Wei-An Chao

Seismologically determined bedload flux during the typhoon season

Continuous seismic records near river channels can be used to quantify the energy induced by river sediment transport. During the 2011 typhoon season, we deployed a seismic array along the Chishan River in the mountain area of southern Taiwan, where there is strong variability in water discharge and high sedimentation rates. We observe hysteresis in the high-frequency (5–15 Hz) seismic noise level relative to the associated hydrological parameters. In addition, our seismic noise analysis reveals an asymmetry and a high coherence in noise cross-correlation functions for several station pairs during the typhoon passage, which corresponds to sediment particles and turbulent flows impacting along the riverbed where the river bends sharply. Based on spectral characteristics of the seismic records, we also detected 20 landslide/debris flow events, which we use to estimate the sediment supply. Comparison of sediment flux between seismologically determined bedload and derived suspended load indicates temporal changes in the sediment flux ratio, which imply a complex transition process from the bedload regime to the suspension regime between typhoon passage and off-typhoon periods. Our study demonstrates the possibility of seismologically monitoring river bedload transport, thus providing valuable additional information for studying fluvial bedrock erosion and mountain landscape evolution.

A first near real-time seismology-based landquake monitoring system

Hazards from gravity-driven instabilities on hillslope (termed ‘landquake’ in this study) are an important problem facing us today. Rapid detection of landquake events is crucial for hazard mitigation and emergency response. Based on the real-time broadband data in Taiwan, we have developed a near real-time landquake monitoring system, which is a fully automatic process based on waveform inversion that yields source information (e.g., location and mechanism) and identifies the landquake source by examining waveform fitness for different types of source mechanisms. This system has been successfully tested offline using seismic records during the passage of the 2009 Typhoon Morakot in Taiwan and has been in online operation during the typhoon season in 2015. In practice, certain levels of station coverage (station gap < 180°), signal-to-noise ratio (SNR ≥ 5.0), and a threshold of event size (volume >106 m3 and area > 0.20 km2) are required to ensure good performance (fitness > 0.6 for successful source identification) of the system, which can be readily implemented in other places in the world with real-time seismic networks and high landquake activities.

Seismology-based early identification of dam-formation landquake events

Flooding resulting from the bursting of dams formed by landquake events such as rock avalanches, landslides and debris flows can lead to serious bank erosion and inundation of populated areas near rivers. Seismic waves can be generated by landquake events which can be described as time-dependent forces (unloading/reloading cycles) acting on the Earth. In this study, we conduct inversions of long-period (LP, period ≥20 s) waveforms for the landquake force histories (LFHs) of ten events, which provide quantitative characterization of the initiation, propagation and termination stages of the slope failures. When the results obtained from LP waveforms are analyzed together with high-frequency (HF, 1–3 Hz) seismic signals, we find a relatively strong late-arriving seismic phase (dubbed Dam-forming phase or D-phase) recorded clearly in the HF waveforms at the closest stations, which potentially marks the time when the collapsed masses sliding into river and perhaps even impacting the topographic barrier on the opposite bank. Consequently, our approach to analyzing the LP and HF waveforms developed in this study has a high potential for identifying five dam-forming landquake events (DFLEs) in near real-time using broadband seismic records, which can provide timely warnings of the impending floods to downstream residents.

Imaging high-pressure rock exhumation in eastern Taiwan

Increasingly detailed studies of out crops of high-pressure rock terranes in combination with rapidly evolving numerical modeling studies have given rise to a number of possible explanations for the processes by which these rocks are exhumed. Imaging actively exhuming high-pressure terranes remains one of the fundamental, but elusive, tasks that could advance the understanding of how these important rocks reach Earth’s surface. Seismic tomography along the active arc-continent collision in eastern Taiwan images a high P- and S-wave velocity zone that extends from the shallow subsurface beneath a high-pressure metamorphic terrane to ∼50 km depth. We present a petrophysical analysis of this high-velocity zone that indicates the presence of rock types common to high-pressure terranes. The high-velocity zone is seismically active throughout. We determine focal mechanisms for 57 earthquakes, and carry out full waveform modeling on 10; these have double-couple focal mechanisms with a compensated linear vector dipole component up to 20.6%. We suggest that the high-velocity zone comprises an exhuming high-pressure terrane. Focal mechanisms for earthquakes within it indicate that shear faulting dominates in the deformation, but high fluid pressure may also play a role.

Centroid Fault-Plane Inversion in Three-Dimensional Velocity Structure Using Strong-Motion Records

The Chia-yi area in southwestern Taiwan is a region of relatively high seismicity. Frequent earthquakes of various magnitudes in the shallow crust provide valuable information for understanding the regional seismotectonic environment. In this study, we use strong-motion records from the 22 October 1999, ML 6.4 Chia-yi earthquake sequence to determine the focal mechanisms based on a criterion combin- ing the P-wave polarity and the cross correlations between recorded and synthetic waveforms. A recent 3D structural model for the study region is used, and to ensure the computational efficiency, a database is established for the Greens functions obtained by the finite-difference method. Optimal focal depths and fault-plane solu- tions are obtained through a grid-search scheme driven by the genetic algorithm to further improve the efficiency. We apply this approach to determine the focal mecha- nisms and centroid depths of the small and moderate events in the Chia-yi earthquake sequence. Results show dominant strike-slip and thrust mechanisms that are in good agreement with previous results based upon P-wave first-motion polarities and moment-tensor inversions. This highly automatic and efficient approach will enable the determination of focal depths and fault-plane solutions immediately following the occurrence of small- to moderate-sized earthquakes.

How Well Can We Extract the Permanent Displacement from Low-Cost MEMS Accelerometers?

Following the recent establishment of a high-density seismic network equipped with low-cost micro-electro-mechanical system (MEMS) P-wave-alert-device (P-Alert) by the earthquake early warning (EEW) research group at the National Taiwan University, a large quantity of strong-motion records from moderate-magnitude earthquakes (ML > 6) around Taiwan has been accumulated. Using a data preprocessing scheme to recover the dynamic average embedded within the P-Alert data, we adopted an automatic baseline correction approach for the P-Alert accelerograms to determine the coseismic deformation (Cd). Comparisons between the Cd values determined using global positioning system (GPS) data, strong-motion records from the P-Alert network, and data from the Taiwan Strong Motion Instrumentation Program (TSMIP) demonstrates that the near-real-time determination of Cd values (>2 cm), which provide crucial information for seismic hazard mitigation, is possible using records from low-cost MEMS accelerometers.

Relationship Between Earthquake b‐Values and Crustal Stresses in a Young Orogenic Belt

It has been reported that earthquake b-values decrease linearly with the differential stresses in the continental crust and subduction zones. Here we report a regression-derived relation between earthquake b-values and crustal stresses using the Anderson fault parameter (Aφ) in a young orogenic belt of Taiwan. This regression relation is well established by using a large and complete earthquake catalog for Taiwan. The data set consists of b-values and Aφ values derived from relocated earthquakes and focal mechanisms, respectively. Our results show that b-values decrease linearly with the Aφ values at crustal depths with a high correlation coefficient of 0.9. Thus, b-values could be used as stress indicators for orogenic belts. However, the state of stress is relatively well correlated with the surface geological setting with respect to earthquake b-values in Taiwan. Temporal variations in the b-value could constitute one of the main reasons for the spatial heterogeneity of b-values. We therefore suggest that b-values could be highly sensitive to temporal stress variations.

A First‐Layered Crustal Velocity Model for the Western Solomon Islands: Inversion of the Measured Group Velocity of Surface Waves Using Ambient Noise

Two earthquakes,Mw 8.1 in 2007 andMw 7.1 in 2010, hit the western province of the Solomon Islands and caused extensive damage, which motivated us to establish a temporary seismic network around the rupture zones of these earthquakes. With the available continuous seismic data recorded from eight seismic stations, we cross correlate the vertical component of ambient-noise records and calculate Rayleigh-wave group velocity dispersion curves for interstation pairs. A genetic algorithm is adopted to fit the averaged dispersion curve and invert a 1D crustal velocity model, which constitutes two layers (upper and lower crust) and a half-space (uppermost mantle). The resulting thickness values for the upper and lower crust are 6.9 and 13.5 km, respectively. The shear-wave velocities (V S) of the upper crust, lower crust, and uppermost mantle are 2.62, 3.54, and 4:10 km=s with VP=V S ratios of 1.745, 1.749, and 1.766, respectively. The differences between the predicted and observed travel times show that our 1D model (WSOLOCrust) has average 0.85and 0.16-s improvements in travel-time residuals compared with the global iasp91 and local CRUST 1.0 models, respectively. This layered crustal velocity model for the western Solomon Islands can be further used as a referenced velocity model to locate earthquake and tremor sources as well as to perform 3D seismic tomography in this region. Electronic Supplement: Figures showing the misfit of inversion process and the comparison between observed and synthetics and the location of experiments in previous studies and tables listing information about the seismic network, parameters of the genetic algorithm (GA), information of earthquakes used in this study, and results obtained from different 1D models. INTRODUCTION The Solomon Islands is located in the southwestern part of the Pacific Ocean. Several tectonic plates, including the Pacific, Australian, and Woodlark plates, subduct beneath the Solomon arc, forming an active subduction zone (Fig. 1). In 2007, an Mw 8.1 earthquake occurred in the western Solomon Islands and ruptured across the Pacific–Australian–Woodlark triple junction (Taylor et al., 2008; Chen et al., 2009; Miyagi et al., 2009). This event generated a hazardous tsunami with a maximum wave height of 12 m that hit the western province of the Solomon Islands, which resulted in 52 deaths and thousands homeless (Fisher et al., 2007; Fritz and Kalligeris, 2008). About 3 yrs later in 2010, a relatively small earthquake with the moment magnitude of 7.1 occurred near the hypocenter of the 2007 earthquake (Newman et al., 2011; Kuo et al., 2016). Despite its size, this event also generated a local tsunami (Newman et al., 2011). Unfortunately, there is a lack of local seismic recording during these two earthquakes. Hence, neither analyzing the source mechanisms of the events in detail nor further developing the tsunami warning system is viable. To understand the seismic activity in the western Solomon Islands, we installed eight broadband seismic stations around the rupture zone of the 2007 earthquake, aiming to provide quantities of records from earthquakes and continuous signals from ambient noise. The velocity structure of neighboring areas has been previously proposed (Cooper, Bruns, et al., 1986; Cooper, Marlow, et al., 1986; Miura, 1998; Shinohara et al., 2003; Miura et al., 2004; Yoneshima et al., 2005); however, there is no available velocity model in our study area. Using a dense seismic network, an Earth structure model can be derived from either the travel-time tomography (e.g., Bording et al., 1987) or the ambient-noise tomography (e.g., Shapiro 2274 Seismological Research Letters Volume 89, Number 6 November/December 2018 doi: 10.1785/0220180126 Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/89/6/2274/4536978/srl-2018126.1.pdf by National Taiwan Univ Lib Serials Dept user on 12 December 2018 et al., 2005; Lin et al., 2007). Because of the large aperture of the station distribution and insufficient stations, we first study a simple 1D velocity structure. We accordingly conducted the genetic algorithm (GA; Holland, 1975) adopted for studying earthquake source mechanisms (e.g., Wu et al., 2008; Chao et al., 2011), to determine a 1D crustal velocity model by minimizing the misfit between observed data and the theoretical dispersion curves. We apply the Computer Programs in Seismology (CPS) package (Herrmann, 2013) to predict the theoretical dispersion curves. The observed dispersion curves herein are derived from the cross-correlograms after applying the multiple filter technique (MFT; Dziewonski et al., 1969), and the averaged dispersion curve is used as the input data for an inversion algorithm. The reliability of the inversion scheme depends on the number of unknown parameters. So, we simplify the velocity model into two layers and a half space to provide a layered velocity model. Because there is no previously published velocity model for the western Solomon Islands, our proposed 1D model is examined by a comparison with the global models iasp91 (Kennett and Engdahl, 1991) and CRUST 1.0 (Laske et al., 2013). To check the deviation in between, the predicted travel time is computed by applying a Python package (Cake; Sebastian et al., 2017) on different 1D models. We select earthquakes those occurred within our study area from theU.S. Geological Survey (USGS) earthquake catalog and pick the first arrival of each event manually to calculate the observed travel time. Thereby, the travel-time residuals between the observed and predicted travel times for each event can be estimated to verify the improvement of our 1D model. The advantage of this study using ambient noise and applying the GA to develop the velocity model is to avoid the trade-off between a velocity model and the hypocenter location. Our new 1D model can hence be a better-reference velocity model for seismic study and further help locate small local earthquakes. Walter et al. (2016) reported the evidence for triggering of tectonic tremor in the western Solomon Islands, indicating slow processes indeed happen in this area. To improve the searching for the triggered tremors, a reliable velocity model is urgently needed. Also, such a model will be essential for further understanding the tectonic details to help seismic hazard mitigation. DATA PROCESSING AND GROUP VELOCITY MEASURMENTS Based on the coverage of the rupture zone observed in the 2007 earthquake, we designed an eight-seismometer network and deployed the instruments in the western Solomon Islands since September 2009 (Fig. 1). The seismic instruments are equipped with the broadband seismometer (Trillium 120PA; Nanometrics Inc., Canada) and the 24-bits digital recorder (Q330S; Quanterra Inc., U.S.A.) with sampling rates of 100 Hz. In this study, the vertical-component continuous seismic data from eight broadband seismic stations are used. Records with time shifting or instrument problems are removed manually. The data lengths from the eight stations are shown in E Table S1 (available in the electronic supplement to this article). The empirical Green’s function between two stations can be estimated from the ambient-noise cross-correlation function (CCF). In the past decades, the above statement has been verified by several studies (Campillo and Paul, 2003; Shapiro and Campillo, 2004; Snieder, 2004; Weaver and Lobkis, 2004; Stehly et al., 2007). Based on the procedure of You et al. (2010), the data processing of continuous records can be summarized as follows: (1) preparing daily records of seismic data for each station; (2) removing the instrument response, mean, and linear trend; (3) applying a bandpass filter with a 2to 50-s period range and decimating the sampling rate to 10Hz; (4) conducting a one-bit normalization scheme (Larose et al., 2004; Shapiro and Campillo, 2004); and (5) computing daily CCFs for each station pair with lag times ranging from −300 to 300 s. To increase the signal-to-noise ratio (SNR) of the CCFs, we stack all possible CCFs for each station pair to compute a stacked CCF (SCCF). Then the group velocity dispersion curves of each SCCF can be measured using the MFT (Dziewonski et al., 1969). For more detailed information about the MFT used in this study, please refer to Corchete et al. (2007).

The Large Greenland Landslide of 2017: Was a Tsunami Warning Possible?

Tsunamis generated by mass movements such as landslides, underwater slumps, and rock avalanches can lead to serious inundation of nearby populated areas. The lack of timely estimations of the moving mass volume, however, makes providing operational early warnings for landslides particularly challenging. In June 2017, a large landslide in Greenland generated tsunami waves of about 1 m high that impacted the small town of Nuugaatsiaq. We show how the seismic analysis of real-time seismic records from the Greenland Ice Sheet Monitoring Network (GLISN) can provide estimates of essential physical properties of the landslide such as collapse mass and sliding velocity shortly after origin time. The estimation of the landslide source parameters can be utilized for tsunamiwave simulations. We demonstrate how the real-time integration of seismic waveform inversion with forward tsunami-wave simulation could have enabled a timely operational warning (about 10 min) before the arrival of the impending tsunami waves at the village of Nuugaatsiaq.

Building Effects on the P‐Alert‐Based Real‐Time Shaking Map Determination

Recently, the P-wave-alert-device (P-alert) network, which is a dense array of microelectromechanical system (MEMS) accelerometers that was developed and installed by National Taiwan University for the purposes of earthquake early warnings, has recorded a large number of strong-motion records for moderate-to-large earthquakes throughout Taiwan. However, many of these stations are mounted on the vertical walls of buildings in ways such that further studies of the sensor-structure interactions on recorded acceleration data are required before the data is used in the production of high-quality shake maps. In this study, we collect the free-field accelerograms recorded by the Taiwan Strong-Motion Instrumentation Program (TSMIP) network that were operated by the Central Weather Bureau (CWB), where MEMS accelerometers were in the vicinity. Then, we compare the peak ground acceleration (PGA) ratio (R-value) between P-alert and TSMIP stations. Finally, we demonstrate how to use the R-value correction on the P-alert data, in order to rapidly produce high-resolution shake maps for relief work to be done soon after major earthquakes. At present, the shake maps produced by the P-alert network are posted automatically in real time on Facebook and are provided to the National Science and Technology Center for Disaster Reduction (NCDR) in order to allow for their relief work. These timely products provide improved information for disaster risk reduction, emergency preparedness, and emergency response.