Divers work on the sensor in a test tank at MBARI. Credit: Todd Walsh / MBARI

Scientists have developed an underwater biological laboratory that can identify DNA fragments in minutes, saving researchers the trouble of going into the field to retrieve samples.

The 'ecogenomic sensor', which is packed into a roughly 1-metre-long canister, can test for proteins released by microorganisms and even run DNA tests match DNA to determine which species are present.

"The instrument is, in essence, a microbiology laboratory in a can," says James Birch, a bioengineer who is helping to develop it for the Monterey Bay Aquarium Research Institute (MBARI) in Moss Landing, California.

Traditionally, he says, the most expensive and time-consuming part of such ocean testing is getting samples into the lab. "They have to send somebody out on a ship or a pier, then schlep the water back," he said on 23 February at the American Geophysical Union's 2010 Ocean Sciences Meeting in Portland, Oregon.

But not with the automated lab. Mounted on a fibre-optic cable, the lab can instantly relay its findings to shore. "Instead of going out and sampling once per week, which public health officials do now, this can give you results in an hour and a half," Birch said.

One potential use for such devices is to warn seaside communities of impending algal blooms, which can shut down beaches and do millions of dollars of damage to tourist economies.

The device can also be used for deep-ocean research. Already, Birch said, scientists testing it have found that deep-sea microbial communities are unexpectedly dynamic. "There are lots of changes in the numbers and kinds of microbes, depending on weather and tides 1,000 metres above the instruments," he said.

Swimming free

Another advance that has recently been tested and was reported at the meeting is a wireless underwater modem with a range of 200 metres that uses optical signals to transmit data at megabit speeds. Rather than having to use an expensive submersible to physically plug a cable into a remote instrument, for example, researchers could simply lower a wireless modem, establish communication, then reel the modem back up when they're done.

The modem is also fast enough to transmit video footage, says Norman Farr, an electrical engineer at Woods Hole Oceanographic Institution in Massachusetts, whose team recently used the device to send such images. Not only does that allow the transmission of video data, Farr says, it also allows humans to pilot underwater rovers remotely.

Rovers could be docked at an interesting seabed site — perhaps a hydrothermal vent field. Optical modems would then be installed nearby, on towers with good lines of sight. These would be connected to shore by fibre-optic cables, allowing scientists to operate the vehicles from the comfort of their labs.

"A user would wake up one of the vehicles, drive it to respond to some sort of event, perform an experiment or look at some critters — whatever is desired — all the while downloading data," Farr said.

Once the job was done, the researcher would return the rover to its 'parking' spot and put it back in standby mode, where it would wait for its next use.

To range farther afield, however, ocean scientists are going to need robotic explorers called autonomous underwater vehicles (AUVs). Such vehicles, which can roam for up to 18 hours without a pilot, for are already in use. Since 2006, MBARI software engineer Thom Maughan has been working to improve AUV autopilots, using artificial-intelligence software borrowed from NASA's Mars rovers, which need to be able to go about their business without being directly controlled from Earth.

The Mars rovers, of course, operate mostly by visual input. Not so AUVs, which need to be able to operate in the dark, 1,000 meters below the water's surface. "Ours is more like taste, smell and feel," Maughan said, "but we're using the same underlying software design."

Rather than sending the AUVs out on preset tasks, the new software is designed for maximum flexibility. "It can change its mission on the fly, prioritize and collect the data the scientist needs," Maughan said.

So if a scientist wanted to study the microorganisms living on each side of a temperature gradient, the AUV would find the boundary, follow it, and pick the best spot to take samples. "Were putting in priorities and letting this robotic system decide what is the best fit to give the scientist optimum use of the vehicle," said Maughan.