next up previous contents
Home: Berkeley Seismological Laboratory
Next: Plate Boundary Deformation Project Up: Operations Previous: Parkfield-Hollister Electromagnetic Monitoring Array   Contents

Bay Area Regional Deformation Network



The Bay Area Regional Deformation (BARD) network of permanent, continuously operating Global Positioning System (GPS) receivers monitors crustal deformation in the San Francisco Bay area ("Bay Area") and northern California (Murray et al., 1998a). It is a cooperative effort of the BSL, the USGS, and several other academic, commercial, and governmental institutions. Started by the USGS in 1991 with 2 stations spanning the Hayward fault (King et al., 1995), BARD now includes 51 permanent stations and will expand to about $\sim $70 stations by July 2002 (Figure 6.1). The principal goals of the BARD network are: 1) to determine the distribution of deformation in northern California across the wide Pacific-North America plate boundary from the Sierras to the Farallon Islands; 2) to estimate three-dimensional interseismic strain accumulation along the San Andreas fault (SAF) system in the Bay Area to assess seismic hazards; 3) to monitor hazardous faults and volcanoes for emergency response management; and 4) to provide infrastructure for geodetic data management and processing in northern California in support of related efforts within the BARD Consortium and with surveying, meteorological, and other interested communities.

Figure 6.1: Operational BARD stations (solid triangles) in northern California (top) and in the San Francisco Bay area (bottom). The oblique Mercator projection is about the NUVEL-1 Pacific-North America Euler pole so that expected relative plate motion is parallel to the horizontal. Circled stations use continuous telemetry. The eight station Long Valley Caldera (LVC) network and eight station Parkfield (PKFD) networks are also part of BARD. Other nearby networks (open triangles) include: Basin and Range (BARGEN), and Southern California Integrated GPS Network (SCIGN).
\epsfig{, width=15cm}\end{center}\end{figure*}

BARD presently includes 51 permanent, continuously operating stations, 14 of which monitor the Long Valley caldera near Mammoth. The remaining 37 stations (Table 6.1) include 19 maintained by the BSL (including 2 with equipment provided by Lawrence Livermore National Laboratory (LLNL) and UC Santa Cruz), 6 by the USGS, and one each by LLNL, Stanford University, UC Davis, UC Santa Cruz, Trimble Navigation, and East Bay Municipal Utilities District. Other stations are maintained by institutions outside of northern California, such as the National Geodetic Survey, the Jet Propulsion Laboratory, and the Scripps Institution of Oceanography, as part of larger networks devoted to real-time navigation, orbit determination, and crustal deformation.

Table 6.1: BARD Stations
ID & Name & Instit.$...
Long Valley caldera near Mammoth.} \\

Between 1993 and 1996, the BSL acquired 5 Ashtech Z-12 receivers from UC Berkeley and private (EPRI) funding, which together with 2 USGS receivers, formed the nucleus of the initial BARD network. Since 1996, the BSL has acquired additional Ashtech Z-12 receivers with Dorne-Margolin design choke ring antennas: 13 in 1996 from a combination of federal (NSF), state (CLC), and private (EPRI) funding, and 4 in the 2000 from federal (USGS) funding. All but 5 of these receivers have been permanently installed to enhance continuous strain measurements in the Bay Area and to consolidate the regional geodetic network. The network includes several profiles between the Farallon Islands and the Sierra Nevada in order to better characterize the larger scale deformation field in northern California (Figure 6.1). Three more of the BSL receivers will be installed next year, and 7 additional receivers will be installed during the next year as part of the NSF-funded mini-PBO project, which will establish collocated GPS/seismometer/borehole strainmeter observatories in the Bay Area (see Chapter 7).

In 1996, researchers from the BSL, the USGS, Stanford University, LLNL, UC Davis, and UC Santa Cruz formed a consortium of institutions involved in studies of the tectonic deformation in the San Francisco Bay area and northern California. Members of the BARD consortium agreed to pool existing resources and coordinate development of new ones in order to advance an integrated strategy for improving the temporal and spatial resolution of the strain field. They agreed in principle to the continued development of the network of permanently deployed GPS receivers, to the development and maintenance of a pool of GPS receivers for survey-mode operations that may be deployed in semi-permanent mode in the Bay Area when not otherwise in use, to archiving of all data at the NCEDC, and to the development of a coordinated data analysis facility that will process permanent, semi-permanent, and survey data.

Figure 6.2: Daily variation in relative position of MODB antenna. The east and north components are arbitrarily offset for clarity. The antenna began to malfunction around 2000.5 and was replaced around 2001.1. The station telemetry was poor around 2000.3.
\epsfig{, width=15cm}\end{center}\end{figure*}

In the remainder of this section, we describe the standard BARD station and some of the BARD-related activities the BSL has performed over the last year, including maintenance to existing stations, preparations to install an experimental single-frequency receiver, improvement in data archive and processing methods, and analysis of the data to estimate deformation signals monitored by the network.

BARD Stations

A BSL continuous GPS station uses a low-multipath choke-ring antenna mounted to a reinforced concrete pillar approximately 0.5 meter above local ground level. The reinforcing steel bars of the pillar are drilled and cemented into rock outcrop to improve long-term monument stability. It uses a low-loss antenna cable to minimize signal degradation on the longer cable setups that normally would require signal amplification. Low-voltage cutoff devices are installed to improve receiver performance following power outages. The Ashtech Z-12 receiver is programmed to record data once every 30 seconds, observing up to 12 satellites simultaneously at elevations down to the horizon.

Tests performed by UNAVCO on low antenna mounts revealed that estimates of tropospheric water vapor from the GPS data are strongly correlated with signal multipath errors, which can degrade the precision of the vertical position estimates. Most of the BSL GPS stations use monuments that elevate the antennas 0.5-1.0 m above the ground surface, which helps to minimize the correlations between multipath and tropospheric parameters.

The stations are equipped with SCIGN-designed hemispherical domes. Domes cover the antennas to provide security and protection from the weather and other natural phenomenon. The SCIGN dome is designed for the Dorne-Margolin antennas and minimizes differential radio propagation delays by being hemispherical about the phase center and uniform in thickness at the 0.1 mm level. It is also very resistant to damage and, in its tall form combined with the SCIGN-designed antenna adapter, can completely cover the dome and cable connections for added protection. All new stations use the adapters and tall domes. Some of the older stations in well protected areas use the short domes.

Data from all BSL-maintained stations are collected at 30-second intervals and transmitted continuously over serial connections. Two sites (MOLA and TIBB) have direct radio links to Berkeley, and MODB uses VSAT satellite telemetry. The remaining 17 sites use frame relay technology, either alone or in combination with radio telemetry. Twelve GPS stations are collocated with broadband seismometers and Quanterra data collectors. With the support of IRIS we have developed software that will allow continuous GPS data to be converted to MiniSEED opaque blockettes, which can be stored and retrieved from the Quanterra data loggers (Perin et al., 1998). This approach preserves GPS data during telemetry outages.

Station Maintenance

During July 2000-June 2001, we preformed maintenance on existing BARD stations and prepared for new stations near the Hayward fault, on the San Francisco peninsula, and north Bay area regions.

In August 2000, Omnistar Corporation discontinued operation of their differential GPS Wide-Area Augmentation System (WAAS) network. This network included a station (ORVB) collocated with the BDSN station near Oroville that the BSL had helped to install in 1996. To maintain continuity of operation at this northern Sierran-Great Valley station, we installed a new receiver at ORVB and traded antennas with Omnistar to keep the existing antenna in place.

In late June 2000, the estimated position of MODB began to change at a significantly high rate, resulting in an apparent displacement of over 30 mm to the northeast over a one month period (Figure 6.2). MODB is one of the newest BARD stations, having been installed in November 1999 in the remote northeasternmost corner of California. Although the tectonics of this region, located in backarc of the Cascadia subduction near the Basin and Range province, are poorly understood, the apparent displacement of the antenna was much larger than could be reasonably expected from geophysical causes. Visual inspection of the concrete pier did not reveal any obvious signs of monument instability. An analysis of the raw phase measurements showed a significant drop in signal-to-noise ratios (SNR) after the anomaly began, particularly for observations at high satellite elevation angles (Figure 6.3), indicative of progressive hardware failure. Antenna swap tests revealed that the antenna was anomalous, and repair of the antenna dipole element by Ashtech restored the observations to their original signal levels. Similar antenna failures have been noted in other networks, such as SCIGN, and SNR analysis provides one method to investigate anomalous behavior.

Figure 6.3: L1 phase signal strength for properly operating (solid circles) and malfunctioning (open circles) antennas by angle of elevation above the horizon. The SNR values are in arbitrary units defined by the receiver manufacturer. Observations are binned in 10 degree intervals. The circles show the median value and the box-and-whisker symbols show the 25% and 75% quantile, and the minimum and maximum values. The malfunctioning antenna has significantly smaller SNR values at high elevation angles.
\epsfig{, width=8cm}\end{center}\end{figure}

During the past year, we converted telemetry at the 12 sites collocated with Quanterra data loggers to take advantage of the methods we developed to store data packets in MiniSEED blockettes. Originally these stations used direct serial connections that would result in loss of data during frame relay outages. The MiniSEED approach provides more robust data recovery from onsite backup on the Quanterra disks following telemetry outages. Our comparisons also show the loss of individual records is fewer when using the Quanterra MiniSEED rather than direct serial method due to the superior short-term data buffer in the Quanterra.

Figure 6.4: Schematic of the L1-system instrumentation and monument.
\epsfig{, width=8cm}\end{center}\end{figure}

L1-system Profile

The BSL staff is evaluating the performance of the UNAVCO-designed L1 system in an urban setting. This single-frequency receiver is relatively inexpensive but is less accurate than dual-frequency receiver systems that can completely eliminate first-order ionospheric effects. Hence we expect the L1 system to be most useful for short baseline measurements where ionospheric effects tend to cancel due to similar propagation paths. The systems are self-contained, using solar power and integrated radio modems (Fig. 6.4). During 1999, the BSL borrowed 2 receivers and a master radio from UNAVCO to perform the evaluation, but persistent hardware and software problems limited progress on this project. UNAVCO subsequently resolved many of the problems and in summer 2000, we received new, improved equipment and software for 4 systems and a master radio. We are currently preparing to deploy the systems on a 10-km profile extending normal to the Hayward fault between the UC Berkeley campus and the permanent BRIB site (Fig. 6.5).

We have completed permitting two of the stations (Fig. 6.5), both located on East Bay Municipal Utilities District (EBMUD) property. BDAM is located just east of the Briones Dam (BDAM) site, a few km west from the Briones (BRIB) continuous BARD station. VOLM is located on the ridge of the East Bay Hills close to Volmer Peak. Permitting is nearly complete at the other two stations, Wildcat (WLDC), also on EBMUD property near San Pablo Reservoir, and Grizzly Peak (GRIZ) on East Bay Regional Park property. Finding suitable stations with line-of-sight telemetry across the East Bay Hills proved challenging. Data from WLDC must pass through all the other stations, with its relay path being (in order) BDAM, VOLM, GRIZ, a repeater on the UC Berkeley Space Sciences Building, and then finally BSL in McCone Hall. We expect to install the stations with assistance from UNAVCO in Fall 2001. This profile, complemented by BRIB and EBMD to the west of the fault, will be most sensitive to variations in locking at 2-8 km depth. We expect that these systems will provide useful constraints on relative displacements near the Hayward fault in 3-5 years, and should help to resolve variations in creeping and locked portions of the fault (e.g., Bürgmann et al, 2000).

Figure 6.5: Location of L1-system (open triangles) and BARD (closed circles) stations. BSL, just southwest of the Hayward fault, is the location of the Berkeley Seismological Laboratory, where data from the 4 L1-system receivers northeast of the Hayward will be telemetered.
\epsfig{, width=12cm}\end{center}\end{figure*}

Figure 6.6: Velocities relative to stable North America for the BARD stations and other stations operated in nearby networks. Data from November 1993 to July 2000 was processed by the BSL using GAMIT software. Ellipses show 95% confidence regions, assuming white noise and $1 mm/\sqrt {yr}$ random-walk noise, with the predicted Pacific-North America relative plate motion in central California shown for scale.
\epsfig{, width=9cm}\end{center}\end{figure*}

Data Archival and Distribution

Raw and Rinex data files from the BSL stations and the other stations run by BARD collaborators are archived at the BSL/USGS Northern California Earthquake Data Center (NCEDC) data archive maintained at the BSL (Romanowicz et al., 1994). The data are checked to verify their integrity, quality, completeness, and conformance to the RINEX standard, and are then made accessible, usually within 2 hours of collection, to all BARD participants and other members of the GPS community through Internet, both by anonymous ftp and by the World Wide Web (

Data and ancillary information about BARD stations are also made compatible with standards set by the International GPS Service (IGS), which administers the global tracking network used to estimate precise orbits and has been instrumental in coordinating the efforts of other regional tracking networks. The NCEDC also retrieves data from other GPS archives, such as at SIO, JPL, and NGS, in order to provide a complete archive of all high-precision continuous GPS measurements collected in northern California.

Many of the BARD sites are classified as CORS stations by the NGS, which are used as reference stations by the surveying community. All continuous stations operating in July 1998 were included in a statewide adjustment of WGS84 coordinates for this purpose; a more recent adjustment is currently underway. Members of the BARD project regularly discuss these and other common issues with the surveying community at meetings of the Northern California GPS Users Group.

In the past 2 years the BARD Project and the NCEDC have collaborated with UNAVCO and other members of the GPS community to define database schema and file formats for the GPS Seamless Archive Centers (GSAC) project. When completed this project will allow a user to access the most current version of GPS data and metadata from distributed GSAC locations. The NCEDC will participate at several levels in the GSAC project: as a primary provider of data collected from BSL-maintained stations, as a wholesale collection point for other data collected in northern California, and as a retail provider for the global distribution of all data archived within the GSAC system. We have produced monumentation files describing the data sets that are produced by the BARD project or archived at the NCEDC, and are during the last year began creating incremental files describing changes to the holdings of the NCEDC so that other members of the GSAC community can provide up-to-date information about our holdings.

Data Analysis and Results

The data from the BARD sites generally are of high quality and measure relative horizontal positions at the 2-4 mm level. The 24-hour RINEX data files are processed daily with an automated system using high-precision IGS orbits. Final IGS orbits, available within 7-10 days of the end of a GPS week, are used for final solutions. Preliminary solutions for network integrity checks and rapid fault monitoring are also estimated from Predicted IGS orbits (available on the same day) and from Rapid IGS orbits (available within 1 day). Data from 5 primary IGS fiducial sites located in North America and Hawaii are included in the solutions to help define a global reference frame. Average station coordinates are estimated from 24 hours of observations using the GAMIT software developed at MIT and SIO, and the solutions are output with weakly constrained station coordinates and satellite state vectors.

Processing of data from the BARD and other nearby networks is split into 7 geographical subregions: the Bay Area, northern California, Long Valley caldera, Parkfield, southern and northern Pacific Northwest, and the Basin and Range Province. Each subnet includes the 5 IGS stations and 3 stations in common with another subnet to help tie the subnets together. The weakly constrained solutions are combined using the GLOBK software developed at MIT, which uses Kalman filter techniques and allows tight constraints to be imposed a posteriori. This helps to ensure a self-consistent reference frame for the final combined solution. The subnet solutions for each day are combined assuming a common orbit to estimate weakly constrained coordinate-only solutions. These daily coordinate-only solutions are then combined with tight coordinate constraints to estimate day-to-day coordinate repeatabilities, temporal variations, and site velocities.

The estimated relative baseline determinations typically have 2-4 mm WRMS scatter about a linear fit to changes in north and east components and the 10-20 mm WRMS scatter in the vertical component. Average velocities for the longest running BARD stations during 1993-2000 are shown in Figure 6.6, with 95% confidence regions. We have allowed $1 mm/\sqrt {yr}$ randow-walk variations in the site positions in order to more accurate characterization of the long-term stability of the site monuments and day-to-day correlations in position. The velocities are relative to stable North America, as defined by the IGS fiducial stations, which we assume have relative motions given by Kogan et al., (2000).

Most of the Sierra Nevada sites (CMBB, QUIN, and ORVB), as well as SUTB in the Central Valley, show little relative motion, indicating that the northern Sierra Nevada-Central Valley is tectonically stable. The motion of these sites relative to North America differs from the inferred motion of the western Basin and Range Province, suggesting 3 mm/yr right-lateral shear across the Walker Lane-Mt. Shasta seismicity trend. Deformation in the Pacific Northwest is generally consistent with interseismic strain accumulation along the Cascadia megathrust, the interface between the Juan de Fuca and North America plates, particularly in Washington where the velocity vectors are nearly parallel to the oblique convergence direction. Greater arc-parallel motion in Oregon and northern California may be due to the influence of the SAF system to the south and clockwise rotation of the southern Oregon forearc (Savage et al., 2000).

Deformation along the coast in central California is dominated by the active SAF system, which accommodates about 35 mm/yr of right-lateral shear. The Farallon Island site (FARB) off the coast of San Francisco is moving at nearly the rate predicted by the NUVEL-1A Pacific-North America Euler pole. Two-dimensional modeling of the observed fault-parallel strain accumulation predicts deep slip rates for the San Andreas, Hayward, and Calaveras/Concord faults are 19.3$\pm$1.8, 11.3$\pm$1.9, and 7.4$\pm$1.6 mm/yr, respectively, in good agreement with estimated geologic rates (17$\pm$4, 9$\pm$2, and 5$\pm$3 mm/yr, respectively). Most of the 46 mm/yr of relative motion is accommodated within a 100-wide zone centered on the SAF system and a broader zone in the Basin and Range Province in Nevada.

In the past year, we have begun to analyze the GPS observations in combination with about 50 additional stations, distributed both globally for independent orbit determination and regionally within the Pacific and North America plates for robust estimation of their relative motion. These additional stations significantly improve the definition of the velocity reference frame (e.g., ``stable'' North America), which is critical for tectonic interpretation of broadscale deformation patterns. Additional details about the deformation results and modeling are include in a later chapter, entitled ``Modeling Broadscale Deformation From Plate Motions and Elastic Strain Accumulation''.

Other Projects

We are also developing real-time analysis techniques that will enable rapid determinations ($\sim $minutes) of deformation following major earthquakes to complement seismological information and aid determinations of earthquake location, magnitude, geometry, and strong motion (Murray et al., 1998c). We currently process data available within 1 hour of measurement from the 18 continuous telemetry BSL stations, and several other stations that make their data available on an hourly basis. The data are binned into 1 hour files and processed simultaneously. The scatter of these hourly solutions is much higher than the 24-hour solutions: 10 mm in the horizontal and 30-50 mm in the vertical. Our simulations suggest that displacements 3-5 times these levels should be reliably detected, and that the current network should be able to resolve the finite dimensions and slip magnitude of a M=7 earthquake on the Hayward fault. We are currently investigating other analysis techniques that should improve upon these results, such as using a Kalman filter that can combine the most recent data with previous data in near real-time. The August 1998 M=5.1 San Juan Bautista earthquake (Uhrhammer et al., 1999) is the only event to have produced a detectable earthquake displacement signal at a BARD GPS receiver.


Under Barbara Romanowicz's general supervision, and with Mark Murray as head guru, André Basset, Bill Karavas, John Friday, Dave Rapkin, Doug Neuhauser, Rich Clymer, and Roland Bürgmann contribute to the operation of the BARD and L1 networks. Mark Murray contributed to the preparation of this chapter.


Bürgmann, R., D. Schmidt, R. M. Nadeau, M. d'Alessio, E. Fielding, D. Manaker, T. V. McEvilly, and M. H. Murray, Earthquake potential along the northern Hayward fault, California, Science, 289, 1178-1182, 2000.

King, N. E., J. L. Svarc, E. B. Fogleman, W. K. Gross, K. W. Clark, G. D. Hamilton, C. H. Stiffler, and J. M. Sutton, Continuous GPS observations across the Hayward fault, California, 1991-1994, J. Geophys. Res., 100, 20,271-20,283, 1995.

Kogan, M. G., G. M. Steblov, R. W. King, T. A. Herring, D. I. Frolov, S. G. Egorov, V. Y. Levin, A. Lerner-Lam, A. Jones, Geodetic constraints on the rigidity and relative motion of Eurasia and North America, Geophys. Res. Lett., 27, 2041-2044, 2000.

Murray, M. H., and P. Segall, Continuous GPS measurement of Pacific-North America plate boundary deformation in northern California and Nevada, Geophys. Res. Lett., in press, 2001.

Murray, M. H., R. Bürgmann, W. H. Prescott, B. Romanowicz, S. Schwartz, P. Segall, and E. Silver, The Bay Area Regional Deformation (BARD) permanent GPS network in northern California, EOS Trans. AGU, 79(45), Fall Meeting Suppl., F206, 1998a.

Murray, M. H., W. H. Prescott, R. Bürgmann, J. T. Freymueller, P. Segall, J. Svarc, S. D. J. Williams, M. Lisowski, and B. Romanowicz, The deformation field of the Pacific-North America plate boundary zone in northern California from geodetic data, 1973-1989, EOS Trans. AGU, 79(45), Fall Meeting Suppl., F192, 1998b.

Murray, M. H., D. S. Dreger, D. S. Neuhauser, D. R. Baxter, L. S. Gee, and B. Romanowicz, Real-time earthquake geodesy, Seismol. Res. Lett., 69, 145, 1998c.

Perin, B. J., C. M. Meertens, D. S. Neuhauser, D. R. Baxter, M. H. Murray, and R. Butler, Institutional collaborations for joint seismic and GPS measurements, Seismol. Res. Lett., 69, 159, 1998.

Romanowicz, B., B. Bogaert, D. Neuhauser, and D. Oppenheimer, Accessing northern California earthquake data via Internet, EOS Trans. AGU, 75, 257-260, 1994.

Savage, J. C., J. L. Svarc, W. H. Prescott, and M. H. Murray, Deformation across the forearc of the Cascadia subduction zone at Cape Blanco, Oregon, J. Geophys. Res., 105, 3095-3102, 2000.

Uhrhammer, R., L. S. Gee, M. Murray, D. Dreger, and B. Romanowicz, The $M_{w}$ 5.1 San Juan Bautista, California earthquake of 12 August 1998, Seismol. Res. Lett., 70, 10-18, 1999.

next up previous contents
Next: Plate Boundary Deformation Project Up: Operations Previous: Parkfield-Hollister Electromagnetic Monitoring Array   Contents

Berkeley Seismological Laboratory, 202 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2001, The Regents of the University of California.