UCAR > Communications > Quarterly > Fall 1999 Search


Fall 1999

Less is more: Low-cost GPS focuses on hazardous locales

by Zhenya Gallon



Sometimes, lower cost is the key to higher resolution. Researchers have long understood the potential of the Global Positioning System for studying movements around volcanoes and fault zones. But the possibility of losing a $20,000-30,000 research-quality GPS installation to a lava flow or fault shift has constrained the number of units investigators were willing to place in harm's way.

"People were putting these expensive instruments, for example, on the side of Montserrat Volcano in the Caribbean, and unavoidably losing them. That inspired us to look for other ways to do it," says Charles Meertens of UCAR's GPS Science and Technology (GST) program. Meertens and his colleagues on the L1 development team (see sidebar) postulated that inexpensive, single-frequency GPS systems like those sold commercially to track vehicles could provide good-enough data to detect relative motion while drastically cutting costs.

The group has developed a low-cost system suitable for obtaining very high-resolution measurements over an area of a few hundred meters to ten kilometers. The system can be installed in areas where there is risk of loss or where a dense array would be too costly with traditional GPS. The L1 system is economical: Meertens estimates the cost of installing one at $3,500 to $4,000. Transport is economical, too: one system can be carried in two or three backpacks. An eight-system L1 network in Long Valley Caldera in California--including a base computer and monitor, solar panels, antennas, monuments, tools, and batteries--fits in the back of one sport utility vehicle.

Expensive GPS units track and compare two frequencies (L1 and L2) to make corrections for the delays radio signals encounter in the ionosphere. The L1 system uses single-frequency receivers that can't make that adjustment, but "by keeping the distance short between units, you don't acquire as much error," Meertens explains. Tests are under way to see if interspersing a few dual-frequency units in an L1 network would be sufficient to calculate and remove the ionospheric effects that accumulate over large distances.

Michael Lisowski (USGS) and Charles Meertens check an L1 installation on Kilauea Volcano in Hawaii. (Photo by Jacob Sklar, GST.)

Another expense in dual-frequency units is onboard data processing and storage. So the L1 team eliminated memory and designed a communications system that relays raw data continuously to a computer based a safe distance off site. Working with Boulder-based FreeWave Technologies to adapt a radio modem used first for military tanks and then for golf carts, the team customized the system to meet research requirements. The radio modem in each station both collects the GPS data and repeats them from one unit through the next to the master collection point, using time delay multiple access (TDMA) technology. This multipath relay system assures that if one repeater goes down, the data can travel along another path to the base computer. Data are sent from the base computer to GST's Data Management and Archiving Group for processing. "We've had to build special data translators for the L1s," says Meertens. The team is developing software now that eventually will allow the base station to automatically process the data on site at hourly or shorter intervals.

Eliminating onboard data storage also simplifies the operation of the L1 systems, which include a GPS receiver, antenna, radio data modem, solar panel, and small, low-cost battery. All the electronics are in a compact package above ground for rapid installation, which is vital on the sides of not-so-sleepy volcanoes. Popocatépetl Volcano in Mexico, for example, is known for its lava bombs, rocks thrown out of the caldera that may take up to 15 minutes to land somewhere below its 6,000-meter (18,000-foot) top. A monitoring team watches seismometers and keeps in touch with the installation team via radio, "so if anything looks like it's going to be threatening, you literally run off the mountain," says Meertens. While the electronics are expendable, the anchoring monuments are steel pins drilled into rock or set in concrete so that if a unit is knocked out, the surveyed location can be reoccupied.

Installations around the world

NASA has provided support for L1 development. A field test in Colorado in the summer of 1998 convinced the team to install the ongoing demonstration project in Long Valley Caldera that fall. The U.S. Geological Survey (USGS) has been monitoring the caldera south of Mono Lake since 1982, and several additional instruments are available there for comparison with L1 data. NASA, NSF, and other agencies are supporting experimental deployment at a variety of locations (see sidebar, p. X). L1s are being used not only to study motions of the land and sea, but for atmospheric tomography as well (see sidebar, p. Y).

In September, GST is sponsoring a volcano geodesy workshop in the Tetons for researchers employing deformation techniques such as GPS to study volcanoes.

"We have to emphasize that it's a development project," Meertens notes. Nevertheless, "for a geophysicist who studies active processes, these volcanoes are pretty much the ultimate experience. And working with all the collaborators has been great. We hope to prove that the technology will work, but you have to wait for something to happen. [The volcanoes are] all throwing off steam and plumes of some sort or another, so we have good odds of seeing something somewhere."

L1 projects: Who, what, where

The L1 development team consists of Charles Meertens, John Braun, David Mencin, Curt Conquest, Barbara Perin, and Christian Rocken, GST; Daniel Dzurisin, USGS; and Timothy Dixon, University of Miami. Progress reports are posted on the Web.

GST is supported by NSF and NASA. A grant from NASA's Solid Earth and Natural Hazards Program has funded L1 development and deployment. Additional support from participating organizations is listed below.

Volcano monitoring

Long Valley Caldera, California: Daniel Dzurisin and Elliot Endo, USGS; Charles Meertens, GST.

Popocatépetl Volcano, Mexico City: Tim Dixon, University of Miami; Enrique Cabral, Universidad Nacional Autonóma de México; Charles Meertens, GST. Additional support from NSF.

Mauna Loa/Kilauea Volcano, Hawaii: Charles Meertens, GST; Michael Lisowski, USGS; Paul Segall, Stanford University. Segall had been using dual-frequency stations in Hawaii for some time as part of his effort to model volcanic activity. He identified areas on Mauna Loa/Kilauea Volcano where a denser array might prove useful for studying dike intrusions and rifting events, and several L1s were installed there early this summer.

Taal Volcano, Philippines: Michael Hamburger, Indiana University; Charles Meertens, GST; Philippine Institute of Volcanology and Seismology. Additional funding from NSF. Installation was completed this August.

Fault monitoring

Hayward Fault, Berkeley, California: Mark Murray and Roland Burgmann, University of California, Berkeley. L1s are filling in a network between dual-frequency stations that are spaced 15 to 20 km apart. The higher resolution should be useful for validating changes in topography observed on satellite-based interferometric synthetic aperture radar (INSAR). If all goes well, the L1s will become a regular part of the fault-monitoring network there.

Tide gauge monitoring

U.S. Pacific Northwest: Megan Miller, Central Washington University; Charles Meertens, GST. For studies of sea-level change, both single- and dual-frequency systems have been placed on or close to tide gauges in California and Oregon this summer to provide a reference between the motion of the gauge and nearby land. "We don't know yet if the low-cost option will do as well in this environment. We need to run for a long time to see," Meertens says.

Bridge monitoring

Golden Gate and San Francisco Bay Bridges, California: Mehmet Celebi, USGS. Celebi needs the precision of dual-frequency sensors to monitor structural changes in San Francisco's famed bridges during temblors or high winds. But he'd also like to use the real-time data streaming the L1 communications system provides. This hybrid system is in the early stages of development.

Tropospheric tomography, too

The low cost of deploying L1s in a dense array inspired Christian Rocken in GST's GPS Research Group and graduate student John Braun (University of Colorado and NCAR's Advanced Study Program, or ASP) to test the L1's usefulness for gathering data on water vapor in the troposphere. "What's neat is that John Braun started out as an associate scientist working on the L1 development team, and now he's doing his dissertation research looking at fine-scale tropospheric conditions using the L1s," says GST's Meertens.

When it comes to data for initializing weather models, "one thing we can't measure very well right now is where the water vapor is," says Braun. Rocken and others demonstrated years ago that the time it takes a signal from a GPS satellite to pass through an air mass to a receiver on earth could be used to estimate precipitable water vapor--the amount of water vapor in the zenith direction above that receiver. More useful still would be dense arrays of ground units measuring the amount of water vapor along multiple paths at multiple angles in the directions of multiple satellites. The resulting slant-path measurements could be used to assemble a three-dimensional view of the water vapor field within an air mass. This view of horizontal and vertical structure would improve the level of detail in small-scale weather models.

Braun's research, supported by ASP, builds on pioneering work by GST's Randolph Ware and colleagues to measure slant-path water vapor (also called slant water vapor). "SWV is still a very new measurement," Braun explains. In fact, there have been no peer-reviewed articles with such measurements to date; the GST team has a manuscript in preparation. Braun is encouraged by a recent simulation of SWV measurements by Sandy McDonald at NOAA's Forecast Systems Laboratory. McDonald suggested that spacing GPS receivers at 40-km intervals and measuring SWV would provide sufficient resolution for mesoscale models. The GST experiment will attempt a much smaller-scale view--storm scale--by spacing L1 receivers one mile apart.

Rocken obtained support for the project from the U.S. Department of Energy's Atmospheric Radiation Measurements program; the experiment will take place at ARM's Southern Great Plains (SGP) facility in Lamont, Oklahoma. ARM researchers are interested in the view that the experiment promises to provide. Detailed definition of the water vapor field is important for the efforts to parameterize atmospheric radiation that are at the core of the ARM program.

Instrument installation was slated for late August at 14 locations within a three-mile by four-mile grid near the facility. Braun plans to collect data while a differential absorption lidar (DIAL) visits the SGP facility during September and part of October. Other lidars and three-hour radiosonde launches at SGP will also provide data for comparison with the L1s.

"It's an exciting project and it's been fun to work on," says Braun. "There's a lot of logistics in it--talking to 14 different farmers and asking them to put a piece of concrete in the middle of their field." Braun has a one-year lease with each farmer. If everything works without a hitch, the site will produce a three-dimensional data product every half hour or 15 minutes.

While one site is too small to be more than a proof of concept, McDonald has suggested that a network of receivers measuring SWV throughout the country would have a significant impact on forecasting. If the Oklahoma project is a success, Braun foresees enlisting existing GPS receivers, like those in NOAA's National Geodetic Survey network, known as CORS (continuously operating reference stations). Combining them with the proposed SuomiNet of GPS receivers at participating U.S. universities (see the Fall 97 Quarterly) might yield enough receivers to provide SWV data over a good portion of the United States one day.

Braun's advisors at the University of Colorado's Department of Aerospace Engineering Sciences are Judith Curry and Kristine Larson. William Kuo is his sponsor in NCAR's Advanced Study Program.


In this issue... Other issues of UCAR Quarterly
UCARNCARUOP

Edited by Carol Rasmussen, carolr@ucar.edu
Prepared for the Web by Jacque Marshall
Last revised: Tue Apr 4 15:11:20 MDT 2000