Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Lessons from the 2019 Ridgecrest earthquakes: a geodetic perspective

02 July 2020 10:04am
By Yuri Fialko, Gareth Funning

The 2019 M7.1 Ridgecrest earthquake was the third major earthquake to occur in the Eastern California Shear Zone since the advent of satellite space geodesy in the 1970's and 1980’s. All three events exemplify how high quality geodetic data can provide new insights into earthquake processes. The 1992 M7.3 Landers earthquake famously showed for the first time how Interferometric Synthetic Aperture Radar (InSAR) can image surface displacements in high resolution, enabling the mapping of fault slip in unprecedented detail (Massonnet et al., 1993). Through improved InSAR data processing and coverage, the surface displacement field of the 1999 M7.1 Hector Mine earthquake was the first to be mapped in 3D at high resolution (Fialko et al., 2001), revealing not only the complexities of subsurface fault geometry, but also variations in mechanical properties of the fault zones themselves (Fialko et al., 2002). Global Navigation Satellite Systems (GNSS) measurements during and after both earthquakes recorded ongoing postseismic deformation at ever improving density and precision, providing a window into the constitutive properties of the crust and the mantle beneath (Pollitz, 2003; Freed and Bürgmann, 2004; Takeuchi and Fialko, 2013).

Even compared to their lofty predecessors, the 2019 Ridgecrest earthquakes (the July 5th M7.1 mainshock and its M6.4 foreshock on July 4th) set new standards for the amount and quality of observations of surface deformation due to large earthquakes.

Figure 1. Co-seismic displacements due to M6.4 and M7.1 Ridgecrest earthquakes derived from Sentinel-1 (interferograms, range offsets), ALOS-1 (interferograms), and Cosmo-Skymed (azimuth offsets) data. Colors denote the amplitude of vertical displacements (in cm), and arrows are the horizontal displacements sub-sampled on a sparser grid. Black wavy lines denote the surface rupture mapped by field surveys. Blue and red stars indicate epicenters of the M6.4 foreshock and the M7.1 mainshock, respectively.

The Ridgecrest earthquakes were imaged by multiple radar satellite missions, including Sentinel-1A/B, ALOS-2, and COSMO-SkyMed. Short revisit times (e.g., 6 days for Sentinel-1 satellites) and systematic acquisitions from both the ascending and descending satellite tracks (providing different lines of sight), combined with nearly ideal surface conditions for radar interferometry, made the 2019 Ridgecrest earthquake sequence one of the best geodetically-recorded seismic events to date. Figure 1 shows the horizontal and vertical coseismic displacements for the pair of the M6.4 foreshock-M7.1 mainshock, derived from a combination of Sentinel-1, ALOS-2, and Cosmo-Skymed data. The data reveal an expected pattern of uplift and subsidence in the compressional and extensional quadrants, respectively, of a strike-slip rupture, and “rotating” horizontal displacements that are predominantly anti-symmetric with respect to the M7.1 rupture. Deviations from anti-symmetry are manifested in larger amplitude of displacements on the western/eastern sides of the fault to the north/south of the mainshock epicenter, and can be attributed to along-strike variations in fault dip. In the south-west quadrant, the displacement pattern is further complicated due to a contribution of the M6.4 foreshock. The data also reveal a mini-graben to the north of the epicenter formed by antithetic normal faults representing shallow splays of the M7.1 rupture.

Early results from the data processing (interferograms, phase gradients, pixel offsets) were shared with teams working in the field in near real time, and were used to identify and map additional surface ruptures.

 

Figure 2. Example campaign GNSS time series (station H701, east component) spanning the Ridgecrest earthquakes in 2019, with photos of the tripod-mounted antenna installed above the survey benchmark (round pillar) at various times. The site was first occupied in February 2019, as part of SCEC-funded research targeting the Garlock fault, providing a robust pre-event position (black dashed line). Following the M6.4 foreshock on July 4, the site was reoccupied the same day. We estimate a westward displacement of ~50 mm from the foreshock on the basis of that measurement. The station was found still standing on July 6, following the M7.1 mainshock on July 5; the station was displaced a further 80 mm to the west in that event. The time series was interrupted in September when the antenna was discovered stolen. A final 2019 measurement from November shows a total westward postseismic displacement of 15 mm at this station for the first four months following the earthquakes. (Time series processed by Michael Floyd, MIT, photos by Gareth Funning.)

 

Another major component of the geodetic community response to the 2019 Ridgecrest earthquakes involved collection of campaign GNSS data. It began immediately following the M6.4 foreshock, and involved a coordinated effort of scientists in the field and remote contributors. Multiple groups of field geodesists (from UCR, SIO, UNR and the USGS) mobilized and conducted campaign GNSS surveys of the existing geodetic monuments. A timely response is of essence because, once the earthquake rupture stops, postseismic deformation ensues at rates that are highest early on. Consequently, to accurately characterize coseismic fault slip and post-event ‘afterslip’, observations need to begin as early as possible. In case of the M6.4 foreshock, the first teams began collecting data within hours of the earthquake, fortuitously capturing displacements from both the M6.4 foreshock and the M7.1 mainshock (Figure 2). These campaign GNSS data supplement, and significantly densify, measurements made at continuous GNSS stations in the region from NOTA (the Network of the Americas).

Figure 3. Slip model of the 2019 Ridgecrest earthquakes derived from a joint inversion of the InSAR and GNSS data. Pink lines denote the geologically mapped rupture trace, black dots denote aftershocks, and black circle denotes the hypocenter of the M7.1 mainshock (from Jin and Fialko, Special BSSA Issue, 2020).

 

These and other near-field GNSS measurements of coseismic displacements were essential for constraining slip models of the M6.4 foreshock and M7.1 mainshock (Figure 3), as InSAR data were able to capture only a combined deformation signal from the two events that were separated by less than 2 days.  Campaign data from the field teams was relayed to colleagues working remotely (at MIT, SIO, UNR and the USGS) for processing and combining with data from NOTA stations (e.g. Floyd et al., 2020). With support from NSF RAPID funding, coordinated by SCEC, campaign GNSS measurements continued for one year following the 2019 Ridgecrest sequence  at over 30 sites, capturing a decaying postseismic transient (e.g. Figure 2). The data show postseismic velocities on the order of up to a few centimeters per month, comparable to the velocities measured after the similar-sized M7.2 El Mayor-Cucapah earthquake south of the US-Mexico border in 2010. In both cases, the pattern of the near-field deformation appears to be consistent with afterslip at the periphery of high coseismic slip areas.

What lessons, then, have we learned from the geodetic response to the Ridgecrest earthquake so far? We have seen that both field scientists and remote contributors can play important roles – the former by capturing ephemeral information (e.g. the displacement contribution from the M6.4 foreshock) that would otherwise be lost, the latter by processing the data collected by field teams and satellites and providing information to assist field teams in a timely manner. Modern radar satellite systems now provide data in a time frame (a few days) that can directly inform the scientific field response to an earthquake. In order to manage a rapid GNSS field response, it is vital to maintain ongoing campaign GNSS data collection efforts – such efforts, funded in the past by SCEC and other sponsors, develop institutional memory as to the locations of critical survey benchmarks, and also provide up-to-date pre-event positions than can be compared to post-event data. Maintaining clear communications between groups participating in the post-earthquake field activities minimizes duplicated or wasted effort, and helps to apportion tasks; for instance, it was agreed early on that USGS scientists should focus on measuring GNSS sites within the National Air Weapons Station at China Lake, given the difficulty of obtaining access for other groups. Finally, the density of measurements that is possible with campaign GNSS, with over 30 additional sites supplementing the continuous GNSS network, will provide a wealth of data for future studies of the postseismic deformation. We note that this style of deformation monitoring is highly cost-effective – the expense required to support dozens of campaign measurements is a fraction of the cost of a single continuous GNSS station, and suggest that similar efforts be supported in future post-earthquake investigations.

About the Authors

Yuri Fialko is a professor at Scripps Institution of Oceanography at the University of California, San Diego. His research focuses on deformation of the Earth's crust due to earthquake faults, magmatic sources, and other manifestations of active tectonics. He has been a member of the SCEC Board of Directors since 2006. 
Gareth Funning is an associate professor at the University of California, Riverside. He studies movements and deformation of the Earth's surface to understand earthquake cycle processes, slow movements of faults ('fault creep'), and human activities such as geothermal power production. He has been co-leader of the Tectonic Geodesy working group of the SCEC Science Planning Committee since 2015.

Acknowledgements

This research was supported by the Southern California Earthquake Center. SCEC is funded by NSF Cooperative Agreement EAR-1600087 and USGS Cooperative Agreement G17AC00047. Additional support was provided by NSF RAPID Awards EAR-1945728 and EAR-1945760. We thank our SCEC and USGS colleagues for collaborating in this research.

References