Tip:
Highlight text to annotate it
X
Of course, my topic today is on the modeling of spatial and temporal variability
of the winter soundscape in the Kenai National Wildlife Refuge
uh... because as if you've seen in that anchorage
paper that uh...
my primary
focus is on snow machine
time periods
the sounds cape is uh... the biological, geophysical
and anthropogenic sounds that emanate throughout
the landscape in the emerging field of sound scape ecology
uh... we give these terms to these components as by biophony
geophony and anthrophony
there are four main components of the
of the sound skape including acoustic composition or the frequency
of sound measured in hertz uh...
the amplitude is sound
measured in decibels
and then there's the temporal variation
spatial variability
and the acoustic interactions between
by biophony, geophony, and anthrophony
sound serves as an important
component of an ecosystem
uh... it's a means of communication among individuals and between species
they can also help us identify
where species are distributed
and the occupancy of those areas
although there is also a huge human component
uh... for those of us who enjoy the outdoors
and the natural sounds that the
uh...they emit
there's also um...
however the anthropogenic sounds
can significantly change uh...
these sounds the natural sound scape
causing animals to uh...
relocate a new areas change their vocalization patterns
and uh... or not vocalize at all
and this can also affect the human experience
the Kenai national wildlife refuge is a two million acre refuge located on the
Kenai Peninsula
it manages one point three million acres of congressionally designated wilderness
uh... but it manages these on conflicting mandates
the wilderness act of nineteen sixty four requires that the refuge manage
wilderness for it's wilderness character
including outstanding opportunities of solitude
uh... which means
enjoying the outdoors in the absence of mechanized transports
and then there is ANILCA which uh...
requires that the refuge provide
motorized access into wilderness areas
uh... such as snow machines
so
the uh...
this project um... was initiated on uh...
based on these conflicting mandates
for the need to study the sound scape as a means of
as a means of
monitoring
human disturbance
the wreckage is far from pristine regardless of its uh... one point three
million acres of wilderness uh... it's located
along the Sterling highway which cuts right through the refuge
which allows one point four million vehicles to pass through the refuge
every year
result it's also a major flyway for commercial airlines passing,
coming to and from anchorage
and uh... a great deal of noncommercial aircraft
that uh...
fly over the refuge on a daily basis
surprisingly there's also oil and gas compressor plants
located in the northwestern portion of the refuge
these emit constant sounds throughout the day
and throughout the year
and more ephemeral
more ephemeral
sounds are uh... boating activities in the summertime
the snow machining activities during winter
the potential of these um... anthropogenic sounds to affect the
wilderness uh...
led me to these objectives
one was to determine the temporal variation of biophony and anthrophony
two was to create a spatially explicit model
identifying areas of anthrophony throughout the refuge
to sample a soundscape I split the refuge into six uh...
regions
uh...
there. I set out six sound stations
uh... at permanent locations to record the temporal variation
and uh... I used sound recorders such as the larson davis eight thirty one
and the uh... wildlife acoustics SM2. Both are
uh... high-quality sound recording devices that can capture
the sounds of the environment
and uh... to capture the spatial distribution
I set out six additional sound stations
uh... at temporary sites and moved them two new locations
every two to three weeks
uh... all in all I sampled a total of twenty two sound station locations throughout
the refuge and you can see the distribution
of those sound stations throughout the winter of two thousand ten two thousand
eleven
i recorded ambient sounds uh... for twenty seconds every fifteen minutes
i collected over sixty three thousand sound files
totalling over thirty five
thirty-five gigabytes
of information
and entered them into the remote
environmental assessment laboratories sound library
which is an automated system for calculating biophony and anthrophony
sound files are converted into spectrograms
and separated and one killohertz frequency bandwidths, or bins
uh... it uses the welch's nineteen sixty algorithm for calculating power spectral
density
which is simply the uh...
the degree of saturation
of sound within a bin
or the amount of energy emitted by a sound
so here's how spectrogram is kinda broken down
into uh...
frequency bandwidths
you have uh...
frequencies and kilohertz on the y axis
and time on the x-axis
and then you have um...
a calculated
power spectral density
for each
each bin
and this gives us
an idea of
where
where most of the sound is occurring in this spectrogram
so in this case you see
better to show it over here
that uh... the four
the three to four
three to five kilohertz bandwidth have the most
sound emission
so in order to calculate biophony and anthrophony
based on preliminary data we find that
uh... most biological sounds at kilohertz
above two kilohertz
and so biophony is calculated as power spectra the total power spectral
density
of all bandwidths before
above two kilohertz
anthrophony occurs more at uh... lower frequencies
and uh...
so is the uh...
the one to two kilohertz
uh... bandwidths
um... of the total power spectral density
for uh... anthrophony
these are example these are two spectrogram examples
of uh...
of a gray jay calling at
two kilohertz above and an airplane
flying over at uh...
one to two kilohertz
I analyzed the temporal variation
of uh... biophony and anthrophony roughly from my six permanent sound stations six per minute sound stations
and i'll give you an example of the two of these sites uh...
for time's sake
and then I modeled the spatial distribution of anthrophony from all
twenty-two sound station locations
so here's uh...
here's uh... some temporal results
this is mile one twenty six
uh... sound station
uh... which is located in designated wilderness
and is open to snow machines
the area has relatively if you pay attention to the up
to the y-axis here
the uh...
area has
relatively higher amounts
of uh...
biophony compared to anthrophony
uh... and but biophony increases
during the time periods of march and april
whereas uh...
anthrophony drops down between that time period
and this coincides with the snow machining season which is from december
to march
Paddle Lake is uh... another station
that uh... is within designated wilderness but is closed to snow machines
and interestingly enough you see that biophony gradually increases
over the month of january
to march
coinciding with uh...
with the coming of spring and the melt of
snow
uh... and but uh... anthrophony has a pretty interesting
line here where it
it's lower in january and march
and peaks in
february
uh... and i found this due to uh... the increase of air traffic
so this is the first spatial model
uh... developed for anthrophony on the Kenai national wildlife refuge
and uh... by my knowledge it is the only
model of this uh... of this type
uh... to date
we see that uh... there are higher anthrophony values localized
in the northern
northwestern part of the refuge
uh... in these are hot spots of Soldotna
in this area
Kenai
uh... the oil and gas field and swanson river area
and then the Sterling highway
and mystery creek road which is an area open the snow machines
when compared to wilderness
we find in the northern most wilderness units
are uh...
bordered and occupied by
uh... high anthrophony values
um the central and southern wilderness uh... units are general have generally
lower anthrophony
to determine what the sound sources were contributing to the soundscape
i found listens over ten thousand sound files
with the help of some field techs
it was long and tedious process but
we got through it
but we found that um...
auto mobiles uh... contribute uh... a large amount to the anthrophony values
of the soundscape
uh... an addition to that we found the aircraft
also contribute
uh... went to a lesser extent uh... snow machining
and uh...
in some cases we had people talking but that's now mechanized transport
and then the biophony values uh...
ranged from mostly uh...
mostly um...
bird species
so some of my preliminary conclusions um... considering that uh... this is an
ongoing study
and this is a very basically results for my first field season
of last winter
uh... we find that biophony
value still dominate
a majority of the refuge
during the course of winter uh... although it is in the presence of
anthrophony
and uh... there could be
implications there that uh...
that bird species may perhaps
being influenced by
anthrophony
uh... there's also anthrophony is localized
to certain areas of the refuge
like i said
and uh... uh... this is intruding into uh... wilderness areas as well
inaccessible areas of the refuge have lower amounts of anthrophony
on which is uh... of course southern region of the refuge
there's no roads out there. there's a few trails that you can get to you but
in order to access it it's it's all fly in
the northern part of the refuge is uh... uh...
more accessible and uh... of course has more anthrophony
associated with it
vehicle traffic and air traffic uh...
in these areas uh...
affect the wildlife in the area as well as um... the outdoor experience of
visitors
and in the accessible areas of the refuge
so the next steps uh...
for uh...
for uh... this coming winter
uh... is to increase my spatial sampling
uh...
that's two million, the refuges is two million acres so
it's a large area too
to cover
but i instead of having twelve sounds
sound recorders, I now have twenty seven
so my uh...
my sampling design should be uh... a little more robust
uh... I'll set out twelve permanent stations to get a better temporal variation
uh... throughout the refuge
and uh... set out fifteen additional
uh... sound stations to new and rotate them to new locations
every ten days
instead of two weeks
because i found that two weeks was um...
was a little too long
and uh...
and the values of anthrophony and biophony kinda tapered off after ten days
uh... I also want to identify quiet areas after listening to over ten thousand
sound files, you're listening to a lot of white noise
and um...
so there's a lot of basically what we call silence out there
and uh...
this is important for
uh... the human experience
out in these remote locations
uh... i'll do this by uh... parsing out the samples uh... with the absence of
sound
from the uh...
their powers based on their power spectral density
uh... i'm also going to create
uh... models of anthrophony as well as biophony
based on the more robust design that i have
and then uh... built a time series models to show the temporal variation
spatially
averaging over december january february and march
uh... for my
for my sampling period
i'd like to thank john morton of the Kenai national wildlife refuge Falk
Huettmann uh... university of alaska fairbanks is a modeling specialist
uh... and stuart gauge of michigan state university who is
a soundscape ecologist and one of the most
innovative
uh... individuals in this field
uh... who helped develop who actually developed the uh...
the real sound library
i'd also like to thank ryan park you probably listened about eight
thousand sound files
and worked with me on the on the refuge, snow machining all over the place
uh... i'd also like to thank the u_s_ fish and wildlife service for their
funding and uh...
and the additional funding they're going to give me this year