CONDITIONAL PROBABILITIES OF STRONG GROUND MOTION IN THE HIMALAYA

Abstract

Seismic hazard Assessment (SHA) is the first step while quantifying the risk posed to a region due to the seismic activities in an area. It depends on the geologic and tectonic setup of the region, comprehensive data sets, understanding of the earth’s inner structure and the physical process, and the use of statistical and mathematical models.In this work, the SHA has been carried out for the Himalayan region using Time-dependent statistical models. Interest in and pertinent studies related to the seismic hazard in the Himalayan region have increased significantly since the instrumental era, and each damaging earthquake encourages its re estimation. In the Himalayan arc, the pattern of stress accumulation and its subsequent release has never been observed to be uniform. The Himalaya is having a discordant type of seismicity, resulting in both seismically very active areas as well as seismic gaps, which continuously accumulate stresses over a long period without any release and hence, a singular mathematical or statistical model cannot describe the seismicity of the entire mountain belt. Hence, in the present workfour Time-dependent models namely, Weibull, Log-Normal, Gammaand Inverse Gaussian distributions, are used to estimate the seismic hazard. For this, the study area is subdivided into four Zones viz. the Northwestern Himalayas (Zone 1), the Central Seismic Gap Region (Zone 2), the Eastern Nepal and Sikkim (Zone 3), and the Eastern Himalayas (Zone 4), by considering approximately 380 years of seismological data (earthquake catalogue). The data has been compiled from year 1250-2018, followed by its sensitive treatments (Homogenization, Declustering and, Completeness), which is a prerequisite for precise estimation of the seismic hazard. The seismological data has been used for estimating the earthquake probabilities. The non-Poissonian (or Time-dependent) probabilities of exceedance of magnitudes in a specified time in the future were investigated for different elapsed times based on the recurrence time intervals of past earthquakes in the Himalaya, using the selected models. The suitability of each stochastic model for each zone was estimated using the Kolmogorov Smirnov (K–S) test, to describe the different physical processes responsible for earthquake occurrence. The results show that the Gamma, Inverse Gaussian, Lognormal, and Inverse Gaussian models were most suitable for Zones 1, 2, 3 and 4, respectively, for Mw≥6.0. The cumulative probability of recurrence intervals reaches up to 90% in 30 years for Zones 1 and 2, 49 years for Zone 3, and 41 years for Zone 4. The estimated conditional probability reaches 90% in 30 years in Zone 1, 35 years in Zone 2, and 50 years in Zones 3 and 4. The most suitable models for Mw≥7.0 were found to be the Log-normal for Zones 1, 2, and 3 and Gamma for Zone 4. The cumulative probability for Mw≥7.0 reaches 90% in 85 years in Zone 1, 86 years in Zone 2, 93 years in Zone 3, and 148 years in Zone 4. The estimated conditional probabilities i reach 90% in 80 years in Zone 1, 90 years in both Zone 2 and 3, and 150 years in Zone 4. The comparative studies with Poissonian Modelling, where the hazard rate is assumed to be constant with time, reveal that model suitability be evaluated in complex regions such as the Himalaya before proceeding with seismic hazard assessments. The heterogeneous and complex tectonics of the Himalayas, differentiated plate motions, different stress release patterns (spatially and temporally), and locking/unlocking of faults/thrusts are some of the reasons responsible for the different probabilistic models describing the earthquake occurrence phenomena in these four zones. Once, the earthquake recurrence probabilities have been estimated using Time-dependent models, the Probabilistic Seismic Hazard Assessment (PSHA) has been carried out for the Himalaya to look the effect of such modelling on strong ground motion. For estimating the PSHA, an integrated approach has been developed, in which the earthquake probabilities for Mw<6.0 is estimated using Time-independent models and for Mw≥6.0, Time-dependent probabilities have been used. The hazard maps showing PGA at 2% and 10% probability of exceedances have been prepared using Sharma et al. (2009). In this study, various types of magnitude recurrence models have been applied according to seismotectonic environments. The seismic hazard vis-à-vis the model applicable for a region has been interpreted in terms of the physical process of earthquake occurrence. The suitability of different types of models for acquiring the earthquake recurrence probabilities in seismically varying sources is informative and useful from an engineering point of view, and most certainly from a SHA perspective. The following are the main conclusions drawn from the present research work: 1. The analyses of Time-dependent seismic hazard assessment reveal discordant patterns of energy release in terms of earthquakes in the Himalayas, which shows that the Himalayan belt is not behaving uniformly over its entire length. Thus, Time independent models are not sufficient to enunciate the realistic seismic processes that generate earthquakes. In such models, the occurrence of large earthquake event is not represented well which restrain this model to reflect the real physical process. 2. A case study for four distinguished regions of the Himalayan belt has concluded that different models fit in different regions for prediction of seismic hazard, which may be attributed to different physical process of accumulation of stress and its releasing pattern. The transverse features have played important roles to dissect the regional features and forcing the individual blocks to behave differently in releasing the strain. ii 3. There are numerous hazard studies carried out in the past specifically for seismic gap region which predicts a major earthquake event in near future. Contrary to such believe, fitting of different models (other than Poissonian) reveals that this might not be the case and can be interpreted as either such large energy is not accumulating or it is getting released in some other form viz. slow or quiet earthquakes.