Development of probabilistic seismic hazard microzonation maps at the surface level for central-east Iran (Kerman region), using a hybrid site condition model

Abstract

This study presents the spatial distribution of a grid-based probabilistic seismic hazard model at the surface level for the Kerman region in central-east Iran, using a hybrid site condition model. The hazard model has provided the variation of the peak ground acceleration (PGA), and 5% damped spectral accelerations (SA) at the fundamental periods of 0.2 and 1.0 seconds (s) on 0.05° grids. The study has used a seismotectonic dataset including ∼400 linear seismic sources to characterize the seismic source models. In order to capture the uncertainties of the seismic hazard sources, the model used a logic tree approach, including eight ground motion prediction equations (GMPEs), three characteristic source models, two different declustering methods, three choices for earthquake recurrence parameters, and three alternatives for the maximum credible earthquake. Besides, various GMPEs have been evaluated against local datasets. For this, a series of GMPEs developed based on local, regional, and global data have been selected to find their compatibility with the studied area through statistical analyses, so-called likelihood (LH), and log-likelihood (LLH) tests. Generally, the results display that a wide range of seismic hazards can be expected in the region. For instance, the horizontal PGA with a 10% probability of exceedance in 50 years, ranges between ∼(0.1–0.5)g in the north-eastern and south-western parts of the region, respectively. Moreover, the results indicate that the ad-hoc 1.5 factor to scale a 475-year return period (RP) spectrum for making the ground shaking similar to a 2475-year RP not only is unrealistic, but also creates substantially an underestimated ground shaking; thus, it seems too crude for the seismic design purposes. We have also developed some relationships, dependent on site conditions, for describing general relations between the obtained results in the return periods of 475- and 2475-years for PGA, SA (0.2s), and SA (1.0s). The computed hazard spectra have also been compared to the minimum acceptable level of seismic hazard in the region, according to the latest version of the Iranian national seismic hazard map. In addition, in order to make the results more transparent to the scientific and engineering community, the range of uncertainty contributed to the hazard modeling has also been explored in terms of a relative uncertainty metric. In total, the outcome of this study can make transparency into the picture of the seismic hazard level in central-east Iran and shed insight into possible developing plans.