's correct.
6 Q. Okay.
7 Do you still have those files for the 129
8 trips --
9 A. Yes.
10 Q. -- 129 station trips?
11 I assume, from looking at those files, I
12 would be able to tell approximately how many readings
13 there were and how close together, et cetera?
14 A. Yes.
15 Q. Okay.
16 Would I be able to identify what
17 satellites you're looking at from those files or no?
18 A. Not from those hard copy files. I have
19 the original digital data and I think there's a way
20 that you can actually look at that and tell what
21 satellites it was using to determine position.
22 Q. Okay. All right.
23 You had actually identified some GPS data
24 for me earlier in the deposition, and essentially, if I
25 understand what you just said, I could then pretty much
126
1 look at everything you just explained, with the
2 exception of what satellites from this hard copy
3 readout; is that right?
4 A. Right.
5 Q. With what you identified earlier in the
6 deposition I can find out whether you're reading from
7 five seconds apart or 15 seconds apart or how many, all
8 that?
9 A. Yes,
10 Q. That's all in here?
11 A. Yes.
12 Q. When you get back say, how many stations
13 do you visit in one trip?
14 A. I would say 30 or 40 in a day was a good
15 trip, depending on whether you're using helicopter or
16 airboat.
17 Q. Okay.
18 And then the down load the 40 stations
19 with the various readings for them; is that right?
20 A. Right.
21 Q. Okay.
22 Is there -- you've mentioned a software
23 package that's done, automatically plugs in to give you
24 where, where you actually were. Is this sort of
25 automatic for each station?
127
1 A. No.
2 Q. How does it work?
3 A. You download the data and you combine that
4 data with the base station data that was collected at
5 the same exact time. And then the software, it's
6 differentially corrected the data, you want to make
7 sure you do each individual point even though they're
8 pretty much the same site, you want to average them.
9 After that's pretty simple.
10 Q. Okay.
11 That computer program, does it take into
12 account whether they're reading the same satellite and
13 all that stuff?
14 A. They have to read the same satellite or it
15 won't work.
16 Q. That part, the program itself?
17 A. Yeah.
18 Q. Okay.
19 After you get the readout it's corrected.
20 Do you then average those 10?
21 A. That's correct.
22 Q. Or if its, for instance an airboat
23 station?
24 A. Correct.
25 Q. All right.
128
1 The data that we have to -- we have the
2 corrected 10 prior to averaging or do we have raw data?
3 A. You have all three. You have the raw
4 data, you have the raw data that's been differentially
5 corrected, and you have an average of that
6 differential, the corrected raw data.
7 Q. Okay.
8 And based upon the your analysis, how
9 close do you believe you were to the spot you were
10 actually looking for?
11 A. Well, I know the accuracy of the GPS unit
12 was in the range of three to seven meters because of
13 our trips. We would go to a known survey site out in
14 the field that we were doing our field reconnaissance,
15 we differentially corrected that data set, so we knew
16 throughout the process we were in a range of three to
17 seven meters.
18 Q. Okay.
19 After you, you know, took the 125 field --
20 went to 129 sites out in the field, at that point, as I
21 understand it, you went from 30 classes down to, I
22 believe, it was 19 is what you said; is that right, and
23 then added periphyton later on?
24 A. No.
25 Q. Okay.
129
1 A. This has been asked and answered again.
2 Q. Sir, you have to understand, you've done
3 this, you're out there, you spent over a year doing it.
4 We, on the other hand, are trying to read something,
5 understand what you did. We do not have the memory
6 that you do. If I asked this morning, I apologize if
7 you stated something I didn't pick up on immediately,
8 but it will happen where I ask a question where perhaps
9 you believe you've given me the information I was
10 seeking. All right. It's perhaps a frustrating
11 process, but it needs to be done. Okay?
12 A. Okay.
13 My answer to that is my first deposition I
14 got asked the same thing over and over again 20
15 different ways and the guy knew exactly what he was
16 doing. I think you're doing it too, but that's okay.
17 If you want my answer, I'll keep answering.
18 MR. CESARANO: If I think he's beating you
19 up I'll let you know, okay. I don't think we've
20 gotten there yet.
21 THE WITNESS: Okay.
22 BY MR. KOBELINSKI
23 Q. Using the phrase, I am not tempted to beat
24 you up, all right. You're not being tested in this
25 case, I'm just out trying to really figure out what it
130
1 does, okay? All right. I'm attempting to.
2 Then it shows that you visited 129 sites
3 is what I have. You then did a supervised
4 classification, at which point you got down to, I
5 believe it says 19, and ultimately you added an
6 additional one class; is that correct?
7 A. That's correct.
8 Q. All right. That's perhaps what I asked.
9 And again, I don't want to misread or try
10 to read into my notes, but after the 129 sites you
11 actually went from 30 down to 19 plus the periphyton?
12 A. After the 129 we used that ground truth
13 information to do supervised classification and --
14 Q. All right.
15 A. -- came up with 19 classes.
16 Q. All right.
17 Now, in the -- your paper there's
18 discussion here with regard to preparing ellipses from
19 sample band combinations?
20 A. Okay.
21 Q. In Exhibit 1, as I understand it, you have
22 three bands for the satellite?
23 A. That's correct.
24 Q. Okay.
25 How do you go about comparing these
131
1 ellipses of sample bands, do you look at all three?
2 A. Yes, and combinations of all the bands, if
3 you find anything unique within any combination that
4 makes it separate.
5 Q. Okay.
6 And you have, if I recall correctly --
7 Drawing your attention, just for a moment
8 if I could, back to what's marked as Exhibit 8, and if
9 you would turn to the back of that, Bates 1220026,
10 actually they skipped a page it's marked 4, but it's
11 between Page 1220026 and 27, and I know this is not
12 your piece, but is this essentially what you were doing
13 for the cattail paper, drawing this type of an ellipse?
14 A. The computer makes -- actually makes the
15 ellipse, but it is a fair analogy, yes.
16 Q. And is the computer able to do it for just
17 two bands at a time, does it do all three bands at
18 once?
19 A. It does two bands at a time.
20 Q. Okay. All right.
21 Do you have a file showing that type of
22 information, the ellipse and comparison of the bands?
23 A. Yes, it's all digital.
24 Q. All right.
25 So we don't have a hard copy of that?
132
1 A. No.
2 Q. All right. Get back to where I was.
3 Once the computer does that, does the
4 computer statistically determine which ones are close
5 enough to do vegetative identification or is that
6 something that the operator or the scientist doing it
7 does? How is that done?
8 A. Well, I think it's a combination of the
9 two, where the computer generates the ellipse, you
10 analyze the data, determine, to see if there's overlap
11 between the ellipses or whether, indeed, we have unique
12 spectral characteristics for selected classes of
13 vegetation.
14 Q. Okay. All right.
15 Who did the review of the ellipses, was
16 that you or Mr. Vilchek?
17 A. That's me.
18 Q. All right. All right.
19 Now with regard to the color infrared
20 photography that you obtained for October 10, '91,
21 okay.
22 A. No -- October 10?
23 Q. Of 1991.
24 A. Color infrared?
25 Q. Yeah.
133
1 A. Oh, maybe it was that day. I said it was
2 within a year, okay. It's fair.
3 Q. Okay. All right.
4 Did you go through and attempt to locate
5 the color infrared photos that had the 129 points, is
6 that how you went about doing this or is it a different
7 method?
8 A. I think we used that as an aid in the
9 process, looking at that photography. It was just an
10 extra piece of detail. We didn't rely on it too
11 heavily. We mainly used -- we drove this process
12 digitally basically, but that was an extra data set
13 that it was -- that it was referred to it, yes.
14 Q. Okay.
15 Was -- just so I understand how you used
16 it, so whether or not you were doing a homogeneous
17 community or were you actually able to do vegetative
18 identification with the infrared photography?
19 A. Basically I think when we used it it was
20 like, for instance, when we had the periphyton problems
21 we looked overall, we could see periphyton in the south
22 end, and I don't think we used it too much in the 129
23 ground truth sites.
24 Q. Okay.
25 Now, with regard to 20 classes then, and
134
1 I'm just going to use the -- where is the one we have
2 marked?
3 Drawing your attention in Exhibit 1 to Bates
4 Page 1220435, Page 17 of the document, you have 20
5 listed right there. The sawgrass dense, is there a
6 percentage coverage that was used for that?
7 A. Yes, it's a table that goes along with
8 each of these classes.
9 Q. Actually let me find -- and that table is
10 at Page 19, also Bates No. 1220437?
11 A. Yes.
12 Q. Okay.
13 Now with regard to that, it refers to
14 greater than or equal to 70 percent evenly distributed;
15 the balance in open water. Okay.
16 What are the types of vegetation you would
17 have out in 2A other than sawgrass and open water?
18 What mixes would you find, if any?
19 A. I think this list here pretty well
20 characterizes the types of vegetation you would see in
21 Area 2A.
22 Q. Okay.
23 That would be sawgrass in open water,
24 sawgrass and cattail interspersed to varying degrees,
25 sawgrass/cattail and brush, sawgrass/brush,
135
1 sawgrass/broadleaf and cattail. And broadleaf, what
2 type of plants were those?
3 A. Sagittaria species or Pontederia species.
4 Q. Is that like a water plant?
5 A. Yeah, I would say it's a -- it's a wetland
6 species, yes. All the -- all these are pretty much
7 wetland species.
8 Q. And again, it may sound silly, I know what
9 sawgrass looks like, I know what cattail looks like, I
10 know what they would look like. Are these, the
11 broadleaf, are these low in the water type plants or
12 are they, again, grass macrophyte-type plants?
13 A. They're -- they could be, I'd say, average
14 two to two and a half feet, maybe 3 feet. And we would
15 call them broadleaf because the leaves on them were
16 rather broad.
17 Q. Okay.
18 Do you have any low water type plants?
19 You know what I mean by that?
20 A. Are there any occurring in --
21 Q. Bladderwort. Did you have bladderwort in
22 2A?
23 A. Bladderwort is a submerged species.
24 Q. Okay.
25 A. It occurs throughout the area, yes.
136
1 Q. All right.
2 Does that, the bladderwort, if it's there
3 with the sawgrass, do the two cohabitate in various
4 communities?
5 A. Yes.
6 Q. Does that affect the spectral signature?
7 A. I don't think so, because it's submerged.
8 Q. Okay.
9 So does the depth of the water affect the
10 spectral signature?
11 A. It could, yes.
12 Q. How could water depth affect the spectral
13 signature?
14 A. Based on whether you have ground exposed
15 and you have periphyton out there, it's to what level.
16 Sometimes periphyton is floating, sometimes it's at the
17 bottom. So, you know, that could really affect the
18 spectral characteristics of what you're seeing. That's
19 why you ground truth what you're actually seeing.
20 Q. Okay.
21 Would the bladderwort, if it's a submerged
22 plant, affect the spectral signature to the same extent
23 that water depth would affect the signature?
24 A. I don't think so.
25 Q. Okay.
137
1 Would you have sawgrass stands, be they
2 dense, moderate or sparse, going with periphyton?
3 A. Yes.
4 Q. Where would that fall in this -- the 20
5 classes?
6 A. We didn't break it down. We took a
7 dominant, if it was dominantly periphyton or all
8 periphyton. Typically, we broke it out and if it was
9 dense sawgrass, could possibly be periphyton in this
10 case or not, dense, sparse sawgrass --
11 Q. And in cases where you had sparse sawgrass
12 and periphyton, what would it be listed as?
13 A. Now say that one more time.
14 Q. In a case where you had sparse
15 sawgrass/periphyton, which classification would it fall
16 into?
17 A. From our field we would -- we would
18 probably call that sparse sawgrass.
19 Q. Okay. All right.
20 Would that hold true also for cattail
21 sparse, cattail and periphyton?
22 A. Yes.
23 Q. With regard to sawgrass sparse, is that
24 equal or to 30 percent coverage, what is the low end of
25 30 percent?
138
1 A. I'm not sure. I think it was like five or
2 ten percent.
3 Q. Are there notes as to that low end, as to
4 what the low end was?
5 A. I'm sure they're somewhere. I -- I would
6 say less than 30 percent means from 1 to 30 percent.
7 Q. Okay.
8 Would that same definition hold true for
9 sparse cattail, 1 to 30 percent?
10 A. Yes.
11 Q. Okay. That's a broad question.
12 What difficulties did you find in mapping
13 the sparse sawgrass?
14 A. (No response.)
15 Q. You mentioned periphyton. Anything else?
16 A. I think the major problem I had in this
17 mapping out effort was the periphyton community,
18 because even though there might have been sawgrass
19 sparse or sparse cattail, the dominant signature still
20 came back as periphyton. That's why this map effort
21 mostly suffered in most of its inaccurate -- in its
22 overall accuracy is because of the periphyton
23 community.
24 Q. Was there any predominant place where you
25 had periphyton throughout WCA-2A?
139
1 A. In the southern region when we have a lot
2 of open slough areas down there, sparsely vegetated
3 areas.
4 Q. Okay.
5 Now, did the fact that you're using the
6 August SPOT imagery data set, compared to your field
7 trips which extended all the way through April of '92,
8 did the changes in water levels or vegetation as a
9 result of the dry season impact at all the study?
10 A. I don't think so.
11 Q. Okay.
12 Was there any way to test whether or not
13 it did?
14 A. No.
15 Q. Is there a difference in the spectral
16 imagery from August to, for instance, January or
17 February?
18 A. If you had satellite imagery, yes,
19 probably.
20 Q. Okay.
21 What would those differences be?
22 A. Basically vegetation goes through various
23 stages, different vegetation browning out or if there's
24 a drought some of it just dies out totally. So
25 hydroperiod and time of year can affect what you're
140
1 seeing in satellite imagery.
2 Q. Okay.
3 Does the time of year for the plant, as I
4 guess you put that, brownness, does that also impact
5 the reflective values and spectral imagery?
6 A. I think so.
7 Q. On the sites for the 241 points, my notes
8 are a bit confused. On that 241, when you're doing the
9 accuracy, was that based again upon a 3 pixel by 3
10 pixel or were those just randomly generated by the
11 computer without that, any of that type of requirement?
12 A. There was a stratified random sampling by
13 class. And yes, they had to meet a 3 by 3 meter, a 3
14 by 3 pixel minimum requirement.
15 Q. Okay.
16 On that, if I could draw your attention
17 back to Rutchey No. 4, that's the laboratory notebook,
18 and I believe you pointed out latter pages in this
19 document, essentially Bates No. 1202674 through the
20 remainder, 1202688, reflect the sites, the pixels based
21 upon the computer generation, and that -- then your
22 field data inspection; is that correct?
23 A. 4, Item 4 in the thing.
24 Q. The fourth item?
25 A. Is the field data.
141
1 Q. Okay.
2 Going through a, just to the first page,
3 1202674, Site 116, all 9 pixels are 4. The next one,
4 Site 196, it's a split of 5 and 4, 5 being 7, 4 being
5 18. And just randomly going through this, only
6 approximately 1 out of every 4 to 1 out of every 6
7 actually has all 9 pixels being the same. You have
8 having the same classification, that varies from page
9 to page. But does -- how would you deal with pixels
10 that did not have all 9 the same vegetation class?
11 A. I stated in the paper we used a minimum
12 of -- required minimum of 5 out of 9 meeting majority
13 status and called it, that pixel cluster that item, or
14 that classification.
15 Q. Okay. Perhaps I misunderstood that.
16 You're saying all nine did not require the
17 same classification, they just had to be a majority?
18 A. Majority.
19 Q. Okay.
20 Then, when you're doing your accuracy
21 analysis, would you be looking to see whether or not
22 the field test matched on all 9 pixels?
23 A. (No response.)
24 Q. In other words -- do you understand what
25 I'm saying?
142
1 A. Um-hum.
2 Q. As opposed to getting it to the point
3 which we believed to be in the middle pixel, would you,
4 as I understood it before, look at 20 by 20 when you
5 were doing your field tests for these 241, would you
6 look at all 9 surrounding, in other words, the -- all 9
7 pixels that you were analyzing?
8 A. We would look at what data we got back
9 from the GPS, correct the data. We would make a
10 similar 3 by 3 pixel representation as you see here,
11 pick the majority out of what it was and then look at
12 that majority value or classification and compared it
13 to what we got from the field. And if it was the same
14 we'd count it as correct, and if it wasn't, we counted
15 it as wrong.
16 Q. All right.
17 Just so I understand how this works then,
18 if you could go back to the second page of this
19 document, this one where we were looking at Site No. 4
20 originally right there.
21 A. Okay.
22 Q. And we had gone away from this because you
23 had thought we could look at another page earlier. You
24 have next to 4 in the 1, 2, 3, 4, 5, 6, 7, you have 7
25 columns, excuse me, rows, okay. And in the first,
143
1 fourth, and seventh you have noted in this case 1, 1
2 and 1 there. Are those the computer-generated
3 classifications for Site 4?
4 A. If I remember correctly -- I don't
5 remember.
6 Q. Okay.
7 To the right of that are then some
8 additional 1's. Again, they're in the first, fourth
9 and then there's one in the fifth and two in the 7th.
10 Are these again dealing with Site No. 4?
11 A. I believe so.
12 Q. Okay.
13 Does that refresh your recollection at all
14 as to how this was done?
15 A. Okay.
16 This is how it was done.
17 Q. Okay.
18 A. That's basically what this initial 4
19 data set was, we actually went and looked at 7 pixels
20 by 7 pixels instead 3 pixels by 3 pixels. What you see
21 here in this first part of this data set is that
22 representation. So let's go back to 4. Basically in 4
23 in the first column you see 1 at each one of the
24 corners.
25 Q. And 1 in the middle.
144
1 A. Right. So this basically, to me,
2 everything was 1.
3 Q. Now, is that computer generated?
4 A. That's computer generated. Now when, this
5 other one next to it, you see 1's in each corner, but
6 two 1's in the middle, you see 1, it's the whole --
7 Q. Right.
8 A. The pixels in that area are something
9 other than 1, so what -- what this is, I went beyond
10 the original 3 by 3 in this particular data set.
11 Q. Okay. All right.
12 Just -- so you would then be looking at
13 the four corners, the middle points on the outside and
14 the center?
15 A. Right. But it was something other than --
16 if the corner was something other than 1 I would put 1
17 here, I would put the next 1 in, so on and so on, if it
18 was other.
19 Q. Okay.
20 So for instance, on No. 4 I'm looking at
21 the right-hand side where this said Old 1. Okay.
22 A. Right.
23 Q. Looking at the upper left-hand, or where
24 it says 1 there, then four pixels in a box that are
25 blank, were those observed at all or or those would
145
1 automatically be taken as 1 unless marked otherwise or
2 just literally looked at 3, 6 points, 9 points.
3 A. No, I looked at all of them. I didn't
4 want to write 1's in every one of these little boxes,
5 so I made my own little method of how this -- how to do
6 it. And the one that you're pointing out --
7 Q. Um-hum.
8 A. -- you see -- can see there's 1's in all
9 corners, but on the bottom one you see a 1 directly
10 below the middle pixel, that means everything behind
11 was something other than that.
12 Q. Right.
13 A. Weren't.
14 Q. Okay. Okay.
15 But, for instance, what about to the right
16 of where that pixel, that 1 would have been if it was
17 sawgrass?
18 A. To the right, if you look to the right and
19 there's a 1, then that means everything in between
20 between was a 1.
21 Q. Okay.
22 So everything blank other than the 1, for
23 instance, below those two double 1's are sawgrass,
24 dense sawgrass?
25 A. That's correct.
146
1 Q. Okay.
2 Is there any way to tell what the two
3 blank ones that we don't know what they are, what they
4 are that goes in those grids, what are the two missing
5 squares, what vegetation are they?
6 A. It's available digitally.
7 Q. Okay.
8 All of this data is available digitally?
9 A. Yes.
10 Q. Does the digital data contain more
11 information than that then?
12 A. Yes.
13 Q. Okay.
14 Just so I understand, the digital data set
15 then was not taken from those sheets, but rather this
16 is a summary of the digital data set?
17 A. This is created from the digital data,
18 right.
19 Q. Okay. All right.
20 When there's a discussion of map accuracy
21 of 81 percent in the paper, okay. And that discussion
22 was found at -- on Page 9 of Exhibit 1 which bears
23 Bates No. 1220427, 80.9 percent.
24 A. Okay.
25 Q. All right.
147
1 Is that an accuracy as to boundaries
2 reflected by the map or an accuracy as to vegetative
3 type within those boundaries?
4 A. That's an overall map accuracy of
5 everything within the boundaries.
6 Q. Okay.
7 Does the 80.9 percent apply to each class
8 of the 12 classes?
9 A. No, it applies to the overall map accuracy
10 of the whole vegetation project.
11 Q. How would you determine it as to the
12 various classes?
13 A. How would you determine the accuracy as to
14 the various classes? That was done for some classes,
15 if you look at 1220439, Table 3.
16 Q. Okay.
17 A. And you can see what -- the column, users
18 accuracy. We did that for some of them.
19 Q. Now, this shows all 241, all the -- all
20 the various samples there?
21 A. Yes.
22 Q. Would looking at this refresh your
23 recollection as to which classes were included in the
24 31 or 32 additional tests or sample sites?
25 A. No.
148
1 Q. All right.
2 Users accuracy tells you the accuracy of
3 whether or not the vegetative class that's actually
4 shown by the map is what's in the field, is that --
5 A. (Shakes head up and down.)
6 Q. All right.
7 What does user accuracy mean?
8 A. I think you had hit a little bit.
9 Basically, if you -- if you point to an area on the map
10 you go out in the field to see if indeed that's what
11 you're seeing on the map is correct.
12 Q. All right.
13 A. Compared to to producer accuracy, the
14 producer accuracy, if you went out in the field and got
15 a GPS point and documented what you saw there, and you
16 went back to the map to see if that indeed was there.
17 Q. All right.
18 Why would that be different?
19 A. It's two different ways of looking at a
20 map. They're going to be different numbers. That's --
21 I can't answer it any better than that.
22 Q. But is it still -- are you still
23 essentially just trying to figure out whether or not
24 the map reflects reality under both scenarios?
25 A. To a point, yeah. It's basically looking
149
1 at commission and omission errors.
2 Q. And the conditional KAPPA on the
3 right-hand side means what?
4 A. This is being used, now a KAPPA statistic
5 is being used to -- in image processing projects to
6 report the error.
7 Q. All right.
8 If I am reading your page, Bates No.
9 1220439 of Exhibit 1, if I'm reading this correctly,
10 your -- let me withdraw that.
11 When you say this is -- I'm looking at the
12 20 class map, 70.9 accuracy, is that average of
13 accuracy a producer accuracy?
14 A. No, I would say not.
15 Q. What does the 70.9 come into as compared
16 to user accuracy?
17 A. The 70.9 represents the overall map
18 accuracy as to the whole entire vegetation project.
19 User accuracy and producer accuracy has to be looked at
20 separately.
21 Q. All right.
22 How do you get then the overall map
23 accuracy? How do you calculate this?
24 A. Well, the KAPPA statistic takes into
25 account the user accuracy, producer accuracy, which is
150
1 really commission and omission errors and you have this
2 in KAPPA.
3 Q. All right.
4 It's -- in other words, it's a statistical
5 manipulation of the two to get what the overall
6 accuracy is?
7 A. What is your definition of overall
8 accuracy?
9 Q. All right.
10 That's on Page 1220439, overall accuracy
11 70.9 percent of a 20 --
12 A. Basically, what that represents is an
13 overall map accuracy. How many out of that 241 points
14 that are visited in the field, what -- what percentage
15 of them were correct.
16 Q. Oh, okay. That's all right. That's fair.
17 Is there a means of determining that
18 accuracy on a class by class basis?
19 A. Yes, that's basically what user accuracy
20 is.
21 Q. Okay.
22 For user accuracy there are only six
23 classes that have a user accuracy number. Why is what?
24 A. You had to meet a minimum requirement of
25 numbers. I could find that for you.
151
1 Q. Okay.
2 (Thereupon, a discussion was held off the
3 record.)
4 THE WITNESS: Yes, yes. Page 7, second
5 paragraph are the results. In order to show any
6 significance or do it statistically right you
7 had to have a minimum of 19 field sites.
8 BY MR. KOBELINSKI:
9 Q. Okay.
10 But that's for user accuracy; is that
11 correct?
12 A. (No response.)
13 Q. Let me just make -- the reason I ask is,
14 for instance, on the producers accuracy, I see numbers
15 there for 15, 15, and 11. If you read that across from
16 on the 20 class map, 15 for Class 5, 15 is for Class
17 10, and 11 for Class 20.
18 A. Okay.
19 Q. Okay.
20 Is there a different minimum number for
21 user as opposed to producers accuracy?
22 A. Un-hum. I would say I'm not sure. I feel
23 the method that people who are experts in the field
24 such as Russel Congalton, Rosenfield, Lins-Fitzpatrick
25 (sic), Michael Story, people who have expertise in the
152
1 field, and I use their methods and --
2 Q. Okay.
3 A. -- that's -- this is state of the art
4 here.
5 Q. Okay. All right.
6 Now, from this particular matrix, is there
7 a means -- I'm talking again on Page 21, Bates No.
8 1220439, perhaps the easiest example would be Map Class
9 1, Class 1, ground truth Class 1, and the number there
10 that says 41. It's on the exhibit. You following
11 where I'm at, upper left-hand corner of that?
12 A. Okay.
13 Q. Does that mean 41 sites, you went to 41
14 sites?
15 A. Okay.
16 Within Class 1 I visited 45 sites.
17 Q. I read it across then to 40 -- of the 45,
18 41 were Classification 1 based upon your field visit?
19 A. That's right.
20 Q. Two were Classification 2, and one was
21 Classification 4, one was a Classification 11?
22 A. Correct.
23 Q. Okay. All right.
24 If I recall correctly, you had stated that
25 you had tested an additional 31 sites dealing with the
153
1 cattail to better determine the accuracy of the model
2 with regard to the cattail?
3 A. That's correct.
4 Q. All right.
5 Given those additional 31 sites, if I read
6 this correctly, sawgrass/cattail dense and sawgrass
7 sparse, you don't actually have an accuracy figure, but
8 it was your recollection that those 31 sites didn't
9 include those two categories?
10 A. I think I said I wasn't sure.
11 Q. Okay.
12 Do you recall now or?
13 A. No.
14 Q. Do you believe that they did or just
15 really have no idea one way or the other?
16 A. I'd have to go back and look.
17 Q. Okay.
18 And with regard to the cattail, cattail
19 dense, cattail moderate, cattail sparse, the only one
20 that has sufficient is cattail dense?
21 A. In the 20 class map.
22 Q. Based upon that, were you able to
23 determine how accurate it was for cattail mapping
24 purposes?
25 A. I'd say, if you were looking just at dense
154
1 cattail, that cattail is -- on that we were right 68
2 percent of the time.
3 Q. Okay.
4 What about the other categories of
5 cattail?
6 A. For the 20 class map sawgrass/cattail
7 dense, 62 percent.
8 Q. All right.
9 And sawgrass/cattail sparse 58 percent?
10 A. Yes.
11 Q. Of the 17 samples of cattail, sparse, No.
12 12, only 7 came in correctly; is that right?
13 A. Yes.
14 Q. If I asked you to, given all the data
15 you've collected, would you be able to give me a
16 percentage of accuracy as to identification of cattail?
17 Not the different various categories, but just cattail
18 itself? Is there any way for you to determine that?
19 A. I would use the 12 class map to answer
20 that.
21 Q. Okay.
22 A. And No. 2, I would say overall for
23 sawgrass/cattail, I'd say that we were 70 percent
24 accurate. And for cattail by itself, we're 72 percent
25 accurate. And for cattail/brush we're a hundred
155
1 percent accurate.
2 Q. Okay.
3 That's where you're collapsing the three
4 cattail categories just into one category of cattail;
5 is that correct?
6 A. Yes.
7 Q. Okay.
8 (Thereupon, a discussion was held off the
9 record.)
10 BY MR. KOBELINSKI:
11 Q. All right.
12 Drawing your attention back to Exhibit 1,
13 Page 21, which is Bates No. 12200439, users accuracy is
14 essentially a -- just a simple percentage looking at,
15 you know, how many field -- how many sites did I check
16 that the computer told me was, for instance,
17 Classification 1, how many times was it correct. Is
18 that essentially correct, user accuracy is just a
19 percentage, right?
20 A. (Shakes head up and down.)
21 Q. Straight, simple. All right.
22 If I looked at map accuracy, overall
23 accuracy, 80.9 percent for the 12 class map, would that
24 be just doing that for every single one of the
25 classifications, again as a simple math on that, or is
156
1 it more complicated than that?
2 A. No, it's pretty simple. What you do is
3 look at the diagonal of the numbers that were correct.
4 Q. Oh, I never thought of that. Look at
5 that.
6 A. All right?
7 Q. Neat.
8 A. All right.
9 So I add those up up and you have 241, you
10 take that percentage of that and that's the overall.
11 Q. Kaboom. That's marvelous, that's great.
12 When we were discussing earlier accuracy
13 with regard to 20 class map, you had stated there was a
14 minimum number of 19 that was necessary, okay; is that
15 correct?
16 A. Right.
17 Q. All right.
18 Does that change if -- when you go down to
19 the 12 class map?
20 A. No.
21 Q. Okay.
22 The only reason -- it could just be that
23 you added this to a class, but Class No. 6 in the 12
24 class map you only had 12 sites and you had run it for
25 that particular class?
157
1 A. Well, you should have proofreaded (sic) my
2 paper.
3 Q. Well, that's just -- that's just pure math
4 anyway, just reading across. That technically can be
5 done for every column. If I understand you correctly,
6 for statistical significance you have to have 19 or
7 more samples. Is that essentially what' you've
8 determined?
9 A. Yes.
10 Q. Okay.
11 Now, with regard to the categories, I see
12 that you have on the immediately preceding page that
13 we've been looking at is where you collapsed 20
14 categories down to 12 categories. But just so I
15 understand the -- either the 20 or the 12, I'm looking
16 at Page 19 of this exhibit which bears Bates No.
17 1220437, and with regard to sawgrass/cattail dense, if
18 I understand the description you have there, which is a
19 mixture of Cladium jamaicense and Typha domingensis,
20 vegetative cover excludes aquatics greater than or
21 equal to 60 percent. In that category if, for
22 instance, the combined cattail/sawgrass was 90 percent,
23 it would fit within Category No. 4; isn't that correct?
24 A. If it was 90 percent?
25 Q. Of cattail/sawgrass, that would clearly
158
1 fit within that category?
2 A. Yes.
3 Q. All right.
4 What if it was 80 percent sawgrass and 10
5 percent cattail and 20 percent open water, that would
6 still be in Category 4?
7 A. Yes.
8 Q. What about the opposite, 10 percent
9 cattail, 80 percent sawgrass, 20 percent water?
10 A. Yes.
11 Q. What about the lower end of that, for
12 instance, what if it's one percent cattail, 89 percent
13 or 95 percent sawgrass and 4 percent water?
14 A. I believe it was 10 percent -- I'd have to
15 go back and look.
16 Q. Where would you look to see that?
17 A. Notes. I'm not -- Les probably has it,
18 that determination.
19 Q. Do you recall, since you have looked
20 through a fair amount, not in great detail, I
21 understand that, but you have looked through a fair
22 amount of documents that were produced, do you know
23 whether or not that would be reflected in what you sent
24 to us or produced to us?
25 A. I don't think so.
159
1 Q. But as you sit here you recall it being
2 approximately a 10 percent cutoff?
3 A. I think so.
4 Q. All right.
5 So, for instance, in either direction, if
6 it was 5 percent cattail and 95 percent sawgrass, that
7 would be marked as a pure sawgrass stand?
8 A. You might almost determine this from the
9 ground truth data sheets what the cutoff was. I think
10 we only -- it might even be down to 5 or 10 percent,
11 but I don't recall.
12 Q. Okay. All right.
13 Does the same analysis hold true for
14 sawgrass/cattail sparse?
15 A. (No response.)
16 Q. And by that I mean if, for instance, only
17 a 50 percent coverage and it's cattail and sawgrass, it
18 could either be 40 percent sawgrass and 10 percent
19 cattail or the opposite 10 percent sawgrass and 40
20 percent cattail?
21 A. Yes.
22 Q. All right.
23 Now, with regard to Category No. 6, that
24 one is, you know, perhaps this might be alone, but
25 that's very specific as to how many -- how that
160
1 sawgrass/cattail/brush is divided up. That's not just
2 any combination of them, if I understand correctly.
3 A. Right.
4 Q. All right.
5 It's greater than or equal to 70 percent
6 sawgrass and less than or equal to 30 percent cattail
7 brush, or cattail or brush?
8 A. Right.
9 Q. Okay.
10 What would happen if you had 60 percent
11 cattail and -- I don't care, let's just hypothetically,
12 so we understand, 60 percent cattail -- excuse me,
13 sawgrass. 30 percent or 20 percent cattail, and 20
14 percent brush. Where would that show up in this?
15 Well, before I ask that hypothetical,
16 would you ever find a community like 60 percent
17 sawgrass, 20 percent cattail and 20 percent brush?
18 A. I'm sure it's possible.
19 Q. Do you recall ever seeing anything like
20 that?
21 A. I've seen vegetation mixed in all
22 different amounts.
23 Q. All right.
24 Given that, where would -- if I came
25 across this out in the field, where would I put that?
161
1 A. I need a pen.
2 Q. I'm sorry.
3 A. What now? 60 sawgrass, 20 cattail, 20
4 brush you want this? 60, 20, 20.
5 Q. Sawgrass/cattail/brush.
6 A. Okay.
7 I don't see a place to put that. That's
8 probably a function of we're not finding that
9 particular combination in our ground truth sites.
10 Q. Okay.
11 In the larger, if you go to larger
12 categories, would it fall in? I mean you obviously
13 have the detailed description of larger -- the 12 class
14 map, under a 12 class map, how would I determine the
15 description of the classes?
16 A. Well, you -- like for sawgrass, you know
17 that it's basically dense, moderate, and sparse
18 vegetation. You just would take the present class map
19 and combining it all together and that would equal
20 sawgrass in the 12 class map.
21 Q. Okay.
22 So for instance, when -- the one I just
23 gave, 60 percent sawgrass, 20 percent cattail, 20
24 percent brush, would that just fall into No. 2?
25 A. Well, I can't put it in anywhere, because
162
1 we didn't have that percentage as an actual ground
2 truth site, those percentages, so we didn't make a
3 category for it.
4 Q. Okay.
5 How did you, for No. 6, again, as I said 6
6 is perhaps the most detailed of your descriptions.
7 MR. CESARANO: On here, on this
8 classification?
9 THE WITNESS: 19?
10 MR. KOBELINSKI: Yes.
11 MR. CESARANO: Okay.
12 BY MR. KOBELINSKI:
13 Q. How did you come up -- for Classification
14 No. 6 on the 20 class, how did you come up with these
15 percentages?
16 A. Basically we would have looked at 129
17 ground truth sites and we -- from that we best
18 determined what we could break out as far as being
19 classes. And basically we used that information to --
20 gathered information to seed the original multispectral
21 imagery to come up with unique statistical data. At
22 that point we didn't know if this was indeed unique.
23 When we actually looked at it using the ellipse command
24 when we ran ellipses, when we do the bands, red versus
25 infrared or red versus green band, we tried to see if
163
1 that would make it unique, and that's -- basically that
2 process drove it to the 19 classes.
3 Q. All right.
4 Now, if -- could you look at, pull out for
5 a moment No. 8, yeah. There it is, right in front of
6 you. And I'd like you to go to Bates Page 1220027.
7 Is that an original one that you have with
8 you?
9 MR. CESARANO: That might be mine.
10 THE WITNESS: That's not the original.
11 Nobody has the original except for me.
12 BY MR. KOBELINSKI:
13 Q. Okay.
14 Let me just -- we have a copy. Have you
15 experienced -- I mean you've got copies of this before,
16 but notice how the colors aren't coming out exactly the
17 same. Is that just a function of the copying?
18 A. Yeah.
19 Q. As far as -- have you experienced that?
20 A. I think so.
21 Q. Okay. All right. It's difficult for us
22 to tell.
23 Looking at -- looking at the 1991 map on
24 this page, where is the brush here? Where did you see
25 brush?
164
1 A. Personally I don't ever look at my data
2 like this.
3 Q. Okay.
4 A. This is -- this for a final representation
5 of something to show because we can't give them a
6 computer, so they can say give me all brush out of
7 this.
8 Q. What, for instance, if you're looking at
9 the actual photo digital set, you just type in brush
10 and it comes on screen without any other colors?
11 A. Basically that's pretty much it, yeah.
12 Q. Okay.
13 To your recollection then, since we don't
14 have that capability right here, on the 1991 map, as I
15 understand this, the 1991 map is based upon your WCA-2A
16 paper; is that correct?
17 A. Right.
18 Q. Looking at the southern end of WCA-2A, is
19 that cattail or is it that brush do you recall?
20 A. On this map it's it's cattail.
21 Q. All right.
22 And if I understand your analysis of your
23 2A paper, that carries approximately an 80.9 or 80
24 percent accuracy of that being cattail all down there?
25 A. No.
165
1 Q. All right. Then I don't understand.
2 What accuracy would there be with regard
3 to cattail in that area?
4 A. Why don't we use the map that I generated
5 as opposed to the one that Jensen generated, because I
6 don't want to speak for his map.
7 Q. All right. Yours has 12 colors, it's
8 harder to see. All right.
9 Well, using yours, which would be on
10 page -- hold on, let me get to that. That's on Rutchey
11 Exhibit 1, and I believe you're directing us to your
12 Page 18, Bates Page 1220436?
13 A. That's correct.
14 Q. All right.
15 Now, are you able to tell from this --
16 again the lower half of that, that's cattail?
17 A. That's what that map is depicting.
18 Q. Okay.
19 And under the overall accuracy, if I
20 understand it correctly, that means that 80.9 or 81
21 percent accuracy of that cattail, that lower half?
22 A. No.
23 Q. What is the accuracy of that cattail?
24 A. Okay.
25 Let's go back to this. I'm just going to
166
1 point something out here going from the 20 class map to
2 the 12 class map. We put a stipulation in here and in
3 our -- in the authors' view that there is on Page 9,
4 page nine, third paragraph down.
5 Q. Okay. All right.
6 But whether or not it's monotypic, there
7 is an 81 percent chance you have cattail in that area?
8 A. That's an 81 percent chance that if I go
9 for any pixel or location within this entire classified
10 image that it will be, correct.
11 Q. All right. Let me do it a different way.
12 Drawing your attention to a couple of
13 pages earlier than that, or is it perhaps just one page
14 earlier, your page 17 of Exhibit 1, Bates 1220435, this
15 actually has the cattail sparse and cattail moderate
16 indicated; is that correct?
17 A. That's correct.
18 Q. All right.
19 And if I understand correctly, what
20 percentage then chance is there this accurately
21 reflects the cattail that you found in the southern
22 half of 2A?
23 A. Well, if you look at user accuracy for the
24 12 class map, overall we have 72 percent so there's a
25 72 percent chance that if you go to an area that says
167
1 cattail that it's going to be correct in a 12 class
2 map.
3 Q. Okay. 12 class map. All right.
4 Is that 72 percent though influenced by
5 the overall accuracy of mapping the dense cattail?
6 A. I would say that that's possible.
7 Q. With regard to the transition zones where
8 you have sawgrass and cattail mixed -- and again, as I
9 understand it, this will not indicate whether we're
10 dealing with, you know, five percent sawgrass and 95
11 percent cattail or 95 percent sawgrass and 5 percent
12 cattail -- you have around a 60 percent chance of that
13 being accurate?
14 A. To which one now?
15 Q. I'm talking about the 20 class map. The
16 sawgrass and cattail mix. And we're talking dense or
17 sparse, either/or.
18 A. It appears to be about 62 for the dense
19 and 58 for the sparse.
20 Q. Okay.
21 Now the dense is located, let's see that's
22 No. 4. Is that, the dense located primarily, if I read
23 your map correctly, up below the 10 structures and then
24 a portion of it below the 7 structures?
25 A. Yes.
††††㈳††⁹潵爠浡瀠捯牲散瑬礬⁵瀠扥汯眠瑨攠瑲畣瑵牥猠慮搠瑨敮ഊഊ††††㈴††⁰潲瑩潮映楴敬潷⁴桥‷瑲畣瑵牥猿ഊഊ††††㈵††††††䄮††奥献ഊ†††††††††††††††††††††††††††††††††‱㘸ഊഊ††††‱††††††儮††佫慹
168
1 Q. Okay.
2 And now the No. 6 -- and again, I'm -- I
3 apologize, but I'm dealing only with colors as opposed
4 to on a computer I'd be doing something different. Do
5 I understand correctly with 6, with the
6 sawgrass/cattail/brush, that is primarily below the
7 S-10D and S-10C, and that then also again to the east
8 of the S-7?
9 A. (No response.)
10 Q. If I'm reading the colors correctly.
11 A. I would say it's below all three
12 structures in this region or here, the mid section
13 going from north to south, south of these structures
14 and then...
15 Q. And you're -- because there's some there,
16 is that what you're saying?
17 A. Now you said we're on sparse, right?
18 Q. I'm looking at No. 6, what I thought I had
19 mentioned was sawgrass/cattail/brush.
20 A. Okay. All right. Yes.
21 Q. Okay.
22 But -- so that's essentially, for the most
23 part, that's below S-10D and C and then to the right of
24 S-7?
25 A. Right.
†††′㐠††匭㜿ഊഊ††††㈵††††††䄮††剩杨琮ഊ†††††††††††††††††††††††††††††††††‱㘹ഊഊ††††‱††††††儮††佮⁴桯獥
169
1 Q. On those areas, if I read the description
2 correctly, are greater than or equal to 70 percent
3 sawgrass and the remainder is a mixture of cattail and
4 brush?
5 A. That's correct.
6 Q. And the gray which surrounds it goes then
7 further down, a mixture of sawgrass and cattail without
8 any means of determining the percentages. Did you
9 attempt to -- I mean, for instance, in going out and
10 doing your sampling, did you attempt to break apart or
11 get more specific in the sawgrass/cattail mixes?
12 A. As stated previously, I can't recollect if
13 the 31, if the additional 31, 32 additional classes
14 that we went out, whether that was --
15 Q. I didn't explain myself well. In the 20
16 classes you have sawgrass/cattail mix dense and
17 sawgrass/cattail mix sparse, these are 4 and 5.
18 A. Okay.
19 Q. But the dense and sparse actually refers
20 to the overall vegetative cover as opposed to the
21 percentages between cattail and sawgrass. Do you
22 understand what I'm saying?
23 A. No.
24 Q. Okay.
25 Well, the description on sawgrass/cattail
170
1 dense says vegetative cover excluding aquatics, is
2 greater than or equal to 60 percent.
3 A. Okay.
4 Q. But it does not get into a breakdown
5 between how much is cattail, how much is sawgrass.
6 A. Okay.
7 Q. All right.
8 So, for instance, on sawgrass/cattail
9 dense, if I was mapping sawgrass it would probably be
10 fair to include that as sawgrass if I look at change in
11 sawgrass, because for all we know, 85 percent of it
12 could be sawgrass with 15 percent cattail, or likewise,
13 if I was looking at mapping cattail I could do it the
14 other way, because I don't know if it's 85 percent
15 cattail or 15 percent sawgrass. Where would you fit it
16 in if you were going to analyze sawgrass and cattail
17 change?
18 A. Fit what in?
19 Q. That category, sawgrass/cattail mix.
20 A. It could be in this -- okay, dense, it's
21 greater than 60 percent and it's a composite of
22 sawgrass and Typha domingensis.
23 Q. Dense is just referring to percentage
24 cover, not percentage cover of cattail or percentage of
25 sawgrass?
171
1 A. That's correct.
2 Q. Just vegetative cover?
3 A. Right.
4 Q. Did you attempt at all to see if you
5 could, spectrally with the imagery, identify
6 classifications within the mix, sawgrass/cattail for
7 instance, greater than 50 percent sawgrass, less than
8 50 percent cattail?
9 A. No.
10 Q. You didn't attempt to do it at all, or
11 just it couldn't be done?
12 A. We used, on the original 129 ground truth
13 data sites, the information from that to determine
14 these classes. And I -- we didn't try to break it down
15 out by density function because, as you can see from
16 the 12 and 20 class map, when you start breaking things
17 out, single species or mixed species by how dense they
18 are, you know, you end up having more error.
19 Q. But, for instance, we were just looking --
20 that's what I tried to see, if we could draw any
21 conclusions in No. 6, the sawgrass/cattail/brush
22 mixture which if there's a large stand of -- it's in
23 the middle of sawgrass/cattail mixture, we know that's
24 over 70 percent sawgrass; right?
25 A. Yes.
††††㈳††⁴桥楤摬攠潦慷杲慳猯捡瑴慩氠浩硴畲攬⁷攠歮潷⁴桡琧猍ਠ†††′㐠††潶敲‷〠灥牣敮琠獡睧牡獳㬠物杨琿ഊഊ††††㈵††††††䄮††奥献ഊ†††††††††††††††††††††††††††††††††‱㜲ഊഊ††††‱††††††儮††佫慹
172
1 Q. Okay.
2 Where would I find the initial 30 classes?
3 Is that somewhere around? The computer originally
4 generated it, I think it was.
5 A. Yes, it's in digital.
6 Q. Okay.
7 Did you -- when the computer generated
8 that 30 it didn't say Class 1 sawgrass dense, I mean it
9 just says here are 30 classes, is that --
10 A. That's correct.
11 Q. All right.
12 Did you -- you went off to the field, you
13 look for these 30 sites, essentially try to pick 150,
14 you got down to 129. Did you literally end up with 30
15 specific classes you identified?
16 A. No.
17 Q. I don't think -- I'm maybe not explaining
18 myself. I guess the question is, did you at one point
19 actually have 30 defined classes which you then
20 narrowed to 19 plus the periphyton, or did you -- did
21 you just determine you could only identify 19 classes
22 then plus periphyton?
23 A. I did an unsupervised classification,
24 broke it down out into 30 clusters, used that ground
25 truth I used for analysis and applied it to the
173
1 supervised classification and came up with a final
2 classified map composed of 19 classes.
3 Q. Okay.
4 So there are 30 -- that's why I was,
5 perhaps when I was asking the question I kept saying 30
6 classes, you use clusters, now that's a distinction you
7 were drawing?
8 A. Yes. There weren't 30 classes, there were
9 30 clusters. Classes, clusters, the word can be used
10 interchangeably in the analysis.
11 Q. But 30 clusters weren't defined as, for
12 instance, those descriptions of what the classes are?
13 A. No.
14 Q. Okay.
15 Based upon your review of ellipses --
16 remember we talked about that -- do you recall whether
17 or not you can actually break apart these mixtures of
18 sawgrass/cattail whether you can break it down into --
19 into whether you have over 50 percent cattail, less
20 than 50 percent cattail sawgrass?
21 A. It's possible, but I just used the 129
22 original ground truth site information, that's what I
23 used. I didn't try to go in and break out density
24 functions within -- well, I take it back. I did try
25 that, but I found that's where a lot of error can occur
174
1 if you try to break out the density function of a
2 single class such as sawgrass dense, moderate, sparse.
3 Q. Okay. All right.
4 Looking at the map, the final map of the
5 12 classes which is on Page 18, again, I'm dealing with
6 colors, it's difficult to tell, tree island there
7 actually appears, I guess from both of our maps, there
8 are only a few communities up in northern 2A that are
9 still classified as tree islands?
10 A. Tree islands, right.
11 Q. And there's no way to -- well, I guess
12 there is. Where is Exhibit 1? This isn't a color one.
13 Well, the ellipsoid or teardrop shaped
14 areas that appear to be tree islands out in the middle
15 of 2A, those are no longer healthy tree islands?
16 A. I would say no.
17 Q. Are those primarily cattail or brush or
18 what are they?
19 A. I would say they're composed of tree
20 island, sawgrass brush mixture 1, where if we were
21 going to use the 12 class map we'll -- you'd say
22 sawgrass brush, possible tree island and those are the
23 two major ones.
24 Q. Okay.
25 Looking at Exhibit 8 for a moment, I
175
1 apologize for jumping back and forth, but your work was
2 the foundation for 8, so I need to at times go back and
3 forth to understand what was done.
4 On Page 20, which is Bates No. 1220020,
5 there is a calculation by hectares of the brush,
6 brush/cattail, cattail/sawgrass and sawgrass/cattail.
7 Is that something you also did in relation to your
8 mapping of of 2A?
9 A. Yes, basically it's Table 2, page 20.
10 Q. On Exhibit 1?
11 A. Exhibit 1.
12 Q. But that --
13 A. The numbers in parentheses represent
14 hectares.
15 Q. Thank you. All right.
16 I need to switch gears for just a moment
17 with regard to the Exhibit 8. And I believe you had
18 stated you had provided Dr. Jensen with the five
19 digital data sets for the five maps, okay.
20 Did you ever do any analysis of those maps
21 to determine vegetative mapping from them?
22 A. Just the 1991, 8-10-91, my data set.
23 Q. Okay.
24 Did you ever review his work to determine
25 its accuracy or understand it?
176
1 A. No.
2 Q. Okay.
3 Have you ever worked with the LandSat MSS
4 imagery?
5 A. Very briefly.
6 Q. Okay. When was that?
7 A. Back initially, probably 1987, when I
8 first started looking at remotely sensed data sets I
9 just basically was looking at it, trying different,
10 various image processing techniques on it.
11 Q. Okay.
12 And am I correct you selected the SPOT
13 imagery over the LandSat MSS for a reason?
14 A. Yes.
15 Q. What was that reason?
16 A. Better resolution.
17 Q. With a 20 by 20 as compared to 79 meters
18 by a 79 meter picture?
19 A. That's correct.
20 Q. You had mentioned that during the period
21 of time, '81 through '85, you had done some vegetative
22 mapping with using helicopters. Did you ever -- I
23 believe you stated that it was contained primarily in a
24 report out here. Have you ever attempted to compare
25 your mapping from the helicopter to the maps in the --
177
1 in your report which is Exhibit 8?
2 A. No.
3 Q. Okay.
4 (Thereupon, a discussion was held off the
5 record.)
6 BY MR. KOBELINSKI:
7 Q. Just a few questions from my notes with
8 regard to the data you sent to Dr. Jensen.
9 You stated you sent to him the data from
10 the 129 site visits or site investigations. Did you
11 also send him the data from the 241 field sites?
12 A. Yes.
13 Q. Okay. There was perhaps an inaccuracy in
14 my notes.
15 Now, this is sort of a broad question, but
16 without going into it heavily, in Exhibit 5, in the
17 laboratory notebook we were looking at -- and there's
18 really no need to pull it out, but how -- the 9 pixel
19 often had more than one vegetative class in the pixel.
20 However, when you went and produced a map similar to,
21 for instance, the one on Page 17 of Exhibit 1, Bates
22 No. 1200435, you don't see that much in the way of
23 variation. There is some, but how does the computer,
24 when it's doing colors, just take a predominant, like
25 within a, for instance, a 9 pixel, how does it do that,
178
1 create the map?
2 A. The final -- all that is going through an
3 algorithm. It's a -- it's a smoothing out of the
4 entire -- it takes out the salt and pepper effect so
5 that visually it looks pleasing, as you know, if you --
6 Q. As opposed to just a scattering of colors,
7 little dots?
8 A. Right, right. And that's perfectly
9 acceptable, that's what everyone does. It's -- we
10 still have the original digital data.
11 Q. Okay.
12 I would assume then that when you
13 calculated the hectares or hectares or whatever the
14 correct pronunciation is of each category, that was
15 based upon the 20 by 20 pixels, not the smoothed or how
16 do you --
17 A. It's based on the unsmooth data, right.
18 Q. Okay. All right.
19 With regard to -- and perhaps this is
20 going off the 2A paper, but with regard to Exhibit 8,
21 Dr. Jensen's collaboration with you, who selected the
22 four additional dates to look at?
23 A (No response.)
24 Q. You had the data, I believe you testified,
25 but who went back and selected those dates?
179
1 A. I believe he did. It wasn't me.
2 Q. Okay.
3 You didn't participate in that selection?
4 A. No.
5 Q. Okay.
6 As a follow-up on the -- on the GPS, when
7 I believe you went -- said you went to a number of
8 known benchmarks, is that what they're called, or known
9 stations with known positioning accuracy; is that
10 correct?
11 A. Right.
12 Q. When you did that would you hover or would
13 you land at those sites?
14 A. Those sites we actually went to physically
15 on the ground.
16 Q. Okay.
17 And in the helicopter would you always
18 hover or would you land?
19 A. I would say that we didn't do it every
20 single trip.
21 Q. You lost me.
22 A. We didn't actually go to a known control
23 every single trip out of all the trips we took. It was
24 intermittently throughout the process.
25 Q. When you were using a helicopter and
180
1 visiting 129 sites or, for instance, 241 sites, did
2 you, in the helicopter, would you ever land or would
3 you always just hover over the site?
4 A. Hover.
5 Q. Drawing your attention to Exhibit 1, Page
6 9 and if you see at the bottom there it says
7 interpretation of final maps, I want to draw your
8 attention immediately preceding that there is a
9 discussion there with the regard to the -- the
10 difficulties caused by the bright signature of floating
11 periphyton. And you conclude that there may be
12 infrequent times when periphyton is either absent or
13 submerged which would mask this problem.
14 Are any particular times you had in mind
15 that would be -- would allow for that problem to be
16 avoided or perhaps lessened?
17 A. Oh, okay.
18 Q. I didn't let you get to the page, did I?
19 A. Yeah. I would say in South Florida
20 periphyton tends to float in the summer months, and in
21 the winter, depending on hydrology of the area, could
22 be totally absent if there -- if the area gets pretty
23 dry, even then though it could, if it goes dry, it
24 forms almost like a calcareous, crusty surface that's
25 white in appearance.
181
1 Q. Would winter months then be a -- perhaps a
2 time when you would have less of the brightness from
3 the periphyton?
4 A. It's possible. If there's water on,
5 covering it, most of the area, then it would be at
6 the -- most likely at the bottom. Not too much of it
7 floats up in the winter months. But the drawback of
8 collecting satellite imagery in the middle of winter is
9 that vegetation isn't at its best as far as
10 differentiation. It will sometimes, if we have a
11 freeze, it tend to all brown out and look the same.
12 Whereas, in spring, and up almost through early fall,
13 you have a unique differentiation in spectral
14 characteristics of vegetation, it's at its most robust
15 growing stages.
16 Q. Are there particular times of the year
17 then that for different vegetative classes it would be
18 better to look at?
19 A. In my opinion, from spring to early fall
20 is the best.
21 Q. For all classes?
22 A. Yes.
23 Q. Other than periphyton?
24 A. Yeah.
25 Q. Okay.
182
1 A. Periphyton is not really a vegetation,
2 it's like an algae.
3 Q. But periphyton can mask during that period
4 is what you were saying?
5 A. Yes.
6 Q. Okay.
7 Spring meaning what?
8 A. I'd say, depending on the previous winter,
9 it's going to depend on freezes, it would be as early
10 as March and go as late as, again depending how cold
11 fronts move through, it gets as late as September
12 November.
13 Q. Okay.
14 And would that hold true with regard to
15 cattail and sawgrass then, that during the -- for
16 instance, you did August, would that be about the best
17 time, in your opinion, for distinguishing cattail and
18 sawgrass?
19 A. I would say in that range. If you look at
20 all the preceding data the hydrology and cold fronts.
21 Q. Okay.
22 A. That happens to be worse time of the year
23 to, also to acquire satellite data in South Florida.
24 Q. Because of the cloud cover?
25 A. That's right.
183
1 Q. Having done this, is there a particular
2 class of vegetation that you believe is the easiest to
3 map in the Everglades?
4 A. Well, I looked at it as a total project,
5 mapping the vegetation. I wasn't looking for any one
6 particular species to map. I wanted a clear picture of
7 what was there on the ground before I started doing
8 this work. And we did just look at specifying things
9 like cattail. This is like back when I first started
10 learning image processing, when we were doing it and we
11 would try to pick out one specific species. People do
12 that. But on this project I looked at everything as
13 equal and I just wanted to do a realistic
14 representation of what was on the ground.
15 Q. Okay.
16 Well, see if we are doing this clever
17 technique, looking at the diagonal, looking at the 12
18 category map, looking at Page 21 of Exhibit 1, Bates
19 No. 1220439. I have to add up these numbers again.
20 If you add up on this wonderful diagonal
21 all the numbers, you come up with 195. Which, of
22 course, then if you divide 241 you actually come up
23 with 80.91 percent, which is exactly what you have,
24 80.91 percent. Now, if I exclude Category 1, in other
25 words, your mapping of sawgrass from that, I take 195
184
1 correct identifications, subtract out 99 which leaves
2 me with 96 correct identifications, and then divide
3 by -- I assume I divide by 140; is that correct? 241
4 less 101 or do I -- is it 241 less 99?
5 A. You'll have to tell me, because I wouldn't
6 do this.
7 Q. Okay.
8 Well, I'm just trying to look at your
9 overall accuracy for all categories, for all categories
10 but sawgrass. How would you tell me --
11 A. For sawgrass?
12 Q. No, for everything but sawgrass. I want
13 to know what the accuracy is.
14 A. I don't think I can tell you just looking
15 at this data that's provided.
16 Q. Well, why not? What else would you need
17 to know?
18 A. Because this data is looking at the whole
19 entire area, I mean not looking at one specific
20 species.
21 Q. Well, user accuracy, you've used simple
22 math to tell me it's 70 percent accurate at mapping
23 sawgrass/cattail and 72 percent at cattail, and that's
24 again just using simple math. I'm just trying to, if
25 you take out sawgrass, determine what would be the
185
1 overall accuracy for all other species.
2 A. I don't know how to determine that.
3 Q. Couldn't you just take out the sawgrass
4 category?
5 A. I'm not sure.
6 Q. Well, let me ask you this, if your
7 instruction was to map every category but Category 1,
8 would you come up with the same map, just with a lot of
9 blanks in there?
10 A. How would I map everything but Area 1?
11 How would I know how to mask it out in the imagery if I
12 didn't know what it was?
13 Q. Well, you can identify it, but I'm saying
14 once you identified it you can mask it out. Could you
15 go to the computer and ask it to give you a map of
16 Categories 2 through 12?
17 A. Yes, that is after the fact, after I've
18 classified it.
19 Q. Yes, I understand that, okay.
20 And once you have that map, could you
21 determine what the accuracy of those categories are?
22 A. I'm not -- I'm not sure.
23 Q. Okay.
24 A. This gets into some pretty heavy
25 statistics. I'd have to go back and read. It might --
186
1 it might be simple statistics, but I'm not going to
2 guess.
3 Q. Well, I'm not asking you to guess, but I'm
4 just saying how would you go about determining accuracy
5 of your mapping of Category 2?
6 A. On the 12 class map?
7 Q. Um-hum.
8 A. My accuracy?
9 Q. Um-hum.
10 A. I would say it's 70 percent.
11 Q. But you did that by saying that out of 50
12 sample you were able to -- 35 were mapped correctly; is
13 that right?
14 A. Right.
15 Q. You just divided it out?
16 A. Right.
17 Q. Okay.
18 What about with regard to Category 5?
19 What's the accuracy?
20 A. 72.
21 Q. Okay.
22 You did that based upon out of 36 samples
23 26 were correct; is that correct?
24 A. That's correct.
25 Q. All right.
187
1 Now with regard to overall accuracy of all
2 12 categories, that's 80.9 percent; is that correct?
3 A. That's correct.
4 Q. And you did that by adding up all the
5 correct hits which were 195 and dividing by 241?
6 A. That's correct.
7 Q. All right.
8 Now if I want to know just Categories 2
9 through 12, wouldn't I follow the exact same thing
10 you've done following the rows,, et cetera, just add up
11 everything but Category 1 and divide by the number of
12 attempts?
13 A. I'm not sure. I'd have to go back and
14 read.
15 Q. Well, what you would read?
16 A. I want to know what's the legitimate thing
17 to do. You're telling me. You're not an expert, so I
18 don't know if your way of looking at is correct. And I
19 don't know in my head if it's correct, so I'm not going
20 to say that you're correct.
21 Q. Okay. All right. Fair enough.
22 If you reclassed 2 through 12 as one class
23 similar to your reclassifying the 20 classes into 12
24 classes, all right, so now you end up with a 2 class
25 map --
188
1 A. Now say that one more time.
2 Q. You reduced your 20 classes into 12
3 classes; is that correct, by adjoining certain classes?
4 A. That's correct.
5 Q. What if we reclass a 12 class map into 2
6 classes, one being Category 1 and other one being
7 Category 2, and Category 2 being what is currently
8 Categories 2 through 12. Would you then be able to
9 determine the user accuracy for your new Category 2?
10 A. I wouldn't do that.
11 Q. Well, you did 12 from 20.
12 A. Because it made sense to take like species
13 or same species that have different densities within
14 the environment and clump it together. It doesn't make
15 sense to take different species of vegetation and clump
16 them together. That just doesn't make sense to do
17 that. Why would I do that?
18 Q. Well, why did you start with 20 classes
19 instead of 12 if there weren't differences between the
20 classes?
21 A. Because I worked with all the ground truth
22 information I had available to me and came up with the
23 initial 19 and added one additional, came up with 20
24 class map.
25 Q. Okay.
189
1 Then why reduce down to 12? If you saw
2 distinctions between 20 different classes why did you
3 just go down to 12?
4 A. Because I wanted to see if the error that
5 was associated, the margin of error that was associated
6 in the 20 class map was a function of trying to break
7 out the densities, characteristics of individual
8 species such as sawgrass, dense, moderate and sparse.
9 I wanted know from that the problem. That, in my mind
10 that made sense, because it's all one species and is
11 the satellite imagery able to tell that boundary of
12 density within that species and --
13 Q. But sawgrass dense, moderate, and sparse
14 had your highest accuracy in the 20 class map?
15 A. That's correct.
16 Q. Well, then, how could that account for
17 your inaccuracy -- let me rephrase, maybe I
18 misunderstand what you were saying.
19 Why would you collapse down sawgrass if it
20 has the highest accuracy in 20 classes?
21 A. I didn't just do it for sawgrass. I did
22 cattail, sawgrass/cattail mixture, whatever the other
23 ones that are shown on Page 20 of Exhibit 2.
24 Q. Okay.
25 Well, look at the 20 class map which is
190
1 the Category 1, it had sawgrass dense, 91 percent
2 density; is that correct?
3 MR. CESARANO: The 12 or the 20?
4 BY MR. KOBELINSKI:
5 Q. The 20.
6 A. The 20.
7 Q. Yeah, the overall accuracy is better than
8 12 class. I'm just talking about overall accuracy of
9 the 12 class map.
10 A. I don't think you can look at it that way.
11 Q. Okay. Why not?
12 A. I'd have to read about that. You're --
13 you're comparing a single species classification
14 accuracy to the overall accuracy of the collapsed map
15 of the same original work. You're trying to make a
16 comparison, it doesn't make sense to me.
17 Q. All right.
18 Would it have improved the overall
19 accuracy of the map if you had combined categories,
20 original Categories 4 and 5, which included sawgrass,
21 with sawgrass in the 12 category map?
22 A. I thought we broke it out based on things
23 looking similar. And sawgrass is unique and we
24 considered sawgrass/cattail to be unique. And when we
25 collapsed them we kept all the sawgrass together and
191
1 when we collapsed sawgrass/cattail we kept it
2 altogether.
3 Q. Well, I understand that, but you also
4 thought it was unique when you prepared the map using
5 20 categories, didn't you?
6 A. Yes, it was unique.
7 Q. But you had three different categories for
8 sawgrass, sawgrass/cattail dense, sparse and then
9 sawgrass/brush, all three of these being unique
10 categories.
11 A. That's correct.
12 Q. All right.
13 Then I'm at a little bit of a loss as to
14 why collapse them if they're unique.
15 A. To see -- I'll answer this again, this has
16 already been asked and answered, but I wanted to see if
17 the inaccuracies that the 20, original 20 class map was
18 trying to break out density functions within a single
19 species such as sawgrass dense, moderate and sparse.
20 And that was where most of my error was associated
21 with.
22 Q. Okay.
23 But you didn't just break out the single
24 species. I guess that's where I'm getting confused,
25 because you also combined other species into one class,
192
1 for instance sawgrass/cattail mixture is combined three
2 different ones which include brush.
3 A. But it's also all similar. So sawgrass
4 dense, moderate, and sparse goes to 1, sawgrass,
5 sawgrass/cattail mixture dense and sparse and
6 sawgrass/cattail/brush mixture go to sawgrass/cattail.
7 It makes sense. I don't understand what your problem
8 is at this point.
9 Q. Okay.
10 Well, what did you conclude then going
11 from 20 to 12 was the inaccuracy of the map due to
12 attempts to distinguish between dense and sparse
13 vegetation?
14 A. Overall accuracy was appreciably increased
15 by 10 percent. But, I have to go back and read this.
16 When you look at KAPPA statistics there's a test that
17 you can see if there's a significant difference between
18 the two maps.
19 MR. CESARANO: Page 7 or 6?
20 THE WITNESS: It's a pairwise test. I'm
21 going read it almost word to word. If you look
22 at Page 6, at the bottom, last two paragraphs,
23 and then top of the Page 7 that explains what
24 I'm -- if you want me to read it, I'll read it.
25
193
1 BY MR. KOBELINSKI:
2 Q. Okay.
3 Is your conclusion that there is a
4 difference, statistical difference between the two?
5 A. No.
6 Q. All right.
7 I guess it's no significant difference?
8 A. That's correct.
9 Q. All right. Okay.
10 I guess my question then is, why did you
11 include the 12 vegetation map?
12 A. I thought it was a matter of interest to
13 people who were interested in this kind of research to
14 see the method that I used to determine that this
15 overall process of looking at the two maps and
16 considering the density function within a species might
17 have been a cause for my error in my initial map. I
18 wanted to show the steps I went through to try to show
19 what I did and I think it was -- well, it got accepted
20 for publication, so I think they thought it was
21 interesting.
22 Q. Is it based upon though -- that there's no
23 statistical difference, does that mean then the 20
24 category map is as accurate as 12 category map?
25 A. I would say that's a fair statement.
194
1 Q. Okay.
2 Why didn't you use the 20 categories to
3 Exhibit 8, the -- Dr. Jensen's study with you?
4 A. You're going to have to ask Dr. Jensen.
5 Q. Let us move on to another area.
6 (Thereupon, Mr. Birch left the room.)
7 BY MR. KOBELINSKI:
8 Q. As I understand from this morning's
9 testimony, you've also been involved in vegetative
10 mapping of the Holeyland; is that correct?
11 A. Yes.
12 Q. When did you start doing that?
13 A. I think initially when we acquired the May
14 1992 SPOT satellite imagery.
15 Q. Had you done any other vegetative mapping
16 or analysis of the Holeyland prior to your obtaining
17 the SPOT imagery?
18 A. No.
19 Q. Let me look through my notes so I don't go
20 over it.
21 According to my notes I don't think we
22 really discussed the Holeyland, so I'm -- I'm trying
23 not to repeat myself, but after obtaining the May 1992
24 SPOT imagery, what did you do with regard to the
25 Holeyland?
195
1 A. I did some very preliminary data analysis
2 of the imagery. I rectified the imagery. I broke
3 something out of the imagery through that preliminary
4 process and then I just sort of stopped at that point.
5 Q. Okay.
6 Are you still attempting to map the
7 Holeyland?
8 A. I've heard recently that I might be put
9 back on that.
10 Q. Did you actually produce any type of
11 Holeyland vegetative maps?
12 A. Just -- just depicting cattail.
13 Q. Okay.
14 What are those maps from?
15 A. Actual field reconnaissance using GPS.
16 Q. Is this in relation to the May 92 SPOT
17 imagery?
18 A. No.
19 Q. Okay.
20 When did you do the mapping that you
21 produced of the Holeyland?
22 A. I don't have an exact date, but it's -- I
23 think our effort was in '92 or early '93.
24 Q. Would that be late '92, early '93?
25 A. I think so. I'm not positive.
196
1 Q. Okay.
2 You had identified this packet which is
3 Bates Nos. 1202773 through 2848 as part of the
4 Holeyland data. Is that related to your SPOT analysis
5 of the Holeyland?
6 A. Yes, this is -- this is the preliminary
7 work that we did.
8 Q. Okay.
9 Now that's different than your mapping
10 just using GPS?
11 A. That's right.
12 (Thereupon, a discussion was held off the
13 record.)
14 BY MR. KOBELINSKI:
15 Q. Did you also produce documents with regard
16 to Holeyland mapping outside of SPOT imagery?
17 A. Yes.
18 Q. Is that something that you identified this
19 morning?
20 A. I'm not sure.
21 Q. All right.
22 MR. KOBELINSKI: Mark that as 9.
23 (Thereupon, the document was marked
24 Rutchey Exb. No. 9 for Identification.)
25
197
1 BY MR. KOBELINSKI:
2 Q. All right.
3 Mr. Rutchey, I'm showing you what's been
4 marked as Rutchey Exhibit No. 9, which is a number of
5 Bates stamped documents bearing 1220441 through
6 1220583, and the first page of which has, appears to
7 again be a file folder stating Holeyland.
8 Why don't we have you put all this other
9 stuff we have in front of -- I guess that's yours.
10 (Thereupon, a discussion was held off the
11 record.)
12 BY MR. KOBELINSKI:
13 Q. And if you could go through this and tell
14 me -- well, just identify it for me, if you would, is
15 this -- are these documents related to your Holeyland
16 mapping just using GPS?
17 A. Yeah, I would say this is.
18 Q. All right.
19 I note that you've pulled out a number of
20 documents, those documents that you pulled out are not
21 part of that mapping effort?
22 A. (Shakes head side to side.)
23 Q. Is that right?
24 A. Yeah.
25 Q. Let me quickly read off the Bates numbers.
198
1 First item you've pulled out bears Bates
2 Nos. 1220491, through 494. And this is a December 6,
3 1993 memorandum from Garth Redfield. There is a single
4 page, Bates No. 1220465 and is a scale indicator of
5 WCA-2A vegetation. I assume this is part of that scale
6 analysis you're doing with Obeysekera?
7 A. That's correct.
8 Q. Next is Bates Page 1220526 and 1220527
9 which states Holeyland soils, 1990 and 1993.
10 Did you ever do a comparison of the
11 vegetation with the Holeyland soils?
12 A. No.
13 Q. Okay.
14 Do you know if anyone else has?
15 A. Yes.
16 Q. Who has done that?
17 A. It's -- it's being looked at, it has been
18 looked at and it's still being looked at by a number of
19 people.
20 Q. Okay.
21 Anyone within the District?
22 A. Yes.
23 Q. Who?
24 A. Sue Mewman, Tom Fontaine, me.
25 Q. Excuse me?
199
1 A. Ken Rutchey.
2 Q. Oh, I thought you said Mike --
3 A. Marie Pietrucha.
4 Q. Can I ask you to spell the last name?
5 A. I'll guess it. I should have said no.
6 P I E T R U C H A.
7 Q. Okay. Anyone else?
8 A. There are other people, but I don't know
9 their names.
10 Q. Anyone outside of the District that you're
11 aware of doing that?
12 A. Yes.
13 Q. Who?
14 A. I don't know their names.
15 Q. Okay.
16 You mentioned that David Lane was looking
17 at transects below the S-10's. Do you know if he's
18 looking at the Holeyland?
19 A. No.
20 Q. Okay.
21 And you say -- are you still currently
22 doing this?
23 A. Yes.
24 Q. Okay.
25 And where does -- where does that project
200
1 stand?
2 A. Parts of the paper have been written. And
3 basically we're in the data analysis part of the paper
4 at this point.
5 Q. Are you working with someone on that
6 project?
7 A. Yes.
8 Q. That is?
9 A. The people I just mentioned.
10 Q. Okay.
11 Sue Newman, Tom Fontaine, Marie Pietrucha
12 and yourself?
13 A. And I guess there's one other person,
14 George Shih is being involved too.
15 Q. Do you have a draft of that paper already?
16 A. I think that you have a draft introduction
17 to the paper and that's all that's been -- I don't know
18 if I saw that here. It's not in this pile. It's not
19 in this pile.
20 Q. All right. We'll try to take a look for
21 it on the break.
22 What are divisional responsibilities
23 between -- the division of responsibilities between the
24 five of you with regard to that particular project?
25 A. What is the division? I think Tom
201
1 Fontaine has taken the lead. Sue Newman's doing a lot
2 of data analysis and looking at all aspects of the
3 project, all components and George Shih is also helping
4 in that effort. And there's also, my latest
5 understanding is that there's a person outside the
6 District that's taking a look at that data.
7 Q. Do you know who that is?
8 A. I don't know his name.
9 Q. Okay.
10 A. My part, I wrote the introduction to the
11 paper. And I've put a lot of the data sets together
12 there that are being used in the analysis. And I'll
13 probably be involved in looking at the results. The
14 discussion session and conclusion, I believe, probably
15 all of us will be looking at them. Marie's mainly
16 looking at GIS aspects of it, making maps.
17 Q. Okay.
18 What are the components to the project?
19 A. Water quality, soil phosphorus or soil
20 chemical constituent data, vegetation, muck fires,
21 water supply and that's hydrology.
22 Q. With regard to vegetation, are you looking
23 at all vegetative types or just some specific species?
24 A. At this point it's just been specifically
25 cattail.
202
1 Q. Okay. Okay.
2 This all started with the two page
3 document 122526 and 527 which you have excluded. This
4 has soil P as part of the data you're using for that
5 project?
6 A. Yes.
7 Q. Okay. All right.
8 And the next document you've taken out, or
9 two documents -- no, the document you've taken out of
10 that Holeyland pack is Bates No. 1223537 which is
11 Holeyland water quality monitoring network. All right.
12 That and that next one you've taken out is
13 Bates No. 122576 and then you have excluded 1220556
14 through 1220572, which are all in a file called
15 Holeyland vegetation map. What are these vegetation
16 maps?
17 A. I have them in the file, I got them from
18 someone, probably Dewey. I think this is an attempt by
19 the Department of Transportation to use, I believe it
20 was using MS data, LandSat data MS in December of 1982.
21 That's all I know about it.
22 Q. You've not used these for anything?
23 A. No.
24 Q. All right.
25 Then we're listing as Rutchey Exhibit No.
203
1 9, Bates No. 1224441 through 1220583 excluding those
2 Bates Nos. I just pulled out as documents that truly do
3 not belong in this file. Is that accurate, Mr.
4 Rutchey?
5 A. Your initial question was to pull out the
6 files that didn't pertain to the mapping of cattail.
7 Q. Okay.
8 A. And I pulled those out.
9 Q. Okay. All right.
10 These were there for a different purpose
11 then?
12 A. Yes.
13 Q. That being?
14 A. That they were part of the overall
15 Holeyland project.
16 Q. Okay. All right.
17 Well, for purposes -- I guess we'll just
18 include them then. I'll just put them at the bottom,
19 we can take care of the Bates Nos..
20 With regard to the Holeyland vegetation
21 project, given your work with the WCA-2A SPOT image of
22 2A, did you attempt to just use the classifications you
23 used there to map out the Holeyland?
24 A. You'd have to look at every scene that you
25 get. It's a unique scene. You cannot apply the
204
1 statistical information from one satellite imagery to
2 another. That depends, but...
3 Q. Well, do you have the same type vegetative
4 communities?
5 A. I would say there's similarities, yes. I
6 consider it to be a disturbed Everglades environment.
7 Q. Would you consider 2A to be a disturbed
8 Everglades environment?
9 A. Would I? I consider 2A to have an
10 impacted zone and a nonimpacted zone.
11 Q. Okay.
12 In the impacted zone, would you consider
13 that disturbed, since that's the term you used for the
14 Holeyland?
15 A. Yes.
16 Q. Okay.
17 And do you find -- you find sawgrass in
18 the disturbed zone in 2A; is that correct?
19 A. Yes.
20 Q. Okay.
21 And do you find sawgrass in the Holeyland?
22 A. Yes.
23 Q. Cattail in the disturbed zone in 2A?
24 A. Yes.
25 Q. Cattail in the Holeyland?
205
1 A. Yes.
2 Q. Sawgrass/cattail mixed in the 2A?
3 A. Yes.
4 Q. Mixed in the Holeyland?
5 A. Yes.
6 Q. Are there any vegetative communities that
7 you find in the Holeyland that are not in 2A?
8 A. Oh, I'd have to go -- I wouldn't say
9 there's what I would consider to be tree islands in the
10 Holeyland and polignum/brush mixture.
11 Q. You would not find that in the Holeyland?
12 A. Well, you might find it, but I don't
13 recall seeing it.
14 Q. Okay.
15 Is there anything in the Holeyland that
16 you wouldn't find in 2A?
17 A. Well, not that I -- they have typical
18 Everglades vegetation.
19 Q. Okay.
20 Then, given all of your work with ground
21 truthing that you did in 2A, could you use just SPOT
22 imagery and statistical analysis to just create cattail
23 map for the Holeyland?
24 A. No.
25 Q. Why not?
†⁎漮ഊഊ††††㈵††††††儮††坨礠湯琿ഊ†††††††††††††††††††††††††††††††††′〶ഊഊ††††‱††††††䄮††䉥捡畳攠
206
1 A. Because every satellite image that is
2 acquired, that satellite imagery acquired from space is
3 on a different day and every day has unique qualities
4 about it. The atmosphere, the angle of the acquired
5 satellite imagery. There's just a number of things
6 that could come into play, you cannot apply it, it just
7 isn't done.
8 Q. Well, okay. Can't you correct for the
9 atmosphere, the angle and all that?
10 A. You can do that over time if the areas are
11 in the same location, but not -- not separate
12 geographically. We're talking about two different
13 areas from each other.
14 Q. Okay.
15 Why would that matter?
16 A. Well, I'm not -- this is all technique in
17 change detection. It's not my expertise, and I'm not
18 going to get into all the -- my opinions of what.
19 Q. Can you try and do it?
20 A. No. You can't. You shouldn't do it. You
21 can't do that.
22 Q. All right.
23 Then looking at Exhibit 9, how did you do
24 your mapping?
25 MR. CESARANO: That's 9.
207
1 THE WITNESS: Okay.
2 BY MR. KOBELINSKI:
3 Q. I'm sorry, Composite 9.
4 A. Basically we used the original May 1992
5 satellite image. All we did was a preliminary analysis
6 of that data. We did not make a final vegetation map
7 from that. Basically what we did, and we're -- the
8 only one date of this whole cattail mapping project was
9 1992, we went out in the field, it was Les Vilchek,
10 myself and Sue Newman. And we encircled areas with the
11 GPS unit within the Holeyland and got basically their
12 boundary and we each independently estimated percentage
13 cover that we saw in a boundary. And then we compared
14 our results. I don't think we have deviated from each
15 other more than 10 percent. So from that we took an
16 average and came up with a percentage cattail within
17 each of those zones.
18 Q. Okay.
19 I see that you are looking at one of the
20 first documents in Exhibit 9, and that bearing Bates
21 No. 1200442 through 446, which is a memo from yourself,
22 Les Vilchek, Everglades Systems Research Division, to
23 Maura Merkal dated October 21, 1992; is that correct?
24 A. That's correct.
25 Q. All right.
208
1 Now, it's -- I'm looking at Bates page
2 1220445 of this Holeyland cattail distribution in 1992,
3 and G has 0.75 percent cattails, that's a large area
4 there?
5 A. Um-hum.
6 Q. What is that? I mean how does it equate,
7 for instance, to -- I don't know, number of cattails or
8 percentage of cattail within a percent acre, but number
9 of cattail within an acre, just a couple of plants?
10 What -- what does that mean?
11 A. The last number there in the table tells
12 us.
13 Q. Okay.
14 I'm just trying to visualize what .75 of 1
15 percent cattail means. Am I looking at an acre, a
16 representative acre, with just a couple sporadic
17 plants, is it just a different type of growth pattern
18 out there which resulted in this?
19 A. You would -- I mean you want to know how
20 much an acre, what --
21 Q. What would I see in area G?
22 A. I would take 43,560 square feet and
23 multiply it by, you know, the point -- I guess it would
24 be .0075, and that would be your answer of square feet
25 that you'd find in an acre.
209
1 Q. Would I find it though just interspersed
2 everywhere, individual plants?
3 A. Yeah.
4 Q. There wasn't, for instance, a growth --
5 clumps of cattail just interspersed?
6 A. That's possible too, but it wasn't big
7 enough to happen such as those other polygons. I mean
8 we're not mapping something less than significant.
9 We're not going to go after the clump of cattail and
10 try to delineate it. In this area it was pretty sparse
11 so we just --
12 Q. All right.
13 A. I'll estimate we -- I think we all came up
14 with like 1 percent or something.
15 Q. Less than one percent is essentially what
16 you called it?
17 A. I think we used the figure that Game and
18 Fish used in the previous mapping.
19 Q. Okay. All right.
20 A. Which map are we looking at?
21 Q. I was looking at this within Bates No.
22 1200445.