Greg Beroza, Wayne Loel Professor of Geophysics at Stanford, and Co-Director of SCEC, has been awarded the prestigious Humboldt Research Award (also known as the Humboldt Prize). The award is granted each year by the Alexander von Humboldt Foundation to recognize up to 100 researchers of all disciplines from around the world, whose discoveries or insights have had a lasting impact. For Greg, this recognition is based on his research and teaching excellence. Congratulations Greg!

Greg has been an active member of the SCEC Community for more than twenty years, and in leadership positions since 2002. He has served in multiple roles: Earthquake Source Physics Group Co-Leader, Chair of the SCEC Science Planning Committee (since 2007), Vice-Chair of the Board of Directors (2002-2006), and Deputy Director/Co-Director since 2007. These varied roles have provided Greg a unique perspective on the evolution of the SCEC Collaboratory and the overall study of earthquakes which informs his vision for the future. He shares his views in response to a few questions.

How and when did you choose seismology as a career?  

I did not follow a direct path to seismology. I decided at a young age to be a science professor but hadn’t chosen a field.  I was initially a chemistry major, but the professors seemed to use calculus without understanding it. That bothered me and led me to physics. I liked physics, but not quantum mechanics, and wanted the opportunity to spend time outdoors. My girlfriend (now my spouse) who I met during my junior year in college told me that her high school English teacher’s husband, who was a geologist, said that there were opportunities in that field. I took three upper division Earth Science classes the next quarter without having taken the prerequisite Geology 1 (I read the book over winter break). I got hooked on seismology because it was interesting, it had the math and physics, and it offered the opportunity to do fieldwork. That summer I wanted to do research but had no support and only a few hundred dollars to my name. So I trained to be a wildfire fighter, and helped Rob McCaffrey with his graduate research while waiting for a fire call.  Before that happened, Karen McNally, who was joining the UCSC faculty, asked if I wanted to do fieldwork in Mexico. I jumped at the opportunity and helped set up and run a 6-station temporary seismic network atop the Guerrero gap, which is now known to host very large slow earthquakes. Karen had faith in me and gave me a chance, despite my lack of qualifications. After her, I have benefitted greatly from other mentors, including my graduate advisor, and former SCEC Director, Tom Jordan as well as Paul Spudich and Bill Ellsworth both of the USGS. [Editor's note: the linked article on the Guerrero gap begins coincidentally with a story about Alexander von Humboldt experiencing an earthquake in what is now Venezuela in 1799.] 

How has the science of seismology changed over the course of your career?

The biggest changes have been enabled by advances in data availability and in computing. I am old enough to have started at the very end of the analog era. As an undergrad I deployed smoked-paper seismographs, measured earthquake arrival times by hand, and located earthquakes using punch cards. Now we’re analyzing millions of earthquakes using machine learning to measure arrival times and we’re locating earthquakes using the cloud. From smoked paper to the cloud, is an immense change, but at its heart, constructing earthquakes catalogs is much the same.

There has been parallel progress in earthquake modeling. I don’t do much of that myself, but I try to follow it closely. When I started my career, earthquakes were portrayed as occurring on planar faults. It was recognized by all that faults weren’t planar, but computational capabilities and even the underlying constructs needed to represent the behavior of realistic faults with complex geometries did not exist. That has changed dramatically in the past decade or two. 

We are also blessed with much more data. Early in my career, every new well-recorded earthquake had some surprising aspect, which I took as an indication of how young the field was. I don’t think that’s so true anymore, but take that as a sign of progress. Perhaps with dense nodal deployments and DAS measurements more surprises will emerge.

What is the focus of your research?

My research has been data-driven and, as is often the case in our field, driven by events (earthquakes). As an undergrad I did network seismology in a seismic gap, as a grad student at MIT my dissertation was on using strong motion data to image the earthquake source, and as a postdoc I worked on measuring free oscillations to detect slow earthquakes. Once I came to Stanford, I got re-interested in network seismology and, in particular, in making precise measurements through cross correlation and empirical Green’s function analysis. Other interests followed with the challenge of understanding slow earthquakes expressed through non-volcanic tremor, and strong ground motion path effects using the weak motion of the ambient field. Most recently I have re-invigorated my work in earthquake monitoring through a combination of data mining and machine learning. I’ve had a lot of interests in earthquake seismology, and that’s been helpful in advising students and postdocs on a broad range of research topics.

What have been SCEC’s main accomplishments during your tenure in SCEC leadership?

There are many: exploring the mechanisms that lead to dynamic weakening, illuminating discrepancies between expected and observed surface faulting, mining seismic wavefields to detect small earthquakes, development of realistic multi-cycle earthquake simulators, developing rigorous testing of earthquake forecasts, and much more.  All are challenging problems and progress has been a combination of reaching an improved understanding and also reaching a greater appreciation that more needs to be understood than was first realized. 

As SCEC Co-Director and Chair of the Science Planning Committe, Greg Beroza provides vision and leadership for the Center’s annual scientific program.

Progress in developing physics-based ground motion forecasting through the CyberShake Project is particularly impressive and emblematic of the SCEC approach. CyberShake requires an integrated understanding of all aspects of plate boundary deformation: ranging from the slow accumulation of strain energy in the Earth’s crust using geodesy to representing the sudden release of that energy using pseudodynamic source approximations to full-waveform seismic tomography for capturing wave propagation effects to structural analysis by earthquake engineers for validating the simulations. CyberShake’s ambition is to improve, and eventually to supplant, empirical approaches to hazard analysis. Hazard characterization is foundational to society’s defense against earthquakes, so this is a big deal. It requires the full spectrum of earthquake disciplines from a diverse team of scientists to succeed.  It also requires computing on an absolutely massive scale. SCEC can undertake CyberShake because we draw on the full spectrum of research expertise represented in the collaboration, because we know how to access and how to use state-of-the-art high-performance computing, and because we have the convening authority to bring scientists and engineers together to solve problems. It’s hard to imagine this happening without SCEC.

Despite progress, there remain many challenges. To realize CyberShake’s full potential, for example, we need to increase the extent, density, and frequency range of ground motion simulations. This will require orders of magnitude more computing power, better information on all aspects of earthquake behavior, and more extensive data to test predictions. Other key challenges arise from the fact that earthquake and ground motion forecasts are intrinsically difficult to test because the probabilities are low, and the time scales are long. We’ve made progress using, for example, fragile geologic features for constraints on ground motion intensities in past earthquakes, but we need to do more.

Another accomplishment to highlight is ShakeOut. ShakeOut started as a community earthquake scenario and preparedness exercise that combined state-of-the-art science with an exceptionally effective outreach program. It began in Southern California, as is appropriate for SCEC, but it went viral, and then it went global. There are now tens of millions of participants from around the world in this annual earthquake preparedness exercise. It serves as a model for outreach and preparedness programs for earthquakes and other hazards, and it wouldn’t have happened without the sustained effort of the SCEC Communication, Education, and Outreach program. [View Greg's interview in an earthquake simulator on ShakeOut Day in 2019]

What new techniques/activities hold promise for the future of seismology? 

Many of SCEC’s notable accomplishments might be described as “model-driven.” Cybershake is a good example, as are earthquake simulators, full-waveform tomography, the Broadband ground motion simulation platform, and pseudodynamic rupture modeling. These are all examples of modeling capabilities that SCEC developed and pursued. While modeling will always be critically important, SCEC is poised for a more data-driven approach to understanding earthquakes. This will come about due to improved technologies, such as better InSAR, widespread lidar, DAS and nodal seismic observations, and improved geochronology. It will be accelerated dramatically by application of state-of-the-art techniques from the rapidly evolving field of data science. This powerful combination has the potential to accelerate progress on understanding earthquakes in the immediate future and beyond. We have exciting times to look forward to in earthquake science, and reaching our full potential will require contributions from a new generation of scientists from diverse backgrounds and with diverse interests.