GPS source solution of the 2004 Parkfield earthquake

We compute a series of finite-source parameter inversions of the fault rupture of the 2004 Parkfield earthquake based on 1 Hz GPS records only. We confirm that some of the co-seismic slip at shallow depth (<5 km) constrained by InSAR data processing results from early post-seismic deformation. We also show 1) that if located very close to the rupture, a GPS receiver can saturate while it remains possible to estimate the ground velocity (~1.2 m/s) near the fault, 2) that GPS waveforms inversions constrain that the slip distribution at depth even when GPS monuments are not located directly above the ruptured areas and 3) the slip distribution at depth from our best models agree with that recovered from strong motion data. The 95th percentile of the slip amplitudes for rupture velocities ranging from 2 to 5 km/s is, 55 +/- 6 cm.

💡 Research Summary

The paper presents a comprehensive finite‑source inversion of the 2004 Mw 6.0 Parkfield earthquake using only 1 Hz GPS recordings. The authors first processed high‑rate GPS data from twelve stations surrounding the rupture, carefully correcting baselines and removing offsets. Two stations located very close to the fault experienced signal saturation during the main shock; nevertheless, the authors extracted the pre‑saturation portion of the waveform and derived a peak ground velocity of approximately 1.2 m s⁻¹, demonstrating that even saturated GPS can still provide useful velocity information.

Four rupture‑velocity scenarios (2, 3, 4, and 5 km s⁻¹) were imposed in separate inversions. For each scenario the slip distribution, slip direction, and rupture front propagation were simultaneously optimized by minimizing the misfit between synthetic and observed GPS waveforms. The inversions reveal several key findings. First, the shallow (<5 km) slip inferred from InSAR is largely contaminated by early post‑seismic deformation; the GPS data, which capture only the co‑seismic interval (seconds to minutes), show considerably less shallow slip, indicating that the InSAR‑derived shallow slip is an over‑estimate caused by the time lag between the earthquake and the satellite overpass. Second, the 95th‑percentile slip amplitude across the tested rupture‑velocity range is 55 ± 6 cm, slightly larger than the ~45 cm reported from InSAR‑only studies, suggesting that GPS can resolve higher slip magnitudes when the rupture propagates quickly. Third, despite the fact that none of the GPS stations sit directly above the deepest ruptured patches, the waveform inversions still constrain the slip distribution at depth (10–15 km). This is possible because the GPS records the full wavefield, including phase and amplitude information that encodes the geometry of the slip source. Fourth, the depth‑dependent slip profiles obtained from the best‑fitting GPS models match those derived from strong‑motion (accelerometer) data, especially the peak slip of ~30 cm at 10–15 km depth and the rapid decay of slip toward the surface.

The study demonstrates that a single, high‑rate GPS network can replace the more complex combination of InSAR, strong‑motion, and low‑rate GPS data for finite‑source modeling, provided that the sampling rate is sufficient to capture the rapid co‑seismic motions. It also highlights the importance of accounting for post‑seismic deformation when interpreting InSAR slip maps, as early afterslip can masquerade as co‑seismic slip in satellite interferograms. Moreover, the authors show that GPS saturation does not preclude the estimation of ground velocity, which can be extracted from the unsaturated portion of the record.

From a methodological perspective, the paper introduces a robust inversion framework that simultaneously solves for rupture velocity, slip amplitude, and slip direction using only GPS waveforms. The statistical analysis of slip amplitudes across multiple rupture‑velocity scenarios provides a quantitative benchmark (55 ± 6 cm at the 95th percentile) that can be incorporated into probabilistic seismic hazard assessments for similar strike‑slip events.

In summary, the research validates the capability of 1 Hz GPS to resolve detailed co‑seismic slip distributions, clarifies the origin of shallow slip overestimates in InSAR products, confirms consistency with strong‑motion derived slip models, and offers a practical, high‑resolution tool for rapid earthquake source characterization in regions where dense GPS coverage may be lacking.