Earthquake Science Center Seminars

The Costa Rican convergent margin is widely considered an 'erosive-type' convergent margin. Its forearc displays several key proxies thought to be linked to basal erosion, including continental basement extending to near the trench, a record of outer forearc slope subsidence and trench embayment. However, recently collected 3D seismic reflection and drilling data offshore the Osa Peninsula reveal forearc structure and composition that conflicts with the forearc basal erosion model. This talk utilizes these new 3D seismic reflection and drilling data to explore Pleistocene shortening, vertical motion and an in-situ mass budget that challenges the basal erosion model, suggesting that proxies used to characterize convergent margins as erosional may need to be revisited.

After an earthquake, individual buildings may be unusable for a long time while waiting for repairs. However, Christchurch, New Zealand's earthquake in 2011 demonstrated that even undamaged buildings could experience downtime if there is concern that nearby, heavily damaged buildings may collapse in an aftershock. This talk will present a framework for assessing the additional downtime induced by the access restrictions associated with safety barricades (or "cordons") around damaged tall buildings. Downtown San Francisco will be presented as a case study for how to collect and incorporate real-world data at scale. Though still preliminary, this example demonstrates what is possible and how the results could inform policy decisions that focus on community resilience.

Recent earthquakes in Hokkaido Japan (M6.6, Sept 5) and Anchorage, AK (M7.0, Nov 30) triggered landslides and lateral spreads of radically different severity and consequences. In Hokkaido, over 6,000 landslides have been mapped to date from the highly destructive Sept. 5 M6.6 earthquake, resulting in damages to ~9,000 residential dwellings. Conversely, ground failures and damage to property appear to be limited from the Nov. 30 M7.0 Anchorage earthquake, where widespread reports of two road damaging landslides reflect exceptional ground-failures from otherwise muted consequences. In this talk I will discuss observations from reconnaissance efforts for both earthquakes focused on cataloging and digital preservation of ground failures and ground truthing of the USGS Ground-Failure product. I will also contrast coseismic ground failures triggered by these earthquakes and the potential trends they represent.

We reexamine the geometry of the causative fault structure of the 1989 moment-magnitude-6.9 Loma Prieta earthquake in central California, using seismic-reflection, earthquake-hypocenter, and magnetic data. Our study is prompted by recent interpretations of a two-part dip of the San Andreas Fault (SAF) accompanied by a flower-like structure in the Coachella Valley, in southern California. Initially, the prevailing interpretation of fault geometry in the vicinity of the Loma Prieta earthquake was that the mainshock did not rupture the SAF, but rather a secondary fault within the SAF system, because network locations of aftershocks defined neither a vertical plane nor a fault plane that projected to the surface trace of the SAF. Subsequent waveform cross-correlation and double-difference relocations of Loma Prieta aftershocks appear to have clarified the fault geometry somewhat, with steeply dipping faults in the upper crust possibly connecting to the more moderately southwest-dipping mainshock rupture in the middle crust. Examination of steep-reflection data, extracted from a 1991 seismic-refraction profile through the Loma Prieta area, reveals three robust fault-like features that agree approximately in geometry with the clusters of upper-crustal relocated aftershocks. The subsurface geometry of the San Andreas, Sargent, and Berrocal Faults can be mapped using these features and the aftershock clusters. The San Andreas and Sargent Faults appear to dip northeastward in the uppermost crust and change dip continuously toward the southwest with depth. Previous models of gravity and magnetic data on profiles through the aftershock region also define a steeply dipping SAF, with an initial northeastward dip in the uppermost crust that changes with depth. At a depth 6 to 9 km, upper-crustal faults appear to project into the moderately southwest-dipping, planar mainshock rupture. The change to a planar dipping rupture at 6-9 km is similar to fault geometry seen in the Coachella Valley.

Ground deformation is usually a hydrologic response to groundwater processes, expressing its diversity in the length and rate in time, as well as the magnitude and distribution in space. Satellite radar interferometry allows us to resolve the deformation with mm-level accuracy in a range of hundreds of kilometers per acquisition, and the repeatability of several to tens of days, which effectively advance the incorporation with the available hydrological data sets such as precipitation and water levels. In this talk, I will focus on a) some hydrologic-triggered landsliding systems in western United Stated. Our results reveal the complex three-dimensional configurations of the landslide mass, how the time variable motion relates to rainfall, how the surface velocity during sliding changes in space, and how the ground surface subsides slightly in weeks prior to downslope sliding in the case of Crescent landslide, WA. And b), the aquifer systems underneath the metropolitan Salt Lake City, Utah. Our results show the time-dependent ground deformation highlighted by a long-term uplift and seasonal variations that can be correlated with water level changes. Based on that, we investigate the sensitivity of the aquifer skeleton in response to the persistent groundwater recharge decades ago, the stain rate of the vertical shearing groundwater reservoir, the amount of water stored in the confined aquifers, and the basin-wide hydraulic diffusivity. Anthropogenic land subsidence has been also characterized using statistical and geoengineering models.

Presentation slides

The Cascadia Subduction Zone (CSZ) can produce long-duration, large-magnitude earthquakes, whose ground motions will be modified by the deep sedimentary basins, which underlie several cities in the Pacific Northwest (e.g., Portland, Seattle, and Vancouver, BC). The effects of these basins on the ground motion duration and frequency content are poorly understood because no recordings are available for large-magnitude earthquakes in this region.

To compensate for this paucity of recordings, researchers from the United States Geological Survey and the University of Washington generated ground motions for numerous scenarios of an M9 event. The simulations were generated using deterministic, finite-difference (T>1s) and stochastic (T<1s) approaches. This seminar will present the impacts of 30 motions simulated in Seattle for 32 archetypical, reinforced-concrete-core-wall structures (4- to 40-stories).

The frequency dependent basin amplification increases the spectral accelerations at periods corresponding to tall structures and results in damaging spectral shapes. The duration of the M9 motions are long, but the basins have little effect on the significant duration of the motions. Tall buildings designed to enhanced building code requirements have collapse probabilities similar to that expected for a 2475-year motion, but the return period of an M9 earthquake is only about 500 years. Tall buildings designed to meet the minimum building code requirements were much more likely to collapse. A newly developed ground-motion intensity measure (effective spectral acceleration) explains these results and made it possible to evaluate the likely regional impacts of these motions.