http://news.bbc.co.uk/2/hi/technology/2305725.stm
Thursday, 10 October, 2002, 07:50 GMT 08:50 UK
Chips offer insights into birdlife
Remote sensors are being used to find out more about birds in their natural
habitat. Molly Bentley went to Great Duck Island in Maine to see how researchers
are using the technology.
On an island off the US coast of Maine, Alan Mainwaring kneels in the duff near
a black spruce pine tree.
At its base is a hole and a pile excavated dirt that suggests animal activity,
probably recent.
Instinctively, Dr Mainwaring, a computer engineer at the Intel Research
Laboratory in Berkeley, California, pulls back. Then he rolls up his sleeve and
reaches in.
"I'm pretty sure I'm bigger than anything that's in there," he says, one cheek
against moss, and his forearm underground.
After exploring more than 40 burrows, experience tells him he will not fair much
worse than a peck from a disgruntled beak. After all, bird-watching is not
without some risk.
Dr Mainwaring retrieves his arm. The nest is bird-free, but now it has a new
occupant: a wireless sensor that he has embedded in the burrow wall.
Real-time results It is one of more than 40 biological sensors - or motes -
buried on the island, each only slightly larger than the two double-A batteries
that power them.
The latest has already begun to collect data on the conditions inside the burrow
and relay it to a laptop computer.
This is birding without a field guide or binoculars. Ultimately, the scientist
will be absent.
Dr Mainwaring is hoping to track one of Maine's most elusive seabirds, the storm
petrel, with a subterranean online sensor network, and do so from his office
3,000 miles away.
He stands up and tucks the remaining motes in his jacket pocket. Ever-shrinking
computers allow one tiny mote to carry a suite of sensors that collect data on
temperature, barometric pressure, humidity and infrared heat.
The sensors then use low-power radios to communicate with each other.
The resulting network averages readings, flags abrupt changes, such as
temperature rise, and feeds information in near real-time to the web.
While remote sensor networks have been tested in machines, the installation on
Great Duck Island, a 250-acre nature reserve, is the first application of a
sensor network to a natural habitat.
The project is a collaboration between the Berkeley Intel Lab and the College of
the Atlantic in Bar Harbor, Maine, that oversees the nature reserve along with
the Nature Conservancy.
Exciting research
College of the Atlantic biologist and president Steve Katona says the sensors
already show promise for environmental research.
"These little gadgets are going to see a real change in how we look at the
world," said Dr Katona, from his office on the mainland.
Dr Mainwaring likes to compare the data received from sensor networks to
satellite photos, the more familiar form of remote sensing, which create
topographical or weather maps.
While a satellite gave an overall picture of a forest, he said, a sensor network
provided information on every tree or, in this case, every bird nest.
It is an exciting prospect for a biologist. "It will enable us to study
ecosystems at a level that has not been conceived," said Dr Katona.
Motion sensors
While Mr Mainwaring puts bird nests online, on the opposite coast, in
California, Intel engineer Robert Szewczyk scans the computer data streaming
from Great Duck Island.
His finger isolates the infrared reading, which suggests how much time the birds
spend in their burrows.
Data from mote 29, for example, tell him that the nest is empty.
"During the day the occupancy sensor reads more or less nothing," says Mr
Szewczyk. "But in the evening, there's a distinct signal from the motion
sensor."
It is a good sign. The first test of the sensor network, he says, is to
determine whether the underground motes can actually collect and transmit data
over the internet.
The mote's infrared sensor has detected the bird's body heat when it entered the
nest.
The motes have a few advantages over human data loggers. Sensor networks collect
the sort of raw data that biologists collect in the field, except that a mote
can go where a human cannot, like down a six-inch-wide petrel burrow.
And the sensors take readings on many variables, continuously. Even the most
diligent graduate students cannot be everywhere at every moment in a research
field.
Nor would you want them to be. The key to the technology is that once it is
installed, grad-students and scientists like Dr Mainwaring can leave the field
and the birds in peace.
Future plans
A remote sensor network has the potential to virtually eliminate human
disturbance in fragile ecosystems, such as a nesting ground, where even a quick
stroll can leave animals stressed.
"Just because an animal settles down and preens its feathers after a few moments
with us doesn't necessarily mean that animal has really recovered," says Dr
Katona.
"The effects of our presence can be subtle and can last for some time."
But the technology does not limit human participation in petrel research. The
data is available to anyone with internet access.
Dr Mainwaring predicts that the network will extend beyond bird nests.
Berkeley's Intel Lab plans to test a sensor network in the James Reserve in
southern California, with the University of California, Los Angeles.
There are also plans to install motes in an Oregon vineyard. Sensors on
grapevines could monitor the conditions of single plants, said Dr Mainwaring,
and lead to greater precision in agriculture.
But it does not stop there. He envisions the day when billions of things become
networked.
"This has the potential to be as revolutionary, if not more so, than what
happened when a modest number of computers formed the primordial internet," said
Mr Mainwaring.
Birding-Aus is on the Web at
www.shc.melb.catholic.edu.au/home/birding/index.html
To unsubscribe from this mailing list, send the message
"unsubscribe birding-aus" (no quotes, no Subject line)
to
|
