Models of strong ground motions for probabilistic detailed seismic zoning of the territory of North Ossetia-Alania. Part 2

Abstract

Relevance. Due to the large number of difficultly controlled natural factors, the characteristics of soil movements during strong earthquakes are random in nature, and therefore it is advisable to model them in a probabilistic form. Epistemic uncertainty is taken into account by the inclusion of alternative models in the probabilistic analysis. Aim. To develop a regional model for the territory of the Republic of North Ossetia-Alania. The object of the study is seismic impact. Methods. Probabilistic analysis of seismic hazard. Results. The article describes the models of spectra and accelerograms of soil vibrations, as well as “non-traditional” (“spectral”) models of shake intensity, developed as the part of studies to create a new version of the probability map of detailed seismic zoning (DSZ) of the territory of North Ossetia-Alania. Fourier spectra and response spectra are modeled as sets of spectral density levels (|S|) and spectral accelerations (SA) at different frequencies (f). These parameters are considered as random variables having a log-normal probability distribution for various combinations of earthquake magnitudes (M) and focal distance (D). The construction of distribution functions is carried out in two stages. First, “average” estimates |S| (M, D, f) or SA (M, D, f), and then by adjusting these estimates are close to the conditions of the considered territory. Corrections are made on the basis of the analysis of the seismic and geological features of the area, affecting the parameters under consideration for strong ground motions (SGM). Basic (“average”) calculations were carried out using more than 2500 acceleration spectra of soil vibrations during earthquakes in different parts of the world with M = 2.5–7.7, D = 2-808 km and I = 3-10 MSK points. A new method has also been applied for reconstructing the spectra of soil vibrations during earthquakes according to their macroseismic field. For this purpose, data from macroseismic surveys of 27 strong earthquakes in the region with M = 3.5–7.0 and shaking strength at the epicenter of 6–9 MSK points were used. Accelerograms were modeled both in the form of time functions generated by random vibrations and by selecting records of real earthquakes with their subsequent scaling in the spectral-temporal region. In both cases, the criteria for the accuracy of the constructions were the proximity of the spectra, as well as the amplitudes, periods, and durations measured on the accelerograms to the corresponding spectral models and models of unit parameters described in the first part of the article. Estimates of all SGM parameters obtained by means of the developed models were compared with each other, as well as with estimates from other models known in the world. A high level of consistency and consistency of all simulated SGM parameters during earthquakes of all the considered M and D is shown. The same tests performed for comparison for some other SGM models recommended for use on the territory of the Russian Federation showed the advantages of our models for this indicator. Practical significance. The developed SGM models can be useful in probabilistic and deterministic (“situational”) analysis of seismic hazard in the territory of North Ossetia-Alania and other areas with similar seismic and geological conditions.