Of chips and chicks Cutting-edge computer technology used in bird study

Envision a remote, uninhabited island off the coast of Maine, and stereotypical images of seabirds, rocky shores and pounding surf probably come to mind.

Most likely not part of the picture is cutting-edge computer equipment being used in a high-tech, scientific experiment.

The natural setting of Great Duck Island has helped attract a project that, according to scientists, could play a role in determining the extent to which computer technology will permeate our planet, both in its urban settings and in the remote expanses of unpopulated terrain.

The island, located seven miles off Mount Desert Island, is an unsheltered chunk of land roughly 1 mile long and a half-mile wide, the vast majority of which is a nature preserve owned by The Nature Conservancy. Because it is mostly uninhabited by humans and other predators, the island is host to seasonal colonies of gulls, guillemots and Leach’s storm petrels. The petrels live and raise their young in underground burrows.

Petrels – small, gray relatives of the albatross that feed on plankton – spend much of their lives far out at sea. When they are on land they tend to be underground during the

day and fly around late at night making scientific observation difficult, according to professor John Anderson of College of the Atlantic in Bar Harbor.

Battery-powered remote computer sensors being developed by researchers at Intel and the University of California at Berkeley, however, are enabling COA students to study the petrels while intruding into the birds’ lives as little as possible.

“How do you effectively monitor wild animals without disturbing them too much?” Anderson asked recently while standing amid a clutch of burrows obscured from view by low-lying bunchberry plants.

A few yards from the petrel colony at the island’s southern tip stands a former lighthouse keeper’s house where researchers from the college live during the summer.

“We don’t really know what they’re like when they’re not being bothered,” he said of the petrels.

Each summer since 2002, computer scientists in California have supplied COA with wireless sensors, or motes, that are placed individually in petrel burrows to collect data on the birds’ behavior and their subterranean environment, according to Anderson. Petrel researchers had 25 motes two years ago and this summer expect to have more than 200 at their disposal.

The devices enable researchers to determine humidity and temperature in each burrow without having to stick a camera or a human hand intrusively down into the burrow’s narrow opening, he said. This information helps researchers determine when birds are in the burrow, and how many birds are in the colony.

The motes have changed over the past two years as scientists have adapted their designs to be more resistant to the elements and to make their battery power last longer.

The collected data are relayed over short distances along a network of motes back to a central computer, which transfers the information live to the Internet by a satellite dish.

Anderson said that not only COA researchers, who have been studying birds on the island since 1998, but anyone with Internet access could examine the data as the information is transmitted to the Web. The extra eyes can help draw conclusions from the hundreds of thousands of readings the motes collect each summer, he said.

“We’re getting a much more detailed picture of what’s going on in the [petrels’] microenvironment,” the ornithologist said.

Alan Mainwaring, a research scientist with computer company Intel in Berkeley, Calif., said recently the idea of using the motes in the petrel study came up in 2001 when he and Anderson were talking about their respective projects. The two scientists had become friends in the late 1980s when Mainwaring worked at COA during his summers off from college in Rochester, N.Y.

Mainwaring, a native of Needham, Mass., said the wireless technology being used to study the birds can be applied to other situations where multiple, remote places need to be monitored closely. Similar sensors are being used to study how structures react when exposed to earthquakelike tremors and to monitor the temperature of grapes growing in California vineyards, he said.

Other possible applications include monitoring motors and equipment in factories to determine when a particular part might be wearing out and require replacement.

“You can do a much more specific job of assessing your machinery,” Mainwaring said. “That kind of early warning can be helpful and save a lot of money.”

Scientists predict that motes someday can be powered by vibration energy instead of by batteries, which limit how small the motes can be and eventually run out of power, according to Mainwaring. Without having to rely on battery power, the motes can be small enough to be measured in millimeters rather than inches, he said.

At that scale, they could be mixed into construction materials such as paint or insulation and used to monitor structural stress on bridges or to alert homeowners about water leaking into their homes.

“Wouldn’t it be nice if your foundation told you about that before it’s too late?” Mainwaring said. “The potential for this kind of technology is huge.”

As far as the role of the technology in ecological studies, Anderson said use of such motes will become more widespread, especially for research projects in hard-to-get locales such as caves and tree canopies. He said he has stressed to the computer scientists that the motes, which have cost several thousand dollars to make, must be durable and cheaply maintained for them to be practical for biological research. Biologists, he said, tend to work with smaller budgets than computer scientists. Already, about $500,000 has been spent developing motes for the project.

Mainwaring said that the use of the motes might be expanded on Great Duck Island as the technology continues to evolve. A small vibration-powered version of the mote could be attached to a petrel’s feather and could keep track of where the bird goes when it flies away from its burrow, he said.

When the bird returns, the information could be uploaded to the larger mote in the burrow and then transferred to the researchers’ database.

“That would keep us busy for at least a year and probably for several years,” he said.