برآورد خطر لرزه ای احتمالاتی برای رومانی و تجزیه و تحلیل حساسیت: بررسی موردی با در نظر گرفتن مشترک از لرزه خیزی ​​عمق متوسط (Vrancea) و کم عمق (پوسته)

The earthquake risk on Romania is one of the highest in Europe, and seismic hazard for almost half of the territory of Romania is determined by the Vrancea seismic region, which is situated beneath the southern Carpathian Arc. The region is characterized by a high rate of occurrence of large earthquakes in a narrow focal volume at depth from 70 to 160 km. Besides the Vrancea area, several zones of shallow seismicity located within and outside the Romanian territory are considered as seismically dangerous. We present the results of probabilistic seismic hazard analysis, which implemented the “logic tree” approach, and which considered both the intermediate-depth and the shallow seismicity. Various available models of seismicity and ground-motion attenuation were used as the alternative variants. Seismic hazard in terms of macroseismic intensities, peak ground acceleration, and response spectra was evaluated for various return periods. Sensitivity study was performed to analyze the impact of variation of input parameters on the hazard results. The uncertainty on hazard estimates may be reduced by better understanding of parameters of the Vrancea source zone and the zones of crustal seismicity. Reduction of uncertainty associated with the ground-motion models is also very important issue for Romania.

The earthquake risk on Romania is one of the highest in Europe and seismic hazard for almost half of the territory of Romania is determined by the Vrancea seismic region, which is situated beneath the southern Carpathian Arc (Fig. 1). The region is characterized by a high rate of occurrence of large earthquakes in a narrow focal volume at depth from 70 to 160 km (e.g., [1]). Besides the Vrancea area, several zones of shallow seismicity located within and outside the Romanian territory are considered as seismically dangerous. Full-size image (143 K) Fig. 1. Epicentral map of the Vrancea earthquakes (1990–2002), depth distribution of hypocenters, and location of permanent digital K2-network stations in Romania. Figure options The information related to expected seismic effect and expressed in terms of earthquake ground-motion parameters, such as seismic intensity, peak amplitudes of ground motion, pseudo-spectral acceleration (PSA) and ground-motion time histories, is necessary for design of buildings and structures in earthquake prone regions, seismic risk estimation and management, and insurance business. The specification of engineering (or design) ground-motion parameters is the goal of seismic hazard analysis (SHA). It involves the quantitative estimation of ground-shaking hazard at a particular site taking into account characteristics of potentially dangerous earthquakes around the site. The relation between deterministic and probabilistic approaches for SHA is a subject of much controversy [2], [3], [4] and [5] (see also discussion in EOS, 2003, 2004, 2005). The decision what approach should be applied depends on (a) the final goal—how and where do we expect to use the result and (b) the parameters of seismicity or likelihood of the worst-case event. Seismic hazard mapping, development of design codes, retrofit design, financial planning of earthquake losses requires mostly probabilistic hazard assessment. Several studies have been carried out to evaluate the seismic hazard in Romania using the probabilistic approach. Among the most recent ones, which consider the Vrancea seismicity, we should mention the following. Lungu et al. [6] and [7] calculated the hazard in terms of effective peak acceleration and PSA based on recurrence of large earthquakes. Musson [8] estimated peak ground acceleration for various return periods (probability of exceedence) using stochastic Monte Carlo modeling of the earthquake catalog. The distribution of peak ground acceleration (PGA) values for return period of 475 years was calculated by Mantyniemi et al. [9] for three characteristic depths of large earthquakes in the Vrancea area. Ardeleanu et al. [10] performed probabilistic seismic hazard assessment in terms of macroseismic intensities. The results of site-dependent PSHA in terms of macroseismic intensity, PGA, and PSA have been presented by Sokolov et al. [11] for particular locations and by Ismail-Zade et al. [12] for the northern, eastern, and southeastern parts of Romania. Mohindra et al. [13] described probabilistic SHA performed for the whole territory of Romania using so-called “stochastic events” technique and considering both intermediate-depth and crustal seismicity. The crustal and intermediate-depth seismicity were jointly considered also by Musson [8] and Ardeleanu et al. [10]. Moldovan et al. [14] and [15] estimated seismic hazard from crustal events for central Romania in terms of macroseismic intensity. Description of development of the codes for earthquake resistance of buildings and structures in Romania during last 60 years, as well as the current standards for seismic zonation, and design provisions, was given by Lungu et al. [16]. The present seismic code of Romania (P100-1/2004) considers the earthquake hazards for return period (mean recurrence interval) of 100 years (40% probability of being exceeded in 50-year exposure time). The assessment of seismic hazard, as it has been noted in Ref. [16], should be revised in accordance to the recent requirements. For example, as recommended in EUROCODE 8, two variants should be considered for ordinary constructions, namely: a probability of exceedance of 10% in 50 years (recurrence period of 475 years), and a probability of exceedance of 10% in 10 years (recurrence period of 95 years). Recently logic trees become a popular tool in seismic hazard studies. The logic trees technique [17] and [18] is used for incorporating the epistemic uncertainty, or uncertainty reflecting the incomplete knowledge of the nature of seismic motion, into the hazard calculations. Keeping in mind that handing uncertainties is a key element of seismic hazard assessment, we applied in this study PSHA with elements of the logic trees approach. In this article we describe results of PSHA performed for the Romanian territory considering the intermediate-depth and crustal seismicity and using the most recent knowledge on seismic hazard in Romania. When possible, we used various available models of seismicity and ground-motion prediction equations, as alternative variants, which will be described below. The PSHA was performed in terms of PGA, amplitudes of PSA and seismic intensity (MSK or MM) for various return periods or probabilities of being exceeded during specified exposure time.

The seismic hazard for the Romanian territory is determined, for small return period, mainly by the Vrancea intermediate-depth seismicity. However, when considering the small values of probability of being exceeded or large return periods, the influence of crustal events may be significant (more than VI–VII MSK and PGA up to 200–250 cm/s2) for some areas located in the western and the southern parts of Romania. For example, the intensity hazard is larger than VII MSK almost for the whole territory of Romania for return period 475 years. We have to note, however, that for Moesian platform the UHS amplitudes may be underestimated at vibration periods more than 1.0 s, because we did not take into account in this work some peculiarities of site response during large earthquakes within this range of periods (see, for example, Ref. [75] for description of recent analysis for the city of Bucharest). The results of PSHA are expressed in the form of so-called hazard curves. The curves show annual probability of exceeding different levels of ground-motion parameter, or probability that the levels will be exceeded in during a specified time (average live period of construction). Every single run of the hazard calculations, i.e. every particular set of the input parameters (source zonation, magnitude-recurrence model, depth distribution, attenuation relationships, site amplification coefficients, etc.) produce a single hazard curve. The epistemic uncertainty, i.e. existence of various models of the input parameters may be considered using the weighted logic tree technique. The technique joins the weighted hazard curves, which were obtained using combinations of particular models, into a single one. Consideration of overall variability, which includes uncertainty of the particular input parameter or the used model (e.g. maximum magnitude, seismic rate coefficients, depth distribution, site amplification function, etc.), requires: (a) knowledge about characteristics of the variation and (b) a lot of hazard computations using, for example, Monte Carlo technique. In this case, the analysis would result in a group of discrete curves. The level of ground-motion parameter for particular return period (or annual probability of exceedence) would be characterized by a probability distribution function showing the range of possible values (see for example, Ref. [57]). Characteristics of such distribution are important when the knowledge of variance of loss for the portfolio is of interest. One other hand, particular values can be extracted from the distribution for the engineering and design purposes, e.g. the median or specified percentiles levels. Sensitivity analysis shows that uncertainties in input parameters affect on results of probabilistic seismic hazard assessment in a complicated manner depending on interrelation between characteristics of seismic zones and the type of ground-motion descriptor (maximum amplitude, macroseismic intensity, response spectra, etc.), on relative location of the site and the zone, and on the chosen return period (probability of being exceeded). Therefore, for the case of complicated distribution of seismicity along the studied territory, the simple interpolation between a few widely distributed sites may give misleading results. On other hand, it is not possible to convert the results obtained for peak acceleration into macroseismic intensity directly. When analyzing the impact of particular input parameters, it can be concluded that for the case of intermediate-depth earthquakes the parameters of the source zone (magnitude–frequency distribution and depth distribution of earthquake sources) are the most important. The shape of source zones is crucial for shallow seismicity. The choice of minimum magnitude may be important when it is necessary to take into account high-amplitude high-frequency peak acceleration. Reduction of uncertainty associated with the attenuation models is also very important issue. We have to note that one of the used ground-motion models has been constructed by Lungu et al. [6] and [7] using records of three large Vrancea earthquakes. Application of the model for small magnitudes may result in overestimation of the peak amplitudes [66].