Engineering Challenges
in the Environmental Biology of Swallows
David W. Winkler
The Problem:
A brood of swallows consists of up to eight birds in the nest who will
leave the nest (fledge) at an age of between 20 and 30 days.
After the chicks are 12 days old, the nest-boxes can no longer be opened
for inspection, as the chicks are at risk of overestimating their flight abilities
and prematurely leaving the box to a dismal fate. We are very interested in the actual dates
that individual chicks leave the nest, as we would like to relate individual
fledging dates to earlier growth rates, etc. of known individuals. To date, we have tried a variety of visual
means for censusing the birds still in the box by looking through the nestbox
hole with a dental mirror and flashlight or a fiberoptic probe, but there
are a number of problems with visual methods: chicks often sit on top of and obscure brood-mates, identifying
marks wear off, or feathers lining the nest obscure identifying marks.
These problems have caused us to search for electronic means to determine
which chicks are left remaining in each nest-box and, if possible, which are
still alive.
Possible Solutions:
We need a device that would generate a 3-4 bit ID (even though we have
hundreds of nest-boxes, we can reuse ID codes on all nests).
The birds are about 15-20 g in mass, and it would be great to develop
a very simple device that could be affixed to the metal band on each bird's
leg. One route to solving this problem would
be to evaluate available RFID technologies to see which tags are sufficiently
small to allow attachment to a bird's band, and then work on development of
that technology in the swallow system. It might also be possible to use a
simpler passive RF technology (possibly harmonic emitters or FBARs), since,
within a single nest, only about six codes are necessary. We need anywhere
from a few cms to a m range. We are hoping that the census of ID devices and
states could be taken from a nest with a very short visit to the outside of
the box. ID is very much more
important than state, as we can estimate the age of death of any chicks left
in the box with their tags when all their brood-mates have gone.
The Problem:
We do a great deal of work on the behavior of parents coming to and
from each swallow brood to feed the chicks, and it would save a tremendous
amount of researcher time and open up possibilities for automatic behavioral
experiments if we could automatically recognize the parents coming to a nest-box.
As in the previous problem, there is no need to be able to recognize large
numbers of individuals, and the challenge is merely to distinguish male and
female at a given box.
Progress toward
one solution: Clearly, any method
developed for the identification of chicks in the nest might work for the
parents as well. But some pilot
work has already been done on another approach. We developed a nest box with Hall effect sensors mounted in
the box wall on either side of the hole entering the nest-box. And we developed an appropriate technology
for affixing rare earth magnets to one leg by epoxying the magnet to a plastic
band on one of the parentŐs legs. This creates a situation in which, with the magnet on the left
leg of one parent and the right leg of the other, the parents can be recognized
by the sequence of left and right signals from the Hall effect sensors on
either side of the nest-hole as the parents enter and exit the nest-box.
Needed research:
The pilot detector requires further developments, with fine-tuning
of the Hall effect sensor sensitivity and placement to increase the accuracy
of parent identification. There
is also a rich field of possibilities to develop automated data capture and
playback of nestling begging calls within individual nest-boxes.
The Problem:
Swallows are subject to a great deal of variation in their energy budgets
over very short time scales: if
the weather gets cool or wet, the swallows' own energy requirements go up
at a time when their aerial insect prey cease flying and become unavailable
for food. The responses of these birds to this short-term
environmental variation are one of the major foci of our work, and we would
gain a great deal being able to monitor the mass variation of individual birds
throughout the breeding season as they respond to the vicissitudes of the
environment. The other aspect
of state that is interesting to monitor is the temperature of the nest and
the surrounding nest-cavity. The
former provides important information on the incubation and brooding behavior
of the female, and the latter provides an integrated measure of the thermal
environment for the parents while they are in the box.
Progress toward Solutions: We have already developed partial solutions to this problem.
We have developed card-board nest-box liners in which the birds build
their nests. This allows us to remove the nest-box floor and place the nest
directly on the pan of an electronic balance beneath the nest, free of friction
with the nest-box walls. We have
also successfully deployed probes for HOBO data loggers to monitor nest temperatures.
To reduce unit cost, we have developed a pilot mass-monitoring device
that uses a strain gauge talking to a HOBO to gather mass data without all
the overhead inherent in a commercial scale.
It would be great to combine the data-logging for both temperature
and mass, and there are some interesting problems in signal processing and
filtering to increase the usefulness of both sources of data. One final possibility is to take the measurement of mass outside
the nest-box by attaching a tunnel to the nest-box and weighing individual
parents as they enter and leave the box.
We have developed the behavioral means to train the birds into a tunnel,
and I believe there could be a great deal to be learned from our local producer
of high-speed weighing equipment, Hi-Speed Checkway.
4.
An Integrated Solution to the Above Three?
One of the limitations of the HOBO data loggers is that
they require repeated visits to the nest to download data and reposition the
temperature probes. We have been working with an INTEL scientist to develop
a mote-based solution to monitoring of nest-temperature, and should have units
out in the field in spring 2005. There
is a rich suite of possibilities for further development of sensors to interact
with the motes, and it is conceivable that sensors could be developed to monitor
mass, RFID presence, etc. simultaneously.
5. Radio Transmitters for Telemetry
Background: One of the biggest challenges of studying
birds is that they can move so far, and it is very difficult for the ornithologist
to keep track of his/her subject.
Problems/Opportunities:
We (mostly the Bioacoustics Research Program at the Lab of Ornithology) have
been working on a new generation of microprocessor-controlled telemetry tags
that allow very flexible transmission scheduling, and thus tremendously improved
tag-life. This opens up all kinds of new questions for the inquisitive biologist.
But the problem is that the majority of birds in the world weigh less than
30 g, and very many weigh less than 20 g. This creates pressure to develop
a tag that is less than 1 g in total weight. Using surface-mount components,
the electronics can now be trimmed down to about 0.7 g, but this creates tremendous
pressure to simplify and miniaturize any sensors on-board and reduce the size
of the battery that can be carried. We are particularly interested now in
the development of light-sensing capability and exploration of alternative
or supplementary energy sources (piezo-electric, thermal gradient, photo-voltaic,
etc.) for these tags.
6. Measurement
of Aerial Insect Abundance
Background:
Swallows feed on flying insects, and insects fly only when environmental conditions
are right. Fine-scale variations in weather thus place limits on the potential
foraging success of swallows.
Problems/Opportunities:
We currently sample flying insects with a 12-m high suction sampler. This
works quite well in Ithaca, where connection to the grid is possible, but
in a sampler in Belize, we have had a great deal of trouble powering a sister
sampler. There are two challenges that we would like to tackle. The first
is to explore the possibility of re-designing the sampler itself to produce
the same catching efficiency with much smaller frictional energy losses. A
more efficient sampler would be very much easier to run, and hopefully to
construct, in remote parts of Latin America. The other challenge would be
to develop a photo-voltaic array to serve the current generation of sampler,
with an eye toward developing a system that delivers the most power per dollar.
Finally, we would be interested in the possibility that insect densities could
be sampled ultrasonically, without the need always to take actual samples
of the insects, but rather that the numbers, sizes and height-distributions
of insects could be derived from the echoes received at a vertically oriented
transducer.
7. Analysis
of flight dynamics
Background: Over the past several years we have developed
a Flight Performance Test Tunnel to obtain standardized measures of the flight
performance of swallows. This
information is part of our push to understand as much as possible about the
biological basis of individual differences in reproductive performance. The
test bird is placed in the tunnel (after a series of measurements on its flight
state are taken), and released into the tunnel. The bird orients toward the light at the
far open end of the tunnel, and it flies down the tunnel and out the open
end after negotiating a pair of obstacles (panels of thin wires placed 1 cm
apart in the birds' escape path). In
its flight the bird is monitored by four synchronized video cameras, two facing
each other from either end of the tunnel, and two taping from above and to
the side of the obstacles. This
design has been developed to produce measures of the reaction times and maneuverability
in the obstacle course and of flat-out acceleration after the obstacle.
Problems and
Opportunities: The largest challenge
of tunnel design is to gather the necessary information without interfering
with the birds' motivation and orientation to get out of the tunnel in as
near a straight line as possible. For
example, we have been using narrow floor-to-ceiling mirrors on the sides of
the post-obstacle run of the tunnel to record when the birds cross certain
points in their flight out. Each of these mirrors is oriented toward the camera
at the open end of the tunnel, and we can thus obtain the times crossing multiple
points in the tunnel from a single camera. We were delighted with this development
until we realized that a sizeable fraction of the birds were departing from
linear flight as they passed these mirrors, and we now think that the reflected
view of the open tunnel end is confusing some birds into exploring for a way
out of the tunnel. An alternative
way of measuring progress down the tunnel (perhaps infrared light screens?)
would be very helpful. Also, the possibility of dynamic obstacles that are
triggered electronically as the bird progresses down-tunnel would be very
interesting. These, together with other problems having to do with the provision
of sufficient light in the tunnel without interfering with the birds focus
on the tunnel's end provide lots of scope for innovative technical solutions.
There is also a rich array of possibilities for any student with an interest
in aerodynamics, as we would welcome someone in the lab with a great deal
more sophistication in the analysis of flight to analyze any part of the birds'
performances around the obstacle or in acceleration out the tunnel.