SCEC Award Number 20097 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Understanding the spatial variation of high frequency radiation from earthquakes in Southern California
Investigator(s)
Name Organization
Rachel Abercrombie Boston University Peter Shearer University of California, San Diego
Other Participants 1 UCSD graduate student (TBN)
SCEC Priorities 1d, 2d, 4a SCEC Groups Seismology, FARM, GM
Report Due Date 03/15/2021 Date Report Submitted 03/11/2021
Project Abstract
INTERIM REPORT: Earthquakes of similar moment are often observed to radiate very different amounts of high-frequency radiation, which limits the accuracy of ground motion prediction equations (GMPEs). These variations are commonly modeled in terms of earthquake stress drop, a fundamental source parameter implicit in many of the science goals of SCEC 5, but hard to measure reliably and well. For several years, the PIs have investigated the origins of the large uncertainties and scatter in stress-drop estimates by comparing different approaches to analysis of P-wave spectra from small to moderate earthquakes in Southern California. We have gained an understanding of the limitations of the methods and how they can be improved. One of our insights is that stress-drop estimates rely heavily on certain modeling assumptions, which are difficult to verify from observations even with high-quality data. Thus, we favor measures of high-frequency radiation that are closer to the data and rely less on specific theoretical rupture models or poorly-constrained empirical Green’s functions (EGFs).
Our work for this award has focused on further method development, and uncertainty analysis as we work towards our longer term goal of improved quantification of the high-frequency earthquake radiation: (1) Using data from a deep borehole in the Cajon Pass “Earthquake Gate” region where data from below the highly-attenuating near-surface layers provides “ground truth” checks on assumed path corrections, and (2) combining the results from multiple previous spectral-decomposition studies and using a spectral-ratio method to distinguish between depth-dependent source and path effects.
Intellectual Merit Earthquake stress drop is a fundamental source parameter, implicit in many of the science goals of SCEC 5. Stress drops are now commonly estimated from seismic data, but are hard to measure reliably and well. The large uncertainties and scatter in results affect strong ground motion prediction, and also limit our understanding of the physics of the earthquake rupture process, including distinguishing induced seismicity from natural seismicity. We are working to combine different approaches, and data to improve our quantification of the high-frequency source radiation. Our current focus is to distinguish more reliably between source and path effects, in obtaining absolute stress drop values, and their dependence on depth. Accomplishments include:
• Applying spectral decomposition to a cluster of 1787 earthquakes of the 1992 to 1993 Big Bear aftershock sequence in southern California, which were recorded by both the Southern California Seismic Network (SCSN) and the Cajon Pass borehole seismometer at 2.5-km depth. The deep borehole records are less affected by site and attenuation effects than surface seismometers and thus provide more direct constraints on earthquake source spectra. We use the Cajon borehole data to provide absolute calibration for the SCSN spectral decomposition results, which can then be used to process all the Big Bear earthquakes and estimate earthquake source properties without the parameter tradeoffs normally associated with purely empirical approaches.
• Analysis of over 50,000 earthquake event spectra from previous studies in multiple regions to investigate whether reported depth dependence in stress drop is real. A carefully designed spectral ratio analysis shows that a significant part of some previously reported depth variation is actually an artefact of the over-simplified corrections for depth-dependent attenuation. We are now working to incorporate these findings to improve future applications of the powerful spectral decomposition approach.
Broader Impacts The aim of our collaborative work is to improve the quality and reliability of earthquake stress drop estimates, which are fundamental to studies of earthquake physics, and also seismic hazard prediction. We have presented our results at SCEC (SCEC Contributions #10725, #10662), and also have a peer-reviewed publication in Journal of Geophysical Research – Solid Earth (SCEC Contribution 10845), and a second in preparation. The results of our analyses are already being implemented as improvements to a number of ongoing analyses, because both PIs are collaborating with multiple groups and analyzing earthquakes in many parts of the world (e.g., Pennington et al., 2021; Zhang et al., 2021; Chen and Abercrombie, 2020). This award also provided partial support to PI Abercrombie for her Invited Review paper for the forthcoming Philosophical Transactions of the Royal Society Special Volume on Fracture Mechanics.
Results of both parts of analysis are promoting future progress by Shearer for his USGS grant: Developing a new unified approach for analysis of earthquake source spectra in Southern California, with the ultimate goal of redoing the Shearer et al. (2006) study of the entire region, but for an expanded time period and testing for consistency with existing 1D and 3D Q models.
Our work therefore meets the both the specific and broader goals of SCEC, to use Southern California as a natural laboratory to improve our understanding of earthquake processes in Southern California and beyond.
Exemplary Figure Figure 4: Effect of correction using depth-specific EGF events for all study regions. The ratios of corner frequencies between the deep and shallow earthquakes calculated using all approaches are plotted against the ratio of the average P-wave velocity at the depths of the respective deep and shallow event populations included. Color is by data set, with alternating symbols to ease distinction. Multiple symbols per data set represent the different magnitude bins, and also the use of the alternative velocity models – their variability represents uncertainties. Left panel is without correction for depth dependent attenuation and right is with correction for depth dependent attenuation. The black dashed line represents a constant stress drop with depth, if rupture velocity is proportional to P-wave velocity. The correction for depth-dependent attenuation, using depth-specific EGFs, is able to remove much of the previously reported depth dependence of stress drop, suggesting it is an artifact.

Abercrombie et al. (2021), in prep for JGR.