Forum on Wildlife Telemetry
Innovations, Evaluations, and Research Needs
Final Discussion Session - 21 September 1997
CHARLES AMLANER: Good morning everyone. My name is Charles Amlaner, I'm from Indiana State University and my background is in physiology and biotelemetry. I'm going to act as your host chairman for the morning. We have a rather exciting agenda this morning; it's a little bit different than we've featured for the last few days. Whereas the last couple days have been structured talks and lectures with some discussion, this morning is semi-structured, in fact, very loosely structured. We have four people up here to help act as catalysts to raise questions with you and for you and on your behalf. We have several questions that you as an audience provided us to use to spark some discussion. We're hoping that many of you out here will have had so many pent-up questions that you're just going to jump to the microphone and ask us and ask other people in the audience. We are not going to be able to answer every question. We will do our best. We are hoping that there are more experts out in the audience than is up front and that you will pop up to our aid when we're stuck.
So to begin with, let's do a few introductions. Stan Tomkiewicz is from Telonics, Incorporated and is the Director of Environmental Programs. We have his outline from the Web which was given to me this morning. He has been a consultant with Telonics from the late 1970's helping other wildlife researchers adapt telemetry equipment to all sorts of study goals and he has been very influential in developing the ethos of telemetry for many of us in the field. Immediately to his right is Jay Rotella. Jay is from the Fish and Wildlife Management Program at Montana State University at Bozeman and he is an Assistant Professor and his background is in Avian Biology. He has been working with the Prairie Habitat Joint Venture of the North American Waterfowl Management Plan. Immediately to his right is Mike Samuel. He is from the U.S. Geological Survey, the Biological Resources Division, National Wildlife Health Science Center in Madison, Wisconsin, and he is officially employed as a statistician and has been doing that for many years. He will be acting as our expert in terms of the mathematical aspects of tracking and the like. He's probably going to have lots of questions about how I'm going to do that.
Okay, with that introduction, Stan is going to begin with a short introduction of his section and then we'll open the floor up for questions.
STANLEY TOMKIEWIEZ: Good morning. It's good to see so many up so early and in attendance.
The section that I was asked to be involved in or summarize was the section on innovation in transmitting and receiving subsystems. I guess we've all spent the past couple days listening to papers that are really about innovation in transmitting and subsystems in one form or another. All of the papers involve some degree of innovation and I think there was an excellent grouping of papers over a broad range of topics but if we harken back to Sunday morning when we started out in the first segment, I'll try and give you just a brief little synopsis and maybe reawaken your memories.
We heard a number of very interesting papers in that segment that kind of provided the range of what telemetry can provide to us as a tool today or as a series of tools today. One of the first things we see is that conventional VHF telemetry that's been around for fifty years, or just about fifty years, has almost reinvented itself with a series of new tools, including micro-processor-controlled transmitters and sophisticated data-logging/receiving equipment. Both of these tools in that particular area have opened up new opportunities; for example, the micro-processors have allowed us to run transmitters for much longer periods of time while still being quite tiny. The sophisticated data-logging systems allow us to perhaps not be present in the field all the time, which does several things. One is that it prevents the disturbance of the animal we are trying to study. In addition to that, it may prevent the disturbance of other species that are in the area that might be influenced by our constant monitoring; and I think it gives graduate students a little shut eye. So those are very good things.
We also saw some developments in the field of satellite telemetry that came on in the 1980's that really first saw the tracking of large mammals with transmitters that were relatively large but suitable for large critters. Then we saw a reduction all the way down to small bird and sea mammal and sea turtle transmitters that have been developed here in the early 90's and deployed and been very successful in allowing that technology to address questions that could previously not even been approached concerning long-range migration and movements of animals over places where it was difficult to follow them for protracted periods to time. So that's an exciting aspect of it.
In addition, Argos itself, as we heard from Jeff Wingenroth, is in the process of undergoing a sort of revitalization as new instrumentation on board the spacecraft is becoming available to increase the band width, essentially allowing us to use more transmitters, PTT's; increase the sensitivity of the instrument onboard the satellite; and perhaps utilize two-way communications with the satellite for some studies, if that becomes practical over the next few years. So the system is sort of adapting itself, whether it's directed or not for applications in the wildlife field, for low-powered transmitters, for example; that increased sensitivity in the on-board satellite instrument will allow smaller transmitters for use on the Argos system.
We see, in addition to that, commercial companies are developing a wide array of transmitters for the Argos system that are pretty intelligent devices. Some of them have full on-board data loggers and others are just capable of really sophisticated control so they can turn the transmitter on and off at various times - making the transmitter able to monitor the animal during periods of interest and go to sleep during periods when perhaps we can just save battery life until the next period when we're interested. So that's an exciting area associated with Argos. Then comes along the field of integration of sensors, new sensors, physiological and environmental sensors that can be incorporated with these intelligent PTT's and input to Argos. One of those sensors may well be a GPS sensor, a position sensor, and that's the way GPS appears to be used. Not only with Argos but as a stand-alone unit where it simply data logs the GPS position, or where it's relayed back across some sort of data link to an observer through an FM data link as we heard Art Rodgers describe, or through other types of data links that may relay information back to an observer as well as possibly using Argos as one of those data links.There is also the same developmental process that I suspect will go through with incorporating GPS into systems as we went through Argos where initially the systems are large but there's every intention. We heard some speculation about how small systems might become over the next few years into the few hundred gram size that gives us hope that those system swill come down in size and be practical for use on smaller animals over the years.
We saw innovative use of the LORAN system. Prior to GPS coming on line, the folks at the Starkey Wildlife Area in Oregon have innovatively adapted LORAN for applications that they had on elk, and we saw several presentations there that talked about their success in that area.There are some systems that are sitting out on the periphery, if not in general use, that are out there for our consideration. A very good presentation was done on a pulse-coded spread-spectrum location system that could be set up for studying presumably small animals, but which may be adaptable to other species as well.
Then there were some technical papers on specialized antenna systems that may be useful and important in the collection of automated data acquisition information. Telemetry data when you're not there and some of the finer nuances of antenna design for those systems.
Then there was also, to kind of break our train of thought away from warm and fuzzy things, an interesting paper on tracking insects using a harmonic radar system, fascinating technology, if not just a sort of off-the-shelf technology, certainly one that has applications in a very specialized field and allow us to track critters that previously could not be done.
I'm sure I've left out important innovations from many of the other papers and maybe I can rely on the audience to bring those out and fill in the gaps, but I think at this point, we've had such a full session in the innovation area that there was little time to really discuss some things, so hopefully that little description will bring back your thoughts on it and we can open it up to the audience for questions that you might have or that are hold-overs from those sessions.
AMLANER: Now this is your opportunity. What we'd like you to do, if you had some questions, if you could come to the microphone, identify yourself for the benefit of the recording so that possibly in the near future, we are hoping that if this all works out right, we may be able to get a transcript of these questions and answers and provide them on the Web possibly in the next few months. I believe the organizers are going to try to make that come true. So if you have some questions, feel free to come to the microphone.
Seeing no one at the microphone, I just happen to have a few questions in front of me from you already so we'll do a little bit of seeding here.
One of you asked the question: "Can VHF receivers be redesigned to filter out interference which appears to be increasing, particularly in the vicinity of urban areas?" Stan, you want to try that and possibly if anybody else out there has some ideas, feel free to come to the mike.
TOMKIEWIEZ: It is an interesting question I think with a lot of potential answers. Obviously, in urban areas there's a lot of radio frequency transmissions. Some of those transmissions for various reasons spill across into frequency bands that we use for tracking animals. If you're working in an urban area tracking a particular species, you probably have encountered a great deal of interference. Some of that interference is sort of out-of-band. It's not exactly where we track, where the frequencies are we're trying to track, and in general, it's possible to filter out-of-band signals. Most of the receivers that have originally been designed for most applications don't have a great deal of filtering for those out-of-band signals, simply because most of us were operating in areas that were very isolated from those signals. We were way out in the boondocks. But as more studies focus on urban areas, I see a greater emphasis on receiver design to handle some of those signals.
Perhaps Larry Kuechle may have some comments in this area about receiver design with regard to some of the things that they've been doing and please maybe some of the others might also have some comments.
LARRY KUECHLE: There really isn't a lot we can do to eliminate most of the noise because if it's broad band, it's going to come in to the channels that we're listening to. If we're talking about automated recording, there are some more things we can do because we can apply codes, I.D. codes or whatever, to the signal which will give us additional noise immunity, and those systems are out there and being used in a number of applications in different studies. For automated recording, we certainly can do some enhancement over the single beep-beep-beep because it is very difficult for any automated system to tell the difference between a beep and a burst of noise but if we add coding to it, that gives us some additional detection elements that we can use to do that. But if the noise is broad band there really isn't a lot we can do.
Audience question: What about matched filters?
KUECHLE: If you're listening with your ear, you've probably got about as good of a match filter as anyone can possibly design. Because if you look at the way the ear operates, the ear operates as about a 50 hertz-wide filter and that's a very narrow band filter; it's going to keep just about all the noise out of there. Plus your brain is very good at figuring this is valid, this is not valid. So if you're listening and I think where most of the people are having a problem with the noise in urban areas, they are tracking using their ear and that's probably certainly as good a system as we're going to come up with in the next many years.
Now there are some things we certainly can do in receivers to make them a little bit better, but it's only going to be a little bit.
RORY WILSON: My name is Rory Wilson. It's not concerned with the previous question, but I thought perhaps for the benefit of those of you who were not at Strasbourg, while Stan was talking about various systems that were being used, there was a presentation in Strasbourg by a bloke called John Bishop and it was entitled "The Last Omega Receiver." What he'd done is he'd looked at various systems that he felt he could use to position or determine the position of animals, and he homed in on the Omega system, a system which is global, and there are I think, correct me if I'm wrong, I think seven masts positioned around the world and these very low frequency radio waves which can be used and are used by aircraft and by people in boats and so on, and particularly in the southern hemisphere by meteorologists, people with weather balloons to determine the position of themselves or of objects of interests. So he sat down and for his Ph.D., he designed very small, I don't know, a 40 gram unit, which used the Omega system to position itself. It was rather like LORAN-C in many respects. What he did is he paid about 130 pounds, popping down to his local electronics store and buying huge and silly components at great expense, and he knocked up a couple of these, and then he stuck them on his head and canoed around the valleys in Wales to find out how accurate this was. He didn't work with animals himself; he just wanted to create a system which he felt was worthwhile, and he found a number of things about it. Firstly was that the unit itself was comparable in size I suppose to the smallest PTT's available today, and the positional accuracy was better than 300 meters. But the real kick came in that his system takes 10.2 seconds to get a fix at any time because of course, this system is transmitting the whole time. So for those of you who are obsessed with marine mammals or marine birds, and you have a problem with animals that only come to the surface or become available to receive positions that come out of their cave or whatever for a short period (the penguin comes up and breathes for 15 seconds and then goes again), you still have an incredibly short response time of this system where it can store the information necessary to get it's positional fixes. And he gave this talk and everyone thought, wow, isn't that amazing: that's just what we need and why didn't someone do this before, and got very excited about it. Then he said, well I'm doing my Ph.D. on this and I've got to tell you the Omega system is up and running and it cost about (I think the figure was) seven million dollars a year to run and the running costs are covered principally by the U.S. but by four or five other countries as well. It's peanuts for them, and everybody's very happy with it particularly in the southern hemisphere, but they're going to turn it off in September this year, we think, but we're not sure. And why are they going to turn it off? Because of the GPS lobby. So, we all screamed around in tight circles for awhile and spoke to a few people; could we possibly get a system like this running at least for the time that the Omega system should still be running. I don't know what the final outcome of this is, whether the Omega system has been turned off or whether it has got a new lease of life or whether it can be maintained. Once they turn it off and pull the masts down, then the cost of setting it up again will be phenomenal, but the cost of running it are really trivial.
I don't know how much other people know about it, but if anyone's got any clout and can make it stay on or anyone has any thoughts about it, it is a system that is potentially of incredible use to people in this room and anyone else who is working on telemetry applications because the receivers, like I say, John paid 130 pounds, really paying through the nose for his stuff. There are companies that would be willing to develop miniaturized versions of this with SMD technology which would be a snip, a couple hundred dollars for each unit and, like I say, in ten seconds you've got your position.
So just thought I'd make you aware of a sad day for us all, maybe they've turned it off already. But if anyone has any clout in this field or any thoughts about it would perhaps be worth considering as well.
AMLANER: Thanks Rory.
TOMKIEWIEZ: If anybody has any comment with regard to that, my comment was on the urban noise thing, but, please come up to the microphone.
CLARK WINCHELL: I'm Clark Winchell; I work for the Fish and Wildlife Service. This may be a question or a comment, but two experiences I've had with noise in my receivers or transmitters has been very predictable and I'm wondering instead of adjusting our system if there aren't knowns out there we can just build around and I'll give two examples.
One was on San Clemente Island tracking goats. I had over forty collars; however, collar 152.670 will always stick in my mind because when I got to the southern end of the island, I picked up Mexican fishing boats predictably every time. The other was in San Diego, using a series of transmitters, and all of the ones between 162 and 164 would always be interfered only with military fighting aircraft. So to me, two very predictable situations, that I knew in advance. I was working near the Mexican border, and I was also working on a military air station. Now when civilian aircraft went by or simply transport aircraft went by, there was no interference. But when fighter jets went by or AWACS went by, my signals would just lock on. It wouldn't be a beep-beep-beep, it would be a b-e-e-e-e-e-e-e-e-e-p, just until the thing went by. But they were predictable was the situation, so giving this to a manufacturer could you say, I'm working near Mexico, I'm working near Canada, I'm working near a situation that has known interference, I'm near a receiver or a microwave transponder for a PCS phone system, whatever. Are those, to electrical engineers, known systems that interfere with ours and therefore when you're in those areas, you exclude them. You say, you're on a military base, therefore you want your frequencies, 168 and above, or you want to drop down to the 152 range or something like that, and you help the end-user in that way, because he has knowns?
MICHAEL SAMUEL: By the way Clark, don't be surprised if you get a visit from DOD next week.
TOMKIEWIEZ: I guess in the way of comment, there are some signals that are obviously, or some bands and frequencies which have allocation. There is getting to be such a crowded frequency spectrum, that what may be true in one area may not be true in the other area. Often your manufacturer, whoever it might be, may have had previous experience in that particular site. But it may also be that they may be able to help you in the way you suggest and say, oh, I know that in that particular place there is a lot of interference in the 150 MHZ frequency band or wherever it might be. In and around urban areas, a lot of the signals that you pick up are essentially signals that are not the main transmission that you're hearing. They are signals that are suppressed. Even on your computer, you'll look on the back, and it says it complies with certain FCC standards where they have suppressed the noise level dramatically because if you're running a very fast processor, you generate RF noise and they suppress it. The thing is, we work with very low power transmitters and we have very narrow band widths and we're sort of listening for those pins to drop in the background and we're hearing other peoples' noise and their systems are unaffected because it's below the threshold that they're attempting to look at. We're very sensitive, which is how we get ranges of 20 miles to an aircraft from a transmitter that puts out a few milliwatts to a receiver in the aircraft. And so we're so sensitive that we end up with a lot of these stray signals that come along and not all of them are predicted. I wish they were.
One thing that I was going to make my comment on was that often you can go out, work with the manufacturer that you're working with and scan the area. It's not an absolute means because it could be a temporal thing: it comes on at the end of the month when they're doing a lot of fighter overflights or something like that. But you can go out in an urban area and spend several days essentially listening for areas that are particularly noisy with a scanning receiver. It's not an absolute answer, but it may tell you, oh, this area is really cluttered in this particular place and I'm going to get frequencies that are not in that particular area.
WINCHELL: Right, I did that actually, ATS was extremely helpful there. It turned out both the problems were solvable, one with a new collar and the other I was working with owls, it was night-time stuff.
ROBERT KENWARD: A related question to the interference one. We talked a little bit about getting background noise from the galaxy as an ultimate bottom line for things, and we worried about things from urban areas, but we're going to get more and more transmissions from space in the future and I'm thinking particularly if the proposal to put a large number of satellites up to facilitate Internet transmission. Do we see this as a problem and if so, should we start adding our voice to any lobby that wishes to prevent that happening and get Internet communication going down fiber optic cables only?
TOMKIEWIEZ: Larry's going to help us out here again.
KUECHLE: Larry Kuechle. I think there's a couple comments I would just like to respond to, Clark Winchell's first of all. That is that all the telemetry applications are considered secondary users, which means that someone else is also using that frequency, and they are the primary user. We're using it, or they are allowing us to use it really, because they figure we're so weak we're not going to hurt anybody else and if they are interfering with us, that's our problem. That's really where it's at in terms of the use of those frequencies. As far as getting interference from satellites, I don't really think we have a particular lot of worry there because most of those transmissions are up high in the gigahertz range and at least, so far, except for Argos which are at 400 megahertz, most of our transmissions are 225 megahertz and below and so the amount of noise that they're going to put in, at least from what I can see, is not going to be a problem because they're going to be up in a the gigahertz range, even the satellites that they're putting up for personal communication are up to 1.6 and higher gigahertz so they're going to be much higher than we're going to operate at.
LARRY PATER: Just a comment on urban noise. A couple of my colleagues, who you've seen as co-authors on my papers here, have studied this urban noise problem rather extensively, and if you're in an urban area and working, and are getting interference, something to look for is garage door openers. They are notorious and they don't have to be, it's not just when they're opening the door; it's all the time. Mine happens to be one of the worst ones. I have to replace it before I can do a study I have in mind.
AMLANER: There was another question from the audience that we received; something like this: "We are experiencing difficulty tracking desert tortoises tagged in mountainous rocky areas. The terrain is very rugged, boulder strewn and slopes of greater than 38%. The problem is locating tortoises due to signal bounce within the study area. Any suggestion to help pinpoint actual signals from reflected signals?" Adding in the questions of multipath, it seemed liked kind of a good, general meaty-type question that maybe a few in the audience or Stan might like to comment on.
TOMKIEWICZ: Maybe I'll make an initial comment while you're considering it, but I know many of you work in steep, dissected terrain.
The classic approach to handling multipath which are these signals that are bounced around and coming off the sides of wet slopes and wet vegetation and cliffs and so forth (we've all, if we'll admit it, walked in the wrong direction at one point or another in our experience), the general approach to that, that we often recommend to people, is to continue to reduce the gain of the receiver.
There's sort of a notch where we work in the mid VHF range, somewhere between, depending on where you are in the world, about 140 MHZ up to maybe a couple hundred MHZ around 220 or so, that most VHF telemetry is operating in. Some of you have assigned bands if you're with certain specific agencies or in certain countries. Some of these frequency bands, generally the higher frequency bands, tend to have more bounce associated with them. There are more things in the environment that tend to reflect those signals, and if you have state-of-the-art receivers from any number or the manufacturers that are out there, you should be able to turn down the gain of the receiver and rather than work with a signal that's very loud in your headphones when you're tracking, you turn it down and work with a signal that is very weak. What you're hoping is that the louder signal is the direct signal coming from the animal. Also, you can help yourself by using smaller and smaller antennas. Really sophisticated large antennas with uncontrolled lobes in the back of the antenna or poorly tuned antennas, poorly matched antennas, that are not essentially tweaked up a little bit can receive a lot of these bounced signals from lobes on the side of the antenna where they have a receiving pattern as well as in the front of the antenna. So, the recommendation I think that I would make to folks is to turn down the gain of their receiver and try and use smaller directional antennas in those kinds of situations. I'm sure others have had experience in that area as well.
AMLANER: Has anyone in the audience ever experienced multipath? I was wondering if all multipath was bad, but maybe that's something we need to detail. Go ahead Larry.
LARRY PATER: I don't want to be lionizing this thing, but something occurs to me that I didn't mention during my talk yesterday. If you're just analyzing the temporal variation of the signal for activity, multipath doesn't matter.
TOMKIEWICZ: That's true.
AMLANER: Good point. Any other questions from the audience? How about if we give Stan just a little break here? Go ahead please. Your name sir?
JAMES REID: Jim Reid. I'm with the USGS-BRD, and we track manatees in urban areas in Florida. We're confronted with some of the worst case scenarios, I think for both noise and bounce. We use a switchable, 101-dB attenuator (Kay Elemetrics, model 839) connected inline between the receiver and antenna. At dB settings of 10 or more, this device improves tracking performance by eliminating noise interference, reducing signal bounce, and enabling us to regain antenna directionality at close range. Trackers can calibrate range with dB setting to provide an indication of relative distance to the transmitter. Attenuators are now standard equipment for our tracking.
AMLANER: Thank you. How about if we give Stan a little break, for a moment. Well one more, go ahead, we're on a roll.
ROGER HILL: I'm Roger Hill from Wildlife Computers. Just a thought on multipath. The signal to your animal will surely be the first signal that's heard; it may not be the loudest. Could one not design automatic direction-finding equipment that triggers on the first received signal to figure the direction rather than the loudest, which would completely obviate that problem?
TOMKIEWICZ: People have looked at the approach. It's a valid approach. In the early days, it was very difficult to make the measurements in a very short period of time, like the 15 or 13 milliseconds that was available to have the receiver do very many sophisticated things. As receiver technology gets better, I see people considering that approach. One of the difficulties is, it's often difficult to determine whether any signal is the direct signal from the animal and in some cases, none of them are. You simply have to follow essentially the strongest one as the main bang as opposed to the main bounce.
AMLANER: Okay. I'd like to go ahead and move over to Jay and we'll give Jay an opportunity to make a short presentation and we'll move into questions on Session II, so be thinking about questions that are related to attachments. We also threw in some of the animal welfare questions, the animal care questions that you raised because of the audience, and we put those into this section.
JAY ROTELLA: I'm happy to be here. We had a large number of very good talks on the session on attachment methods and I guess I was chosen to do this because I'm somewhat famous in certain circles for having suffered dramatically a few years back with some waterfowl problems and trying to get birds to behave normally using radiotelemetry, and fortunately we have. We've made a lot of great advances and a lot of people have contributed to that so I'm happy to review some of the things from the talk and I put some slides together. But before we throw those up and turn the lights down, I often think and work with students who are going to use telemetry for the first time. I always try to tell them if you're about to go on a back-packing trip, before you get too casual about throwing these on the birds, just think about letting me pack your backpack and letting me adjust the straps and then you can't ever change it. And if you can always remember that when you're putting these things on these different animals, you might be a little more aware of some of the things they go through and if you have my friends, you definitely don't want them packing your backpack; you'll be carrying all their food, etc.
So with that, I think I'll just turn the lights down and we're going to switch over to the computer; I have the audacity to try the computer again after yesterday. So I was asked to summarize Session II, which was on attachment methods, and if you could just flip down to the next slide, the birds sort of dominated the session to a fair extent, about two-thirds of the talks, on this dealt with birds, and I think clearly, the reason is that birds are dealing with flight, and are typically quite small; I always like the fact that the large mammal people get to use radios that are a lot bigger than the birds I work on.
The marine species were also represented quite nicely as were amphibians, reptiles and some other small mammals and I think it's quite obvious, things that have to deal with high speeds and drag, etc., in water and air, we have a great deal of concern over. I wanted to just sort of review and put this in a context that overviews some of the work that is done, and there is sort of a variety of stages and it seems to me that one of the first stages is just try to get a tag that you can keep on the darn thing for as long as you want, and we've gone through some pretty major trials. I know a lot of you in the room have, especially those with passerines and small mammals and amphibians and reptiles, just trying to keep a tag on, and so one of the first things people have to do, and we saw a number of presentations along these lines, is come up with an attachment method in a package shape that you can just keep on the animal for the desired length of time and actually collect the data you want.
Then it seems like, and this is usually done hand in hand, the second thing here, is trying to come up with a tag shape and design of the attachment, that just doesn't blow the animal away. I worked with Mike Sorenson on the redheads; his paper was mentioned the other day and I was his technician the year we started following redheads. They worked fine in captivity, the birds behaved beautifully with these small neck collars on, the radios were small, they preened them in, and then we put them on the wild birds and within two days we knew we had major problems. Birds were getting their bills caught; birds were flopping around in the water, quite dramatically at night; owls were having a feast. It just didn't work and yet the birds would behave beautifully in captivity, so some of this business about trying to find attachments to stay on and then avoiding overt behaviors is not as simple and luckily you can usually lean on your peers but if you're starting new work, you really can't, it just is difficult. It takes a lot of our energy.
Then the final thing, eventually, once you get one of these packages that stays on, then we get into testing whether it really is impacting the parameter we're actually interested in and that's often not that easily done.
The next thing is developing this working attachment and I think just boring in on these a little bit. We see that, for many species, this is the state of the art. There was a lot of talk here in that session about just, we tried this kind of attachment with Teflon, and we tried this with something under the skin and so on, and we see that a lot of people are really at this first stage of just trying to get a radio to stay on the animal. It's really heavily done in the passerine world. We've worked on some passerines recently and on some woodpeckers recently and it is difficult to keep radios on those things. And then trying to have small tags that give you the life and range, etc., is quite difficult. So it's a problem and for many species if you're going to begin to work on those, you can just be aware that that's going to be what you're up against early on in the studies and I would caution to the graduate students and potential graduate students that don't be surprised if your thesis topic changes quite dramatically from some ecological question to some techniques question about a fifty page thesis on how to tag species A and we've seen it over and over and I don't think it's going to go away.
Then this second stage, boring in a little bit on this one, trying to avoid overt effects. It's typically done as you work out the attachment methods. Again, I mentioned that the captive studies don't always reveal the problems that you'll see in the wild and often you're at the stage where you're having small sample sizes either in captivity or in the wild because you're saying, there may be five different techniques I want to try and so you end up with a handful of animals with each and it's sort of the early stages, but you do see the major behavioral effects, hopefully, and then you go from there, and this is sort of the second stage. I think what we saw from the presenters that you really do have to be careful and thorough out there. You have to really consider which parameter you're interested in. Somebody else maybe studied survival and the radios have worked fine for that parameter as far as we know, and yet when you go to study reproductive behavior, or you go to study physiology, it may have a subtle to fairly large effect. So you have to think carefully about that issue.
The species: we've seen this quite dramatically in waterfowl and in passerines, where it works fine in one species, the bird handles it beautifully, and another species, there are massive impacts and the birds will not tolerate it. It can be to do with how radios are fit and different abilities of researchers, but I think it's probably often times species specific behaviors.
And then there are several other things that came up about sex, season, the conditions; I think in the ball-chain talk, somebody asked about frostbite and heat leaks and things like this and it was made quite clear that we haven't had to deal with that where we worked; that's a good point.And again at the bottom here, I can't say it strongly enough how much I've leaned on my peers over the years, the last eight years of doing this stuff myself, and we all need to always remember that maybe someone out there has done it out there and this forum is great experience for a lot of us that way.
The evaluation of impacts is interesting. There are a variety of methods that people have used and the first method of course is captive studies. They are very convenient and you can have control and treatment groups and you can actually test the null hypothesis of no effect of radio against a treatment group and a control that has no radios and that's wonderful and is quite useful but we do always worry about how valid those inferences are going to be to wild birds where they have a great number of additional stresses on them.
The wild studies are a little more difficult. It's rare that we can have true controls and so you cannot test the hypothesis of no effect. You go and you look for the transmitter with the least effect. We've struggled with that question. If you could have good controls in the wild, then you probably wouldn't be using telemetry in the first place. We've seen over and over that it's quite a pain to use. But you can at least identify a method that you believe has the lowest effect or the least impact on the parameters of interest by comparing multiple techniques.
The evaluation of the evaluations: I would just comment that we don't very often have the luxury of using large samples in the studies of the impacts of transmitters. It's quite difficult to do. We don't see a lot of discussion of power or how large an impact on the parameter of interest is acceptable. We all want there to be no impact, but can you tolerate a certain level of impact, perhaps, given the population models you want to use the data in, etc. That's really not often discussed and we put out a handful of transmitters and then we say we didn't find an impact. Well, you had probably no power to find an impact so there is some concern there that we really need to move further on and yet we all want to get back to the ecological questions and so you want to sort of get past this, but if you really are concerned, we do have people in the room that can help design the studies quite nicely. Clearly, more evaluations are needed, we probably want a combination of field and captive studies. The peer consultation again I think is always needed and typically when large-sample studies are done they'll find some effects. They'll either find that there is abrasion, extra preening, there was extra drag, on cetaceans, etc., and so you do need to deal with how large an impact you are willing to tolerate. I just wish you all good luck because this stuff is quite difficult. With that we can swing it over to questions.
AMLANER: If we could have the house lights on. Thank you.
RORY WILSON: There are three things I would like to say about this and they are basically unrelated. The first is for those of you who work on cetaceans or marine mammals or marine birds, birds that dive a lot, is that some years ago I tried harness designs on penguins and it worked very well. First of all, I tried them in captivity and it was fabulous. Then I tried them on birds in the wild and they seemed to be able to cope with this, and then we tried them in a swim channel, which is a thing where you can determine energy expenditure by respirometry (birds swim up and down a 20-meter length covered hideously with perspecs and then they can breathe in chambers in the end), and we discovered that these harnesses, which were invisible because they rode up under the feathers, increased energy expenditure by 240% of swimming under water at any particular speed and that happens, speaking to my hydrodynamics friend, because even though the harness is invisible, it pushes the line of feathers up. It was a body and a neck harness and it pushes the line of feathers up so that it breaks the contour of the bird and these things, as I said, are suppository shaped, and what happens is the laminar flow over the first part of the suppository is okay until it gets to the harness and then it becomes highly turbulent and, in fact, subsequent to this, you know, just watching them swim through the water, it was quite clear that they were really having to work hard to do it. Anybody who's working on, cetacean people generally don't do it anyway, but who's considering cetacean-type harness stuff, don't. It really is the kiss of death unless you're working on something that moves so slowly, like a manatee, that is not going to be a problem.
The second thing is elasticity in harnesses with respectto potential effects and something that I built into an original penguin harness, and I don't know how many birds I lost but probably five or six. I had a harness with an elastic section in the neck and, you know I put it on (it was just a marvelous fit and it really looked terrific) and the birds just disappeared. I was getting really worried about them; I thought, is that just coincidence that they are gone, then you try another one and eventually you really begin to worry. One particular bird went away and came back and demonstrated what the problem was. Incidently, this wasn't a problem in captive birds at all, and that is, that if there's an element of elasticity within it, then, in my case, in the neck section, what the birds tend to do is they put their beak down and they pull it and if there is give in the system, then it sort of makes them pull it more; they get obsessional about it, can't stop ("oh, it gives a little bit"). What happens is, it has a metastable position; while it's around the neck it's okay, and they can pull and pull and pull and then "ping," harness is caught in the beak. Then of course, it makes swallowing tricky unless you've got a straw or something to swallow around the side. And I had a bird come back and there it was, the thing stuck in its beak and I realized what the problem was. There were two ways in getting around it: one is not to use elasticity at all and that works very well. The systems we used then were very thin ultra-bird-friendly leather, which was sewn on to size on the bird; you did need to put in a few cut-away sections, so you can just cut it and fit it to the exact size of the bird. The other is to have some sort of long-term elastic, not the sort of boing-boing-boing stuff that people use, but something that gives over time. So the bird doesn't feel it's getting something when it pulls at it.
And finally, the other thing is that the talk on geese and tail attachments and it takes thirty minutes, which was yesterday or the day before, (my time sense is completely out of the window). We attach loggers to a variety of seabirds, not for very long periods, six weeks or a couple of months at the most really (in the case of penguins, it's up to eight months), but the first line of approach is always to use this Tesa tape. I don't have shares in Tesa, it's a German company; I have no interest in the fact they sell a lot of tape, but it is terribly quick and it is terribly easy, and if you look at it by the poster over there, you think, this will never work, but try it out, and as long as the feathers will hold, the tape will not let go, unless you work on macaws or something. So, any of you who feel, I need three weeks study time, if the feathers don't pull out, then you don't need a harness and then you don't have all the problems associated with harnesses, callouses, and the birds can't fly properly. Thank you.
AMLANER: We do have a couple of slides. Bob Kenward, I believe, is going to make a comment and had some slides to back up the comment. Could we mount those slides right now?
KENWARD: Thanks. If we go with the slides, I'll keep it quick. For the first fifteen years of my radio tagging, I wouldn't use anything but tailmounts because I was horrified by all the stories that there were about problems with harnesses, so I stuck with them and we've got a lot of data from comparisons with ringing large numbers of goshawks showing there's no impact. However, and this is only to remind me to say it, we now have tags which last routinely for four years. Birds don't normally hold on to their tail feathers that long, so you've got to go to a more long-term method and so we had to start using harnessess and we had to try to develop a technique that would work and be safe. And we developed a method that was reasonably easy to show assistants where we could measure the length of straps. Although I'm not going to show it here, there is in fact a relationship between the weight of the bird, at least in falcons, with the back (the hindmost) strap. And there's quite a good relationship there and you can predict even from a large falcon down to a small falcon what the size should be. I think we need to start moving particularly with harnesses to developing methods of assessing what it should be because, you know, then people can get it right a bit more. Certainly our comment would be that you could use harnesses but if you're more than 5% out on the backstrap, it will either rub or cut. It'll rub if it's too loose, it'll cut if it's too tight. So I'm not surprised that there are not some studies that have shown bad effects of harnesses.
And finally, we always at the start of a study, and this is just preaching, but we always test whether there is effect, starting with a base method, and we've used tailmounts as a base method, small tail mounts because nobody has yet managed, I think, to show a serious effect with a small tailmount. I'm not talking about quite heavy ones, but I'm talking about 1% type ones, and then you can test it. It seems abdominal mounts in some species are a similar comparison, but the point I really wanted to make about this is some attachments you can do without too much training. You can do from reading about it in a manual. You can put necklaces on, you can put tailmounts on or many of them. But nobody, no biologist in their right mind, would think about putting an implant in without at least consulting a vet and I think probably no vet would, bearing in mind their welfare training, would think of doing that implant without going and working with someone who has done similar implants elsewhere and we ought to use the same thing for harnesses. We should not just be putting on harnesses willy-nilly, and I'm terrified about publishing things saying no effects of harnesses with very extensive tests but people will then, willy-nilly, go over to them without looking at this. In Britain, we do have actually a system which controls what you do and you have to have some training, and I would recommend that generally that nobody should put on difficult attachment techniques without going and doing it with someone else first.
And while I'm up here, I would like to take the opportunity as one of the foreigners, to say how much I've enjoyed being here and I'd like to thank particularly the organizing committee because they've really done a lot to help us. Debbie and Wayne Norling, particularly; Jerry Godbey has worked wonders helping us get computers set up; Pam Pietz, Jane Austin, of course. And I would like to thank all of them and perhaps you'd like to join me in thanking them too because I think they've helped a lot of people from North America as well.
AMLANER: Robert and Jay, just as a bit of a segue, there is a question that came from the audience that is somewhat related: "Can we better coordinate/consolidate efforts to conduct evaluations of transmitter attachments?" this individual asked. "Perhaps this group could develop a white paper to address this issue or develop a committee to brainstorm and come up with a game plan." Other than the developing of a committee, I've always been a little bit worried about more and more committees, but I like the idea of, say, maybe a white paper or possibly a central area that could help disseminate information on attachment techniques and effects. Jay, would you comment?
ROTELLA: I think it would be good to have a central repository and I'm sort of looking at Jane. Jane, you don't know it yet, but I'm curious what they think. It seems like it would want to be someone at one of the labs and maybe you have a comment, Pam or Jane on that. I think this has been a great start in that direction and so I guess I'm sort of leaning on you folks.
JANE AUSTIN: We really don't know where this meeting may go in future years. I can see some potential mechanisms. Certainly mission of our agency, the Biological Research Division, is to provide non-advocacy science and information to scientists, especially within the U.S. but certainly around the world as much as we can. With Pam's deep interest and experience in attachments, she certainly is perhaps (I'll put her on the spot) the appropriate person to help out. Perhaps one of the other places that might provide a central point for such a white paper or place to put it out, one of our sponsors is the TWS GIS and Remote Sensing Working Group which includes telemetry. I've been a chairman of a telemetry subgroup within that group, perhaps by default, but that may be a mechanism by which we could put forward a white paper which would be outside the government hassles we sometimes have. But we'd certainly welcome anyone who would be interested in participating or leading such an effort. I know there's a lot of ideas and there are a lot of concerns, which Josh Dein and Karen Machin will addressing later this morning, and which will probably cue a lot of peoples interests. Again, if anyone is interested in working on that idea, we welcome you to get in touch with us.
AMLANER: I presume we could probably use the E-Mail as as central source to get feedback. Go ahead.
PAM PIETZ: I was just talking yesterday to Mark Fuller, Kirk Bates, Chris Kochanny, and Kevin Kenow, about the fact that many of us have tried to accumulate all the literature we could find on various attachment techniques (and effects of) and, rather than everybody reinventing the wheel, could we put together our stockpiles and serve it someplace. ATS has started a Web page; we're considering asking them to serve up the literature portion, the published stuff, but we still have the question, even if we can pull that off and have someone keep it up, they don't have the access to our library resources so we'd have to have people feeding them the information. And we're not all that masochistic, so we haven't decided quite how we're going to pull that off. But the unpublished stuff, that's why we organized this, because we're all fighting the same problems and the failures don't get in the literature, so trying to keep each other from reinventing the wheel. Potentially the website could serve this also but it would have to be an interactive thing where people could suggest and give us input on the unpublished stuff. Maybe we can pull it off, we'll try, but we'll need a lot of help.
ROTELLA: I'll make one more comment. The other is the way that we normally do business, which is contacting peers, because it really is about the only way you're going to hear about people's failures, that we've tried two transmitters of such a type. Gee, you know, when I started, we had someone warn us, gee, don't put superglue on that harness. We put 25 of them out and five days later, when the superglue ate through the tubing, we had all the radios back and it was too late to trap the birds and the students got to an extra field season. So really, chatting with folks is one of the most important things you can do.
AMLANER: Yes sir, your name?
GARY WHITE: Gary White at Colorado State. It's really difficult to get most of these little studies of effects published because we don't have power, and none of us, it's like Jay pointed out, none of us want to do the technique studies. We want to get on to the bigger question. So given that we're not going to be able to publish a lot of these -- it's too difficult because they lack power, we haven't got a long term study; they're not going to get published, and they're not going to show very much by themselves -- but there's a technique and statistic that's being used a lot more now, called meta-analysis where we bring these things together. One of the big problems of meta-analysis and statistics is that there's a censoring process in the literature. Only studies that show effects are going to be published for the most part. So one way to avoid this is, sure, you keep track of your peers, but probably if we could get Jane and Pam to take over this website, and everybody put their little bit of a result out there and a description of how they did it, at some point, we're going to have enough information and it's going to become clear that we've got a real problem even though no one study is particularly important.
AMLANER: Very good point. Very good point. Yes sir. If you could give us your name as you come to mike.
KENNETH BEVIS: Hello, I'm Ken Bevis, Yakama Indian Nation; I work with spotted owls and something that's occurred to me in the discussion is that the interface with the engineers is very important. And especially when we start working more and more with endangered species, I think we might want to devote some thought to materials, technologies, and de-attachement techniques. I'm thinking of many of the animals that we attach these things to and then we wish them well, and I haven't really heard a satisfying discussion of about how long it takes, for example, the thread you use for such and such to decay. So this is more of a thought for the group. But I think that gets into the realm of engineering. Fortunately when you tail mount, for example, you know the bird takes care of that for you. So, a thought for the group.
TOMKIEWIEZ: Just in the way of that comment, often being a manufacturer, sometimes it's not always information that's relayed back and very often with some of the corrosible or degradable sorts of lengths, our experience has been that we see a tremendous amount of variation depending upon the climate that it's been deployed in, and the specific environment, and I think it is a good idea to try and get as much information in that regard as possible. Sometimes, these expansion collars, for example, that have been around for years, where three rows of stitched material of a certain kind of thread are used, or a latex tubing. There's no control on the part for the manufacturer of the telemetry for going back to the latex tubing manufacturer and saying to the latex tubing manufacturer, oh, by the way, your tubing has changed, it doesn't last on bears the same length of time it did before and they kind of look at you and go, it doesn't last on bears? They have no idea what you're talking about because the thickness has changed just a bit in the tubing or the composition of the materials or the amount of IR inhibitors that they've put into the polymer have changed, and so these are very subtle sorts of things that can occur when you buy materials. Often what happens is you buy the big roll of thread so that you hope that in your career, you'll never run out of it.
AMLANER: That's true. Very True.
PIETZ: Pam Pietz again. Back to Jay's comment about getting to talk to your peers. We have the luxury of having a lot of colleagues working in waterfowl that we know, who to go to to ask questions. A lot of first-time telemetry users don't have that, and the Biotelem listserve has served a really large number of studies in the last couple years, but it's only as good as the people who are willing to respond to these poor student's dilemma's and I do encourage as many people as possible to get on to it so that if the student is out there looking for who does have experience with this kind of a critter, these kinds of materials, this kind of attachment, that they'll be able to get some help.
AMLANER: Well, with this as just a little bit of a lull, there are a couple questions that we still have left in terms of animal welfare and I'm going to be a little bit short in terms of questions on physiology. I noticed there were only two or three that came in, so I'm going to use a little bit of chairman's prerogative and push a couple of these animal welfare questions into this next session if that would be okay.
I'd like to begin this session briefly by saying that I'm going to act as a bit of a provocateur as it were. We had a session on physiology and behavior and activity, as you are all aware, and after that session and sometimes before that session, I was asked and others were asked questions about what types of things were available and where are we going with biotelemetry. You've noticed that there's been a very strong emphasis here in terms of the technologies, especially in the first day of this session, and I notice as I go from telemetry meeting to telemetry meeting, that more and more people are interested in knowing what's going on inside the animal in terms of it's internal physiological parameters. We even heard from one or two of the speakers during the physiology session and questions asking about what might be available.
A few years ago, we asked this question of some of the manufacturers and I believe several of the companies in telemetry are represented here and I would like to challenge you to give serious consideration to the possibility of developing for us, on our behalf, because we need it desperately, some multi-channel biotelemetry devices that could be utilized on a broad spectrum of animals.
First of all, of course, we would need to develop some form of a bioamplifier, and whether or not the manufacturers go through this particular procedure, whatever comes out of that design in terms of a bioamplifier, we really do need to end up with on-the-ground animal experimentation. We need to have multi-channel because it is not convenient really to answer questions in a biological framework when you only know heart rate but you don't know, for instance, stroke volume in an animal. You don't know blood flow in an animal. You don't know brain temperature or EEG, the brain waves, or the muscle activity of an animal. And too frequently we find that yes, we can come up with a quick and dirty one-channel transmitter. Several of us have done it. We've even done two-channel transmitters and of course NASA, with their wisdom and pocketbook, have developed multi-channel devices but which are not available to us on the ground. So I would like to challenge the construction people here if they could just build us some, we'll be more than happy to do some testing for you, and I believe we will provide a market for these devices if they become available to us.
The simplest idea here would be an implant that could in fact be used as a repeater system for a larger animal, a collar could actually retransmit the physiological parameters. That implant might, in fact, act alone so that the system that you develop could have a two-fold purpose for smaller animals; all you might need is the implant for short distance studies and for larger animals you could use the repeater to increase range.
Another simple approach would be to use some of the voltage controlled oscillators with designed preamplifiers for brain waves and electromyograms or muscle activity and various other types of physiological parameters and come up with standard bioamplifiers that could be either connected together by the user or ordered directly over the shelf, rather than develop brand new bioamplifiers for everything. We do know what basic physiological voltages are necessary to accomplish recording of these brain waves and EMGs and as such, it should be a reasonable process for an engineer to be able to develop these bioamplifiers for us.
The other area that I found that seems to be lacking in biotelemetry is proper electrode development. If any of you get into this field, you're going to find very quickly that you have to develop custom-designed electrodes for your animals. I've talked with many in the biomedical industry and usually when you go to the human biomedical field, the electrodes are extremely expensive and highly tuned to things like pacemakers and other electroshocking devices and they're often way too big for the animals that we work with, and they're awfully way too expensive. Probably the most convenient type of electrode that most of us use is the loop of an end of a wire. This can be easily made from a Teflon-coated stainless steel wire from Kuner, but it does not have a very good frequency response. And if you're looking at brain waves or a muscle activity and need a high frequency response, you're going to have to go into gold-plated electrodes. These are not very easily available and you usually have to make them, going through the plating process yourself. We need that type of resource.
Another area, I think, and this may actually be talked about a little bit later in this morning's session, is the aspect of anesthesiology. If any of you have gotten involved with animals that you were unable to find any published information whatsoever on the recent levels of anesthesia, for instance, halothane and isoflurane are quite commonly used now in veterinary practices, but if you have an exotic animal and you try to find the volume or the percentage of isoflurane needed to induce anesthesia or maintain anesthesia say for a long period of time (due to an implant surgery lasting for two hours), you won't find this published very readily.
Now several people asked me the question, what about the zoo environment and the veterinary practices within zoos, and I think this may be a resource but we haven't actually tapped into that yet. We need to have a common area where we might be able to find some of these measurements and values.
We rarely ever, also back up what we're transmitting with hard-wire technology. I'd just like to point out that while these devices here, these polygraphs are rather expensive, most physiology laboratories at universities do have them. I would challenge you that if you get into the business of trying to telemeter information and physiology, you need to also back up some experiments with hard-wired measurements so that you can do cross comparisons to make sure that your biotelemeter, if you were custom making it, is actually transmitting the signals that you expect to receive. One thing that I'd like to point out right up here -- this is a pulsoxymeter. It's designed for human work, but we have been in our laboratory, taking the pediatric sensors used for babies in neonatal critical care programs and adapting them to bird work. We are able to take these pediatric sensors and wrap them around the legs of our birds and during all the surgical procedures, we're able to measure the oxygen level in the blood, we're able to measure heart rate and the like. Also, we use standard biomedical strain gauges to measure respiratory responses during the surgeries. These are things that can be adapted at reasonably low costs. That pulsoxymeter right there can store data for us up to 24 hours and it also costs less than a thousand dollars. And in the grand scheme of things, if you're needing to monitor the physiology, a thousand dollars is a small investment.
One other thing too that we find, is that the placement of electrodes in animals is clearly not very well cataloged. If I were to ask you to give me an idea of what the standard three-lead or four-lead or five-lead ECG would be for, say a duck, we might not know the answer. But if I asked you that for a human, it would be highly tuned to a standard of the industry. And I would like to challenge all of us to really consider this as a serious question. If we're going to get into the biotelemetry business, we need to think about where we're placing our electrodes and, if we're going to do cross-comparative studies, each of us is going to have to consider putting our electrodes roughly in the same place or else the data that we look at will not be very comparable, especially if we're looking at wave forms or frequency analysis.
One thing that we can use, and I saw something like this in one of our papers, in fact, the squirrel paper, where the SCN was lesioned, used a stereotactic manipulator made by Culfins Industries. Culf makes the stereotactic manipulators for a broad spectrum of animals and you can use these to implant electrodes very very accurately throughout the brains of animals.And lastly, we need to be very considerate of where we put the electrodes and at the same time we have to do the necropsies to ensure that where we thought we put the electrodes and where the electrodes were placed, either in thebrain or in the animal's cavity, the electrodes were maintained. If you do a long-term study, as we do often times for a year or two or three, on a bird, the migration of an electrode, theelectrodes in the skull cap right here placed on the cerebral cortex of the bird, if those electrodes migrate, then that's going to modify the results of our experimentation.
So with that, I would just like to throw out these ideas to you. Hopefully, the industry will listen to us and that possibly this will spark some interest in developing technologies that will allow us to do these measurements.
With that, I would like to have the house lights up and throw out some questions right now. I've been told that the microphones in the audience areas loud as they're going to get and if you could just speak directly into those microphones, it will help us greatly. Thank you.
This was a question that came from the audience: "Will implants become common on small birds, similar to waterfowl, for radiotracking purposes or biotelemetry purposes?" I noticed that several of you did do implants and I'm very pleased to see that. What is the feeling of the audience? Have you actually found that implants are becoming more popular and that you see this as reducing the effect of stress on the animal?
ROTELLA: I can comment on that. The folks from the Institute for Wetland and Waterfowl Research and I did one of the papers on this and several others in the room have worked on this topic as well.In the reproductive arena, with breeding birds, the implants are working quite well and the Prairie Habitat Joint Venture teams are using those. It's not going to become common place with current performance. Those radios don't transmit very far and the folks that read the paper that we put out called in and said, "My God, these things transmit like a few hundred meters under many circumstances and we're dealing with migratory birds and we're dealing with wintering birds, etc." So it's not going to be a universal solution in turn because you just can't get data. So while we would like to do them, they're logistically difficult to deal with, that you can overcome, but the data needs, that's not going to be overcome unless the technology advances.
AMLANER: One thing that I found, Jay, is that when you're working with multi-lead biotelemetry, an implant often times is impractical, even in a duck, or in geese. We have to actually bring the electrodes through the skin in some fashion. The problem that I've been faced with bringing these wires out to a transmitter that will be harnessed to the back of the animal, is that, at that point, where the wires come out, unlike a couple of the papers that showed us antennas coming out and very very cleanly having the tissues adhere to the antenna. If I bring nine or ten or twelve leads out in that particular spot, there is a lot of abrasion and movement and the like and the wound is not going to heal properly and we've noticed that there is tissue necrosis at that point.
I was very pleased to hear Kevin Kenow talk about this Dacron cloth that is actually somehow heat-sealed or bonded to, was it teflon or solastic? Teflon tubing. This looks to me as a very very promising way of exteriorizing wires that need to be implanted in an animal and brought through the skin where the teflon will bond or I guess not necessarily bond, but provide an abrasion-free area and the dacron cloth can be used actually to adhere to the tissues either through suturing or through connective tissue redevelopment in that site. That to me looks very promising. Now Kevin told me that that's all custom made, I believe in a Canadian company, did you say? And so we're going to have to find out the address and see whether or not they can make various bore sizes for different size wires coming through the animal.
Are there any questions from the audience in this area? In terms of your interest, I'm kind of curious, do you see biotelemetry as something that is important to you? Rory, go ahead.
RORY WILSON: I don't do biotelemetry in that sense, but I do have a question. Does anyone have experience with a situation where you have a sensor placed in a particular position on an animal which leads then to either a logging or a transmitting unit via a series of leads or a lead, leading over the body of the animal to that transmitter which is somewhere else? Any ideas that might help the lead become less irritating, noxious, or anything else to your animal, because you've got the elasticity problem and you've got the fact that the bird, I don't know, might pull it off. It's completely new to me and we're playing with it for something and if anyone does this sort of thing and has any tips, it would be fun.
AMLANER: Do we have any kinds of data logging experiments going on like that, or any phyisological monitoring? I know, Rory, in my own experience, we do similar kinds of things -- multilead brainwave analysis projects where the transmitters typically are on the back of the bird and we have to exteriorize the wires. We bring the wires down the neck of the bird, using a trocar under skin and still come through the skin right at the base of the neck and then plug into the transmitter and then the transmitter will run for "n" days or "n" weeks and then we will recapture the animal and often times replace the batteries in the transmitter since we're really very power hungry in terms of our transmitter design. We are not data logging on the bird and I see this as a feature that could be very very useful if I could do some post processing of brainwaves and actually do FFTs possibly with the device and just store the results of the FFT or the power spectral analysis of the brain waves. I don't need all of the brainwaves per se. I don't have to see that for 24 hours. I just need information of the power or the energy in the brain. I would imagine that we find that there is a bit of irritation to the birds at the base of the neck as we mentioned with the way in which we exteriorize the wires and I don't see any way around that until we get to a better way of sealing that point or implant the whole device and use a transcutaneous or subcutaneous way of transmitting information across the tissue of the animal. Tom Frier, some of you may know, who is with NASA for many many years, has developed devices like this. They've been published in biotelemetry books of ten and fifteen and twenty years ago. The designs, while the electronics may be out of date, the concept, I think, is still there, where you could actually have a coil underneath the skin of the animal and transmit through that coil, provide energy to the transmitter through that coil, and also get information. That might be a way of minimizing any exteriorization of wires.
TOMKIEWICZ: A comment in that regard: One of the things that often has been a problem in terms of making a very complex system has been having it end up weighing several hundred grams rather than several grams, and most of the devices even today which are being deployed on wildlife in free-ranging situations on the small critters, people are shooting to weigh a few grams and obviously there has to be some give there. We've been involved in a project recently which kind of leads in this direction that Charlie was just talking about in that we have some people who would like to monitor the presence or absence of cubs relative to a female bear. We're going to relay that information from Argos and determine the presence or absence of cubs, by the development of an interactive system between the Argos transmitter and a controlling receiver in there that monitors the presence or absence of cubs. It's our intention at that same time to go ahead and take that same technology, for example, and collect up information on heart rate over a similar link or body temperature over a similar link and then be able to relay that through a satellite system. It's not exactly what you're describing but it tends to lead in that direction. I think the key things that tend to make that happen is, there is a greater range of available receiver technology becoming available in small chip form to allow that to occur. There is also smarter and smarter microprocessors that run at low voltages and can store the information necessary to make these sorts of things practical. And often even though the electronic circuitry may be quite small, we have to consider that the majority of the weight in most telemetry systems ends up in a situation where you're trying to power them and the batteries and the packaging are critical. If the packaging itself fails, obviously, things will not work well and penetrations like this looking for small voltages are subject to all sorts of in a saline environment if you go back to Tom Frier's work where the real crux of the issue seems to occur. We've developed heart-rate units over the past few years where we feel like our packaging has reached a level where it is reliable over a protracted period of time. I think some of these things will come about. In that direction, I'm not sure how multichannel it will proceed to, but the more intent the heart rate unit is a very different thing, for example, than a wave-form EKG, and the amount of information to be transferred is very different and so it becomes an information transfer problem as much as anything else.
AMLANER: I know I'm probably stressing the minds of the engineers when I ask for a six or eight channel EEG transmitter. The technology there is very difficult. If you have a Holter monitor and you're in a hospital, that's no big problem. You have this great big package that you can put around your belt and have the wires connected to you, but when you put it on a bird, the bird will never fly.
I have a question here from the audience which I think might be an interesting potential segue right now. We're asking questions about doing these invasive techniques on animals, in say, physiology, and one of the members of the audience asked: "Under what circumstances should telemetry be restricted as an inappropriate technique?" Now I showed you a little bit of a gory slide up there of a necropsy. I can honestly tell you that we do surgeries on a fairly regular basis in my laboratory and I have trained graduate students over the years to do surgeries. I've worked very closely with vets in my laboratory and I feel very strongly about the close association of graduate students with the animals that they work with and a veterinarian and the professor, but I do not necessarily believe that we should completely give over to the veterinarian every detail of the surgical procedure. I believe the graduate student should be involved intimately with their animal so that they know what is going on and what is going in. That is not to say that they should be oblivious of animal welfare issues but when we had this question come up, I'm wondering, do some of you feel that doing invasive measurements on animals or invasive surgeries might be inappropriate? Any comments? Please go ahead, come up to the microphone, give your name.
PATRICIA DE COURSEY: I'd like to point out relevant to the earlier comments that we don't just turn over a new process without a lot of practice. While I showed some fairly, looked like fairly simple surgery yesterday, in practice, that was worked out on about a hundred hamsters. And we were learning from people who had done the technique before us. I think that certainly it would be appropriate to have a substitute animal preferably a lab animal that you could purchase from a supply house and use that to practice many many times or to work out the stereotaxic coordinates. For example, take a large number of animals, and you don't really want to use your wild populations, so that learning from other people, but very often it's not practical to hire a vet to do your tasks, nor desirable. I think the use of this surrogate practice animal may be an answer.
AMLANER: Other comments? Go ahead, Bob.
KENWARD: I'd like to address this question on the welfare side of things and also with regard to exteriorizing antennas because it's relevant. In the UK, it would be extremely difficult to get permission to do something like that. We have a very active animal welfare system. We have to have licenses for anything like this. It's widely known that the animal welfare extremists can tap into the systems and get the names and you're liable to find letter bombs coming through your mail if you do this sort of thing. There was a question a little bit earlier on saying we're not getting reports of the failures, the things that aren't working. Well I was thinking about this, and thinking, well, maybe not through the scientific press, but there is a real risk of them coming through the newspapers. We really do need to think about these issues. My personal view is that for some species, for instance, diving ducks, and for the marine mammals, then you've really no option but to go for a technique like this. If as you suggested, one could go for a loop under the skin, then that's great, because if one can get away with, but bringing the antenna through is very dodgy, because if a member of the public then finds that animal, and we get plenty of our animals found, the automatic assumption is that it's the tag that's screwed that animal up. If that animal has been hit by a car, or something like that, I mean the car driver feels guilty, they're very happy to lash out and suggest it was the tag that made them hit the animal. This is not trivial; this happens. And just as we have actually now when we get bomb threats, we have a list of questions that have to be asked so now, we have to address the question when people ring in, saying we found this animal, what's this harness doing on it? There are a list of things that we can say from our experience because we have data on this, that there is no effect, no the animal is not suffering from this, we've compared against lots of controls. What worries me, is with an exterior antenna, it's fine if it's interior, no one will see it. That's grand. No one will know and I'm delighted that this is happening mainly to marine animals because the chance of them being found by the public is reduced. They're likely to sink, frankly. So, there is a reduced risk. But with these things, if someone finds that antenna, they'll say, "Hey, what's going on here?" The next thing, it's in a vet's lab asking for a full report; it's with the newspapers and, wow, the balloon will go up. So I don't think we're going to be doing that in Britain at all probably.
AMLANER: Good point, Robert. Go ahead. Jerry.
JERRY GODBEY: My name is Jerry Godbey. I was sitting herebecoming a little concerned about . . . [tape change]
AMLANER: Isoflurane and the like can be worked out in a laboratory. In fact, what we have routinely done is taken a similar species of bird, that was not necessarily something that we were actually deciding to do all of the implant on, and all we did was work up a dose response. Then what we would take would be some of our study animals and just do nothing but dose response curves on those and compare before we ever even used a scalpel. We often would do this over a period of weeks to work out very carefully the percentages. In particular, we look at these long-term experiments where we're implanting for two and three and four hours and we need to be able to monitor that animal very carefully.
If you're just putting an animal under an anesthesia for a few minutes, there are probably some guidelines that would come from the veterinary literature, but very rarely will you find these long-term type guidelines like you mentioned.
Okay. We're at a point now where I'd really like to move over to our last session. Michael, here, has been verypatient with us for the last hour and half, and I'd like to turn the microphone over to Mike and let him do the presentation on the data-processing session and then we'll field a couple questions from the audience and/or the papers we have here, Mike.
MICHAEL SAMUEL: I have a couple of slides I wanted to go through.
First I would like to join Bob Kenward in thanking and complimenting the organizers. It's been a really good session. And also the presenters; all the presentations have really been great and just kind of observing some of the discussion going on at the poster session, I gather that most of you think the same, and it's created a lot of things to think about.
I guess I view sort of my role, or data processing and analysis, in a bit of a broader perspective maybe than the individual session was. So I think some of my comments are probably going to bleed over to some of the other sessions. Also, I have perhaps a biased perspective on some of this, and I think a lot of you have some different and probably a lot better insights on some of these topics than I do, and I hope you will share those insights as we move on to the discussion.
Basically, I think there are probably about four areas in terms of data analysis and data processing that we probably want to key in on or at least to discuss, I'd like to discuss. The first broad topic is sort of animal location and movements. That has been a focus of a lot of papers and a lot of posters that are here. Particularly things like accuracy and how accuracy relates to scale issues and some of those problems.
Animal use patterns I tend to view that as sort of a broad perspective. That involves things like habitat use and resource use, home-range kind of applications, but also can be things like food use or some of those aspects as well.
Survival, I think has been an important topic for radiotelemetry, although it didn't receive much attention in this particular conference. There are a couple of papers that dealt with survival, mostly related to other techniques perhaps.
And then finally, the sort of glue that holds a lot of this together, makes it more feasible for us to do a lot of these things, is the data processing and the computer software aspects and the computer hardware aspects and I would like to touch on some of those as well.
Accuracy and scale issues. I think we're all familiar with some of those concepts. We've had a range of those presented. Accuracy, I think depends a lot on the scale of the question you're asking. We saw, for example, in goose migration studies, where when we're dealing with migration patterns or large scale movement patterns for marine animals, that probably scale is not a very important issue.
However, when we're dealing with elk habitat use, moose habitat use, some of those questions, accuracy becomes a much more important issue and I think a lot of us are beginning to struggle with how we do accuracy or what we're going to do about accuracy with some of the methodology. I think perhaps that we had a handle on how to address some aspects of accuracy for the conventional telemetry methods but I think we need to spend some time now looking at accuracy from the Argos systems and from the GPS systems. We have various levels of quality of data that are coming across from those. I'm not sure we have a good handle on how those different levels of quality relate to accuracy of the locations that we know about and how those then in turn relate to the objectives of our study.
Another issue on that, that I think surfaced during the meeting for me, was there are some places related to topography or perhaps habitat where you're not always going to get a location on the animal. That raises some potential bias issues I think which need to be considered as well. I don't think there are answers on these yet but I think they are issues that people need to be working on.
Methods for data capture. I sort of view that as the first step in terms of data processing, if you will. There were a number of aspects and a number of papers that talked about ways to do that. They range all the way from these automated systems, like Argos or GPS, where you can download the data to remote data loggers to actually data loggers that ride right on the animal itself and collect lots of information about where the animal has been and what the environmental conditions are like.
I think frequency of collection is something that's really becoming much more of interest and much more feasible for all of us. I guess I sort of feel like we've been in a data poor environment for quite awhile. It's been expensive to get locations and time consuming. It seems like we're about to go to a very data rich environment now where we can get lots of information through these data-logging processes, information that's pretty accurate. I should say lots of data. The question is, does that equal information and how do we make it equal information? How do we learn from that data when we've got 400,000 observations on moose. How do we convert that into information about what's going on with those animals. I think there are some real challenges that are going to come up in that regard.
The movement models I think that we've looked at in general have been probably fairly simple. Animal moving at various rates from point A to point B perhaps, average rate of movement, some of those kinds of things. I was really glad to see some very different kinds of approaches to that, particularly in some of the manatee work, where people are beginning to look at how the biology might tell you about what movement patterns are likely to be and then how you use those movement patterns to learn more about what pathways animals are using or where they are likely to be.
I want to offer some guesses about where we might be going in the future as well, if I can, and I hope that a lot of you will jump in on this as well.
Obviously, we're headed towards real frequent data capture. That's really feasible now. We saw applications that are in fact real-time data capture as well, if you want that capability. That brings up, I think, some of the problems that I've mentioned already in terms of so much data and frequent data that we need to figure out how to deal with it effectively in terms of our analytical methods. We obviously now have remote applications. We can work on animals that we couldn't even get to before. We can get locations from areas where we couldn't get to before. It really removes some of the problems that we've had in the past and some of the biases that we may have had in the past of locating animals only when they're near roads or some of those problems.
We have now the capabilities from what I would loosely call anytime data. We can get data under poor weather conditions; we can get data during the nighttime; we can get data during the daytime. We have lots more flexibility now than we've had in the past. I think that opens up lots of areas for biological exploration or biological questions that are a lot more interesting than we've addressed in the past. I think one of the dangers, or one of the issues here that I think we need to be careful of, is that the loss of biological insight that we gain by being in the field and seeing what kind of environment the animals are experiencing and what's going on in their environment. There is, I think, there is a danger here in terms of this remote applications and automatic data collection that we should at least be aware of.
In terms of the resource use patterns, I guess one of the major resource use patterns that wasn't discussed here was habitat use. I think our present use of that, I think maybe people feel pretty comfortable about that. Typically we're looking at frequency; we're looking at frequency of use kind of data compared to some kind of availability data. That's typically what we've done for radiotelemetry. If you're interested in that topic there are some other papers later on in the TWS meeting that are going to look at habitat use issues.
Spacial pattern: I think I would like to thank Bob Kenward for dragging us back to the biology of what we're doing with home range. Up front in his talk, I think sometimes we get pretty focused on the methods that we're using and lose insights as to what we're really trying to do with those methods.
One comment I would make about the methods, I think using the sophisticated methods that we have been, these volume methods, the kernel methods and whatever: In my view, we tend to under-utilize the information that's probably in those kind of estimators. Basically, we tend to look primarily at boundaries only and, more recently, we began to look at things like core areas which may at least divide up some of that home range into more intensively used areas, but I think there are other aspects that we need to be thinking about with biological information there that we might glean from that analysis.
Typically, we've also looked at temporal patterns and how resource use might change temporally. Typically, we've done that in very simple scale like seasonal patterns, seasonal patterns in home-range shift, for example, or seasonal patterns in habitat use.
Well I think the future, with now the enormous amount of data that may be collected, brings a lot of challenges in that regard. In fact now, I think we're not only going to have to look at spacial issues but temporal issues will be important as well. When we're out collecting frequent observations of animals, the temporal relationship of those should be considered in our analysis and what we can learn about what's going on with that animal, how the animal may move from one habitat type to another, for example, in what periods of time. Whether those changes occurred during the daytime or the nighttime and how those patterns may change from day to night now are more feasible.
One of the opportunities I think that is coming up are multiple-animal studies. We can look at things now more readily perhaps like predator/prey interactions than we could before. We can look at multiple animals that are associated with each other more easily, perhaps and look at the interactions that those animals have.
Life-stage type of studies: That was mentioned in some of the grizzly bear work. We can now perhaps monitor animals long enough that we can begin to think about gaining biological information and how that biological information changes as the animal's life history changes. I think there are lots of opportunities there for good biological insights.
I think the real challenge is to intergrate all these aspects. To integrate how the animals are using their habitat and their environment. How they are doing that in terms of what behavior is going on and how they are doing that in time as well. I think we're beginning to see a little bit of those, peeks of those kind of models, perhaps in some of the manatee work. There are aspects of that going on in these data-logger systems that Rory talked about in terms of things that he's putting on animals and collecting information about behavior in the environment. So there are lots of opportunities I think coming up for gaining lots of really good insights.
I guess I will just touch on survival a little bit. Again, it wasn't mentioned here too much. Pretty much the standard I think now has become sort of the continuous time or what you might want to call time-to-failure models. Those include things like the Kaplan-Meier, Weibull, the Cox proportional hazards model, some of those kinds of models I think that now have become pretty standard use.
We need to really begin to look and I think people are beginning to look, at how we integrate survival estimates with things like habitat use. We sort of have an underlying paradigm in wildlife ecology that habitat, quality, and quantity of habitat, are the keys to managing populations, but yet it's pretty difficult to show that there is in fact a direct link between survival and those aspects of habitat. I think there are opportunities to do that in the future that integrate radiotelemetry and some of the survival and habitat work that go with that. Also just to maybe comment a little bit about where some of that may be going. There was a poster that talked about automatically determining when an animal might die. I think there is another option for you that may give you some insights as to what time of day or under what circumstances an animal might have died that might give you some information about what the cause was.
I think we need to develop some probably more realistic analytical methods, ones that take into account the lack of independence that we have between brood mates or litter mates in some of the data that we're looking at. Some methods that maybe look at how risk changes seasonally perhaps with animals rather than the continous time or constant kind of models that have been typically used for medical studies. I think the real challenges are how we determine what causes of mortalities are rather than necessarily rates and how will the causes of mortality interact with eachother and what risks those animals are facing out there as they relate to environment and some of the aspects that they're dealing with.
Lastly, I guess I'd like to touch on some things about the data processing and again I think the front end of that is data capture. People have talked about a lot of ways to do that and there are lots of options now and in the future. Another aspect that's been touched on is once you get the data, what do you do about quality control issues? What do you do about errors in the data? What do we do about potential outliners? How do we address those kinds of issues? I think those are important issues in terms of analyzing and processing your data and making decisions about what you're going to do about those.
I think the key to sort of help us do a lot of this is the integration of different components of software. I think you're seeing a lot of that at this meeting. A lot of software that is becoming available now that integrates the data management aspects that we need to manage the data we collect, to help us do some of these quality control aspects, to give us mechanisms to present the data graphically and with maps andGIS so we can look at the data, and try to understand what it means, and try to do some analysis on it. And then also that provides the sophisticated analytical tools that we need to look at things like survival, habitat use, and some of those aspects as well, that are almost impossible for any of us to do without some kind of computer software and hardware.
Well I'll admit that I'm pretty confused, even now, by where some of the computer hardware is, so I'll just offer some even guesses about where the hardware and software may be going or may need to go. I suspect we're going to continue to see increases in speed of hardware, increases in sophistication of software in terms of graphical aspects. I think you're seeing a lot of that represented now in the software that is available. All the Windows software really helps the user interface.
The Internet: We have a great resource now to share software, to share ideas about how to process data. I suspect that's going to become an even larger tool in the future in that regard. And I think this idea about software integration, integrating various components of analysis and data collection, is still a key one. I think it's one we need to work on some more, and I think people are doing a really good job on working on that and making these programs much more user friendly and easier for people to deal with. I just wanted to maybe offer a couple of comments at the end. I think the trends in the analysis and some of the methods tend to get away from us on occasion, but I guess I sort of view this as a positive step. I think we're making good progress towards quantifying and, both qualitatively and quantatively, increasing the amount of science that we're putting into wildlife management and understanding what's going on with animals through the methods that we have.
I want to also stress though the importance of doing more than simply observational studies. We really need to beg in to think about doing experimental work, to begin to formulate hypotheses with the data that we're collecting, to try to test those hypotheses, and integrate ways that we can understand how animals are reacting to their environment both behaviorally, and how they are reacting to management strategies that may be going on in the landscape and changes in the landscape as well.
I think those are some of the big things that we need to be doing in the future and pay more attention to. I guess just one last point as I think, also, we tend to get a little focused sometimes on the fact that we have a thousand observations on one particular animal and what do we do about it? I think we need to also keep our eyes open for the fact that really the questions that we're asking tend to be population-level questions. What's the difference between one population and another? What's the difference among animals within a population? To do that,the individual animal is really the unit that we want to focus on. Yes, we need to estimate what their home range is, or yes, we need to estimate what the habitat use is. But we need to replicate those individual animals and we can't lose track if that is the basic unit that we need to focus on in some of those studies about hypothesis testing and populations. Thanks.
AMLANER: Okay. I have a few questions from the audience regarding some of the data processing issues. That was a fabulous introduction to this; I appreciate those comments, Mike. Just as a bit of food for thought here: This writer indicates that there have been many impacts of radio tags on animals; that goes without saying. Most of the studies reviewed here at the meeting, and including waterfowl (they mentioned that specifically) appeared to have poor statistical power. Now that was a fairly straight statement and I think we need to review it. "Relatively few animals are in test versus control groups, and impacts on behavior are monitored for relatively short times." What do you think about that comment? Are there ways in which we could address those criticisms?
SAMUEL: I think Jay hit on that in his comments and also Gary White did as well. It's a really difficult problem, I think. The problem is that we typically need lots of sample size to do that. It's really the opposite thing that we tend to do when we look at hypothesis testing. Typically we want to reject the null hypothesis. In this case, we want to actually say that the null hypothesis is true, that there is no effect, and it's very difficult to do. I'm not sure there is a good way to do it. I think we're stuck because I think a journal, in particular, is getting a little more rigorous in what they are willing to accept now for validating those kind of studies. I think Gary's idea is an excellent one of accumulating evidence and using that.
Another idea is to look at some of the indirect methods, perhaps related to behavior, or the study that was done on doves, at least at that point, captive, but looking at other indicators of how the techniques and the attachments may be affecting animals as well. I'm sure other people want to offer some comments too.
AMLANER: Other comments from the audience? Do you see careful hypothesis formulation being very critical to answering this issue? You've mentioned in your talk about the hypothesis formulation. How does it play out?
SAMUEL: I think we saw a couple good examples of how people addressed that issue in terms of the quail study that was done, in terms of different attachment methods, and in terms of the mallard study that was done on effect on broods. In those cases, people had sufficient data to look at whether there were differences or not.
That's the first step. The difficulty is planning a study where you in fact want to show that there is no difference and you want to use that method and that requires a lot of planning. It requires you to decide how big of an effect you're willing to live with in order to set sample size. So yes, it is very integrative with hypothesis testing.
AMLANER: Any comments? I bet I know what you are after. It's probably a cup of coffee, very soon, and we guarantee you within about five minutes, we're going to go for that.
Another question from the audience. "The big debate (for home range, habitat use, etc.): Are telemetry observations really independent? At what point do we consider telemetry location recordings as independent observations? Or is it necessary that they be "independent"? Are we kneeling to a statistical god that really doesn't care? Are telemetry positions, by nature dependent?" Any comments?
SAMUEL: And the audience should join in on this one. I hope the audience will join in. Erran Seaman said the other day in his talk that he thought auto-correlation was basically an overblown issue. I agree with him. I think lots of other people do too. Particularly when it comes to home-range analysis and I think the issue that he touched on is representation, particularly with some of the methods that we have now that look at utilization distributions. I think that representation is a much more important issue than auto-correlation.
When we come to things like habitat use, where we've typically done frequency of our proportional use based on frequencies, I think we have some different problems there. I think, Gary disagrees, so good; come on up, Gary. But I think we're going to move to a time when we're going to have to take, we have enough data, and sequential data with habitat use, that we're going to have to more vigorously consider the time frame between locations in habitat use as well.
AMLANER: Gary, go ahead.
GARY WHITE: I feel very strongly about the auto-correlation issue, that it's been way overblown and that Swihart and Slade (1985; Ecology 66: 1176-1184) put the whole profession back ten or fifteen years because of this. It's the auto-correlation of the data that is the home range. That's what we're interested in, and the points obviously are not independent because that's why it is a home range. The issue gets overblown because we worry about making emphasis for that individual animal. As Mike correctly pointed out, it's inferences about that individual, it's among differences of the individuals that we're after. It's the population. The sampling unit is the animal, not the individual data points. So we want to get the best possible home range we can. You don't do that by taking less data. The more data you have, the better the home range estimate is going to be.
Mike, I feel the same way about use data. Again, the sampling unit is the individual animal, so we shouldn't be looking at the individual locations and their use as what we're sampling. It's the estimates of habitat use across individuals. So Nicholas' [I can never say his last name], his compositional type analysis correctly reflects that the individual is the sampling unit. That's where we need to be headed. Therefore, this issue of auto-correlation is in my mind, we've gone the wrong direction. We don't throw away data. That's crazy. How do you get better estimates? You take more data; you don't take less data. Any statistician is going to tell you that this is crazy. You've got to take more, not less. So this concept of auto-correlation and home range has just gone the wrong way for the last ten years.
AMLANER: Good. Yes. State your name.
DEAN CLUFF: Dean Cluff. I did my thesis work on coyotes and I did that as a chapter of my thesis. I looked at auto-correlation and home ranges and I didn't have the actual home range. I thought I would use real data, field data, and compare home range techniques, and I guess I've been fairly delayed in getting it published. Isubmit it, and Gary White knows about it, and we had some changes and I never got around to making those changes. But I would just concur that the basic result was that as long as you have reasonable sample sizes, then I didn't have any problems with auto-correlation.
AMLANER: Other comments? Go ahead.
ROTELLA: Gary you need to come back up, if you would; I have a question. It's just a question, and it's because what you just said ( we run into this a lot with graduate students and in my own work) is: how often can you go get that next observation? Certainly you're sitting in a telemetry vehicle or the antenna tower, and I think a lot of people are going to walk away wondering, "Okay, so a lot is really good. How often can I go get them?" People have different thoughts on that. I'm just kind of curious what you think about that?
GARY WHITE: I didn't stress the point that you've got to have a valid sample of the animal's movements and so suppose you want to sample an elk for the month of January. It's not fair to go out there on January 15th and sit there for 24 hours and monitor movements and claim that's representative of the month of January. If you want a valid sample of those movements, for some time interval, and home ranges have to be put into some time perspective, then you've got to sample, some kind of a valid sample over that time interval. And, typically, we know the best way to get a valid sample is from a random sample, so we go out there and take random points in time. Well, the logistics of that are just infeasible. Still, if the sample is unpresentative of the the animal's movements, you're not going to get a valid home range esimate, no matter what.If we could start on January 1st and sample every ten minutes, the entire month of January, great; we've got that animal's home range defined as well as we could ever expect to do it. Auto-correlation is not an issue because we've got our estimate. If you want to start talking about how to put confidence intervals on that individual home range, and nobody puts confidence intervals on home ranges because the methods are not really there, only by making some weird assumptions can you do it. I shouldn't say that. There are a lot of good ways of doing that; we just tend not to do that in the profession. It's really not that big of a deal because it's the fact that we're looking at the variation across individuals, not within an individual, that we're really interested in. So anyway, the real point is, take a valid sample of the time interval, and that was alluded to here by our other speaker; so that's the key point.
AMLANER: Any other comments? Yes sir.
MARTIN MAIN: My name is Martin Main. The comments that have been made are, in my opinion, terrific, but I think what needs to be done now is a panel, and maybe Gary and whoever else feels strongly about this, needs to get something published in the Journal so that that mass of unknown group of people we call peer review sees that, because this is an area where you get seriously beat up on during peer review. It causes a tremendous amount of time; papers get backlogged, and we've heard a little bit about that. So I would recommend that something short and sweet and to the point gets published in the Journal about this.
AMLANER: It sounds to me like the statisticians need to get together on a white paper to the editors? Certainly we biologists are not going to convince them.
SAMUEL: You haven't taken enough heat on the power issue yet, Gary, that you want to tackle another one?
AMLANER: Fair enough. The sample paper then.
We're at 10:05, it's time for a coffee break. I really do want to thank the audience for their participation, your questions, your written questions, and your comments right now. I also want to thank the three members of the panel that have worked with me on this.
[break]
AMLANER: Continuation of the "Current Issues and Future Directions in Wildlife Telemetry." There are several key issues, key topics that have arisen over the past few years. In particular, we're going to begin with an overview of radio frequency allocation. This is a hot topic right now in biotelemetry and especially since the government has involved itself in a big way in dealing with frequency allocations and some very high-dollar sales. We have a second session on animal care and welfare issues which I know all of you will have significant interest in. We have two key speakers there and we'll have some opportunity for questions. Thirdly, we're going to end with a future directions, where we are going to have several engineers come on up and talk to us a little bit about some of these future directions including a fourth speaker, Robert Sheldon, whowill be representing part of the battery industry and will give us some ideas about lithium chemistry and some future chemistries in battery development.
So with that as an introduction, I would like to introduce you to Noel Newberg, from the U.S. Fish and Wildlife Service, Information Resource Management Division, from Denver, Colorado. He's going to be talking to us a little bit about radio frequency allocation coordination and then we will have some opportunity to have discussion right after that. Thank you.
NOEL NEWBERG: Some of the things I would like to talk about today are some of the issues with radio frequencies for wildlife telemetry. I want to make a point that in the past, a lot of people have been operating their telemetry transmitters without proper frequency authorizations. In the past, they have gotten away with this with not too many problems but there have been some problems that have been arising recently. For instance, in Yellowstone Park a couple years ago, they came in for some frequency requests for the wolf studies up there, and for the reintroduction. We got to looking, we turned up ninety unauthorized telemetry frequencies in Yellowstone Park. And it turns out that one of these frequencies was right on the input frequency of the main repeater in Yellowstone for the landmobile radio operation. Because of this frequency conflict, they had had an interference problem going on for three or four years there in the park and it had been driving them nuts. They had spent a lot of man-hours travelling up and down the mountain looking, trying to figure out what the problem was, where the interference was coming from, and they never really could nail this thing down. After they did this study to see what all kinds of telemetry projects were going on in the park there, they realized that they had a problem that was being generated by an unauthorized radio transmitter. That is the real reason for the need for radio frequency authorization for wildlife telemetry use.
Some things are changing. We have the people in Congress at this point in time that have decided to auction off the radio spectrum as fast as they can in order to raise revenue to bail out the national debt. I wasn't here yesterday but I understand there were comments made about the costs of radio frequencies and that is indeed true. U.S. Fish and Wildlife Service's charges this current year are like $87,000 for frequency support. And each different agency has associated costs depending on the quantity of radio frequency authorizations that they have.
Other things that are changing, we are in the process of narrow-banding the VHF and UHF radio spectrums in the federal government and also it is starting to take place now in the civilian site which is governed by the FCC. Doing so is done to relieve more frequencies to be available for other radio services, to be used for new development, new technologies. One of the things that was mentioned earlier here today was the technology about Internet access, wireless internet access, and that is one of the things that is happening there. They are out trying to grab frequencies as fast as possible for lots of different things, including the PCS cellular and Little Leo satellite systems and so forth to develop some of these new technologies.As this goes forward, this creates other problems and what we're doing in the federal government right now, as we narrow -band, we're going to find that there is less and less radio spectrum available for wildlife telemetry. In fact, I hate to say this, but as wildlife telemetry users, we are a very very small quantity of people as far as demand for the radio spectrum and as a result, there is very little impact as far as getting radio spectrum in the future.
Knowing this, a couple years ago a couple people from the Department of Interior and the Department of Agriculture got together, and we were talking about the need for radio assignments for wildlife telemetry use. As a result, some of you are already aware of this and some of you are already using it, we had developed a new set of frequencies available for wildlife telemetry for the federal government users. In the past, we had thirty VHF radio frequencies available. They are in a 164-167 MHZ range, and these frequencies are going away very rapidly. In fact, right now if you would ask for an assignment on those frequencies, you would probably be denied by the Department of Interior frequency management people because these frequencies are now becoming on-channel land mobile radio assignments. But what has taken their place is that we have now got assignments for 830 wildlife telemetry frequencies in the 162-174 MHz range for federal government use. Doing this a little bit differently than in the past, and one of the problems in the past has been, as you as the users had requested their frequency, it would takebetter than two years to get a piece of paper back from the department saying that you were authorized to use the frequency and I understand the dilemma you were in because you could not get a timely return on a frequency request in order to accomplish a mission you had to do. So most of you were out boot-legging radio transmitters. Well this is kind of changing now. We've turned this around and all the telemetry assignments have been given to the U.S. Fish and Wildlife Service to administer for the rest of the federal government. So we have blanket authority now for these 830 frequencies in the U.S. Fish and Wildlife Serivce.
We have developed a scheme of things in order for other agencies to obtain these frequencies. If you want to put that first slide on there, we have a very simple form to fill out, and if you look at this, this is a little different than what they've done in the past, but we're just asking very simple questions. We basically need to know the nature of the study, the location of it by counties, and the start and stop dates, the contact person for the study, and whether the animal is endangered or not, and the type of animal. Very simple information.
How we're doing this now is that we would like the different agencies .... and some agencies are already aboard: We have been supporting in the last year APHIS, BRD,National Park Service, and Fish and Wildlife Service people, and also some DOD. There will be other agencies coming aboard. I believe the National Marine Fisheries is looking into jumping on the bandwagon with this; we've had some contact with them. Other federal agencies may have need in the future. The requirements are that the frequencies be used for wildlife telemetry, and there is a monetary cost involved because we are paying for these frequencies for administering of this data base where we keep all this information at. So you can get hold of me and I will give you my phone number here and we can tell you from other agencies how to contact us and so forth.
As far as the people from the private sector, or the different universities that are doing research under contract with the federal government, these frequencies are available to you also through contracting to the federal sector for research projects.
This data base that we have on this, we maintain this on Internet; it is through a firewall and it is accessed by the different radio liaisons for the different bureaus which you would belong to. If you have a frequency request, what we would request you do is that you request your radio liaison for your specific bureau that you are part of to submit a request for these frequencies and we are able to turn around support on these things most of the time in 24 hours to 48 hours. So you have very fast turn-around on frequency requests.
We do some coordination on these, and I heard one gentleman earlier talking about their problem with interference in particular areas. If you would put a slide there, I'll show you what we do. We have access to GMS, the master files for all the frequencies for the federal government. This is kind of a pick out of this thing and it shows kind of some information (since this classified, I can't put much up there on the screen), but this shows, on the frequencies, like this right here, there is nothing right there. That is a wildlife telemetry frequency just adjacent to it. It shows adjacent to it, who each of the users are. Like"J" is for Justice, "A" is for Agriculture, "I" is for Interior, "C" is for Congress and so forth. So we can tell each adjacent channel who is available in that area, who is using the radio frequencies. I can actually go into the GMF and pull up and tell you exactly the information of every transmitter on every adjacent channel. So when somebody comes in from an area like San Diego or somewhere where there is military installations, we would try to go down this list and assign you frequencies other than what are being used by adjacent users in that particular area. So we do some coordination to try to protect you.
Now these frequencies, if you realize, are located interstitially in between all the landmobile radio assignments in that frequency range of 162 to 174 MHz. As a result, some of the adjacent channels have high-power transmitters on them; some of them may be as much as 250 watts. So that's another reason we try to stay away from them because they are going to interfere with you. You're not going to interfere with them running ten milliwatts or less transmitter power; the chances of you bothering them are very slim. So we try to coordinate that very closely.
Some of these assignments, for instance this one right here, AGA means "all government agencies" can be assigned that particular frequency and it has a slash on the backside which means that the Treasury is using that for some purpose in that area. To tell exactly what the state or location is, I have to go into the GMS and pull up exactly who is using it in a particular area.
I can do a radio study of a county. If you would come in with a telemetry project in a particular county, I'll go in to that county and I'll do a study and pull a printout and see who all is using that frequency in that county on adjacent channels if there need be. Or if you're going to a National Park area, we'll try to avoid Interior frequencies, or a Forest area, we'll try to avoid Agriculture frequencies. So it gives you an idea kind of how we coordinate that.
Be aware that the thirty frequencies that are out there that have been used for many years are going to disappear very rapidly as this narrow-banding process goes forward. The narrow-banding process, the first round will be completed by January of 2005. It is starting in California, moving to the four corners area and then the rest of the west coast will be caught up and then it will jump to the east coast. Basically what they're doing is they are narrow-banding the areas where the highest use, the VHF and UHF radio spectrums are located at.
As this narrow-banding process goes forward, you're going to find a lot more radio stuff on adjacent channels. Also, you're going to find the landmobile services as going to a digital mode now so it'll be a different kind of interference. All you're going to hear is white noise; basically it sounds like an open squelch on a hand-held radio. It's nothing intelligible on it but it is just a lot of noise there.
They are going to narrow-band the spectrum again come the year 2005, and start going to 6.25 kilohertz channel assignments which is going to make things even tighter. And one of the other things mentioned earlier today, about adjacent channel interference, is going to require that there be closer coordination of these frequency assignments and also that the receiver IF band widths be tightened up to exclude adjacent channel interference, because it's going to get more severe as times goes forward. The number of users per channel will probably also increase.
Something else that is happening at this point in time, in the FY98 budget that's coming out, there is legislation in there that is going to eliminate the low-band radio assignments that we have in the Department of Interior and in the federal government. They're going to be given up for auction. I don't know who they're going to or who's after them, but that means that all the users we have on low-band now and also telemetry is going to have to relocate to another radio spectrum, probably in the near future. So there's a letter coming out asking impact information from us on that in the next few weeks I understand from IRAC, which is the Interagency Radio Advisory Committee, which is under the Department of Commerce. These are the people that do frequency assignments for the federal government users, and they are pretty much at Congress's whim to do what they want and find more radio spectrum.
Right now they're trying to find another 120 megahertz of radio spectrum to auction off, and they're asking the FCC for 100 megahertz of it and the FCC is basically telling them they don't have it. So, I don't know where they're going to find all this radio spectrum, but they're after the federal sector, and we're going to lose a lot more spectrum in the near future, I believe. And I don't know where it's going to stop. Somewhere down the road, we may run out of radio spectrum and it may be really difficult to find a