Earthquake Hazard Models Research Articles

Magneto-based earthquake hazard models for Absheron Peninsula

Magneto-based earthquake hazard models for Absheron Peninsula

San Andreas Fault Earthquake Hazard Model Validation Using Probabilistic Analysis of Precariously Balanced Rocks and Bayesian Updating

Abstract The Mojave section of the San Andreas fault is the closest section to the megacity of greater Los Angeles. A major issue for the population is that the life-threatening hazard estimate of a future rare, large earthquake on this fault section is highly uncertain and untested at timescales and ground motions beyond limited historical recordings. Of relevance to this issue is that the nearby precariously balanced rocks at Lovejoy Buttes have survived these ground motions, despite the past tens of thousands of years of San Andreas fault earthquakes. Therefore, the fragility and age of these precariously balanced rocks provide crucial ground-motion constraints over the timescales of rare, large San Andreas fault earthquakes. We rigorously validate and update an earthquake hazard model for the Mojave section of the San Andreas fault using the independent observational data of precariously balanced rock survival at Lovejoy Buttes. The joint probability of survival of all five studied precariously balanced rocks was used to validate the hazard estimates and reweight the estimates using new Bayesian updating methods to deliver an improved, precariously balanced rock-informed earthquake hazard estimate. At an annual frequency of exceedance of 1×10−4 yr−1, equivalent to a mean return period of 10,000 yr, the precariously balanced rock survival data significantly reduced the mean hazard ground-motion estimate by 65% and the 5th–95th fractile uncertainty range by 72%. The magnitude of this inconsistency provides striking evidence for the need to reevaluate both the source and ground-motion components of our earthquake hazard model for the southern San Andreas fault.

Damage modelling of a bridge pier subjected to multiple earthquakes: a comparative study

This paper discusses and compares two recently developed methodologies for the prediction of damage accumulation in structures subjected to multiple earthquakes within their lifetime, one based on a regression model and one based on a Markov-chain based approach. A stochastic earthquake hazard model is considered for generating sample sequences of ground motion records that are then used to estimate the probabilistic distribution of the damage accumulated during the time interval of interest using the various methodologies. A simulation-based approach provides a reference solution against which the other methodologies are compared. Besides assessing the effectiveness and accuracy of the two methodologies, some improvements of the regression model are proposed and evaluated. The comparison between the methodologies is carried out by examining a reinforced concrete (RC) bridge pier model and using the Park–Ang damage index to describe the damage accumulation. The study results demonstrate the importance of considering the possibility of occurrence of multiple shocks in estimating the life-cycle performance of structures and highlight strengths and drawbacks of the investigated methodologies.

Exploring the impact of spatial correlations of earthquake ground motions in the catastrophe modelling process: a case study for Italy

Catastrophe models are important tools to provide proper assessment and financial management of earthquake-related emergencies, which still create the largest protection gap across all perils. Earthquake catastrophe models include three main components, namely: (1) the earthquake hazard model, (2) the exposure model and, (3) the vulnerability model. Simulating spatially distributed ground-motion fields within either deterministic or probabilistic seismic hazard assessments poses a major challenge when site-related financial protection products are required. In this framework, we develop ad hoc correlation models for different Italian regions (specifically northern, central and southern Italy) and thereafter we perform both deterministic scenario-based and probabilistic event-based hazard and risk assessments in order to advance the understanding of spatial correlations within the catastrophe modelling process. We employ the OpenQuake engine for our calculations. This is an open-source tool suitable for accounting for the spatial correlation of earthquake ground-motion residuals. Our outcomes, albeit preliminary, demonstrate the importance of considering not only the spatial correlation of ground motions, but also its associated uncertainty in risk analyses. Although loss exceedance probability curves for the return periods of interest for the (re)insurance industry show similar trends, both hazard and risk footprints in terms of average annual losses feature less noisy and more realistic patterns if spatial correlation is taken into account. Such results will have implications for (re)insurance companies evaluating the risk to high-value civil engineering infrastructures.

Resilient electrical distribution grid planning against seismic waves using distributed energy resources and sectionalizers: An Indian's urban grid case study

The urban and suburban advancement has increased the consumption of energy per capita gradually. Electrical distribution companies have expanded their network to cater to the demand, making the system more vulnerable to low probability high impact events. The extreme events occurred in last five year had shown us complete to partial blackout, which calls for resiliency enhancement. The presented study concentrates on the Dehradun, India, which listed as the most earthquake-prone region by Union Nations Development Programme. Hence, this paper proposes three main strategies to make a resilient distribution grid against an earthquake; (i) Hardening approach at grid side and demand side against earthquake disaster (ii) Monte-Carlo methodology for earthquake hazard model (iii) K-means algorithm with clustering quality index to identify the multiple vulnerable zones. Grid side hardening is performed using the optimal unit placement of the Solar PV plant (SPP) with storage and parallel cables with sectionalizers. For demand-side hardening, electric vehicles are considered storage after an event, duly taking their future growth rate into account. Furthermore, a Mixed Integer non-linear problem is formulated to identify the optimal sizing and sitting of SPP with storage. The proposed methodology is developed and applied to the distribution system (156-bus) of Dehradun, India.

Techno-economic hardening strategies to enhance distribution system resilience against earthquake

The electrical distribution grid is unremittingly vulnerable to natural disasters. Many researchers propose strategies mainly based on grid-side solutions to improve critical load's survivability during the targeted emergency period. However, the main agenda of resilience enhancement is to improve the overall system resilience. Therefore, this paper proposes a proactive framework that combines the grid-side and demand-side solutions to enhance the overall system resilience. Here, the grid-side approach presents optimal hardening of the distribution grid by using resilient energy storage units (ESUs), underground cables (UCs), and the demand-side by using home battery inverters (HBIs) & its communication infrastructure. For resilient hardening against earthquakes, it is essential to identify it's all possible occurrences. Therefore, a Monte-Carlo-based probabilistic earthquake hazard model is developed through which the vulnerability is characterized using the peak ground acceleration (PGA) model and fragility curves. For optimized hardening investments, the vulnerable zones of the system are identified via clustering algorithms. With the formulated mixed-integer nonlinear problem, the optimal ESUs and UCs are identified for each cluster. The proposed methodology is tested on a real-world 156-bus distribution system of Dehradun district, India, under seismic zone IV.

Forecasting Induced Earthquake Hazard Using a Hydromechanical Earthquake Nucleation Model

AbstractIn response to the dramatic increase in earthquake rates in the central United States, the U.S Geological Survey began releasing 1 yr earthquake hazard models for induced earthquakes in 2016. Although these models have been shown to accurately forecast earthquake hazard, they rely purely on earthquake statistics because there was no precedent for forecasting induced earthquakes based upon wastewater injection data. Since the publication of these hazard models, multiple physics-based methods have been proposed to forecast earthquake rates using injection data. Here, we use one of these methods to generate earthquake hazard forecasts. Our earthquake hazard forecasts are more accurate than statistics-based hazard forecasts. These results imply that fluid injection data, where and when available, and the physical implications of fluid injection should be included in future induced earthquake hazard forecasts.

Development of risk-targeted seismic hazard maps for the Iranian plateau

This study aims to provide the risk-targeted seismic design maps and corresponding adjusting coefficients for different intensity measures on the Iranian plateau, using the iterative risk integral approach. The main purpose of this study is to provide the risk-targeted hazard maps with a uniform level of risk based on an acceptable collapse probability and risk threshold. The investigation applies the most recent and notable seismic hazard model for the studied region, namely the Earthquake Hazard Model of the Middle East (EMME). Having done the preprocessing procedures for characterizing the possible bias in the input data, hazard curves on 0.1° grids were extracted from the median values of the logic tree of the EMME model. Utilizing the deliberately rational values of the required parameters by the risk-targeting approach, i.e. 2% exceedance probability in 50 years as the 10% quantile of the fragility function, leads to reach some multiple uniform seismic risk maps with 1% probability of collapse in 50 years for each intensity measure. The imaginable inconsistency of the maps with historical data (as a picture of reality) has also been assessed in terms of underestimations or overestimations, using a two-sided statistical test at a 95% confidence level. The outcome of this study can be incorporated in the next revision of the national seismic code for a more reliable and robust seismic design of building structures.

Seismic micro-zonation framework for earthquake risk in Medan-North Sumatera Indonesia using geotechnical engineering survey

The micro zonation is very essential to identify the tectonic and geological formation in the study area, and still fundamental for determining the seismic source and establishing a realistic earthquake hazard model. Indonesia is located in the Pacific Ring of Fires, it is constantly at risk of earthquakes, tsunamis, and the eruption of volcanic. In this case, we focused to review the previous study of micro zonation in Indonesia and making the framework of micro-zonation for earthquake risk in Medan-North Sumatra, Indonesia. The goal of this study to illustrate the diversity of methodologies presently in use for the preparation of seismic hazard maps and to evaluate basic principles of zonation for different purposes and at different scales in new planning in Medan, North Sumatera. We select the micro zonation methods and looking for the opportunity for planning the micro zonation based on the literature review. The result of this review, the plan and framework of micro zonation in Medan, North Sumatera will be important and needed for mapping the seismic hazards. The output of seismic micro-zonation map based on the geotechnical engineering survey with the Multichannel Analysis of Surface Waves (MASW), Spatial Auto Correlation (SPAC) and Horizontal to Vertical Ratio (HVSR) for dominant period.

Comparison between outcomes of the 2014 Earthquake Hazard Model of the Middle East (EMME14) and national seismic design codes: The case of Iran

Regional, multi-country seismic hazard models provide a comparison basis for national seismic hazard models that are generally used to underpin the seismic design prescriptions of national building codes. Our study presents an attempt to formalize a framework for performing such a comparison. This comparison consists of sequential steps for identifying and understanding similarities of the key elements informing the seismic hazard models and the code design ground motions in addition to their numerical comparison. The challenge may arise, on one hand, from the lack of transparency of some national seismic codes and, on the other hand, from the intrinsic difficulties in comparing seismic hazard models. In this study, as an example we compare the seismic design spectrum of the Iranian national design code with the uniform hazard spectra from the recent, fully-harmonized, cross-borders Earthquake Hazard Model for the Middle East region (EMME). This comparison focuses on the two 10% in 50-year exceedance probability maps for PGA and on the pairs of design spectra for four cities with different seismicity levels. While, in general, the two reference maps for PGA on rock seem similar, the comparison of the uniform hazard spectra and design spectra for the four selected cities for different soil conditions show large differences. We offer some plausible causes for these differences as well as generic recommendations for overcoming them.

Holocene geologic slip rate for the Banning strand of the southern San Andreas Fault, southern California

AbstractNorthwest directed slip from the southern San Andreas Fault is transferred to the Mission Creek, Banning, and Garnet Hill fault strands in the northwestern Coachella Valley. How slip is partitioned between these three faults is critical to southern California seismic hazard estimates but is poorly understood. In this paper, we report the first slip rate measured for the Banning fault strand. We constrain the depositional age of an alluvial fan offset 25 ± 5 m from its source by the Banning strand to between 5.1 ± 0.4 ka (95% confidence interval (CI)) and 6.4 + 3.7/−2.1 ka (95% CI) using U‐series dating of pedogenic carbonate clast coatings and 10Be cosmogenic nuclide exposure dating of surface clasts. We calculate a Holocene geologic slip rate for the Banning strand of 3.9 + 2.3/−1.6 mm/yr (median, 95% CI) to 4.9 + 1.0/−0.9 mm/yr (median, 95% CI). This rate represents only 25–35% of the total slip accommodated by this section of the southern San Andreas Fault, suggesting a model in which slip is less concentrated on the Banning strand than previously thought. In rejecting the possibility that the Banning strand is the dominant structure, our results highlight an even greater need for slip rate and paleoseismic measurements along faults in the northwestern Coachella Valley in order to test the validity of current earthquake hazard models. In addition, our comparison of ages measured with U‐series and 10Be exposure dating demonstrates the importance of using multiple geochronometers when estimating the depositional age of alluvial landforms.

Statistical analysis of earthquakes after the 1999 MW 7.7 Chi-Chi, Taiwan, earthquake based on a modified Reasenberg–Jones model

We investigated the temporal–spatial hazard of the earthquakes after the 1999 September 21 MW=7.7 Chi-Chi shock in a continental region of Taiwan. The Reasenberg–Jones (RJ) model (Reasenberg and Jones, 1989, 1994) that combines the frequency-magnitude distribution (Gutenberg and Richter, 1944) and time-decaying occurrence rate (Utsu et al., 1995) is conventionally employed for assessing the earthquake hazard after a large shock. However, it is found that the b values in the frequency-magnitude distribution of the earthquakes in the study region dramatically decreased from background values after the Chi-Chi shock, and then gradually increased up. The observation of a time-dependent frequency-magnitude distribution motivated us to propose a modified RJ model (MRJ) to assess the earthquake hazard. To see how the models perform on assessing short-term earthquake hazard, the RJ and MRJ models were separately used to sequentially forecast earthquakes in the study region. To depict the potential rupture area for future earthquakes, we further constructed relative hazard (RH) maps based on the two models. The Receiver Operating Characteristics (ROC) curves (Swets, 1988) finally demonstrated that the RH map based on the MRJ model was, in general, superior to the one based on the original RJ model for exploring the spatial hazard of earthquakes in a short time after the Chi-Chi shock.

Performance assessment and optimization of fluid viscous dampers through life-cycle cost criteria and comparison to alternative design approaches

The performance assessment and optimal design of fluid viscous dampers through life-cycle cost criteria is discussed in this paper. A probabilistic, simulation-based framework is described for estimating the life-cycle cost and a stochastic search approach is developed to support an efficient optimization under different design scenarios (corresponding to different seismicity characteristics). Earthquake losses are estimated using an assembly-based vulnerability approach utilizing the nonlinear dynamic response of the structure whereas a point source stochastic ground motion model, extended here to address near-fault pulse effects, is adopted to describe the seismic hazard. Stochastic simulation is utilized for estimation of all the necessary probabilistic quantities, and for reducing the computational burden a surrogate modeling methodology is integrated within the framework. Two simplified design approaches are also examined, the first considering the optimization of the stationary response, utilizing statistical linearization to address nonlinear damper characteristics, and the second adopting an equivalent lateral force procedure that defines a targeted damping ratio for the structure. These designs are compared against the optimal life-cycle cost one, whereas a compatible comparison is facilitated by establishing an appropriate connection between the seismic input required for the simplified designs and the probabilistic earthquake hazard model. As an illustrative example, the retrofitting of a three-story reinforced concrete office building with nonlinear dampers is considered.

Seismic hazard across Bulgaria and neighbouring areas: regional and site-specific maximum credible magnitudes and earthquake perceptibility

A probabilistic seismic hazard assessment is developed here using maximum credible earthquake magnitude statistics and earthquake perceptibility hazard. Earthquake perceptibility hazard is defined as the probability a site perceives ground shaking equal to or greater than a selected ground motion level X, resulting from an earthquake of magnitude M, and develops estimates for the most perceptible earthquake magnitude, MP(max). Realistic and usable maximum magnitude statistics are obtained from both whole process and part process statistical recurrence models. These approaches are extended to develop relationships between perceptible earthquake magnitude hazard and maximum magnitude recurrence models that are governed by asymptotic and finite return period properties, respectively. Integrated perceptibility curves illustrating the probability of a specific level of perceptible ground motion due to all earthquakes over the magnitude range extending from −∞ to a magnitude Mi are then developed from reviewing site-specific magnitude perceptibility. These lead on to achieving site-specific annual probability of exceedance hazard curves for the example cities of Sofia and Thessaloniki for both horizontal ground acceleration and ground velocity. Both the maximum credible earthquake magnitude M3 and the most perceptible earthquake magnitude MP(max) are of importance to the earthquake engineer when approaching anti-seismic building design. Both forms of hazard are illustrated using contoured hazard maps for the region bounded by 39°–45°N, 19°–29°E. Patterns are observed for these magnitude hazard estimates—especially MP(max) specific to horizontal ground acceleration and horizontal ground velocity—and compared to inferred patterns of crustal deformation across the region. The full geographic region considered is estimated to be subject to a maximum credible earthquake magnitude M3—estimated using cumulative seismic moment release statistics—of 7.53 Mw, calculated from the full content of the adopted earthquake catalogue, while Bulgaria’s capital, Sofia, is estimated a comparable value of 7.36 Mw. Sofia is also forecast most perceptible earthquake magnitudes for the lowest levels considered for horizontal ground acceleration of MPA(50) = 7.20 Mw and horizontal ground velocity of MPV(5) = 7.23 Mw for a specimen focal depth of 15 km.

Locking depths estimated from geodesy and seismology along the San Andreas Fault System: Implications for seismic moment release

[1] The depth of the seismogenic zone is a critical parameter for earthquake hazard models. Independent observations from seismology and geodesy can provide insight into the depths of faulting, but these depths do not always agree. Here we inspect variations in fault depths of 12 segments of the southern San Andreas Fault System derived from over 1000 GPS velocities and 66,000 relocated earthquake hypocenters. Geodetically determined locking depths range from 6 to 22 km, while seismogenic thicknesses are largely limited to depths of 11–20km.Theseseismogenicdepthsbestmatchthegeodeticlockingdepthswhenestimated at the 95% cutoff depth in seismicity, and most fault segment depths agree to within 2 km. However, the Imperial, Coyote Creek, and Borrego segments have significant discrepancies. In these cases the geodetically inferred locking depths are much shallower than the seismogenic depths. We also examine variations in seismic moment accumulation rate per unit fault length as suggested by seismicity and geodesy and find that both approaches yield high rates (1.5– 1.8 ×1 0 13 Nm/yr/km) along the Mojave and Carrizo segments and low rates (∼0.2 × 10 13 Nm/yr/km) along several San Jacinto segments. The largest difference in seismic moment between models is calculated for the Imperial segment, where the moment rate from seismic depths is a factor of ∼2.5 larger than that from geodetic depths. Such variability has important implications for the accuracy to which future major earthquake magnitudes can be estimated.

Fault-Based Earthquake Rupture Forecast in Central Italy: Remarks after the L'Aquila Mw 6.3 Event

Abstract In 2006 we published an earthquake hazard model called LASSCI (LAyered Seismogenic Source model in central Italy). In October 2008 we began to update the model for use in 5- and 10-year forecasts. The LASSCI–2006 model is supported by good fault-based definitions of seismogenic sources and simple physically motivated models of earthquake occurrence; the LASSCI–2009 model has been improved by revision of the error propagation assumptions and increased accuracy of the earthquake probabilities. The 6 April 2009 earthquake that struck L’Aquila occurred on the model fault having the highest probability of occurrence in the 2009 revised LASSCI forecast: it is therefore consistent with our model assumptions. Furthermore, peak ground accelerations were in reasonable agreement with the values having 90% probability of not being exceeded in 50 yr. In the revised 2009 model, the aggregate probability of having a maximum-sized event in the next 5 yr on at least one of the neighboring sources (less than 25 km distance away) decreases in L’Aquila from 10% to 7% after the earthquake occurrence, but still remains a maximum there along the central Apennines. The LASSCI models 2006 and 2009, featuring characteristic fault sources and time dependence, seem to be suitable for guidance in reconstruction and seismic retrofit in the central Appenines.

Volcanic ashfall in New Zealand – probabilistic hazard modelling for multiple sources

The airfall ash component of the New Zealand Probabilistic Volcanic Hazard Model (PVHM) estimates the likelihood of volcanic ash deposits of any given thickness at any site, based on the frequency-magnitude relations of all the significant volcanic sources in New Zealand and the wind distribution statistics. The main source of error in these models is the uncertainty in the historic ash production records of the volcanoes. Another source of error is that the current wind pattern may not be the same as during the period that the observed ash record was deposited. The spatial distribution and magnitude/frequency curves of volcanoes are very different from those of earthquakes, and this is reflected in the hazard models. Whereas a typical probabilistic earthquake hazard model will have the expected acceleration increase by a factor of 3 to 5 between 500 and 10,000 year return periods, the expected ash thickness can change by up to a factor of 1000 over these periods. The Auckland Volcanic Field is now occupied by New Zealand's largest city, so the comparatively small and infrequent eruptions potentially represent a major social and financial hazard. Our modelling of lake core ash records in the Auckland area suggests that the known vents and eruptions probably represent less than half of the total volcanic activity from the field, and that the hazard from local events is comparable to that from more distant sources. An area needing improvement is the hazard in Auckland from local eruptions, with more ash thickness data and better basaltic eruption models.

ALM: An Asperity-based Likelihood Model for California

In most earthquake hazard models, the b -value of the Gutenberg-Richter law plays a central role in forecasting future seismicity based on the observed history. The cumulative earthquake-size distribution is commonly described by a power law: log10( N ) = a - bM , where N is the cumulative number of earthquakes of magnitude M or greater, a is the earthquake productivity of a volume, and b is the relative size distribution (Gutenberg and Richter 1944; Ishimoto and Iida 1939). The slope of this power law, the b -value, is a critical parameter in seismology that describes the size distribution of events. A high b -value indicates a relatively larger proportion of small events, and vice versa. In earthquake forecasting projects such as the source-related probabilistic seismic hazard assessment (PSHA), an underlying fundamental question is: What do the (numerous) smaller earthquakes tell us about the (infrequent) larger ones? Embedded into this question is also the question of stationarity: Can one trust recent small earthquakes to convey accurate information about infrequent large ones? It is these questions of scaling and stationarity that our model fundamentally addresses. In PSHA projects, the b -value is either chosen as a regional constant or varies spatially between local zones. However, there is currently no obvious scientific basis for choosing either of these approaches. The model we propose, the Asperity-based Likelihood Model (ALM), assumes that the earthquake-size distribution, and specifically the b -value of recent micro-earthquakes, is the most important information for forecasting future events of M 5+. Below we first briefly review the evidence that leads us to propose the ALM model, and then we describe the actual steps involved in deriving the model parameters. Most of the evidence in support of ALM stems from observational data from a variety of tectonic regimes as well …

Rates and uncertainties on seismic moment and deformation in eastern Canada

On the basis of earthquake statistics from the Canadian earthquake hazard model, we estimate the rates of seismic moment and deformation in seismic zones of eastern Canada. We use a logic tree approach to derive median values and 66% confidence intervals for models based on earthquake cluster distributions and geological and tectonic structures. Typical moment rates are low, (0.1–1) × 1017 N m yr−1, equivalent to a magnitude Mw = 7 earthquake every 350–3500 years in each seismic zone. In a few zones, such as Baffin Bay, moment rates reach (2–10) × 1017 N m yr−1. Expressed in terms of relative velocity between zone boundaries, seismic deformation rates are typically 0.2–0.5 mm yr−1. For a few regions of high seismicity, such as Baffin Bay, Charlevoix, or Grand Banks, the relative velocities reach 1–2 mm yr−1. Strain rates are typical of the intraplate environment (10−13–10−10 yr−1), with values locally as high as 10−8 yr−1. These rates give some insight into geodynamic aspects of the northeastern North America plate. Assuming that the seismic deformation rates represent long‐term deformation of the lithosphere, they provide some constraint to the amplitude and variability of intraplate strain. On a shorter timescale, seismic strain rates can be compared with postglacial rebound predictions and might help constrain some aspects of these models. Because they reflect the seismic energy released over the last few hundred years, our estimates, in combination with geodetic measurements, can also help constrain models of earthquake dynamics in eastern North America.

A multi-tiered earthquake hazard model for Australia

Earthquakes result from tectonic processes, and their distribution is strongly influenced by large-scale geology and the tectonic stress field. However, earthquake hazard estimates, particularly ground motion recurrence, have traditionally been computed using source models based primarily on instrumental and historical seismicity. In areas of low to moderate seismicity such as Australia, large earthquakes commonly occur in areas which have experienced little or no recent activity, making it difficult to develop source models based solely on seismicity. The seismotectonic model developed for Australia that is presented here (AUS5) is based on geology, geophysics, tectonics and seismicity. The model was developed using a number of tiers of information, so that new information can easily be incorporated. The information used includes, but is not limited to, tectonic provinces, basins and ranges, gravity, magnetic, topography, and seismicity, all on a regional scale. On a local scale, for a site-specific earthquake hazard study, active faulting can be incorporated to provide fault source zones. An earthquake hazard map showing peak ground acceleration with a 10% chance of exceedance in 50 years for southeastern Australia using the geologically defined seismotectonic model AUS5 is presented as an indication of how the model performs.