The 2019 Ridgecrest earthquake sequence occurred on two orthogonal faults, rupturing the surface in a complex pattern that included surface cracking, fault splays, and stepovers. The earthquake sequence began with a series of small events that preceded a M6.4 foreshock on July 4, 2019 on a northeast striking left-lateral fault followed by a M7.1 southeast striking right-lateral mainshock ~34 hours later on July 5, 2019. The extent of surface rupture in an arid region and the availability of new optical imaging platforms made it possible to image the ruptures, associated surface cracking, and surface change spanning the earthquake sequence. We used small Uninhabited Aerial Vehicles (sUAS) or drones to target specific areas after the earthquake and commercial spaceborne imagery to measure the entire ruptures and separate deformation from the M6.4 and M7.1 earthquakes (Figure 1).

Figure 1. Image differences from Planet Labs and from the small UAS. Center top shows the location map with the Ridgecrest earthquake sequence rupture area outlined in blue. Left two images show deformation maps of surface motion projected into fault parallel direction. Fault parallel directions are shown by black arrow. The top left images show the deformation map from correlating images spanning just the foreshock and the bottom left shows the surface displacement estimated from images spanning just the mainshock. Black rectangles show the sUAS survey areas. Top right shows the two target areas outlined in red for the small UAS surveys with inferred surface cracking shown in blue within the areas. Bottom right shows an example of cracks that can be observed in the stereo photogrammetric reconstruction.

One of the questions regarding this earthquake sequence was whether there was observable triggered slip on the M7.1 fault before the mainshock occurred and whether faults involved in the foreshock experienced additional slip caused by the mainshock. Most of the geodetic data, such as Interferometric Synthetic Aperture Radar, light detection and ranging, and optical satellite imagery, were acquired after both events had occurred making it difficult to discern which surface fractures happened when and their possible triggering mechanism. Optical images from the Planet satellites, a constellation of ~175 cubesat satellites acquiring images at least once a day were acquired before the M6.4 foreshock, after the M7.1 mainshock and between the two events. The image collected between the two events allowed us to separate the surface deformation between them (Milliner and Donnellan, 2020). We found that there was no observable significant triggered displacement at the surface along faults involved in the mainshock before it ruptured in the subsequent Mw 7.1 event, or re-rupture of faults involved in the foreshock. Due to the coarse resolution of the imagery we could not rule out smaller amounts of triggered slip or fault re-rupture less than 15 cm, but the results do provide a clear upper bound on the amount of possible fault displacement and interaction that occurred.

Figure 2. Comparison of before and after images of the M7.1 rupture using Google Earth imagery before the earthquake and the small UAS reconstructed image after the earthquake. Creation of the surface cracking associated with the rupture and offset of features such as road can be observed between the images.

Small UAS can produce extremely high-resolution imagery with a ground sample distance of 1 2 cm. We targeted small areas of both the M6.4 and M7.1 ruptures just south of highway 178 (Figure 1). We have re-observed each area multiple times since the earthquake to search for transient afterslip along the ruptures or other postseismic deformation near the fault ruptures (Donnellan et al, 2020). The surveys cover approximately 500 × 500 m areas just south of Highway 178 with an average ground sample distance of 1.5 cm. The first survey took place five days after the Mw 6.4 foreshock on 9 July 2019 and we have resurveyed the area with decreasing frequency since then. We collected hundreds of images over the areas and reconstructed stereo photogrammetric images using the Pix4D software package, constraining the surveys with ground control points. Differencing images from before the earthquake and from the small UAS measurements shows clear fault offset (Figure 2). We have not detected postseismic motion in post-earthquake image differences, suggesting that afterslip did not occur on these sections of the fault in the first few months after the events. We are continuing drone imaging of the area to search for longer term postseismic motions. The imagery proved useful for identifying surface cracking (Figure 1) and guiding field observations.

Both the near-surface and space-based observations show large rupture along the faults involved in the foreshock and mainshock. There is no clear evidence of triggered slip prior to the mainshock rupture along the northwest orientated faults, no sign of significant re-rupture (> 15 cm of slip) along the southwest orientated faults involved in the foreshock after the mainshock occurred, and little evidence of postseismic slip on the ruptures in the first few months after the rupture sequence. These results will ultimately help constrain crustal rheology, and improve understanding of how faults slip, how fault zones deform, how stress changes with large earthquakes, and fault interaction and mechanics.

About the Authors

	Andrea Donnellan is a principal research scientist at NASA's Jet Propulsion Laboratory. She studies earthquakes and crustal deformation by integrating geodetic imaging data and computational infrastructure with modeling and analysis tools. Her current research focus is on analysis and modeling of geodetic observations in California using UAVSAR, GPS, and Stereo Photogrammetry.
	Chris Milliner is an NPP NASA postdoctoral fellow using space-based geodesy (SAR interferometry, lidar, and optical pixel tracking), GPS, and numerical inverse methods, to constrain how faults release strain throughout the earthquake cycle, and to track large water variations following hurricanes.

Acknowledgements

The research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004). We thank our SCEC colleagues for collaborating in this research.