High-speed transit (1300 km=hr) using pods traveling in evacuated tubes has been proposed. This S... more High-speed transit (1300 km=hr) using pods traveling in evacuated tubes has been proposed. This Short Note addresses how earthquake ground shaking is changed when it is experienced by a high-speed pod that is confined to a track. In particular, earthquake motions can cause lateral deformations of the tube that cause centripetal forces in the pod. I discuss the nature of these forces for the cases of (1) a tube that crosses a fault offset, (2) a tube that is deformed by traveling waves in the Earth, and (3) a tube that resonates between fixed points (e.g., a simple bridge). I suggest several schemes to control the peak centripetal accelerations of the pod.
The 2019 Mw 6.4 and Mw 7.1 Ridgecrest earthquake sequence occurred in the eastern California shea... more The 2019 Mw 6.4 and Mw 7.1 Ridgecrest earthquake sequence occurred in the eastern California shear zone (ECSZ). The mainshock ruptured the Little Lake fault zone, and aftershocks extended from the Garlock fault in the south to the southern end of the 1872 M 7.5 Owens Valley earthquake rupture in the north. We present data from the Southern California Seismic Network (SCSN) and partner seismic networks recorded by the SCSN in the region. These time-series data and related products such as the SCSN earthquake picks and catalogs, available from the Southern California Earthquake Data Center, provide the most comprehensive seismic datasets for the 2019 Ridgecrest earthquake sequence.
The Community Seismic Network‐Los Angeles Unified School District is a network of 300 low‐cost mi... more The Community Seismic Network‐Los Angeles Unified School District is a network of 300 low‐cost microelectromechanical systems accelerometers located in schools in the Los Angeles, California, region. They are capable of accurately recording strong motion up to ±2g and are sufficiently spatially dense that they provide unaliased measurements of strong motions up to 1 Hz following a major earthquake. They are used to provide state‐of‐health monitoring for the schools and surrounding communities to guide the emergency response. As a research tool, they can be used to provide estimates of the site response at the schools and, therefore, provide a much denser set of site responses for ground‐motion prediction than is currently available.
Independent constraints on the ground motions experienced at Point Reyes station during the 1906 ... more Independent constraints on the ground motions experienced at Point Reyes station during the 1906 San Francisco earthquake are obtained by analyzing the dynamic response of a train which overturned during the earthquake. From this analysis, we conclude that the P GA and P GV at Point Reyes station would have been at least 4 m/s 2 and 0.5 m/s, respectively. This lower bound is then used to perform simple checks on the synthetic ground motion simulations of the 1906 San Francisco earthquake by Aagaard et al. [1]. It is also shown that the hypocenter of the earthquake should be to the south of Point Reyes station for the overturning of the train to match the description provided by Jordan [7].
... Engineering 11 wcBE ISBN: 0 08 042822 3 The TriNet Project Thomas Heaton (1), Robert Clayton ... more ... Engineering 11 wcBE ISBN: 0 08 042822 3 The TriNet Project Thomas Heaton (1), Robert Clayton (1), James Davis (2), Egill Hauksson (1), Lucille Jones (3), Hiroo ... This data will be stored on an optical jukebox (600 Gbyte), and data can be retrieved through the Internet. ...
Historic earthquake sequences on subduction zones that are similar to the Cascadia subduction zon... more Historic earthquake sequences on subduction zones that are similar to the Cascadia subduction zone are used to hypothesize the nature of shallow subduction earthquakes that might occur in the northwestern United States. Based on systematic comparisons of several physical characteristics, including physiography and seismicity, subduction zones that are deemed most similar to the Cascadia subduction zone are those in southern Chile, southwestern Japan, and Colombia. These zones have all experienced very large earthquake sequences, and if the Cascadia subduction zone is also capable of storing elastic strain energy along its greater than 1000 km length, then earthquakes of very large size (Mw > 8½) must be considered. Circumstantial evidence is presented that suggests (but does not prove) that large subduction earthquakes along the Cascadia subduction zone may have an average repeat time of 400 to 500 yr.
We compare teleseismic P-wave records for earthquakes in the magnitude range from 6.0 to 9.5 with... more We compare teleseismic P-wave records for earthquakes in the magnitude range from 6.0 to 9.5 with synthetics for a self-similar, ~2 source model and conclude that the energy radiated by very large earthquakes (M,, > 8~) is not self-similar to that radiated from smaller earthquakes (Mw < 81). Furthermore, in the period band from 2 sec to several tens of seconds, we conclude that large subduction earthquakes have an average spectral decay rate of ~-,5 This spectral decay rate is consistent with a previously noted tendency of the J model to overestimate Ms for large earthquakes.
We study scale dependence of strength in dynamical systems failing at multiple length scales and ... more We study scale dependence of strength in dynamical systems failing at multiple length scales and its physical basis. We show that, for a spring block slider model, the strength defined as the energy per unit slip per unit rupture length decreases as a power law as the rupture length increases with the scaling exponent varying between 0 (the plastic limit) and -0.5 (the brittle limit). We conjecture that the physical basis of this scaling is in the pulse-like nature of the propagating rupture on velocity weakening (VW) interfaces which results in larger events having higher slip rates and lower dynamic friction. As a consequence, large amplitude slip pulses have less frictional work per unit slip (our definition of strength) than do small amplitude slip pulses. Since large amplitude pulses are also associated with long ruptures, the strength decreases with the physical size of the system. We investigate the energy budget in numerical simulations of the highly nonlinear multi-degree-o...
The Virtual Seismologist (VS) algorithm is a Bayesian approach to earthquake early warning (EEW) ... more The Virtual Seismologist (VS) algorithm is a Bayesian approach to earthquake early warning (EEW) being implemented by the Swiss Seismological Service at ETH Zurich. The application of Bayes’ theorem in earthquake early warning states that the most probable source estimate at any given time is a combination of contributions from a likelihood function that evolves in response to incoming data from the on-going earthquake, and selected prior information, which can include factors such as network topology, the Gutenberg-Richter relationship or previously observed seismicity. The VS algorithm was one of three EEW algorithms involved in the California Integrated Seismic Network (CISN) real-time EEW testing and performance evaluation effort. Its compelling real-time performance in California over the last three years has led to its inclusion in the new USGS-funded effort to develop key components of CISN ShakeAlert, a prototype EEW system that could potentially be implemented in California. A significant portion of VS code development was supported by the SAFER EEW project in Europe. We discuss recent enhancements to the VS EEW algorithm. We developed and continue to test a multiple-threshold event detection scheme, which uses different association / location approaches depending on the peak amplitudes associated with an incoming P pick. With this scheme, an event with sufficiently high initial amplitudes can be declared on the basis of a single station, maximizing warning times for damaging events for which EEW is most relevant. Smaller, non-damaging events, which will have lower initial amplitudes, will require more picks to be declared an event to reduce false alarms. This transforms the VS codes from a regional EEW approach reliant on traditional location estimation (and it requirement of at least 4 picks as implemented by the Binder Earthworm phase associator) to a hybrid on-site/regional approach capable of providing a continuously evolving stream of EEW information starting from the first P-detection. Offline analysis on Swiss and California waveform datasets indicate that the multiple-threshold approach is faster and more reliable for larger events than the earlier version of the VS codes. This multiple-threshold approach is well-suited for implementation on a wide range of devices, from embedded processor systems installed at a seismic stations, to small autonomous networks for local warnings, to large-scale regional networks such as the CISN. In addition, we quantify the influence of systematic use of prior information and Vs30-based corrections for site amplification on VS magnitude estimation performance, and describe how components of the VS algorithm will be integrated into non-EEW standard network processing procedures at CHNet, the national broadband / strong motion network in Switzerland. These enhancements to the VS codes will be transitioned from off-line to real-time testing at CHNet in Europe in the coming months, and will be incorporated into the development of key components of CISN ShakeAlert prototype system in California.
In the absence of actual building accelerometer data, the linear response of a structure to stron... more In the absence of actual building accelerometer data, the linear response of a structure to strong ground motion is estimated by the convolution of the dynamic response of the structure with an input ground motion. The input motion is usually provided by a local `reference' station record. In this study, we look at whether actual recorded ground motion at two instrumented buildings with well studied dynamic properties can be satisfactorily modeled using a local ground station. All stations record continuous 24-bit data streams on the CISN network, so analysis of a variety of weak earthquake motions, as well as ambient noise, is possible. Our buildings are the 9-story reinforced concrete Millikan Library (CISN Station MIK) and the 3-story braced steel frame Broad Center (CBC), both on the Caltech Campus. Motions recorded on their upper floors are compared with motions from ground stations located in the basement of a lightweight wood-frame house (GSA), and in a subsurface vault (CRP). All stations are within 200m of each other. Recent work using the new continuous datastream indicates that the natural frequencies of these structures can vary by up to 5% during normal ambient conditions, due to such factors as changing building usage, diurnal temperature variation, and wind/rainfall events. These shifts can be sudden, and models of building motions are sensitive to these previously un-documented changes. Further, during stronger motions, such as forced vibration testing, and minor earthquake shaking, natural frequencies are shown to drop by up to 10% (2003 M5.4 Big Bear Earthquake, Δ = 119km), with near-instantaneous recovery once the excitation is over. Moderate earthquakes can temporarily reduce frequencies by up to 30% with no apparent structural damage (1971 M6.6 San Fernando Earthquake, Δ = 31km). Post-event permanent reductions of about 10% have been observed. The ability to monitor these evolving dynamic characteristics makes a re-evaluation of the application of linear transfer functions timely.
High-speed transit (1300 km=hr) using pods traveling in evacuated tubes has been proposed. This S... more High-speed transit (1300 km=hr) using pods traveling in evacuated tubes has been proposed. This Short Note addresses how earthquake ground shaking is changed when it is experienced by a high-speed pod that is confined to a track. In particular, earthquake motions can cause lateral deformations of the tube that cause centripetal forces in the pod. I discuss the nature of these forces for the cases of (1) a tube that crosses a fault offset, (2) a tube that is deformed by traveling waves in the Earth, and (3) a tube that resonates between fixed points (e.g., a simple bridge). I suggest several schemes to control the peak centripetal accelerations of the pod.
The 2019 Mw 6.4 and Mw 7.1 Ridgecrest earthquake sequence occurred in the eastern California shea... more The 2019 Mw 6.4 and Mw 7.1 Ridgecrest earthquake sequence occurred in the eastern California shear zone (ECSZ). The mainshock ruptured the Little Lake fault zone, and aftershocks extended from the Garlock fault in the south to the southern end of the 1872 M 7.5 Owens Valley earthquake rupture in the north. We present data from the Southern California Seismic Network (SCSN) and partner seismic networks recorded by the SCSN in the region. These time-series data and related products such as the SCSN earthquake picks and catalogs, available from the Southern California Earthquake Data Center, provide the most comprehensive seismic datasets for the 2019 Ridgecrest earthquake sequence.
The Community Seismic Network‐Los Angeles Unified School District is a network of 300 low‐cost mi... more The Community Seismic Network‐Los Angeles Unified School District is a network of 300 low‐cost microelectromechanical systems accelerometers located in schools in the Los Angeles, California, region. They are capable of accurately recording strong motion up to ±2g and are sufficiently spatially dense that they provide unaliased measurements of strong motions up to 1 Hz following a major earthquake. They are used to provide state‐of‐health monitoring for the schools and surrounding communities to guide the emergency response. As a research tool, they can be used to provide estimates of the site response at the schools and, therefore, provide a much denser set of site responses for ground‐motion prediction than is currently available.
Independent constraints on the ground motions experienced at Point Reyes station during the 1906 ... more Independent constraints on the ground motions experienced at Point Reyes station during the 1906 San Francisco earthquake are obtained by analyzing the dynamic response of a train which overturned during the earthquake. From this analysis, we conclude that the P GA and P GV at Point Reyes station would have been at least 4 m/s 2 and 0.5 m/s, respectively. This lower bound is then used to perform simple checks on the synthetic ground motion simulations of the 1906 San Francisco earthquake by Aagaard et al. [1]. It is also shown that the hypocenter of the earthquake should be to the south of Point Reyes station for the overturning of the train to match the description provided by Jordan [7].
... Engineering 11 wcBE ISBN: 0 08 042822 3 The TriNet Project Thomas Heaton (1), Robert Clayton ... more ... Engineering 11 wcBE ISBN: 0 08 042822 3 The TriNet Project Thomas Heaton (1), Robert Clayton (1), James Davis (2), Egill Hauksson (1), Lucille Jones (3), Hiroo ... This data will be stored on an optical jukebox (600 Gbyte), and data can be retrieved through the Internet. ...
Historic earthquake sequences on subduction zones that are similar to the Cascadia subduction zon... more Historic earthquake sequences on subduction zones that are similar to the Cascadia subduction zone are used to hypothesize the nature of shallow subduction earthquakes that might occur in the northwestern United States. Based on systematic comparisons of several physical characteristics, including physiography and seismicity, subduction zones that are deemed most similar to the Cascadia subduction zone are those in southern Chile, southwestern Japan, and Colombia. These zones have all experienced very large earthquake sequences, and if the Cascadia subduction zone is also capable of storing elastic strain energy along its greater than 1000 km length, then earthquakes of very large size (Mw > 8½) must be considered. Circumstantial evidence is presented that suggests (but does not prove) that large subduction earthquakes along the Cascadia subduction zone may have an average repeat time of 400 to 500 yr.
We compare teleseismic P-wave records for earthquakes in the magnitude range from 6.0 to 9.5 with... more We compare teleseismic P-wave records for earthquakes in the magnitude range from 6.0 to 9.5 with synthetics for a self-similar, ~2 source model and conclude that the energy radiated by very large earthquakes (M,, > 8~) is not self-similar to that radiated from smaller earthquakes (Mw < 81). Furthermore, in the period band from 2 sec to several tens of seconds, we conclude that large subduction earthquakes have an average spectral decay rate of ~-,5 This spectral decay rate is consistent with a previously noted tendency of the J model to overestimate Ms for large earthquakes.
We study scale dependence of strength in dynamical systems failing at multiple length scales and ... more We study scale dependence of strength in dynamical systems failing at multiple length scales and its physical basis. We show that, for a spring block slider model, the strength defined as the energy per unit slip per unit rupture length decreases as a power law as the rupture length increases with the scaling exponent varying between 0 (the plastic limit) and -0.5 (the brittle limit). We conjecture that the physical basis of this scaling is in the pulse-like nature of the propagating rupture on velocity weakening (VW) interfaces which results in larger events having higher slip rates and lower dynamic friction. As a consequence, large amplitude slip pulses have less frictional work per unit slip (our definition of strength) than do small amplitude slip pulses. Since large amplitude pulses are also associated with long ruptures, the strength decreases with the physical size of the system. We investigate the energy budget in numerical simulations of the highly nonlinear multi-degree-o...
The Virtual Seismologist (VS) algorithm is a Bayesian approach to earthquake early warning (EEW) ... more The Virtual Seismologist (VS) algorithm is a Bayesian approach to earthquake early warning (EEW) being implemented by the Swiss Seismological Service at ETH Zurich. The application of Bayes’ theorem in earthquake early warning states that the most probable source estimate at any given time is a combination of contributions from a likelihood function that evolves in response to incoming data from the on-going earthquake, and selected prior information, which can include factors such as network topology, the Gutenberg-Richter relationship or previously observed seismicity. The VS algorithm was one of three EEW algorithms involved in the California Integrated Seismic Network (CISN) real-time EEW testing and performance evaluation effort. Its compelling real-time performance in California over the last three years has led to its inclusion in the new USGS-funded effort to develop key components of CISN ShakeAlert, a prototype EEW system that could potentially be implemented in California. A significant portion of VS code development was supported by the SAFER EEW project in Europe. We discuss recent enhancements to the VS EEW algorithm. We developed and continue to test a multiple-threshold event detection scheme, which uses different association / location approaches depending on the peak amplitudes associated with an incoming P pick. With this scheme, an event with sufficiently high initial amplitudes can be declared on the basis of a single station, maximizing warning times for damaging events for which EEW is most relevant. Smaller, non-damaging events, which will have lower initial amplitudes, will require more picks to be declared an event to reduce false alarms. This transforms the VS codes from a regional EEW approach reliant on traditional location estimation (and it requirement of at least 4 picks as implemented by the Binder Earthworm phase associator) to a hybrid on-site/regional approach capable of providing a continuously evolving stream of EEW information starting from the first P-detection. Offline analysis on Swiss and California waveform datasets indicate that the multiple-threshold approach is faster and more reliable for larger events than the earlier version of the VS codes. This multiple-threshold approach is well-suited for implementation on a wide range of devices, from embedded processor systems installed at a seismic stations, to small autonomous networks for local warnings, to large-scale regional networks such as the CISN. In addition, we quantify the influence of systematic use of prior information and Vs30-based corrections for site amplification on VS magnitude estimation performance, and describe how components of the VS algorithm will be integrated into non-EEW standard network processing procedures at CHNet, the national broadband / strong motion network in Switzerland. These enhancements to the VS codes will be transitioned from off-line to real-time testing at CHNet in Europe in the coming months, and will be incorporated into the development of key components of CISN ShakeAlert prototype system in California.
In the absence of actual building accelerometer data, the linear response of a structure to stron... more In the absence of actual building accelerometer data, the linear response of a structure to strong ground motion is estimated by the convolution of the dynamic response of the structure with an input ground motion. The input motion is usually provided by a local `reference' station record. In this study, we look at whether actual recorded ground motion at two instrumented buildings with well studied dynamic properties can be satisfactorily modeled using a local ground station. All stations record continuous 24-bit data streams on the CISN network, so analysis of a variety of weak earthquake motions, as well as ambient noise, is possible. Our buildings are the 9-story reinforced concrete Millikan Library (CISN Station MIK) and the 3-story braced steel frame Broad Center (CBC), both on the Caltech Campus. Motions recorded on their upper floors are compared with motions from ground stations located in the basement of a lightweight wood-frame house (GSA), and in a subsurface vault (CRP). All stations are within 200m of each other. Recent work using the new continuous datastream indicates that the natural frequencies of these structures can vary by up to 5% during normal ambient conditions, due to such factors as changing building usage, diurnal temperature variation, and wind/rainfall events. These shifts can be sudden, and models of building motions are sensitive to these previously un-documented changes. Further, during stronger motions, such as forced vibration testing, and minor earthquake shaking, natural frequencies are shown to drop by up to 10% (2003 M5.4 Big Bear Earthquake, Δ = 119km), with near-instantaneous recovery once the excitation is over. Moderate earthquakes can temporarily reduce frequencies by up to 30% with no apparent structural damage (1971 M6.6 San Fernando Earthquake, Δ = 31km). Post-event permanent reductions of about 10% have been observed. The ability to monitor these evolving dynamic characteristics makes a re-evaluation of the application of linear transfer functions timely.
Uploads
Papers by Thomas Heaton