Legislature(2007 - 2008)BARNES 124
02/13/2008 08:30 AM House FISHERIES
| Audio | Topic |
|---|---|
| Start | |
| Overview: Alaska Department of Fish & Game Camera Sled Research/habitat Assessment/seabed Mapping Program | |
| Adjourn |
* first hearing in first committee of referral
+ teleconferenced
= bill was previously heard/scheduled
+ teleconferenced
= bill was previously heard/scheduled
| + | TELECONFERENCED | ||
ALASKA STATE LEGISLATURE
HOUSE SPECIAL COMMITTEE ON FISHERIES
February 13, 2008
8:43 a.m.
MEMBERS PRESENT
Representative Paul Seaton, Chair
Representative Craig Johnson
Representative Gabrielle LeDoux
Representative Peggy Wilson
Representative Bryce Edgmon
MEMBERS ABSENT
Representative Kyle Johansen
Representative Lindsey Holmes
COMMITTEE CALENDAR
OVERVIEW: ALASKA DEPARTMENT OF FISH & GAME CAMERA SLED
RESEARCH/HABITAT ASSESSMENT/SEABED MAPPING PROGRAM
- HEARD
PREVIOUS COMMITTEE ACTION
No previous action to record
WITNESS REGISTER
GREGG ROSENKRANZ
Alaska Department of Fish and Game
Kodiak, Alaska
POSITION STATEMENT: Provided an overview of the seabed mapping
program.
ACTION NARRATIVE
CHAIR PAUL SEATON called the House Special Committee on
Fisheries meeting to order at 8:43:21 AM. Representatives
Seaton, LeDoux, Johnson, and Edgmon were present at the call to
order. Representative Wilson arrived as the meeting was in
progress.
8:43:33 AM
^OVERVIEW: Alaska Department of Fish & Game camera sled
research/habitat assessment/seabed mapping program
8:45:01 AM
CHAIR SEATON announced the committee would hear a presentation
by the Alaska Department of Fish and Game (ADF&G) scallop
biometrician, Gregg Rosenkranz, on the camera sled mapping of
seafloor habitat and stock assessments. Chair Seaton said he
watched the equipment be installed on the vessel in Homer and
wanted the entire committee to learn about this cutting edge
technology. He then announced the committee would hear an
overview on mariculture permitting on Friday and discuss a
committee substitute to HB 257. As originally introduced, HB
257 transfers mariculture permitting from ADF&G to the Division
of Agriculture in the Department of Natural Resources. The
committee substitute removed the transfer but allows the Alaska
Grown label to be applied to mariculture products, a major
change from the original legislation. He told members to ask
questions as the presentation is in progress and asked Mr.
Rosenkranz to begin.
8:46:24 AM
GREGG ROSENKRANZ, Biometrician, Division of Commercial
Fisheries, Alaska Department of Fish & Game, introduced two
people who have also worked on the seabed sled project, Rick
Sheppard, an analyst programmer in Kodiak, and Mark Hoffman
(ph), a boat officer from Homer.
8:46:57 AM
MR. ROSENKRANZ presented a slide show and told members the
following:
I put this slide up to remind myself to give a little
bit of a disclaimer. That is Kodiak there ... Kodiak
is a fishing town and I'm an ex-fisherman so a lot of
my perspective on these things is influenced by my
past experience as a fisherman and I'm liable to say
some things that aren't official department policy. I
don't have anything to do with official department
policy. I'm a researcher so I just wanted to throw
that out right from the start.
I'll talk a little bit about the history of this
scallop video project. I'm not going to really be
talking a lot about the scallop fishery but a few
things will come in. I'm going to describe this new
system that we put together that's based on a GigE
Vision camera and I will show you with maps where we
sample and I'll show you underwater images of what you
see, what we see, and I'll talk a little bit about the
data analysis and where we're trying to go with this
research.
This is the map that shows Alaska weathervane scallop
fishing areas and actually the red areas are where
fishing has occurred since 1993 when the observer
program started. And actually, to make those red
polygons visible, I had to make them bigger than they
actually are. The footprint of the fishery is fairly
small. It's confined to these pretty compact areas
and then there's a lot of black. There are a lot of
areas that are closed to fishing and a lot of them we
know that there are scallops in and that's probably
our best conservation measure is the fact that we have
all of these closed areas. ... For example, out here
we know there are quite a few scallops in and we know
there are scallops down here.
I don't know how familiar you folks are with the
history of this but a lot of these areas were closed
to protect tanner crab habitat and that's really the
main reason that you see the closed waters there.
Why imaging for scallop stock assessment? Well, on
the left there you see a sampling dredge that ADF&G
deploys off of one of their own vessels. It's kind of
a smaller model of what is used in the commercial
fishery with a liner inside of it so you retain the
smaller scallops but these things are very finicky.
Sometimes you catch what's there. Sometimes you only
catch one-half of it. Sometimes you might only catch
one out of ten scallops and it's just about impossible
to estimate that efficiency. The sled is also non-
intrusive, which is nice in case, say, you want to
sample in some closed areas. It provides habitat
information ... underwater images do. The dredge does
in a way but it integrates it. Basically you bring up
the habitat and dump it in a pile on deck, sort
through it, and then throw it back over. And then the
bottom one, cost effective, that's always a concern
for Fish and Game so there is no way that I'd be here
talking about this if what we were doing was not cost
effective.
8:50:10 AM
MR. ROSENKRANZ continued:
And this is one reason why this technology can be cost
effective is because you don't need as many people at
sea. This is a bunch of ADF&G biologists sorting
through a dredge catch near Kayak Island where there
is some pretty heavy scallop fishing but it's a very
labor intensive way to sample.
Factors affecting catch per unit effort - this is sort
of my laundry list of reasons why I think that the
dredge sampling is not a particularly good way to go
and also what makes it difficult to manage fisheries
with just catch per unit effort data, which we do a
lot of in Fish and Game. If there's no survey there,
you'd have fishery data like catch per unit effort and
that's what we manage a lot of stuff on. But, there
[are] all of these little things that could affect the
measure of CPUE like the captain, the vessel, the sea
state, the electronics, etcetera and the list goes on
and on and there's a similar list for trawl surveys as
well. I mean I don't think a trawl is a particularly
effective sampling device either. So, for those
reasons, we started working on doing some underwater
imaging as a possible replacement for dredge surveys
back in 2000 and the first thing we had was this
little drop camera.
And there's just a little camera pointed straight down
right there and there's a recording device inside
there and some batteries and we're taking pictures of
the bottom. Well, you can't cover very much ground
with a drop device like that and so it wasn't too long
before we thought well, we need to drag this thing
along the bottom so this is what we had in 2001. And,
again, there's the little underwater camera and
there's just batteries and a camcorder where the data
is recorded and a couple of lights pointing down. Oh,
and that was when we had the camera pointed forward.
We were worried about were the scallops swimming out
of the way. Scallops can swim and we noticed that was
a very rare occurrence so we changed to point the
camera straight down.
So we kept modifying the sled and towing it around and
we did a lot of surveys with it and we got a bigger
battery, more lights, we had a little fancier housing
with the camcorder in here but still, you're really
limited by that battery and you're recording the data
on tape so you have to haul this thing back every 15
minutes, which takes a lot of time. And, in fact, you
spend most of your time hauling it up, getting the
camera out of there, making sure you got data, putting
it back down, etcetera. And also there are really
some safety issues too. We were setting this over the
side on a fairly small boat and you can see how banged
up that thing is there. We'd be doing this like 50
miles offshore in the Bering Sea. So I finally - I
kind of got tired of it. It wasn't very efficient and
it was dangerous and I was like well, we're going to
keep doing this. Someone's going to get hurt. That's
not going to be good. So, I had to kind of sell the
people around the state on the fact that we needed to
upgrade. We needed to do something different and the
way that we financed it was simply by not going on a
survey for one year. I should note that the funds for
this particular project have always come from CFEC
[Commercial Fisheries Entry Commission] since day one.
It's not a large budget but that's the money we used
so we took the money we would have used to go to sea
one year and saved it up and decided to do this
upgrade. That was about the time I met Scott
Gallagher. The one semi-smart thing that I did while
we were towing that other sled around was I did write
a paper on that and got it published in a fisheries
journal so people sort of knew who I was and I was
able to go to this meeting in Seattle about underwater
video stuff and Scott Gallagher was one of the people
there talking about his HabCam system. I realized
pretty quickly that this technology addressed all of
the concerns that I had about what we were doing: one,
it had live feed to the surface so you could see what
was going on down there; two, much higher resolution
than putting a camcorder in a pressure housing and
pointing that at the bottom; and three, the lighting -
using strobe lights. I just was like oh, wow, that
really makes a lot of sense Scott, you're a great guy.
So we went and visited them in November of 2005 and I
should note the HabCam is actually a joint project
with Woods Hole and the fishing community and these
guys are actually ex-fishermen that are also really
smart guys and they approached Scott and said hey, we
need to do something better here than what's getting
down as far as scallop research on George's Bank and
that's how the project got started - their project.
And I really think I was delusional looking back on
it. But we actually had the basic pieces of the
system ... mounted on the old sled, which wasn't very
optimal but we had that going that next June and I
don't see how we ever did that looking back on it but
we did.
So, what are the parts of the system? I should first
explain that GigE Vision is a standard from the
Automated Imaging Association, which is a machine
vision industry trade group. Scott Gallagher was the
one who was smart enough to see that in a way towing
something along the bottom and emptying the bottom is
a lot like machine vision where you basically have a
camera pointed at a conveyor belt - is often the
scenario in machine vision - where you're inspecting
something. So he was able to see the connection and
start using technology from the machine vision
community in fisheries research. So we had design
assistance from the HabCam project. They basically
just showed us here's the stuff we use. They didn't
say here's how you ought to do it or anything. They
said here's how we do it. You could try this.
8:56:26 AM
MR. ROSENKRANZ continued:
Of course we were looking for a low cost approach
because we didn't have a lot of money. And it also
needed to be pretty simple because we're not Woods
Hole. We don't have the technical expertise that
those people have there. So really, kind of the
subtitle of this whole talk could be taking Woods Hole
to the - can Woods Hole technology be used by a state
agency? It seems like the answer to that is yes.
So the system components are this GigE Vision camera
I'll show you, strobe lighting - and that's really
important because when you're towing this thing across
the bottom what you really want to do with the data is
look at a still image. People will show you some nice
video and you'll see a fish swimming along but when
they stop that video playback, the still image will
always be a little blurred and the reason why is
because you traveled a certain distance across the
bottom while the shutter was imaging so there's always
motion artifacts. Well [with] the strobe lighting
you've got about 20 milliseconds of light and you
don't travel very far during 20 milliseconds and so
you get a much, much sharper image than you'll ever
get with video with analog lighting.
The other really nice thing about this system is it
uses commercial off-the-shelf Ethernet hardware. It's
the same type of stuff they use to wire up the Capitol
Building probably, so you all have Internet
connectivity at your desks. These are, you know, you
can go to 10 different suppliers and buy a lot of the
pieces that we use. There is also armored fiber optic
tow cable, which isn't widely available, one of the
more spendier parts of it, wench with a slip ring and
fiber optic rotary joint and I'll show you pictures of
this. This oil filled junction box and computers and
data storage are on just regular computers and hard
drives.
8:58:21 AM
MR. ROSENKRANZ continued:
And I do have to acknowledge the people that helped,
Mark Hoffman of the Pandalus and Rick Sheppard. And
that was on our recent trip just a couple of weeks
ago, we were in Woods Hole and you can see them there
checking out the (indisc.). It was really nice. We
were very warmly received there. They thought it was
great that we were using technology that they've
developed there and, surprisingly, not that many
people are trying to do it. And then the third
person, Mark Blakely, we did hire a consultant to help
us with the first - he actually was on contract the
first couple of years and he's a local Kodiak guy who
has his own [remotely operated vehicle] ROV and he
helped us build some of the components of our system.
So this thing is basically a computer network. That's
really the beauty of it and on the top side and the
tow vessel you have a few computers and an Ethernet
switch, which is a very common Ethernet device.
Then you have the two cables and on the sled there's a
bottle that has another Ethernet switch in it, a
serial device server, and then the camera connects
directly to that switch.
I don't want to go into a lot of detail on that but
the single mode fiber is - it's kind of mind boggling
because I had to learn how to terminate this stuff.
When you get down to the business end of it, it's - I
don't want to misstate the - I cannot think of the
measurement. It's about one-tenth the size of a human
hair is the fiber diameter. And really, with fiber
you have - I won't say unlimited bandwidth but
compared to what you could do over copper, over
coaxial or twisted pair, the old style of
"telemetrying" data from an underwater vehicle to the
surface, there's just no comparison. That was the big
choice, are we going to jump ahead or are we going to
be satisfied with 30-year old technology and, I guess
at that point in my life I guess I wasn't afraid of
failure. I was like well, let's try it. If it
doesn't work, what do we have to lose? So we went
ahead and tried the new technology rather than
investing money in something that is 30 years old.
9:00:27 AM
MR. ROSENKRANZ continued:
There was also good timing with this. This is
actually serial number blah, blah, blah zero one of
this model of camera. The standard was just coming
out when we were starting to put the system together
but the main feature of this camera is that right
there. It's a regular RJ 45 plug and this attaches
directly to a computer network. This camera will run
up to 15 frames a second. We run at about four - you
don't really need to know about the sensor size. We
just have it mounted in a little pressure housing with
a dome fork pointed straight down and this is parts
and pieces we had from the old system. And then the
strobe lights - it's a machine vision strobe. We just
put it in an underwater housing that's made out of a
piece of aluminum pipe. The telemetry bottle that
fits on the sled has this Ethernet switch in it. I
mean this is very simple. That's a power supply, an
Ethernet switch, and a serial device server. That's
really all that's in the bottle and this is that
Ethernet switch and the one difference between this
and what you probably have a bunch of around the
building here is this. It's a little fiber optic
media converter.
Now there are standards to send Ethernet, Internet
over fiber optic. So, once again, it's not that
expensive. There's one like this, it's also on the
topside, one in the bottle. It's a pair that talks
back and forth to each other and I think it's like
$1,000 or something. It's not hugely expensive
whereas if you want to buy a ROV or have some company
build you a system to telemeter data from a vehicle to
the surface, it will cost you a lot more than that.
We had to get a wench and we had this wench built in
Kodiak and it's taken us three years, really. This
being the third year, we finally have this thing
working really pretty well now. We had some problems
with it but we got it for dirt cheap by having a local
company build it. If we couldn't have got it for that
price, we wouldn't have been able to do the project.
But the one piece of hardware in it that's essential
is this slip ring and it fits right in the hub of the
wench and it allows electrical and fiber signals to
pass around. Basically there's a fiber - there's an
interface there where two ends of fiber, one is
stationary and the other is spinning against it and it
transmits the data across there so that's kind of a
specialized piece. They're about $6,000.
The junction box - you don't need to know a lot about
that but that fits on the end of the telemetry bottle
and that's where the actual fiber and electrical
terminations occur on the sled end and that's
something if you - you know, when we were at Woods
Hole, everything they have, they have oil filled
junction boxes on everything. It's something they've
used a lot and it works well.
And then on the topside, we have computers. I mean
I've literally pulled the computer out from under my
desk and taken it to sea for the last few years. I
think I might get some dedicated computers now and
then the storage, this box, hot swappable array, is
not an expensive component. What fits in there are
regular hard drives like you have in your computer and
we just fill up a lot of them so, you know, it's
cheap. It is $75 for a 300 gigabyte hard drive these
days so that's not a huge expense.
And, of course, we have our favorite low cost
oceanographic platform and that's the Pandalus. And
again, if it wasn't for the Pandalus, we wouldn't have
done any of this.
9:04:11 AM
MR. ROSENKRANZ continued:
Mark Hoffman basically took it on himself to cut the
bulwarks off the stern, get this U-frame built by Bay
Welding - they're in Homer. We used a lot of Homer
and Kodiak vendors to build most of this stuff. So
inside - this is actually a borrowed shack that they
use for their ROV that they have there but we put our
computers and what-not in there and you can see the
sled in there. So, it's kind of an undersized vessel
but we've taken this all the way from Cape Fairweather
up to the Bering Sea sampling so we have quite a bit
of confidence in the boat. I have utmost confidence
in Mark Hoffman's skill as a boat operator. It's
cheap compared to going out on a larger boat or a real
oceanographic vessel. This doesn't really cost very
much.
Then we use a four person crew. We maybe could take
five but you don't need to have that large group of
people to sort through a catch on deck because you're
not really getting a catch on deck. So, this is the
new sled. We just got this built by Bay Welding in
Homer this last spring and used it for the first time
last summer and that's right off of Cape Douglas there
at the end of Shelikof Straits.
So what does this thing do? Well, it takes four - it
streams four 1360 by 1024 mega pixel basically images
and writes them on a hard drive on the boat in real
time. So the data adds up in a hurry. You get 14,400
images an hour or 56 gigabytes an hour so you fill up
one of these hard drives in a matter of five hours.
The other really nice thing about it is with the
strobe lighting, you can tow faster. I see now the
HabCam guys are going to try and go with five frames a
second and tow at five knots and that might be what we
want to shoot for. But, in the meantime, when we were
using our old system we were trying to tow at a knot
and a half, which is almost impossible to drive a boat
that slowly, plus you don't cover that much ground.
But going at four knots you can actually cover a
pretty good area during a day, one kind of Alaska
invention the Woods Hole guys got a big kick out of.
That's our safety release there.
There's a hundred fathoms of quarter inch line in
there - that really strong synthetic line and a hard
float with a galvanic release so if we ever did break
our cable, about 24 hours later that hard float would
pop up hopefully to the surface and we'd have a chance
to recover the whole sled. And then the one other
kind of interesting part, this green cable you see
right there is what actually connects the camera to
the telemetry bottle and we were the first ones to use
that and even the Woods Hole people were actually
really impressed with the way that we did that and
they are using the exact same thing on a bunch of
stuff now. The other thing is our vehicle actually
rides on the bottom. The HabCam is towed off the
bottom, which has some advantages and also has some
problems with it but we have these replaceable wear
strips on here. It's a material called Treks that
they use like on crab boat deck boards. It's part
plastic, part wood. And so on harder bottom, a lot of
the bottoms we go on are pretty soft, but on harder
bottom sand and stuff it wears it off and you can see
how it wore down in the back.
9:07:39 AM
MR. ROSENKRANZ continued:
Well we have a lot of weight in the back of this
because it overhangs in the front. The basic idea of
this, besides having a clear imaging area in the
front, is we wanted something that would go over a
crab pot because there are lost crab pots everywhere
you go in Alaska and if you get hung up on those,
you're going to lose your sled in a hurry. This thing
is actually built to go up over it and we've been able
to go over a lot of stuff because all of the weight is
in the back so the front is kind of light. When we
hit rocks and stuff, we just kind of go over them.
I've been really happy with the way that worked out.
So the first year that we had this in 2006 on the old
sled we spent a long time at the dock in Homer getting
this all to work but we finally did, steamed over to
Yakutat, put it in the water, and were just amazed.
It just worked.
It just kept working. The images weren't the greatest
but as far as being able to get the data, the network
was good. The data storage worked. I was just really
happy with the results of it. As long as we were
looking at a map I did want to just touch on something
that is of interest to this committee, I believe, and
that is the difference between state and federal
waters in relationship to where scallops get fished.
Now you may not know, but that is the border right
there between Central Region and Southeast so these
are managed by Central Region while these Yakutat beds
are managed out of the Douglas office. That state-
federal water line, it runs right through these beds.
I mean Kayak Island; people are talking a lot about
how Kayak Island might be a good state water fishery
but really not that much of where it's fished. Again,
there are scallops outside of these red lines. These
red lines enclose about 98 percent of effort and
harvest since the observer program started in 1993.
So I just wanted you to get a little idea of what
you're talking about. I mean there is one little area
there and one there. These beds are really outside of
state waters.
I'm guessing you folks know your Alaska coast a little
bit at least. Yakutat Bay is right there so this is
from Yakutat up to Cape St. Elias and then between
Yakutat down to Fairweather and, of course, Cape
Spencer is right down here, same thing. While there's
more - this bed is actually inside state waters. Some
of this one, a little of this and then that one is in
state waters. So, it's really - it's hard for me to
think about trying to have a different fishery going
on depending on which side of that line you're on
because they're a contiguous biological population
and, just to hammer that point home, this is an area
boundary between Area D and District 16. That was put
there for salmon or some other fishery because it
basically bisects a contiguous bed that we know about
and it causes all sorts of problems, believe me. That
line is actually defined ambiguously in regulation.
So, depending on where you read, that line is actually
going to move a little bit around and it's enough to
confuse the fishermen. So that's Mount St. Elias and
actually the beach they surf on is right there, right
outside of Yakutat, just so you can see.
So we got this thing in the water and what do we see?
Well ... scallops live on a lot of really muddy
bottoms in Alaska and suspended sediment is a
problem....This is actually a very unusual picture and
there's actually a smaller scallop there too but
there's four scallops in one frame. Usually, the
usual count is zero. Anyway, but that's what we could
see and then, of course, we found out that the water
isn't always silty.
This is the kind of resolution we were getting with
the old sled and it had some very bad lighting
problems. There wasn't physically room on the sled to
get four lights over our field of view. So, the data
was still actually pretty good for manual review but
obviously there was a lot of room for improvement.
Just to show you some of the other stuff we saw from
that survey, that's some skate egg cases on kind of a
sandier bottom with shell ash. That is an eel and I
don't know if you can see it there but you can
actually see the little fin marks. I mean that's the
resolution that you get. There are 1360 pixels going
across this way and it's just over a meter in width so
you have more than a pixel per millimeter on the
bottom. So you can really actually talk about
millimeter resolution on the bottom or being able to
identify things that are 10 to 20 millimeters and on
up, depending on the water clarity of course.
So that was the first year. Then we got out and got
the better sled built and then this last spring we
headed over to Shelikof Strait. What you see here,
again, this is a polygon that encloses about 98
percent of the effort since our observer program
started in 1993. The blue lines are the tow pass of
one vessel during one season.
9:13:14 AM
MR. ROSENKRANZ continued:
What I've shown here is the 100 meter or 50 fathom
curve and, in a lot of areas of the state, scallops
seem to be concentrated around that depth and you can
see how the effort really follows that as well.
Well, let's go ahead and look at some boundaries on
that picture. So you had the state-federal water,
there's another island, a little rock off of here,
with scallops around there but this is the federal-
state water line in Shelikof Straits so, obviously, a
state water fishery here would be very difficult to
enforce. Most of the scallops I think, well, the
majority of the scallops are probably in the federal
water outside.
Now this is also a statistical area lying here and the
way the fisheries managers - actually separate quotas
for the north part of the bed and the south part of
the bed, so it gets pretty confusing pretty fast when
you take these non-natural boundaries and start trying
to manage fisheries with them.
Well, so we went over there and last June, I guess
that was, we spent five days, not full days, at sea
and that's what we were able to cover doing these
kinds of zigzag transects over the bed, 82 nautical
miles towed and a few hundred thousand images.
9:14:39 AM
MR. ROSENKRANZ continued:
Really, the main thing we've learned so far is that
the system is really pretty dependable and works
pretty dang well. A view of Shelikof Straits and, so,
this is what the bottom looks like over there. Once
again, there are scallops on this really, really
sedimentary bottom and you can see right there - maybe
you can't - there's a little portion of the shell that
doesn't have sediment on it and you kind of see the
natural color of the scallops. I was really kind of
amazed to realize that these animals live and thrive
in such a silty environment. I mean you have the
[indisc.] Bay, huge glacial outflow right there, but
these animals live there. It's surprising because
they are filter feeders and you would think it would
be inefficient for them to filter all of this sediment
to get some food but that's where they live.
Just to give you an idea, ... we do see other stuff.
This is just a little bit shallower water on the side
of the bed on the harder bottom and we do see colors
down there. A lot of what we see is kind of the color
of silt. You can see around the outside we're still
working on our optics. We have that dome port lens
that we shoot the camera through and it's very
difficult to get the whole edges of the field of view
into focus with the amount of light we have. So
that's something we're working on. Hopefully we'll
have it solved by this upcoming season.
And then I mentioned that this is a non-intrusive
sampling. Well that's not quite true. In a really
soft bottom, that sled weighs about 1,000 pounds and
it will penetrate a few inches into the bottom. So
that's a sled track that we then cross back over and
got a picture of. And these dots here are 10
centimeters apart to give you an idea of the size of
stuff in the field of view.
Arrowtooth flounder is something that we just see a
lot of down there and an occasional cod, although I
think most of the cod are scared away by the sled and
we don't see a lot of them. After we worked in
Shelikof we went around over to Augustine Island,
Kamishak Bay and we're working on a comparative survey
there where - they also did a dredge survey and we're
going to be able to compare the estimates between the
two methods.
9:17:10 AM
MR. ROSENKRANZ continued:
Just to show you the stuff on the sled, there is the
tube with the camera pointing straight down. That's
sonar that looks ahead. Hopefully if we were going to
run into a sunken crabber we would see that. There
are four strobe lights arranged around the sides and
then the telemetry bottle is right back there and,
really, that's about all there is to it.
So we got on to some harder bottoms over there and,
low and behold, there are a bunch of scallops on these
harder bottoms. I believe these are clams. There has
been some dispute about that and it's one of those
things that yes, you do sometimes need to get samples
in your hand to actually look at and ID but it gives
you an idea. These scallops are probably, you know,
20 millimeter shell height so they're pretty dang
small and we can see them quite well if the water is
clear enough. These six-arm sea stars seem to be a
pretty big predator of the scallops over in Kamishak
Bay. Just some of the other [indisc.] vertebrates
that you see there - there are sea pan (ph), different
kinds of sea stars, and one of the animals that we see
quite a bit of that people wish there was a lot more
are tanner crabs. Well it turns out I'm an old crab
researcher from my days in graduate school so I'm very
interested in tanner crabs as well. We assess these
with trawl nets, which doesn't work particularly well
because a lot of them like to get buried down in the
mud so it's hard to catch them with the trawl. We do
that and I think we're not sure - well for example,
I'll show you on a map, but we could image the same
crabs twice quite easily on a survey [with] the way
that we use the camera sled.
But, I think there's the potential there working with
different sampling methods to just try and find the
spatial extent of these animals. I think that's the
direction we're probably going to want to go over the
long term rather than having point estimates of
abundance, looking at things in more of a spatial
context - ecosystem approaches to management, which
I'll talk a little bit more about.
This is another thing that you see this and it just -
because I have to tell you, I have been about really
the only one that's looked at a lot of this data just
because there isn't anyone else to do it. That's a
little juvenile tanner crab with some, we call these
emergent epifauna.
So you get on these harder bottoms and you see a lot
of stuff and it looks to a lot of people like these
types of things might be extremely important for young
animals of different species: crabs, cod, what have
you.
So, I guess just from my own experience of looking at
this data, I'm realizing hey, we need to look at the
whole picture. We did this because we're interested
in scallop stock assessment but it looks like we've
got to think about other things besides scallops when
we use the data. And there are still a few king crabs
over in Kamishak Bay, not very many of them but there
are a few and we got a couple of pictures of them.
Another kind of interesting thing, there are a lot of
flat fish over there and who knew. There are only two
in this one. There were some frames that had like
five flat fish just lying right next to each other. I
didn't know they were that family oriented. Like I
said, we managed to go over stuff maybe we shouldn't.
This is a big rock we bounced over. An octopus lived
right under there obviously. Octopus, they just go
out into the scallop beds and find their lunch and
then they bring it home, I guess, is the way that they
do it but that's definitely an octopus midden.
So I mentioned that I spent a lot of time looking at
this data because we're doing this comparative survey
so what we did there was - there are 88.1 square
nautical mile stations and we just went up one row and
down the other. Well that took two days so, you know,
like crabs that are pretty mobile, if you saw them
here on one day they could be over here the next.
That's why I say you could image the same animals
twice, a potential problem.
9:21:40 AM
MR. ROSENKRANZ continued:
But this just shows the abundance of scallops. This
is just kind of an example of a data product you could
get out of this. What I did was looked at the data
and then every two-minute clip I would stop and record
the count. I didn't have frame accurate review, which
is what you should have. You should basically know
where exactly each scallop was counted but I just
didn't have the time or software to do that yet. So I
aggregated on 200 and made this little map and it
gives you a really good idea of what's there. I mean
you could convert that to a mean weight estimate
fairly easily and what it shows is their survey grid
there does have the population captured pretty well.
You see the black circles around the outside are zero
counts so mostly, yea, they did find where the bed was
with the dredge and that's where it is.
These are for small scallops less than 50 millimeter
shell height. You can see they're not distributed the
same at all. They're over here and I have not had
time to go ahead and overlay the habitat data but I
can tell you that these scallops are very strongly
associated with these harder bottoms, more pebbles,
more emergent epifauna. And this is the tanner crab
results from that same survey and then, again, that's
why I said we did this on two different days so I
don't know, maybe these tanners that we saw here on
day one moved over here by day two and we took their
pictures twice. I don't know but you get a really
good idea of where they are located in that area with
the camera.
Then the last - one of the things that we've been
trying to do - well, that I've been trying to do
besides just having opportunities to use the sled
more, because we're just learning how to use it.
Basically it's been a one time a year, though it gets
crazy just trying to get everything to work, get out
to sea, people have expectations about what needs to
get done. Well, Dan Doolittle now is over at the
Kachemak Bay Research Reserve and he is a friend
because he was in the Kodiak office and he's very
technologically oriented so he said hey, how would you
like to come over and practice using the sled in
Kachemak Bay? I said great. He had this project
they're talking about expanding the Homer boat harbor
there. ... Army Corps of Engineers funded this and
that's high resolution multibeam data that really
shows the bathymetry in excruciating detail and then
they had some diver surveys. We came along and in
about 20 minutes were able to do all of these
transects.
This was literally a stone's throw off the breakwater
there. You looked out in the water and it just looks
like mud and you're going man, we're never going to
see anything down there. Well, actually it was a lens
of silty fresh water on top and there was some clear
water on the bottom and that's what is right outside
of the Homer boat harbor, a lot of eel grass that gets
blown or carried in there by tide, sea pens, and sea
stars. And there are a few fish. There's a big
sculpin there too. After we got done with that we had
plenty of time so we went out in front of Homer off of
the bluff there and that's just an example of some of
the harder bottom habitat. I just love these flat
fish, the way they can camouflage themselves with the
coloration.
So I just had to say a couple of things about the
HabCam project. This is their vehicle and they tow it
off the bottom, which we're kind of thinking about
doing with ours and then they've also added - this is
a second, separate towed vehicle that's just doing
acoustics so they're basically trying to widen the
swath that they see through acoustics. They have the
sled going and then this thing is back farther and it
has I guess a couple of different kinds of sonar on
it.
9:25:41 AM
MR. ROSENKRANZ continued:
That's sort of the next step that they're taking,
which looks like a pretty good one to me. And then
notice this is on a commercial scallop vessel and
there's their dredge in the water at the same time
they're using the sensing imaging and sonar equipment.
This is another slide that's ... lent from Scott
Gallagher and this just kind of shows what they are
trying to do. These are like real time data products
that you could get out of this but really, we are
stuck right up here. We've got down a little bit of
manual classification. We haven't even done this step
with image processing yet. They're working on that.
They're working on automated classification. I was
able this year - we were able through our near shore
project to get some funding for them, for the HabCam
people to work on Alaska data. Well that project is
going away so that's one that I see a huge need is to
keep helping them. I mean we need to get money for
them to look at our data and we need to just keep on
their good side because we want to get all of this
stuff and they're going to develop this a lot faster
and better than anybody else because they are the
experts. Believe me, this is a huge job. Talking
about being able to display the data in usable forms
where some fishery managers could have access to the
data without having to look through frame by frame but
trying to distill usable data products of it is just
another huge challenge that we need to do. And then
you get over to okay, what are we going to do? Well,
patch dynamics, like looking at the patchiness of
different species, habitat, predator/prey multi-
species management models, and this is where ecosystem
type of thinking would probably come in. Really the
hardest challenge for them right, all of this stuff is
hard but the biggest challenge from a computer
processing point of view right now is this
segmentation where you want to extract regions of
interest from image and, especially now with our muddy
bottom, with not that many scallops it might not be
bad, but when you get on a complex bottom, it's hard
to get the computer to accurately pick the targets out
of it. But they are working on that and not only are
they working on it but they are working with the
people that are the real experts at Los Alamos
National Labs. Well, that's a wonderful thing except
they're DOD and it's hard for anybody to do
cooperative work with DOD so that's kind of slowed
things down and Scott is a little bit frustrated with
it but there are a couple of people there that have
some ideas and some algorithms that ought to be
helpful for this.
So, my "to do" list. I have a big "to do" list. We
want to improve the optics and get that whole image
into sharp focus and we have some new hardware. We're
going to work on that. We're going to shoot through a
flat port rather than a dome. We're building a new
deck house here, hopefully within the next month or
two since the one we've been using was borrowed.
9:29:18 AM
In June and July we're going to be doing surveys,
Kayak Island, Yakutat, and then the Pandalus is going
up north to Norton Sound to do a trawl survey up there
so I am trying to convince people that we really ought
to put the camera on this boat - well we'll still have
it on the boat, leave it on there and do some imaging
on the way up there as well. Additional sensors - we
don't even have a CTD on our sled but we ought to.
That and, we've been looking - we recently had a
meeting to discuss this acoustics, what acoustics
should be added to the package to gather more data at
the same time when we're towing it around to get the
images. That led us to flying off the bottom. Well
most of the acoustic type of things that you would do
would be much better off if the vehicle was not hard
on the bottom just because you'd have more space below
you but also the vibration of bouncing around could be
very problematic.
And then, finally, partnerships. People are just
starting to find out about this now. The Department
of Fisheries and Oceans in Canada wants to build a
system and has been calling me up. Someone from NMFS
[National Marine Fisheries Service] in Kodiak has
called me a couple of times and wanted to talk about
the data and ways they might be able to use some of
the data. So there's just a huge list of things for
me to do and then the thing to remember about, that is
I have my job as statewide scallop biometrician as
well.
So, really, our biggest need for this project is
people and we actually have the people. Rick Sheppard
- I can't say enough good things about him. He won
the employee of the year award six years ago but he's
kind of one of our most valuable players out in Kodiak
because he is the database person there. And
everybody in Kodiak, every project collects data.
Most of it is lying around. Most of it is on a bunch
of CDs or in files there. He's the one that's making
the data usable. Well, he's also the one that wrote
the code to let our camera talk to our network for our
system. He's the one that would be most qualified to
start working on ways to display the data, share the
data, etcetera, but he is more than fully tasked and,
just to let you know, he is an Analyst Programmer 4
and he's been trying to hire an assistant for a couple
of years now. They downgraded the position ... it was
an Analyst Programmer 3, a 2 and a 1. They finally
got one person to apply for it at the 1 level and then
that person didn't take the job. So, within Fish and
Game the people that do understand technology are
very, very few and far between and they are very hard
to hire because we do not pay competitive prices.
People that know how to do that stuff can go and get a
lot of different jobs and the State of Alaska is
pretty low on the list. I'm sure that's true in other
departments where they need technical expertise -
computer networking, database, and etcetera.
9:32:26 AM
MR. ROSENKRANZ continued:
Dan Doolittle has a lot of [autonomous underwater
vehicle] AUV experience and he looked at this and said
this instrument pod, if you will, basically is
attached by 10 bolts. So we're talking about
decoupling that, move the telemetry bottle on there,
add a tail for stabilization and a different bridle
and we might be able to just tow this off the bottom.
We're going to hopefully get a chance to experiment
with that maybe next fall. It's probably the way we
want to go for the future, although having it on the
bottom is kind of easier. Then you don't have to
worry about it. That's one of the problems with the
HabCam is it's always doing this. ... There's an
altimeter, an acoustic device that measures height off
bottom but it's very difficult to fly it at a truly
level height. Ours, where it's on the bottom, one,
you don't have to sit there watching it every minute,
going oh, I've got to pull in a little cable, oh, I
need to let a little bit out because it stays about
the same so the way we're doing it is more user
friendly from that perspective. It seems to work okay
on soft bottoms but if you had it flying up off the
bottom, you wouldn't have to worry about, you know,
unless you're in pinnacles, you could go over a lot
harder bottoms and it would probably be better for
acoustic sensing as well so we're probably going to
try that.
And then one of the other "to do" things is just - I
keep looking at all of these closed areas. We have
not yet deployed the sled in a closed area. Here's an
example right here, Unimak Bite. This was closed to
protect tanner crab habitat 20 or 30 years ago. No
one even knows if that's working. I hear the old
scallopers talk about deck loading out there so -
here's a tool where you could actually go and look in
some closed areas and see what's there, maybe evaluate
whether they are an effective tool. With tanner crab,
I think it's mostly wishful thinking. There's just a
lot less tanner crab around now than there was 20 or
30 years ago. Really, I think for the future we're
really going more towards a monitoring type of outlook
- I am. I think as far as environmental monitoring,
the sled is a much better tool than a trawl or a
dredge.
So, here's where I will stick my foot in my mouth a
little bit. I guess I didn't mention that after I got
done being a commercial fisherman I went to college up
in Fairbanks and mostly took math and statistics. The
red boxes are kind of the way things get done. You
have commercial fishery data, you have trawl and
dredge survey data.
9:35:18 AM
MR. ROSENKRANZ continued:
This is an issue that sometimes gets, well often gets
overlooked but each species has a different set of
catchability and selectivity parameters with each
different kind of gear. Basically you cram all of the
data you got into a mathematical and statistical model
and then what comes out is a number and managers like
that and I give the people that have to manage the
fisheries a lot of credit because they're the ones
that have to answer the phone when irate people call
up and want an explanation about the way things are.
I mean there are reasons why we do things the way we
do and a lot of it is managers want something that
they can point at when they're asked why did you set
the quota at this level. So that's basically the way
that we do it but the problem is we don't know
anything about these recruitment processes, no matter
how good your data and your modeling, you really are
left with hoping for recruitment because if you don't
get any recruitment for a period of years, there is
not going to be anything there no matter how
conservatively you manage the fishery. So what we're
doing here with the camera sled is we're starting to
bring spatially explicit direct observation data into
the picture. Now how we're going to incorporate that,
nobody knows yet. But this is what it is and it's
fundamentally different from trawl or dredge data or
fishery data because you don't have the catchability
and selectivity problems with that type of gear. You
see what you see. There are some limitations with
water clarity and size but it avoids this problem,
which is a huge problem. And then I'm getting the
feeling that there's sort of enough of a wow factor
with the imaging to call a little more attention to it
and get over to these two the fact that we need to
look at more oceanographic data and we need to start
thinking about ecosystem approaches to management,
i.e. looking at the whole picture, using more spatial
kind of statistics rather than just taking data,
running it through a model, getting out a number and
then managing the fishery off of it because really I
can't sit here and argue that this is working all that
well. There's not a lot of tanner crab around now.
There just isn't and I don't care how many surveys you
conducted. That's not going to bring more tanner crab
back so I think we need to add more data to the
picture is probably the best way in my mind to move
forward.
That's really all I have thanks to all of those people
on that list. I also have some just short clips, i.e.
movies of data if you'd like to look at a few of
those. Oh, no I don't. Sorry, it's on my computer.
I put my presentation on this computer. I could fire
this other one up while people could ask me questions,
whatever your preference would be.
9:38:43 AM
CHAIR SEATON said he thought members would be interested in
seeing the clips. He then asked Mr. Rosenkranz if he has used
an ecosystem management model with habitats of concern in the
federal regulations or whether that is a possibility in the
future.
MR. ROSENKRANZ said he has not used an ecosystem model. He said
this has been frustrating for him because any time one gets
involved with council processes, such as the North Pacific
Fishery Management Council, it is talked about but nothing
further happens. He related his opinion that no one is taking
the necessary steps to figure out how to do it. Mr. Rosenkranz
commented that oceanographic information is expensive to
collect, which is one benefit to the seabed sled. It is a new
development to the ecosystem approach of management. The HabCam
people are defining the field and, by maintaining a good
relationship with them, Alaska will benefit by learning from
them.
9:40:45 AM
CHAIR SEATON pointed out that fisheries managers want more data.
Alaska is on the forefront of this technology, which is one
reason why the committee wanted to hear from Mr. Rosenkranz. He
is pleased that ADF&G has taken this initiative.
9:41:28 AM
MR. ROSENKRANZ said he met Scott Gallagher at a meeting in
Seattle and when he asked Mr. Gallagher whether anyone else was
working on underwater video techniques, the answer was no. He
said the technology has come a long way; for example, he opined
that people on the East Coast will have fiber optic to their
homes in five or ten years. Additionally, hardware is easier to
use and getting cheaper.
9:42:28 AM
REPRESENTATIVE EDGMON asked whether funds other than general
funds contribute to this effort.
MR. ROSENKRANZ stated that his position is federally funded
because the scallop fishery occurs in both state and federal
waters. The camera sled project has been funded by the CFEC
from the start.
9:43:07 AM
REPRESENTATIVE EDGMON surmised there is no cost recovery
component tied into this project.
CHAIR SEATON said not to his knowledge. Alaska's congressional
delegation believes Alaska needs to step forward in this arena
rather than ask. He opined that the legislature should let the
delegation know that the state is doing independent projects
with state funds.
9:44:02 AM
REPRESENTATIVE EDGMON remarked that he thoroughly enjoyed the
presentation.
9:44:20 AM
CHAIR SEATON, regarding acoustics and mounting different kinds
of sonar on sleds using fiber optics to provide enough capacity
to get the data through, said avoidance data could be collected
in many directions depending upon the orientation of the sonar.
MR. ROSENKRANZ said that is a rapidly developing area and
different types of acoustic sonar exist. Right now multibeam
sonar provides very accurate depth maps, which show every ripple
and contour along the bottom. Side-scan sonar views the side
and extends to about 50 meters wide. The potential is big but
he does not know the best way to go at this point and he doesn't
believe it would be wise to buy an expensive system at this
time. He informed the committee that he is cultivating some
relationships with people who might allow ADF&G to try their
products. He said he is excited about working with Dan
Doolittle in Homer because he would like to use that area and
the research reserve to help develop more technology. The
visitor center could be used for public outreach. He said the
bottom line is that he needs help.
9:46:51 AM
MR. ROSENKRANZ showed members his video clips and said the
following:
This is at two frames a second. We record data at
four frames a second, twice as fast as this. This is
just some hard bottom right there in Kachemak Bay off
of the bluff there. There are a bunch of clams,
barnacles, things like that in that particular
habitat. This is also in the same area where we just
went into shore so we started going over bigger and
bigger rocks. That's some kind of steel something
there. But this is right off the bluff in Homer.
9:47:53 AM
CHAIR SEATON asked whether the water was clear when the images
were taken.
MR. ROSENKRANZ said he did not recall the water clarity for
those slides but said in general, freshwater floats on top and
it is often silty. He said you never know what you'll find
until you put the camera down. However, on softer bottoms in
areas like Shelikof Strait, suspended sediment can be seen when
the tide runs hard. He noted the current slide was taken in an
area of a depth of a few fathoms right off of the Homer
breakwater in December. He said the research reserve in Homer
will allow research on seasonal changes at the same place. [Mr.
Rosenkranz continued to point out marine life in the slides,
such as barnacles.]
9:50:41 AM
CHAIR SEATON thanked Mr. Rosenkranz for his presentation.
9:50:50 AM
REPRESENTATIVE JOHNSON asked if the camera could be put on a
dredge to look at damage.
MR. ROSENKRANZ said that has been done with some older cameras.
Although the purpose was not to look at damage but to observe
how many scallops went in and how many did not, he said he has
not found that data to be very useful. He said he has been on
dredge surveys. The material looks very different when it is in
its natural environment than when it is hauled on deck.
9:51:56 AM
ADJOURNMENT
There being no further business before the committee, the House
Special Committee on Fisheries meeting was adjourned at 9:52
a.m.
| Document Name | Date/Time | Subjects |
|---|