1998/1999

Wired for Weather

You're stationed atop Arctic ice.
Your data are circling the globe.

Photo © Marty Mulhern

One of two Canadian ships that trekked to the Arctic for the SHEBA field project, the Louis S. St. Laurent shimmers in this nighttime view from base camp.

Evidence of the bear's work was seen by Maclean and colleague Steve Semmer at NCAR's Boulder headquarters. Each day Maclean and Semmer inspected the quality of the data fed to Boulder from the Arctic. The portable automated mesonet (PAM) stations they monitored are called Flux-PAMs because they sense the energy exchange, or flux, between ground and atmosphere. They also collect the standard weather data (such as temperature, wind speed, and humidity) measured by other PAMs. The complete data set from the Arctic was transmitted in real time to the project's base camp, then relayed once a day to Boulder via satellite and the Internet.

From Maclean's vantage point in Boulder, "We could see in the data that a few sensors had gone out, and it wasn't a typical failure mode." Semmer sent an e-mail to technicians aboard Des Grosseilliers, a Canadian Coast Guard ship locked in sea ice for the winter. Des Grosseilliers housed participants in NSF's Surface Heat Budget of the Arctic Ocean experiment (SHEBA). Several dozen Americans and Canadians lived on Des Grosseilliers at any one time from late 1997 to late 1998, working three- to six-week rotating shifts. SHEBA's ambitious goal was to document a year's worth of energy exchange between the Arctic atmosphere and sea ice in unprecedented detail, with the help of instruments aboard Des Grosseilliers and out on the ice. (The SHEBA project office is at the University of Washington's Polar Science Center in Seattle.)

Charlie Martin (left) and Gordon Maclean inspect Arctic data while at NCAR's Boulder headquarters.

It is one of the ironies of post-Internet life that Semmer noticed the polar-bear damage before anyone at the SHEBA site did. Data from field projects like SHEBA can now reach users across the world with lightning speed--often on the day they're collected. Yet, as with actual lightning, the path is seldom a straight line. The satellite that linked SHEBA to the world sat over the western Pacific; it relayed signals from the Arctic to a company in Perth, Australia. The signals then traveled via fiber-optic cable to St. John's, Newfoundland, where the Internet provider for the Canadian Coast Guard is based. Finally, the data were funneled to Boulder, the SHEBA project office in Seattle, and other locales. "On the Internet, Boulder could be between Melbourne and Cape Grim [Australia]," notes Charlie Martin, a systems manager for NCAR's Atmospheric Technology Division (ATD).

Whether direct or not, such connections are helping NCAR to do more with less. For over a decade, NCAR's PAM stations have been placed in some of the most remote places ever observed for atmospheric science. Powered by the sun (and a backup generator for cloudy days or the Arctic winter), these sentries gather data and store them on a disk for retrieval by a technician. Since the mid-1980s, PAM station readings have also been relayed by satellite every five minutes when satellite coverage permits.

A recent step forward has been to fix software remotely. Maclean gave this a try in early 1997 while riding 30-foot seas aboard a U.S. research vessel involved in studying winter storms that traverse the Atlantic. A French research ship was nearby, and both ships had ATD's Integrated Sounding Systems on board. These profile the atmosphere vertically--much like a weather balloon, but using upward-pointing radar beams as well as a PAM station. Maclean came across a software bug in the system and fixed it on his ship; however, the changes needed to get onto the other ship's system as well. When the ships came within two miles of each other, Maclean made contact via hand-held radio with NCAR colleague Larry Murphy. "We sent the software fixes via satellite and the Internet to Boulder, and then the French ship dialed in to get them. Meanwhile, I was on the radio with them, helping them patch in the software." The upshot: a job that would have taken days without the satellite link took only a couple of hours.

Can visualization technology help scientists collaborate across great distances? A new system created by the University of Wisconsin, with assistance from UOP's Unidata program, uses the Internet to set the stage for such interactions. Under development by UW scientist Bill Hibbard, a new software system called Java VisAD establishes the foundation for manipulating and viewing many types of numerical data. Using three-dimensional extensions to Sun Microsystems, Inc.'s popular Java computing environment, VisAD produces sophisticated 3-D images that correspond precisely to data that otherwise would be difficult or impossible to display. VisAD is designed for remote data access and image manipulation (as shown here) by people at multiple sites, all of whom see the results at once. With VisAD, "People all over the Internet can pop up the same interface and work together as if they were on the same workstation," says Hibbard. Unidata--a provider of weather data and related software to universities since 1985--will incorporate VisAD in future offerings, and Unidata's Steve Emmerson is collaborating with Hibbard on VisAD development. Details on VisAD can be found on the World Wide Web. Visualizations courtesy Bill Hibbard, University of Wisconsin

As the use of the Internet in the field takes off, UOP's Joint Office for Science Support finds its services in demand. JOSS specializes in planning and executing field experiments, either from scratch or as part of a multi-institutional team. A keystone of their services, and an essential tool for researchers, is the CODIAC online archival and retrieval system available from the World Wide Web that allows scientists to access field-program data on a daily basis from around the world.

UOP is now looking at what the newest globe-spanning satellite communication networks (such as Iridium, operated by Motorola, Inc.) have to offer. "We don't yet know what kind of capacity these systems will have [for scientific data]," says Jim Moore, field operations manager for JOSS. Moore says the obstacle to a more sophisticated communications link between field programs and laboratories has always been "the bandwidth issue." One adaptation for the field catalog has been to assemble preliminary data sets from a sampling of each day's observations, then publish a more complete data set on the Web after a program has wrapped up.

As the global marketplace burgeons, the constraints on scientific field communications are as much financial as technical. Moore visited the Maldives in early 1998 to help JOSS prepare for the Indian Ocean Experiment, a major field program scheduled for later that year. He found that the islands had "all the phone and computer-modem connections you could ever want--but you have to pay a premium for the connection." However, with the next generation of satellite links, Moore says, "You'll be able to make a phone call from anywhere to anywhere, and the supply of these global communications capabilities should force the costs down."

Communication to and from the field is making automation increasingly feasible. One of NCAR's Integrated Sounding Systems helped lead the way in the summer of 1998. It was deployed in a forest in northern Wisconsin for a study of carbon dioxide exchange in the atmosphere's lowest kilometer coordinated by Kenneth Davis (University of Minnesota). After being fitted with new software for solo work, the system carried out its duties quite well on its own, sending full observations twice daily, says Martin. "It was virtually unstaffed. It was set up without an engineer on hand, and the researchers could be wherever they wanted. The data were taken in Wisconsin, Ken was in Minnesota, and we were in Boulder."

Other issues of UCAR Highlights