Surface Creep Measurements from a Slow Earthquake on the
San Andreas Fault Using InSAR

Ingrid A. Johanson and Roland Bürgmann

Introduction

The ubiquity of slow earthquakes (SEQs) on faults (both dip-slip and strike-slip), suggests that they are a fundamental mode of strain release. The fact that they characteristically occur in the transition region between steadily slipping and locked faults may allow us to draw general conclusions about the mechanism by which faults transition between locked and creeping. To date, the San Andreas Fault near San Juan Bautista is the only location where slow earthquakes have occured on an accessible strike-slip fault. The slip in these events was much closer to the surface than in the typical subduction zone events, making it a unique and potentially very effective location to study the mechanics of slow slip.

Figure 28.1: Location map of the Central San Andreas Fault showing local background seismicity and creepmeter locations. Also shown is the main and aftershocks of the 1998 San Juan Bautista earthquake (in blue). The grey box outlines the profile shown in Figure 28.3.

Three slow earthquakes occured along the SAF in 1992, 1996 and 1998; each with moment magnitude close to that of the largest seismic earthquakes in the area ($M_{w}$5.5). In this report we focus on the 1998 SEQ, whose shallow depth (less than 5 km) and compact size of the slow earthquakes (relative to subduction zone SEQs), provides the opportunity to apply the high spatial resolution and good precision of InSAR to observing the deformation pattern of SEQs (Johanson and Bürgmann, 2002).

1998 SEQ

The 1998 $M_{w}$5.0 slow earthquake was immediately preceeded by the $M_{w}$5.1 San Juan Bautista earthquake (Uhrhammer et al., 1999) (Large blue circle in Fig. 28.1). The two events were located in the same region, with the slow earthquake rupturing the portion above 5km (Gwyther et al., 2000). The earthquake and SEQ ruptures were of comparable size and resulted in similar amounts of slip. An interferogram spanning from Aug. 18 1997 to Oct. 12 1998 contains adequate coherent data along the central San Andreas to observe near fault movement (Fig 28.2).

Figure 28.2: Subset of an interferogram from Track 299 Frame 2861 corresponding in area to the location map in figure 28.1. The interferogram contains significant topography correlated signal which are probably atmospheric errors. However, creep occurs on a scale much smaller than most atmospheric errors

Figure: Comparison of near field motion measured using InSAR and creepmeters XSJ, XHR and CWN. Green triangles are the total amount of creep measured by creepmeters during the time span of the interferogram and includes secular creep, the 1998 slow earthquake and the San Juan Bautista earthquake. Blue triangles are the amount of creep expected without the 1998 slow earthquake or San Juan Bautista earthquake. The interferogram shown is a subset of Track 299 Frame 2861; its location is shown in Figure 28.1.

The interferogram contains a sharp phase gradient aligned with the fault trace. If the range-change signal is attributed to purely strike-slip motion on the San Andreas Fault, then the interferogram indicates right-lateral slip of about the same amount observed by creepmeters. It is expected, however, that the measured range change will contain some vertical motion and may also contain atmospheric errors. We are working on incorporating ascending track frames and increasing the total number of available interferograms to account for these effects.

Near Field Fault Motion

Fig. 28.3 compares surface creep measurments made using InSAR to those from creepmeters. Red points were obtained by averaging phase values every 350 meters along strike and  50 meters away from each side of the fault. Pairs of average phase measurements were differenced and converted to San Andreas parallel strike-slip movement. Error bars are the standard deviations of values in each bin. Though the InSAR and creepmeter data match well, the InSAR data suggest that fault movement is much more variable than would be inferred from creepmeters alone. However, the variability may be the result of a poorly defined fault trace. Figure 28.3 also makes it apparent that the contribution of the slow earthquake to surface slip during the time spanned by the interferogram is small; the difference between the blue and green triangles. The contribution of interseismic creep must therefore be accounted for before slip is attributed to the SEQ. The variability in the InSAR data makes it a requirement that we use interseismic creep rates with similar spatial sampling. We are working on creating interferogram stacks using a patch-work method that will allow us to define average creep rates along the San Andreas.

Conclusions

InSAR is capable of measuring small scale movement such as creep. The interferogram shown here suggests that creep along the Central San Andreas varies considerably along strike. Future modelling of the fault system will benefit from the dense spatial sampling of InSAR. However, the contribution of the 1998 San Juan Bautista slow earthquake to the total range change is very subtle and will require careful removal of the interseismic signal.

References

Gwyther, R. L., C. H. Thurber, M. T. Gladwin, and M. Mee, Seismic and aseismic observations of the 12th August 1998 San Juan Bautista, California M5.3 earthquake,3rh San Andreas Fault Conference, Stanford University, 209-213, 2000

Johanson, I. A. and R. Bürgmann, An investigation of slow earthquakes on the San Andreas Fault using InSAR,EOS Transactions, AGU, 83(47), 357, 2002

Uhrhammer, R., L. S. Gee, M. H. Murray, D. Dreger and B. Romanowicz, The $M_{w}$5.1 San Juan Bautista, California earthquake of 12 August 1998, Seismological Research Letters, 70(1), 1999.