Multiple-site seismic hazard assessment
Final Report Abstract
Seismic hazard and risk analysis for lifelines and distributed critical structures requires estimates of the level of earthquake shaking that are likely to occur concurrently at multiple locations. It may be necessary to know whether the vulnerable elements of a lifeline system are likely to be simultaneously affected by shaking of sufficient strength to disable them, e.g. simultaneous damage to several bridges in a transportation network or to several substations of electric power system. There may be also interest in knowing how large may be the strongest shaking in any, i.e. at least one, of critical points (e.g. important highways and bridges, power distribution systems, oil and wastewater storage tanks, etc.) and whether the shaking may be sufficient to cause failure of the whole system or to amplify negative impacts caused by the earthquake directly. The multiple-location estimations could not be performed directly using the results provided by classical point-wise Cornell-McGuire probabilistic seismic hazard assessment (PSHA) method, because the results contain no information about simultaneous ground motions at different sites during one earthquake. The rates of simultaneous and at least one exceedances of reference ground motion at several sites differ from the individual exceedance rates for the same ground motion level, i.e. estimated separately for every site using point-wise PSHA. The difference depends on number and distance between the sites (i.e. total area), characteristics of ground-motion correlation, parameters of seismicity, and level of reference ground motion. We studied specific features of multiple-location PSHA, as compared with the classical point-wise PSHA, using Monte Carlo simulation. The following characteristic of the multiple-location hazard is considered: level of ground motion, which will be exceeded in at least one site or in several sites simultaneously with reference annual probability (or return period). The analysis has been performed for regions of South-Western Germany, Northern and Eastern Taiwan, which represent different levels of seismicity (low, moderate and high, respectively). The relationships between the multiple-location and point-wise estimations are analyzed and quantified. These relationships allow estimating peak amplitudes of ground acceleration, which will be exceeded in at least one site or simultaneously in several sites, depending on number of sites N (i.e. critical facilities) located within an area of a size SAREA with reference ground motion PGAPNT obtained by ordinary (point-wise) PSHA, and depending on selected model of the within-earthquake correlation. The analysis will provide confidence whether it may be possible to utilize the procedure of pointwise PSHA in particular cases of multiple-location hazard assessment, i.e. for estimation of maximum level of ground motion among several sites (i.e. at least one exceedance), or for providing a lower safety level when considering simultaneous exceedances. For example, the maximum level of ground motion among several sites may be considered as an estimation of the importance factor γ, which should be applied to ensure that none of the critical facilities would experience ground shaking above the design level, i.e. PGADSG = PGAPNT x γ, in one earthquake in a finite time period. We have shown that the importance factor recommended in building codes may lead to underestimation of maximum level of ground motion among several sites. We also compared the multiple-location probabilistic seismic hazard estimations and the data obtained during large earthquakes. On example of the 2008 Wenchuan earthquake, we have shown that the multiple-location hazard assessment, when being performed for standard return period 475 years for relatively small areas (e.g., territory of industrial development or a city district, areas of 10 km2 – 25 km2), provides reasonable estimations of the upper limit of possible ground motions. Consideration of the larger areas (e.g. area of a city or large district of a city, areas of more than 100 km2), as well as low level of the within-earthquake correlation, allows estimating the worst case ground motion, i.e. high amplitudes resulted from peculiarities of rupture propagation along the extended earthquake source and local site effect with high amplification. The multiple-location hazard estimations are extremely useful for analysis of damage for critical elements of lifelines (e.g., hospitals, bridges, electrical substations, gas and water supply stations, etc.) located along large territories and for analysis of lifelines performance. Such estimations may be calculated directly from the standard point-wise assessments using the developed relationships, or they may be obtained using Monte-Carlo approach and considering between-earthquake and withinearthquake variability of ground shaking for urban or industrial areas, or zones of particular economic and social importance.
Publications
- “On the modeling of ground-motion field for seismic loss assessment of urban areas and lifelines for regions with a lack of strong-motion data”. 2013 World Congress on Advances in Structural Engineering and Mechanics (ASEM13), 8-12 September 2013, Jeju, Korea
Sokolov V. and F. Wenzel
- “On the relation between point-wise and multiple-location probabilistic seismic hazard assessments“. Bulletin of Earthquake Engineering, May 2015, Volume 13, Issue 5, pp 1281–1301
Sokolov V. and F. Wenzel
(See online at https://doi.org/10.1007/s10518-014-9661-6) - “Seismic hazard from instrumentally recorded, historical, and simulated earthquakes: Application to the Tibet-Himalayan region”. Tectonophysics, Volume 657, 30 August 2015, Pages 187-204
Sokolov V. and A. Ismail-Zadeh
(See online at https://doi.org/10.1016/j.tecto.2015.07.004)
