Earthquake Science Center Seminars

For much of Oklahoma, augmentation of the seismic network with new public stations in the activated areas has followed rather than preceded the spread of seismicity across the state, and consequently the network geometry is often unfavorable for resolving the underlying fault structures. With this study, we reanalyze the existing earthquake catalog with additional data from two industry operated networks for the period May 2013 to November 2016. These networks include 40 seismic stations and cover seismically active north-central Oklahoma with a station spacing on the order of 25 km. Relative locations obtained from waveform cross-correlation reveal a striking pattern of seismicity illuminating many previously unmapped faults. Absolute depths are usually well constrained to within 1 km. Relative locations provide about one order of magnitude better precision for resolving the structure of seismicity clusters. Relocated epicenters tend to cluster in linear trends of less than 1 km to more than 20 km in length. In areas with stations closer than about 10 km, we can resolve fault planes not only by strike but also by dip. These are generally in agreement with surface wave-derived moment tensor solutions.

We identify 93 sequences with at least 30 events that we use for detailed analysis of the spatio-temporal evolution and clustering of these sequences. For most awakened faults, seismicity tends to initiate at shallower depth and migrates deeper along the faults as the sequence proceeds. We find that no sequence starts with the largest earthquake and many sequences initiate months before they rise to peak activity. We study temporal clustering as a means to quantify earthquake interactions. Some sequences show no temporal clustering similar to Poissonian background seismicity but at much higher rate. Other sequences exhibit strong temporal clustering akin of mainshock-aftershock sequences.

We conclude that once initiated by anthropogenic stressing, faults in the Oklahoma/Kansas area are close enough to failure to sustain rupturing through earthquake interaction. Furthermore, we show that the largest earthquakes occur as a result of continued stressing, rather than by growing the reactivated part of a pre-existing fault. Therefore, improve monitoring can provide us with warning time to actively mitigate the occurrence of larger earthquakes.

Seismic data recorded by surface arrays are often contaminated by unwanted noise. In many conventional seismic methods, the reliability of the seismic data and accuracy of parameter extraction, such as onset time, polarity, and amplitude, are directly affected by the background noise level. As a result, the accuracy of event location and other attributes derived from seismic traces are also influenced by the noise content. Therefore, there is a great need for developing suitable procedures that improve signal-to-noise ratios allowing for robust seismic processing. In this presentation, I introduce four different methods for automatic denoising of seismic data. These methods are based on the time-frequency thresholding approach. The efficiency and performance of the thresholding-based method for seismic data have been improved significantly. Proposed methods are automatic and data driven in the sense that all the filter parameters for denoising are dynamically adjusted to the characteristics of the signal and noise. These algorithms are applied to single channel data analysis and do not require large arrays of seismometers or coherency of arrivals across an array. Hence, they can be applied to every type of seismic data and can be combined with other array based methods. Results show these methods can improve detection of small magnitude events and accuracy of arrival time picking.

Note: Due to technical issues, the speaker's slides are difficult to see clearly. You can download the speaker's slide deck (Microsoft PowerPoint .pptx) and view it on your computer.

Estimating the expected size of the largest earthquake on a given fault is complicated by dynamic rupture interactions in addition to geometric and stress heterogeneity. However, a statistical assessment of the potential of seismic events to grow to larger sizes may be possible based on variations in magnitude distributions. Such

variations can be described by the b-value, which quantifies the proportion of small- to large-magnitude events. The values of b vary significantly if stress changes are large, but additional factors such as geometric heterogeneity may affect the growth-potential of seismic ruptures. Here, we examine the influence of fault roughness on b-values, focal mechanisms and spatial localization of laboratory acoustic emission (AE) events during stick-slip experiments. We create three types of roughness on Westerly granite surfaces and study AE event statistics during triaxial loading of the lab-faults. Since both roughness and stress variations are expected to influence b, we isolate roughness contributions by analyzing AEs at elevated stresses close to stick-slip failure. Our results suggest three characteristics of seismicity on increasingly rough faults: (1) seismicity is spatially more distributed, (2) b-values are higher, and (3) focal mechanisms are more heterogeneous, likely caused by underlying stress field heterogeneity within the fault zones. Localized deformation on smooth faults, on the other hand, promotes larger rupture sizes within the associated homogeneous stress field which is aligned with the macroscopic stress orientation. The statistics of earthquake magnitude distributions may help quantify these fault states and expected rupture sizes in nature.

Current models for dynamic rupture have led to much insight into earthquake propagation, the generation of high-frequency ground motion, and the influence of plasticity on rupture propagation. Although the inclusion of a plastic material response has been shown to reduce stress and slip velocities at the rupture front to reasonable values, little work has been done to understand the evolution of a damage zone (and its impact on rupture) over multiple event sequences. In particular, most dynamic rupture models currently make the assumption of a uniform background stress and are limited to single-event simulations. In this talk I will share the method we have developed for earthquake cycle simulations within heterogeneous, inelastic material, where ruptures nucleate under self-consistent initial conditions. The method is developed for the classical antiplane problem of a vertical strike-slip fault governed by rate-and-state friction, with inertial effects captured through the radiation-damping approximation. Both rate-independent plasticity and viscoplasticity are considered, where stresses are constrained by a Drucker-Prager yield condition. The off-fault volume is discretized using finite differences and tectonic loading is imposed by displacing the remote side boundaries at a constant rate. We quantify how viscosity, isotropic hardening, and cohesion affect the magnitude and off-fault extent of plastic strain that develops over many ruptures. In all cases, coseismic slip in the shallow sub-surface is diminished compared to slip accumulated at depth during interseismic loading. The evolution of this slip deficit with each subsequent event, however, is dictated by the plasticity model. Integration of the off-fault plastic strain from the viscoplastic model reveals that a significant amount of tectonic offset is accommodated by inelastic deformation (about 10% of the tectonic deformation budget).

Repeating earthquakes are a nearly ubiquitous feature of the seismicity at active and erupting volcanoes worldwide. A seismic source is persistently reactivated, producing nearly identical waveforms that can be identified and clustered by cross-correlation methods. Changes in the timing, location, and waveforms of these earthquakes can be informative about ongoing and evolving processes within a volcano. In this talk, I'll highlight several different examples of the utility of repeating earthquakes to understand a "screaming" volcano in Alaska, changes in the shallow velocity structure of Mount St. Helens, and swarms of deep migrating earthquakes under Mammoth Mountain.

The goal of earthquake warning is ground-motion early warning: to predict expected ground shaking intensity so that action can be taken at various locations to prepare for the imminent ground motion. The usefulness of early warning depends both on the accuracy of the ground motion prediction and the timeliness of the warning. (We take timeliness to be the time difference between when the warning is received and when the expected ground motion arrives at the user's location, minus the time needed for users to take precautionary measures.) The time required to issue a warning depends on several factors including the telemetry latency associated with receiving real-time data, the computational time needed to analyze those data, and, most crucially, the time needed to characterize the earthquake source in order to make accurate ground motion predictions. In this study we explore, both theoretically and observationally, the conditions necessary to make an accurate ground motion prediction and the temporal evolution of the expected accuracy of early-warning systems. We consider the limits of rapid magnitude estimation given the non-deterministic nature of earthquake rupture and the band-limited, Gaussian white noise properties of high-frequency (i.e., PGA) ground motion. We also study the best-case warning time scenario given the limits on magnitude estimation and the finiteness of large faults. Finally, we consider the accuracy and effectiveness of real-time ground-motion prediction given the uncertainties and limitations in magnitude estimation and the range of potential warning times.

Myanmar is in a tectonically complex region between the eastern edge of the Himalayan collision zone and the northern end of the Sunda megathrust. Until recently, earthquake monitoring and research efforts have been hampered by a lack of modern instrumentation and communication infrastructure. In January 2016, a major upgrade of the Myanmar National Seismic Network (MNSN; network code MM) was undertaken to improve earthquake monitoring capability. Funded by USAID and OFDA, the USGS installed five permanent broadband and strong-motion seismic stations with real-time data telemetry using newly improved cellular networks. Data are telemetered to the MNSN hub in Nay Pyi Taw and archived at the IRIS DMC. I will present preliminary results from analysis of station noise characteristics and site response. Background noise characteristics vary across the array, but indicate that the new stations are performing well.

In their first six months of operation, MM stations recorded more than 20 earthquakes of M > 4.5 within Myanmar and its immediate surroundings, including an Mw 6.8 earthquake located northwest of Mandalay on 13 April 2016 and the Mw 6.8 Chauk earthquake on 24 August 2016. We use this new dataset to calculate horizontal-to-vertical spectral ratios, which provide a preliminary characterization of site response of the upgraded MM stations.