Uncertainty In Ground-motion Prediction Research Articles

Comparison and selection of ground motion prediction equations for the Sichuan–Yunnan area, southwest China

The Sichuan–Yunnan area is one of the most seismically active regions in China. As ground-motion models form a key component of seismic hazard analysis, it is important to select (or develop) appropriate models for this area. The increasing number of digital ground-motion records of earthquakes in this area has allowed the development of local ground motion prediction equations (GMPEs). This study compares and, later, recommends appropriate GMPEs for the Sichuan–Yunnan area. We first evaluate the inherent quality of local GMPEs, with respect to their underlying datasets, the variables used and their functional forms, to determine a set of candidate GMPEs. Then we investigate how well the predictions from the GMPEs match observations computed from strong-motion records of recent earthquakes in this area. The fit between predictions and observations varies significantly amongst the GMPEs. The results suggest that some recent local GMPEs would lead to biased ground-motion estimates due to limitations of their underlying datasets and functional forms. Based on both evaluations of inherent quality and compatibility with observations, only one local GMPE is recommended. A comparison of the predictions from three widely-used non-local GMPEs indicates that ground motions in the Sichuan–Yunnan area appear more variable than those in other regions but that predictions from these non-local GMPEs are generally unbiased. We recommend use of a mixture of robust local and non-local GMPEs within seismic hazard analyses to capture the epistemic uncertainty in ground-motion prediction for this area.

Comparison of conventional and Monte Carlo simulation-based probabilistic seismic hazard analyses for Shiraz city, southern Iran

Estimation of ground-motion amplitudes of different hazard levels is of paramount importance in planning of urban development of any metropolis. Such estimation can be computed through a probabilistic seismic hazard analysis (PSHA). This paper concentrates on the PSHA of an area located in Shiraz city, southern Iran. The area includes whole of Shiraz city (i.e., one of the largest and most populous cities of Iran) and its outskirts. Conventional and Monte Carlo simulation-based approaches are utilized to perform the PSHA of the studied area. Two areal seismic source models are delineated, and thence seismicity parameters of all zones associated with their corresponding uncertainties are computed. Uncertainties in ground-motion prediction are accounted for via three ground-motion prediction equations (GMPEs) within the logic tree framework. These GMPEs are applied to estimate bedrock ground shaking (Vs30 = 760 m/s) for several return periods (i.e., 75, 475, 975, and 2475 years). In general, the results of the two abovementioned PSHA approaches show relatively similar results. However, the Monte Carlo simulation-based approach overpredicts bedrock spectral accelerations at periods of 0.4–2.5 s compared to the conventional PSHA approach for return periods of 475, 975, and 2475 years.

Site characterisation of GeoNet stations for the New Zealand Strong Motion Database

The New Zealand Strong Motion Database provides a wealth of new strong motion data for engineering applications. In order to utilise these data in ground motion prediction, characterisation of key site parameters at each of the ~497 past and present GeoNet strong motion stations represented in the database is required. Here, we present the compilation of a complete set of site metadata for the New Zealand database, including four key parameters: i) NZS1170.5 site subsoil classification, ii) the time-averaged shear-wave velocity to a depth of 30 m (Vs30), iii) fundamental site period (Tsite) and iv) depth to a shear-wave velocity of 1000 m/s (Z1.0, a proxy for depth to bedrock). In addition, we have assigned a quality estimate (Quality 1 – 3) to each numerical parameter to provide a qualitative estimate of the uncertainty. New high-quality Tsite, Vs30 and Z1.0 estimates have been obtained from a variety of recent studies, and reconciled with available geological information. This database will be used in efforts to guide development and testing of new and existing ground motion prediction models in New Zealand, allowing re-examination of the most important site parameters that control site response in a New Zealand setting. Preliminary analyses, using the newly compiled data, suggest that high quality site parameters can reduce uncertainty in ground motion prediction. Furthermore, the database can be used to identify suitable rock reference sites for seismological research, and as a guide to more detailed site-specific references in the literature. The database provides an additional resource for informing engineering design, however it is not suitable as a replacement for site-specific assessment.

Assessing the impact of ground-motion variability and uncertainty on empirical fragility curves

Empirical fragility curves, constructed from databases of thousands of building-damage observations, are commonly used for earthquake risk assessments, particularly in Europe and Japan, where building stocks are often difficult to model analytically (e.g. old masonry structures or timber dwellings). Curves from different studies, however, display considerable differences, which lead to high uncertainty in the assessed seismic risk. One potential reason for this dispersion is the almost universal neglect of the spatial variability in ground motions and the epistemic uncertainty in ground-motion prediction. In this paper, databases of building damage are simulated using ground-motion fields that take account of spatial variability and a known fragility curve. These databases are then inverted, applying a standard approach for the derivation of empirical fragility curves, and the difference with the known curve is studied. A parametric analysis is conducted to investigate the impact of various assumptions on the results. By this approach, it is concluded that ground-motion variability leads to flatter fragility curves and that the epistemic uncertainty in the ground-motion prediction equation used can have a dramatic impact on the derived curves. Without dense ground-motion recording networks in the epicentral area empirical curves will remain highly uncertain. Moreover, the use of aggregated damage observations appears to substantially increase uncertainty in the empirical fragility assessment. In contrast, the use of limited randomly-chosen un-aggregated samples in the affected area can result in good predictions of fragility.

Consistency of ground-motion predictions from the past four decades: peak ground velocity and displacement, Arias intensity and relative significant duration

Due to the limited observational datasets available for the derivation of ground-motion prediction equations (GMPEs) there is always epistemic uncertainty in the estimated median ground motion. Since the quality and quantity of strong-motion datasets is constantly increasing it would be expected that the epistemic uncertainty in ground-motion prediction (related to lack of knowledge and data) is decreasing. This article is a continuation of the study of Douglas (Bull Earthq Eng 8(6):1515–1526, 2010) for ground-motion parameters other than peak ground acceleration (PGA) and elastic response spectral acceleration (SA). The epistemic uncertainty in the prediction of peak ground velocity and displacement, Arias intensity and relative significant duration is investigated by plotting predictions from dozens of GMPEs for these parameters against date of publication for three scenarios. In agreement with the previous study, all ground-motion parameters considered show high epistemic uncertainty (often even higher than previously reported for PGA and SA), suggesting that research efforts for the development of GMPEs for these parameters should continue and that it is vital that this uncertainty is accounted for in seismic hazard assessments. The epistemic uncertainty in the prediction of relative significant duration, however, appears to be much lower than any other strong-motion parameter, which suggests that currently available GMPEs for this intensity measure are sufficiently mature.

Toward a ground-motion logic tree for probabilistic seismic hazard assessment in Europe

The Seismic Hazard Harmonization in Europe (SHARE) project, which began in June 2009, aims at establishing new standards for probabilistic seismic hazard assessment in the Euro-Mediterranean region. In this context, a logic tree for ground-motion prediction in Europe has been constructed. Ground-motion prediction equations (GMPEs) and weights have been determined so that the logic tree captures epistemic uncertainty in ground-motion prediction for six different tectonic regimes in Europe. Here we present the strategy that we adopted to build such a logic tree. This strategy has the particularity of combining two complementary and independent approaches: expert judgment and data testing. A set of six experts was asked to weight pre-selected GMPEs while the ability of these GMPEs to predict available data was evaluated with the method of Scherbaum et al. (Bull Seismol Soc Am 99:3234–3247, 2009). Results of both approaches were taken into account to commonly select the smallest set of GMPEs to capture the uncertainty in ground-motion prediction in Europe. For stable continental regions, two models, both from eastern North America, have been selected for shields, and three GMPEs from active shallow crustal regions have been added for continental crust. For subduction zones, four models, all non-European, have been chosen. Finally, for active shallow crustal regions, we selected four models, each of them from a different host region but only two of them were kept for long periods. In most cases, a common agreement has been also reached for the weights. In case of divergence, a sensitivity analysis of the weights on the seismic hazard has been conducted, showing that once the GMPEs have been selected, the associated set of weights has a smaller influence on the hazard.

Consistency of ground-motion predictions from the past four decades

Due to the limited observational datasets available for the derivation of ground-motion prediction equations (GMPEs) there is always epistemic uncertainty in the estimated median ground motion. Because of the increasing quality and quantity of strong-motion datasets it would be expected that the epistemic uncertainty in ground-motion prediction (related to lack of knowledge and data) is decreasing. In this study the predicted median ground motions from over 200 GMPEs for various scenarios are plotted against date of publication to examine whether the scatter in the predictions (a measure of epistemic uncertainty) is decreasing with time. It is found that there are still considerable differences in predicted ground motions from the various GMPEs and that the variation between estimates is not reducing although the ground motion estimated by averaging median predictions is roughly constant. For western North America predictions for moderate earthquakes have show a high level of consistency since the 1980s as do, but to a lesser extent, predictions for moderate earthquakes in Europe, the Mediterranean and the Middle East. A good match is observed between the predictions from GMPEs and the median ground motions based on observations from similar scenarios. Variations in median ground motion predictions for stable continental regions and subduction zones from different GMPEs are large, even for moderate earthquakes. The large scatter in predictions of the median ground motion shows that epistemic uncertainty in ground-motion prediction is still large and that it is vital that this is accounted for in seismic hazard assessments.

Prediction of spatially distributed seismic demands in specific structures: Structural response to loss estimation

AbstractA companion paper has investigated the effects of intensity measure (IM) selection in the prediction of spatially distributed response in a multi‐degree‐of‐freedom structure. This paper extends from structural response prediction to performance assessment metrics such as probability of structural collapse; probability of exceeding a specified level of demand or direct repair cost; and the distribution of direct repair loss for a given level of ground motion. In addition, a method is proposed to account for the effect of varying seismological properties of ground motions on seismic demand that does not require different ground motion records to be used for each intensity level. Results illustrate that the conventional IM, spectral displacement at the first mode, Sde(T1), produces higher risk estimates than alternative velocity‐based IM's, namely spectrum intensity, SI, and peak ground velocity, PGV, because of its high uncertainty in ground motion prediction and poor efficiency in predicting peak acceleration demands. Copyright © 2009 John Wiley & Sons, Ltd.

Subevent structure of large earthquakes—A ground‐motion perspective

Our ability to predict ground motions from future earthquakes hinges on the accurate modeling of the radiation from the earthquake source. A successful approach to modeling earthquake radiation has been to represent faults as a series of discrete, independently‐rupturing subfaults, although their physical interpretation has remained largely unclear. Our simulation of ground motions from twenty‐six well‐recorded moderate‐to‐large earthquakes substantiates the hypothesis that large ruptures are made up of a sequence of smaller subevents. The size of a subevent follows a simple linear relationship with the size of the earthquake fault in an apparently deterministic manner. The strength of the high‐frequency radiation is controlled by the maximum slip velocity, which varies stochastically over a small range. Both the characteristic subevent size and the slip velocity appear to be region‐independent, indicating a remarkable uniformity in earthquake source properties. This observation helps to reduce the uncertainty in ground‐motion prediction by constraining its essential and previously ambiguous parameters.

Validation of a procedure for calculating broadband strong-motion time histories with empirical Green's functions

Abstract We quantitatively evaluate the validity of an approach to calculating strong ground motion time histories that involves the kinematic modeling of earthquake rupture by using recorded empirical Green's functions to constrain the propagation path and site contribution in the period range 0.05 to 2.0 sec, leaving only the contribution of the source to be specified. In addition, we use numerical Green's functions for long periods (> 10 sec) where the average crustal structure is assumed to be well known. In the intermediate band (2 to 10 sec) where site and path effects are still important but empirical Green's functions have small signal-to-noise ratios, we use an arbitrary smoothing method to connect the numerical and empirical Green's function synthetics. The evaluation presented here consists of comparing actual recordings of the Loma Prieta earthquake at 26 sites to syntheses from a simple source model similar to what is thought to have occurred during the Loma Prieta earthquake. We show that the standard error between observed and predicted response spectra is less than or equal to errors from other methods for periods between 0.05 and 0.4 sec and is significantly less than regression methods based on pre-Loma Prieta empirical strong-motion data at periods between 0.5 and 5.0 sec. The strong ground motion prediction methodology that we are validating provides an accurate, defendable means to characterize site and path effects and, therefore, allows the uncertainties in the predicted hazard to be due to unresolved issues about the earthquake source such as the geological constraints of a particular fault and details about the physics of earthquakes. This methodology produces accurate and realistic acceleration and displacement time histories, which are becoming increasingly important for modeling the nonlinear response of large, distributed structures and soft soils. The implication for strong ground motion prediction is that a range of representations of an earthquake source based on physically realistic parameters can be used to generate a suite of possible time histories that represent the range of ground motion for that specific site and earthquake. In addition, in the future, as more information about a particular fault and earthquakes in general becomes available, the uncertainty in ground-motion prediction should decrease.