Earthquake Science Center Seminars

Large damaging earthquakes have traditionally been the focus of earthquake studies, understandably so because they are recorded well by a host of sensors. With recent explosion of the number of seismic stations and computational power, we are able to focus on small and even tiny earthquakes, previously not detectable using conventional methods. Aftershock studies are particularly well suited for taking advantage of this opportunity as fault typically becomes most productive right after a relatively large earthquake. Detail studies of aftershock activities, aided by dense seismic network, detection of tiny earthquakes and precise source locations have revealed a number of interesting facets of earthquake physics and fault architecture. I am going focus on two recent examples -- 2015 Mw 7.8 Gorkha earthquake and 2019 Mw 7.1 Ridgecrest earthquake. I was involved in rapid deployment of temporary dense seismic network immediately following both earthquakes. Gorkha aftershock sequence illuminates faults in the Himalayan arc in unprecedented details. We are able to map out a duplex structure between an upper and lower subhorizontal planes indicating a much complex network of fault compared to a canonical flat-ramp-flat model of the Main Himalayan Thrust. In addition, fault structure and its complexities appear to control rupture bifurcation and its subsequent arrest. In case of Ridgecrest aftershock sequence, studies using existing permanent network reveal a number smaller faults including ones orthogonal to the main structure ruptured by the mainshock. Moreover, fault geometry seems to be the key to interaction and triggering of earthquakes in this area. These studies suggest that the details presented by recent aftershock studies are critical to advance our understanding in fault architecture and its implications on earthquake rupture.

Indonesia is one of the most seismically active region in the world, consequently, hazards associated with earthquakes is common and threaten the population. There are major faults that are clearly visible in the landscape and are well studied but many of the recent devastating earthquakes in Indonesia are associated with unidentified inland faults, especially within the volcanic arc region of the archipelago. Faults in this particular setting mostly buried by thick sediments or soil and their characterization is challenging. In this talk, I will provide an overview of the tectonic setting of Indonesia, provide introduction to the many active faults in the region, as well as insight into the completed and ongoing active fault studies especially in the more challenging areas of Java arc. Some of the highlighted results including localization of active tectonic regime through seismic strain analysis, as well as paleoseismology and tectonic geomorphology analysis.

Because seismogenic faults are typically buried at great depth, high spatial resolution observations of interactions between slow slip, fast slip, and locked asperities are limited. The Whillans Ice Plain is a large (~100x150 km2), thin (< 1km), stick-slipping glacier in West Antarctica where these interactions are much easier to observe. I will present analysis of many repeat glacier stick-slip events, or ice stream earthquakes, which I show are usually triggered by periods of accelerated slow slip in creeping regions surrounding a primary locked asperity. Low effective stresses make this system a valuable natural analog to tectonic faults by magnifying and decelerating slip processes before and during daily fault-wide ice stream earthquakes, in which the ice slides ~0.5-1 m in ~25 min. Kinematic GPS and broadband seismometers on the ice surface readily distinguish spatial patterns of background locking and creep, revealing a large, locked, and isolated central asperity, heterogeneous creep in surrounding un-locked regions, and abundant evidence for migratory nucleation of ice stream earthquakes.

I use data from 9 GPS and 76 ice stream earthquakes in 2014 to investigate timing of failure of the locked central asperity in relation to tidally-controlled heterogeneous slip rates in surrounding creeping regions. I find that the main ice stream earthquake, in which the central asperity fails and the entire fault ruptures, is slip-predictable and likely involves complete stress drop, with total slip, peak velocity, and peak acceleration all scaling with elapsed time since the previous event and therefore with accumulated stress from an approximately constant loading. However, event recurrence times range from 8-28 hrs, and for >80% of events, GPS reveal that failure of the central asperity is triggered or nucleated within minutes to several hours by migratory accelerated slow slip caused by falling tide in a region between the asperity and the floating ice shelf. In this simple system, ice stream earthquake moment depends on large scale stresses on the asperity from constant upstream loading, while timing depends on local stress concentration from heterogeneous slip in surrounding regions. These results are relevant for understanding interactions between locked and slowly slipping regions on tectonic faults.

Slip on faults occurs as a spectrum ranging from continuous sliding, where motion is too slow to radiate seismic waves (aseismic creep), to rapid movement during earthquakes (seismic slip). Within the fault slip spectrum lies the transient release of stress through aseismic slip, commonly referred to as slow slip events (SSEs). SSEs are too slow to generate seismic waves so we generally rely on surface geodetic systems (e.g., cGPS, strainmeters, creepmeters, InSAR) to detect transient deformation at depth. However, SSEs located offshore or small aseismic slip transients at depth often generate surface deformation signals beneath instrumental noise levels. To overcome this observational limitation, we use related microseismicity, such as tremor, low-frequency earthquakes (LFEs), and repeating earthquakes, to infer the location of aseismic slip. Tremor is thought to be made up of repeating LFEs that are driven to failure by aseismic slip on the surrounding fault. LFEs are grouped into families that repeatedly rupture a small patch, with individual LFEs likely rupturing multiple sub-patches within the family patch. We propose that similar to families of LFEs, sequences of near-repeating earthquake families (i.e., highly-correlated earthquakes rupturing closely-spaced or partially overlapping asperities) concentrated in time can similarly be indicators of transient aseismic slip and provide another way to indirectly detect and locate aseismic slip transients. Using template matching and land and ocean bottom seismic data, we identify families of near-repeating earthquakes related to two known SSEs at the northern Hikurangi Subduction Margin, New Zealand and four triggered aseismic slip transients at the Anza segment of the strike-slip San Jacinto fault, southern California. These findings indicate that families of near-repeating earthquakes are a reliable proxy for transient aseismic slip, similar to families of LFEs. This is particularly important for determining the full interplay of seismic and aseismic slip in areas that are not tremorgenic or well-instrumented geodetically.

With ever-improving geophysical observations, how do we better understand large earthquakes in the past and foretell their potential behavior in the future? Unraveling the connections between small and large earthquakes across timescales is the key to addressing this question. I will use both physics-based and observation-driven approaches to explore how small earthquakes are influenced by---and hence hold clues to---large earthquake processes and fault zone conditions across seismogenic depths. I will first use laboratory-based fault zone models, along with geophysical observations, to argue that the microseismicity patterns at greater depths indicate the depth extent of large historical earthquakes on major faults in Southern California. The seismic-aseismic interactions offer key constraints on the dynamic rheology of the brittle-ductile transition. I will then use physical considerations and seismogeodetic data to explore the early postseismic phase of the 2004 Mw 6.0 Parkfield earthquake. I will show that early afterslip on fault segments evolved in a complex, multistage manner, and likely controlled aftershock production at fine scales. The new observations shed light on the post-rupture dynamics of the 3D shallower fault zone and suggest different rheological behavior of the deeper fault zone.

In this talk, Joanna L. Dyl, Ph.D., will discuss recurrent patterns across the history of major earthquakes in the San Francisco Bay Area, focusing on social impacts and effects on the built environment. Two key trends have been a consistent refusal to implement measures to reduce risk from future seismic activity and the unequal impact of earthquakes on marginalized social groups, which has been exacerbated during recovery and rebuilding. These patterns emphasize the importance of thinking of earthquakes not as discrete and unprecedented events but as integral components of local histories and local environments. Dyl will draw on her research from her book, Seismic City: An Environmental History of San Francisco's 1906 Earthquake (University of Washington Press, 2017), as well as ongoing research for an article on disaster history and environmental risk in the Bay Area.

We continue to gradually improve upon the current suite of near-real-time USGS earthquake information systems and products that provide estimates of the shaking distribution, casualties, and economic impacts. These products are, of course, aimed at improving situational awareness, targeted response, disaster financing, and earthquake-engineering forensics. Yet, intriguing technological advances on the horizon might allow us to significantly improve upon our current post-earthquake information content.

Advancements in remote sensing, rapid in-situ impact reporting, and machine learning---combined with new datasets such as global building footprints---allow for innovative data-fusion strategies that integrate with existing models and should greatly improve the accuracy and spatial resolution of our shaking and loss estimates. Essentially, we are trying to fill gaps in our current analyses including (1) the lack of direct incorporation of the contributions of ground failure to loss estimates (currently shaking-based only), (2) variable quality and availability of worldwide building inventories and fragilities, and (3) the inability to rapidly update loss estimates despite readily available "ground-truth" observations.

Our vision is a comprehensive, integrated loss-estimation system that incorporates ground-truth observations and imagery, and computes increasingly accurate losses that converge with reality over time. The next generation PAGER system "g-2 PAGER" will assimilate currently disparate yet complementary efforts that will ultimately result in a more fully integrated, spatially accurate, series of PAGER-related products. My presentation will provide an overview of the status of and science behind these efforts, and outline both the challenges and the strategies we plan to employ to contend with them.

Identifying and characterizing cryptic, distributed faulting in high-relief topography can be challenging due to vegetative cover and high rates of hillslope erosion that may mask surficial evidence of active faulting. A multifaceted approach to interrogate the landscape at different spatial and temporal scales commonly illuminates regions of active or recent deformation and helps to characterize faults for seismic hazard analysis in low-strain tectonic settings. This talk will focus on the Pit River region in northeastern California between Mount Shasta and Lassen Peak. In this region, the tectonic regimes of Basin and Range extension, Cascadia subduction zone convergence, and Walker Lane dextral shear converge at the newly characterized transtensional Pondosa fault zone. The Pondosa fault zone is located north of the previously recognized Hat Creek and Rocky Ledge fault systems. We characterize the Pondosa fault zone via analysis of high-resolution lidar topographic data combined with bedrock mapping, 40Ar/39Ar geochronology, and river network analysis with existing geologic and fault mapping. This fault system trends WNW-ESE and disrupts the regional N-S structural fabric. Faults vertically offset lava flows as old ~9.6 Ma and as young as ~12.4 ka, indicating late Quaternary fault activity despite a lack of seismicity along many of the faults. Most of these faults are not currently included in active fault databases, and mapping and understanding the extent, rate, and style of deformation on these faults is important because of their proximity to critical facilities and hydropower infrastructure in addition to interpreting regional tectonic processes. We also discuss the possible role of the Pondosa fault zone in accommodation of relative motion between the Sierran-Oregon Coastal Block microplates.

In this talk, a framework for modeling the effects of disasters on households' well-being will be presented. The framework extends traditional seismic risk analysis to evaluate changes in the output of economic sectors and the effect of losses on the consumption of individual households. The model considers post-disaster changes in utility of consumption (i.e., well-being losses) as a metric to better capture the impact of disasters on individuals, allowing for a more fair comparison between different socio-economic groups. A case study of the San Francisco Bay Area following a hypothetical magnitude 7.2 earthquake on the Hayward fault will be presented. In addition, an adapted version of the framework that evaluates the impacts of COVID-19 measures on the Bay Area households will be shown.

In some subduction zones, the plate boundary fault slips at rates faster than tectonic plate rates, but orders of magnitude slower than typical earthquakes. These 'slow slip events' have been particularly well-documented along subduction plate boundaries, where they are observed at the downdip and updip limits of the earthquake-producing seismogenic zone. A number of plausible hypotheses have been proposed to explain their occurrence; however, that there is no agreed upon cause of these events suggests our understanding of the physics, rock properties, and/or physical conditions along the subduction plate boundary are incomplete. Here, I will summarize some of the factors limiting our understanding of the phenomenon of slow slip, and present results from ongoing studies in our group to constrain the geology, physical conditions, and rock deformation processes in the slow slip environment.