The Bay Area Velocity Unification (BAVU); Bringing Together Crustal Deformation Observations from throughout the
San Francisco Bay Area

Matthew A. d'Alessio, Ingrid A. Johanson, Roland Bürgmann, and the UC Berkeley Active Tectonics Group

Figure 26.1: Map of the San Francisco Bay area in a Pacific Plate-Sierra Nevada block projection with GPS Velocities from 1994-2003 relative to station LUTZ in the Bay Block (yellow square). Velocities consistent with a small circle path predicted from the Euler pole of the Pacific Plate-Sierra Nevada block rotation show up as horizontal arrows.

Introduction

In an effort to put together the most comprehensive picture of crustal deformation in the San Francisco Bay Area, The UC Berkeley Active Tectonics Group has begun work on the Bay Area Velocity Unification (BAVU ``Bay-View"). This dataset unites campaign GPS data for nearly 180 GPS stations throughout the greater San Francisco Bay Area from Sacramento to San Luis Obispo. The BAVU dataset includes data collected from 1991 to 2003 by U. C. Berkeley, the U.S. Geological Survey, the California Department of Transportation, Stanford University, U. C. Davis and the Geophysical Institute in Fairbanks, AK. These are combined with continuous GPS data from the BSL's Bay Area Regional Deformation (BARD) network. The BAVU dataset will form a consistent velocity field that will serve as the basis for monitoring fault slip and strain accumulation throughout the greater San Francisco Bay region.

Technical Overview of GPS Data Collection and Processing

Data Collection

At UC Berkeley we occupy each benchmark in our campaign GPS networks yearly. When possible, we collect data for at least two continuous 24 hour sessions, with some occupations spanning as long as seven days. However, much of the study area is in urban or suburban settings, making it impossible to leave GPS equipment unattended and limiting the occupation time to the logistical limits of the human operator. For these sites, occupations may be as short as 6 hours or as long as 12 hours, depending upon the time it takes to travel to the site and the efficiency of the operator. We usually repeat surveys of these sites at least once. Other agencies contributing data to the BAVU dataset generally follow the same guidelines and provide at least 6 hours of data per site per day.

Processing Baselines

We process campaign GPS data using the GAMIT/GLOBK software package developed at the Massachusetts Institute of Technology, which uses double-difference phase observations to determine baseline distances and orientations between ground-based GPS receivers. Along with campaign data, we process five global stations from the International GPS Service (IGS) network and four to six nearby continuous stations from the BARD network. Cycle slips are automatically identified and fixed using the AUTCLN routine within GAMIT. We use standard models for satellite radiation pressure and tropospheric delay. Ambiguities are fixed using the widelane combination followed by the narrowlane, with the final position based on the ionospheric free linear combination (LC or L3). For baselines shorter than 500 meters, we calculate an additional solution using only L1 data. Baseline solutions are loosely constrained (100 m) until they are combined together.

Combining Solutions

We combine daily ambiguity-fixed, loosely constrained solutions using the Kalman filter approach implemented by GLOBK. Within a given day, we include data processed locally as well as solutions for the full IGS and BARD networks processed by and obtained from SOPAC at the University of California, San Diego. During this combination, we weight each solution file relative to the other solution files for that day proportionally to the prefit chi-squared for the file when run through the Kalman filter independently. We uniformly scale the covariances of the entire combination so that the prefit chi-squared for the combined daily solution is approximately 1.0. Using the Kalman filter, we combine all daily solutions with equal weight to estimate the average linear velocity of each station in the network. We fix the final positions and velocities into the global reference frame using the GLOBK stabilization routine, allowing for rotation and translation of the network. Our final covariance matrix and the uncertainty estimates derived from it are scaled by the aposteriori chi-squared for the full combination. GLOBK also allows for the modeling of uncertainty using a “benchmark wobble,” but our current solutions have not included this effect.

GPS Results

Figure 26.1 shows GPS velocities for the entire BAVU dataset relative to station LUTZ on the Bay Block. Between UC Davis and the Farallon Islands, we observe $33 mm yr^{-1}$ of relative displacement between the Pacific plate and the Sierra Nevada-Great Valley block of the North American plate. The Bay block shows up as a relatively undeformed block with many of the stations having velocities relative to LUTZ that are so small that they plot as dots in the figure.

Figure 26.2: Map of deformation near the transition zone of the San Andreas Fault. Velocities relative to LUTZ on the Bay block.

Southern Bay Area

The southern section of the BAVU dataset includes the region affected by postseismic deformation following the Loma Prieta earthquake (Fig. 26.2). To avoid "contamination" of the regional deformation pattern by transient processes, we have not included data in this area collected before 1994. The southern Bay Area exhibits mostly fault-parallel right-lateral motion, with no indication of the fault-normal compression observed in the Foothills thrust belt immediately after the Loma Prieta Earthquake (Bürgmann, 1997). Other transient processes such as several slow earthquakes on the Central San Andreas fault are captured by the dataset, but their effect is likely small when spread over several years. The future inclusion of InSAR data in the BAVU dataset will allow the identification of some transient deformation events that the GPS networks are too sparse to capture. It should also improve our ability to measure surface creep.

Figure 26.3: Map of deformation near the Hayward fault. Velocities relative to LUTZ on the Bay block.

Hayward Fault

BAVU includes 64 stations within 15 km of the Hayward fault distributed along strike and in profiles perpendicular to the fault (Fig. 26.3). Creep along the Hayward fault allows the Bay block to slide past the East Bay Hills block with only minimal internal deformation and strain accumulation within either block. The BAVU model allows us to quantify the exact creep rate and place relative displacement across the fault in a regional context to quantify strain accumulation. Figure 26.4 shows observed variations in creep rate along strike of the Hayward fault.

Figure 26.4: Variations in surface creep rate as a function of distance along strike of the Hayward fault. GPS data from the BAVU model agree fairly well with the terrestrial geodetic results of Lienkaemper, 2001 (``L2001"). Note how the creep rate from L2001 is different for different time periods. Because BAVU and L2001 cover slightly different time periods, some of the differences between the two results likely reflects actual temporal fluctuations in creep rate.

These variations reflect both the spatial distribution of strain accumulation and temporal variations in creep rate. This is highlighted by comparisons between the BAVU GPS data and trilateration and triangulation collected by the US Geological Survey for the past several decades (grey lines in Fig. 26.4). Places where BAVU deviates most from the long-term observations by the USGS are places where the USGS indeed shows different creep rates on shorter time scales (grey circles and open squares in Fig. 26.4. Observations from five to ten year periods (such as the nine-years covered in BAVU) can deviate from longer-term observations by more than $2\sigma$. These fluctuations in creep rate must therefore be considered when using observations from BAVU to estimate long-term elastic strain accumulation and probabilistic earthquake magnitudes.

References

Bürgmann, R., L. P. Segall, M. Lisowski, and J. Svarc, Postseismic strain following the 1989 Loma Prieta earthquake from GPS and leveling measurements, J. Geophys. Res., 102(B3), 4933-4955, 1997.

Lienkaemper, J.J., J.S. Galehouse, and R. W. Simpson, Long-term monitoring of creep rate along the Hayward fault and evidence for a lasting creep response to 1989 Loma Prieta earthquake, Geophys. Res. Lett., 28 (11), 2265-2268, 2001.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2004, The Regents of the University of California