bbystrek wrote:
>
> Hello,
>
> In regards to frog survey work...
>
> Are there acoustic techniques which would allow one to map locations
> of calling individuals within a given environment? Say with an array
> of microphones elevated to give some geometric perspective above the
> breeding area. It sure would be interesting to graphically be able
> to see the group dynamics over a period of time, even days.
> Individuals and species could probably be easily identified if the
> spectral content could be represented. You would think some of the
> experience gained from marine electronics could yield sort of "frog
> finder". All of the DSP (digital signal processing) integrated
> circuits on the market today sure do make for a lot of possibilities
> for sound processing applications.
>
> Any comments? Does this technology already exist?
There has been some work on this in conjunction with the use of
frogloggers. Usually in conjunction with identifying the species. Many
places where frogs are calling, many species are calling. Along with
insects, birds, alligators, etc. So it's necessary to do a kind of voice
recognition system.
It's making some slow progress, but a reliable off the shelf setup is
probably a way off. Not all frogcalls are that far apart in simple
criteria like spectral content. Most of the ones that are a ways along
are specific for one species, like looking for Ivorybilled Woodpeckers.
And use custom software for the task.
It's a fairly hot topic as frogloggers are available, but can crank out
large volumes of recordings quickly. Which are only useful if listened to.
I think you'd find less movement than you think. From calling
populations I've observed, the distribution is pretty static, even from
year to year. They pick out the good spots and stay with them.
Also frog survey is much more macro. I spent 5 years on it and only
managed a very light coverage. Many, many bodies of water never listened
to at all. And even the ones I did it was often only a few minutes one
night. There are literally hundreds of thousands of potential locations
to check.
The fixed site survey systems like the Frogwatch setup are generally
trying to get a handle on population numbers. I don't participate in
those, not because I think it's not a good idea to get numbers, but
because my experience surveying has shown just how flawed their
methodology is. It hangs too much importance on what's calling during a
specified few minutes. In some cases only repeated a few times a year.
Frogs simply are not little automatons calling at a nice even rate. And
definitely not all species call in the same way.
For instance the protocols often specify getting the count from a couple
minutes or so that occur at a fixed short time after you arrive. One I
know will be missed nearly all the time by this method is the Pine
Barrens Treefrog. It's typical calling pattern is small calling groups,
often only one group at a time. And they call for maybe up to 5 minutes
and then make no calls for up to a half hour or so and then repeat the
pattern. Arrive just after they call make your count and be long gone
before the next call set would be normal. And even if you waited around
and counted them, the calling group for that evening may only be one of
many at the site. Next evening a different group may be calling. BTW
this is a treefrog that may be in Georgia, but has never been officially
documented. It's one of my long term goals to document them in Georgia.
With about a third of the state potential for them it's a big task.
Computerized logging systems could do this population stuff fairly well
for individual sites. But it requires a fair investment in equipment and
manpower to do it over any sizable area. So I expect it will be limited
mostly to just a few high profile species. And it's unlikely to be frogs
much as birds have a much bigger following.
Walt
________________________________________________________________________
________________________________________________________________________
|