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!

Why do some faults appear to slip backwards?

Researchers from SIO and SDSU surveying fractures surrounding the 2019 Ridgecrest Earthquakes. Left to right: Brook Tozer, Andrea Fabbrizzi, Drake Singleton, Yao Yu, and David Sandwell. Photo courtesy of Scripps Institution of Oceanography at UCSD.

Rarely can a two-day group field trip provide new information about the physics of earthquakes. Guided by satellite measurements of fault deformation, SCEC researchers from Scripps Institution of Oceanography and San Diego State University took to the field and found evidence for a fault damage zone that is critical to the understanding of how faults can appear to slip backward.

A wealth of remote sensing data, collected before and after the 2019 Ridgecrest earthquakes, revealed hundreds of small-offset fractures surrounding the main ruptures (see video). Most of these fractures appear to slip in the direction of the prevailing tectonic stress (prograde, or in the direction of the fault rupture). However, many also appeared to slip backwards (retrograde) as reported in Science by Xu et al. (2020).

A large earthquake rupture close to a pre-existing fault can produce a stress change that is opposite to the prevailing tectonic stress (retrograde).

This retrograde slip behavior was also observed along several pre-existing faults surrounding the 1999 Hector Mine earthquake. Fialko et al. (2002) studied the small deformations surrounding the faults in Hector Mine region seen in ERS-2 interferograms. They were puzzled by the apparent retrograde slip since the magnitude of the stress change from the earthquake was far less than the prevailing tectonic stress. They proposed an elegant solution whereby the strain, caused by the main rupture, is amplified in the damage zone of pre-existing faults. In areas of retrograde stress change, both the fault and the surrounding crust have a similar stress. However, the lower shear modulus in the damaged zone produces a retrograde strain concentration resulting in retrograde deformation in the phase gradient maps.

The newer phase gradient maps provided by the twin Sentinel-1A/B synthetic aperture radar satellites, which revealed hundreds of very small fractures in the Ridgecrest area, suggested they were activated by stress changes from the main ruptures. These satellite observations prompted us to ask: Can the sense of offset of retrograde fractures be mapped in the field? Are these fractures new or reactivated older faults? The Fialko et al. (2002) model could explain the retrograde fractures surrounding the Ridgecrest ruptures—provided these fractures are associated with pre-existing faults with significant damage zones. 

To answer these questions, we needed a ground survey. And we needed trained geologists to help search for evidence of offset prior to the 2019 Ridgecrest earthquakes. With SCEC and NASA funding, SIO and SDSU hosted a joint two-day field trip to the Ridgecrest Earthquake area to discover if the small fractures were related to any pre-existing faults. 

Left-lateral offset of a vertical dike in the location of retrograde fracture seen in satellite image data provides evidence for a pre-existing fault.

Following the 2019 Ridgecrest earthquakes, dozens of field geologists—guided by the remote sensing data from spacecraft and aircraft—made detailed measurements of the main rupture and several subsidiary ruptures (Ponti et al., 2020) where the offsets were greater than a few centimeters. They did not observe the small fractures seen in the phase gradient maps because these fracture offsets are typically less than 15 mm and the deformation spread over several hundred meters (Xu et al., 2020). Only a few very minor cracks were found and their sense of offset (prograde or retrograde) could not be determined. Therefore, deformation from the retrograde fractures is usually too small to be observed in the field.

Guided by the phase gradient maps, the group from SIO and SDSU was able to complete the mapping of a relatively large offset fracture that was partly mapped by the prior field campaigns. More importantly, they documented several meters of offset of a dike crossing one of the areas of retrograde strain. As shown in the image, the left-lateral offset of the vertical dike in the location of retrograde fracture provides direct evidence of a pre-existing fault. Questions remain whether that fault has a wide enough damage zone to explain the observations from satellite imaging. After COVID quarantine is lifted, geologists will need to return to the field and search for additional evidence of pre-existing damage zones. Knowing this is critical for accepting or rejecting the prevailing hypothesis for the apparent retrograde slip.

About the Authors

David Sandwell is a professor of geophysics at Scripps Institution of Oceanography, where he studies seafloor tectonics and crustal deformation using the tools of satellite geodesy. He has been a long-time member of the SCEC community and enjoys field work in Southern California and Northern Baja Mexico.
Xiaohua (Eric) Xu is a postdoctoral researcher at Scripps Institution of Oceanography, where he studies crustal deformation with InSAR, including imaging earthquakes, understanding strain and moment accumulation, crustal deformation and plate tectonics. He is one of the developers of the open source InSAR processing software GMTSAR.
Thomas Rockwell is a professor in the Department of Geological Sciences at San Diego State University, specializing in paleoseismology and neotectonics. He serves as an expert for local and global consulting companies on earthquake issues related to specific construction or modernization sites. He has been a member of the SCEC Board of Directors since 2015 and served on the SCEC Science Planning Committee before.


This research was supported by the Southern California Earthquake Center (Contribution No. 10864). SCEC is funded by NSF Cooperative Agreement EAR-1600087 and USGS Cooperative Agreement G17AC00047. Additional support was provided by NASA Grant 80NSSC19K0739. We thank our SIO, SDSU, NASA and other SCEC colleagues for collaborating in this research.