A disaster in Chile creates a challenge to gather accurate data for seismic research. The result is a new GPS network for surveyors and scientists.

After five weeks in South America, Dr. Mike Bevis was looking forward to getting home. A professor of geodynamics at The Ohio State University, Bevis had been in the mountains of Patagonia conducting research on the use of GPS for “weighing” glacial ice sheets. He was in the airport in Lima, Peru, on February 27, 2010, when he learned about the earthquake in Chile. From Bevis’ standpoint as a geophysicist, it would be the most important earthquake he had ever encountered.

The quake was a monster. It struck in the early morning hours, just off the coast of the Maule region in south-central Chile. The enormous shock (magnitude 8.8) drove the Nazca plate down and beneath the South American plate. In Concepción, less than 160 kilometers (100 miles) south of the epicenter, the surface of the earth moved westward by roughly 3 meters (9.8 feet). The earthquake’s rupture zone was more than 600 kilometers (370 miles) long and 130 kilometers (80 miles) wide. Nearly 500 people were killed in the quake; most were victims of the massive tsunami that it generated. In addition to the immense human needs, the quake created a demand for immediate action by GPS specialists--for two important reasons. 

Eric C. Kendrick views the damage to Route P-22 to the west of Arauco. Hundreds of bridges and bridge approaches were damaged in the earthquake. Inset: In Caleta Tumbes north of Concepción, the tsunami caused extensive damage. Most fatalities were due to the tsunami. Photo by Leonardo Perez 

Rapid Reaction

Chile and Argentina have been significant for geophysicists for years. The countries sit on the Pacific Ocean “Ring of Fire,” and earthquakes are not uncommon. A research effort named the Central and Southern Andes GPS Project (CAP) uses dozens of continuous GPS stations (CGPS) to monitor the tectonic plates. Begun in 1993, CAP also uses hundreds of survey markers for episodic tectonic monitoring. Following any major earthquake, geophysicists need to gather as much GPS data as possible--and as soon as possible--after the initial shock. The data are used to compute accurate measurements of displacement caused by the quake and to create baseline information for subsequent measurements of aftershocks and the more subtle but persistent post-seismic deformation.

The second reason was for surveying and geodesy. According to Héctor Contreras, professor of satellite geodesy at the University of Santiago, many surveyors and engineering firms in Chile use GPS for their precise surveys and positioning projects. This work, which typically uses real-time kinematic (RTK) or post-processed static methods, usually covers baselines on the order of 5 to 20 kilometers (3 to 12 miles). The surveyors rely on reference markers whose coordinates are provided by the Chilean Military Geographic Institute (Instituto Geográfico Militar or IGM). There are about 300 IGM markers in Chile, Contreras said, and probably more than 100 were affected. “It’s not only of interest to science,” Bevis explained. “The quake was a geodetic and surveying catastrophe.”

IGM had worked closely with CAP for more than 15 years to develop the network of monuments. While most of the IGM survey markers were intact, the ground on which they sat had moved. Their previously published positions were worthless, and they were continuing to displace. At the time when it was needed the most, Chile’s geodetic reference framework was unusable. The only solution was to build a lot of new CGPS stations in the affected region--and do it very quickly. So instead of getting on a plane for the U.S., Bevis headed south.

Remeasurement of a survey marker. While most such markers were unscathed, the coseismic and posteismic motions left their coordinates unreliable for conventional survey activity. Damage to the building gives evidence to the quake’s power. Photo by Mike Bevis

Project Phoenix

Bevis, together with Jeff Genrich, PhD, from the California Institute of Technology, as well as Research Scientist Eric C. Kendrick, PhD, and Senior Engineer Dana J. Caccamise II from The Ohio State University, were among the first foreign scientists to arrive after the quake. They were soon followed by a team from France. (In addition to the CAP effort, researchers from France, Germany and other nations had also been collecting data for years.) While Bevis was en route to Chile, he was in constant communication with his colleagues. He and his American colleagues coordinated with the National Science Foundation (NSF) in the U.S. to arrange funding under the NSF’s Rapid Response Research (RAPID) mechanism. To obtain GPS equipment, Bevis called on Jim Normandeau at UNAVCO, a consortium that provides GPS and GNSS equipment and expertise to universities and research organizations. He also worked with Ben Brooks, PhD, at the University of Hawaii and Robert (Bob) Smalley Jr., PhD, at the University of Memphis, to devise the overall plan. Bevis was designated as the coordinator of the American effort in Chile, while Smalley would manage work in Argentina. Brooks and Mark Simons, PhD, from the California Institute of Technology (CalTech), were assigned to coordinate the data processing and would provide logistical support in the U.S. The objective was to rebuild the Chilean reference frame as quickly as possible. They decided to name the effort “Project Phoenix.” The fieldwork in Chile involved two long-term CAP partners: surveyors and engineers from IGM based in Santiago, and Juan Carlos Baez, PhD, and his students and colleagues at the University of Concepción.

People and equipment began to move. UNAVCO prepared 25 campaign kits, each consisting of a GPS receiver, solar panels and accessories for remote, long-term data collection. The kits would be loaned to the project for a year. To enable permanent installations, Trimble donated nine Trimble NetRS GPS Reference Stations with Trimble Zephyr Geodetic Antennas. Other companies also supplied equipment, and Bevis borrowed equipment from Ohio State’s inventory of receivers, as well. IGM contributed technical personnel and the use of their facilities as staging areas. Some IGM installations would also serve as locations for CGPS stations.

Normandeau said that one of the early challenges was just getting into the country. “Space on air freight for nonhumanitarian cargo simply was not available,” he explained. “We sent two people with 25 systems, each weighing roughly 40 pounds (18 kg) apiece. They called the airline and made arrangements to take the equipment as excess baggage. It was literally half a ton of equipment.”

Part of Mike Bevis’ situation map shows CGPS stations around Santiago. Colored pins indicate the organization responsible for each station. Photo by Mike Bevis

While they were in a hurry to get the new stations operating, the teams knew they needed to install solid, reliable CGPS stations. Although they preferred to set stations in bedrock, it was almost always necessary to place them in soil. To do so, they devised a simple, durable tripod mount for the GPS antennas by using a design developed in Bolivia in the previous five years. The new mount--a simpler and less expensive version of the mount developed by UNAVCO for its Plate Boundary Observatory (PBO) in the U.S.--was constructed using materials and tools available locally. After a bit of practice, the field teams could install a permanent mount in roughly five hours.

Even for an experienced project manager like Bevis, the scope of work was daunting. He was coordinating the work of 12 to 14 people in the field in Chile and a smaller group in Argentina, and maintaining contact with the French researchers to prevent gaps or overlaps in coverage. He was also collaborating with the team in the U.S., which was analyzing and modeling early results of the GPS observations. Bevis created a large situation map, on which he marked the location of planned and existing stations in three operational centers in Chile. He updated it frequently at the nearest center and sent digital photos of the map to the teams across the country, who then updated the other maps.

In order to move people, construction materials, tools and GPS equipment, the Phoenix team established logistical centers in Santiago, Los Angeles, Concepción and Talca to supply the local teams. “It was like being in a giant vortex and everything is spinning around,” Bevis recalled. “It was hard to get trucks and personnel. You didn’t know where everything and everyone was. You had frontline logistical centers where it was crazy, and you had centers that were calm. Part of the area was under martial law, and a lot of the communications were down. I was reminded about a general who said, ‘Amateurs talk about tactics; professionals talk about logistics.’ That was how it felt.” And through it all, the earth was still moving.

To get the observations started, the teams installed 25 CGPS stations in less than a month. Then they went back and hardened each station by adding durable instrument housings and permanent frames for the solar panels. Data communications from the stations remained spotty. For communications, many stations used Broadband Global Area Network (BGAN) satellite modems built by UNAVCO Senior Project Manager Frederick Blume, PhD, and the CAP team to send data to the UNAVCO servers. As cellular service improves, Bevis expects that many locations will be converted from the costly satellite service to conventional cellular modems. But in the first few weeks after the quake, many stations had to be downloaded manually, and the data carried to a location where it could be sent to Brooks. 

Tsunami damage to buildings in Llico near the tip of the Arauco Peninsula south of Concepción. Photo by Eric C. Kendrick

Fast Answers

As the receivers began to come online, the data flowed to Brooks in Hawaii. Within two weeks, he had developed preliminary estimates on the quake’s displacement. Using software developed by the Massachusetts Institute of Technology (MIT), Brooks and James Foster, Ph.D., processed pre-quake GPS data against reference stations in Brazil, differencing among reference points and survey points to get an array of baseline vectors. They then performed network adjustments to obtain best-fit positions. They repeated the process with post-quake data and took averages of pre- and post-quake positions to compute simple displacement vectors at the measured points. For most individual positions, they achieved an accuracy of better than 3 millimeters (0.01 feet), with two-sigma errors for the displacement vectors of 3 to 5 millimeters (0.01 to 0.02 feet). “For the initial post-earthquake estimates, we had to use a combination of ultra-rapid or rapid orbits,” Brooks said. “As the final orbits became available we reprocessed everything. Our preliminary estimate looked pretty good, and now we’re refining it.”

The computations revealed complex coseismic motion. Near Concepción, the earth’s crust had moved by more than 3 meters (9.8 feet) in a matter of seconds. Closer to the epicenter, the displacement was more than 4.5 meters (15 feet). The deformation affected much of central Chile and caused measurable displacement in Buenos Aires, Argentina, more than 1,300 kilometers (800 miles) from the epicenter. Near the port city of Lebu, Chile, the crust moved upward by 1.5 to 2 meters (4.9 to 6.6 feet), leaving boats stranded far from the new shoreline. 

Soon after the quake, local surveyors recovered IGM monuments. They began to document the damage and assist in reconstruction efforts. Photo by Eric C. Kendrick

New Reference Points for Surveyors

Like most regions of the world, Chile’s existing spatial reference system relies on physical survey markers. There is no real-time GNSS network (RTN) in Chile. Because the previously known coordinates of the markers were unusable, the damage to country’s survey infrastructure was analogous to the damage inflicted on the physical infrastructure of the country. The aggressive approach taken by CAP and IGM will pay off quickly. There are 40 receivers in Chile now, with more to come. The University of Liverpool has joined the CAP consortium and brought six GPS receivers of its own. “We will eventually have more than 40 geodetic reference stations available,” Bevis said. “We hope to build at least 35 continuous GPS stations in an area about 1,000 kilometers (600 miles) long in Chile, and another five to ten in Argentina. Of course, the challenge will be keeping the network going after UNAVCO’s one-year loan period expires. This is why vendor equipment donations are so crucial.”

Although it will take years for Chile to repair the damage from the earthquake, the Chilean surveying community will be better off because of it. The country will be covered with dozens of new monuments; many will be occupied by continuously operating receivers. As more receivers connect to the Internet, data availability will improve. Hector Parra of IGM said that all the new stations would become part of the Chilean National Spatial Reference System and be tied to SIRGAS (Sistema de Referencia Geocéntrico para las Américas, or Geocentric Reference System for the Americas) datum. The connection to SIRGAS will provide a single, well-managed reference frame that connects Chile with the rest of South America.

“Because the quake struck in a heavily-populated region of Chile, it would be a good location to create an active network of reference stations,” Contreras said. “But because of the mountainous terrain and inconsistent cellular coverage, that will take some more time. Having the CGPS stations in place is an important first step.” Contreras said that most RTK surveys will continue to use UHF radio communications.

From the researchers’ point of view, the February quake was the “perfect” earthquake. The faults were well known, and sufficient pre-quake data existed to accurately determine the quake’s deformation. Because of the GPS infrastructure in place before the earthquake and the number of stations that were constructed very quickly after the event, it will probably be the most carefully studied very large earthquake in history. Although the 2004 magnitude 9.2 event in Indonesia was larger, Bevis said that Chile is easier to access and has a better infrastructure for collecting and managing the GPS measurements.

“In terms of the purely scientific impact, I have little doubt this will be the most important earthquake ever,” said Bevis. “And it’s important to realize how the needs of the nation and the survey and engineering community are in parallel with the needs of the scientists. The reference frame in Chile is now bigger and denser than it was before the quake. In the face of the disaster, that was a positive outcome.”

For more information about UNAVCO, visit www.unavco.org. Additional information about Trimble equipment can be found at www.trimble.com

John Stenmark, LS, is a writer and consultant working in the AEC and technical industries. He has more than 20 years of experience in applying advanced technology to surveying and related disciplines.