PROBABILISTIC SEISMIC HAZARD ASSESSMENT OF NORTHWEST AND CENTRAL HIMALAYAN REGION

Abstract

The Himalayan arc occurring at the northern most part of the Indian continent is a part of an extensive Alpine-Himalayan convergent tectonic belt and remains very active seismically. Very strong to moderately strong earthquakes have already occurred in this mountainous belt. In view of its active seismic status and continued developmental activities in mountainous states it is imperative to update knowledge of seismic hazard incorporating latest knowledge on earthquake occurrence and attenuation process. Seismic input is essential for the engineering design of the significant structures in terms of expected ground motion at a site along with the return period. This can be achieved through carrying out seismic hazard assessment employing Probabilistic Seismic Hazard Assessment (PSHA) approach. This approach provides quantified seismic hazard assessment along with the information on uncertainty. Further, the knowledge of seismic hazard assessment assumes more importance due to the high level of natural seismicity and future earthquake potential as indicated by several research works. This research work has been carried out with the aim to deliver new generation PSHA information for northwest and central Himalayan region. Adopted procedures include improved magnitude conversion techniques, updated attenuation relations and implementation different hazard computation approaches. PSHA for northwest and central Himalayan region constitute three methods. The first method is known as Cornell’s (1968) zoning, second is Bungum’s (2007) moment slip and third is Woo’s (1996) zone free. The seismic activity rate are estimated employing different techniques as suggested by Cornell (1968), Bungum (2007) and Woo (1996) and in the later part of processing standard probabilistic empirical relations are employed. In Cornell’s (1968) zoning method it is required to carry out identification and delineation of seismogenic zones. For this purpose at first detailed and comprehensive understanding of seismotectonic set up of the study area has been obtained. The Himalayan orogenic belt of the world is endowed with significant tectonic features viz., thrusts, faults and folds and distinct pattern of earthquake occurrence. These tectonic features have formed in response to convergent tectonism prevailed for several million years and associated seismic activities. Detailed information about the earthquake source mechanism and quantified ii information are only available since last few decades. Based on the current knowledge on tectonics and seismicity seismogenic zones are delineation however, always remains challenging as it involves an element of human judgment. In the research work due consideration was given to ever changing tectonics along the length and breadth of the Himalayan belt. Seismogenic zones were delineated on the basis on earthquake clusters and nature of tectonic features combined and the boundary of zones are drawn following the natural break in seismicity and tectonic features. However more emphasis was given either to tectonic features or seismicity when the adopted criteria were not very clear. In general the Himalayan belt exhibits compressional tectonic features but with exception in the area of Kaurik Fault System which is marked by extensional tectonism and delineated as separate zone. After the delineation of seismogenic zones, earthquake catalog was extracted from respective of seismogenic zones and were used to estimate seismicity parameters and mean annual rate of exceedance. An attempt has been made to assess the effect of size and dimension of seismogenic zones on the hazard result and for this purpose three different situations have been considered such as smaller, intermediate and larger seismogenic zones. In order to know the effect of seismic catalog duration only for smaller sized seismogenic zones another attempt has been made by considering (1) total dataset duration including both historical and instrumental and (2) dataset only from 1964 to 2012 i.e. only instrumentally recorded. Whereas, for intermediate and large sized seismogenic zones both historical and instrumental dataset have been considered. Ground Motion Prediction Equations (GMPEs) plays significant role as the predicted intensity level at a site is governed by the suitability and applicability of the derived empirical relations. After careful examination of the available set of GMPEs, selection of the GMPEs suggested by Zhao et al. (2006), Boore and Atkinson (2008), Campbell and Bozorgnia (2008), Chiou and Youngs (2008) and Akkar and Bommer (2010) have been done to accomplish this research work. These equations are up to date as they incorporate large data set and regressions and also are found to be suitable for the Himalayan region. These global attenuation equations are for active shallow crustal regions and equal weights were assigned to all equations for computation purpose. iii The study area of northwest and central Himalaya and its surrounding region between latitude 27° 2´ N to 35° N and longitude 72° E to 88° E has been divided into small grids of size 0.2° X 0.2°. Peak Ground Acceleration (PGA) and Peak Spectral Acceleration (PSA) at different time periods have been computed adopting approach used in the CRISIS program as discussed in each section of that chapter. PGA has been computed at the center of all the grid points for return periods of 475, 975 and 2475 years (i.e., 10%, 5% and 2% probability of exceedance in 50 years) and ground motion distribution is shown in the form of zones considering varying b-value for each zone. For the small sized seismogenic zones involving 22 zones earthquake dataset from 1964 to 2012 were used and computed seismicity parameters for each zone. In this case, estimated PGA values for return period of 475 years vary from 0.06g to 0.36g for northwest and central Himalayan region. Secondly for the same small sized seismogenic zones seismic dataset from 1552 to 2012 including both historical and instrumental time were considered. In this case, estimated PGA values vary from 0.05g to 0.31g for the study area. Intermediate sized seismogenic zones include the dataset from 1552-2012 in which 13 seismogenic zones have been considered. For return period of 475 years, the PGA values vary from 0.07g to 0.27g. Large sized seismogenic zones incorporate the dataset from 1552-2012 and six large zones identified in this case and the estimated PGA values vary from 0.02g to 0.25g. In the second method the mean annual rate of earthquake occurrence calculated from the moment slip rate. Small sized seismogenic zones (22 zones) have been considered for seismic hazard assessment. In this case, for return period of 475 years, the PGA values vary from 0.1g to 0.45g for northwest and central Himalayan region. The third one is the Kernel Density Estimation method also known as zone free method employed first time for the Himalayan region. This methodology uses complete earthquake catalog (from 1552 to 2012) irrespective of foreshocks and aftershocks and does not require zoning. As a result this method is not marred by the problem correctness in estimation of magnitude, location, depth, rupture mechanism and judgment used in seismogenic zone delineation. PSHA of northwest and central Himalaya has been performed using the kernel density estimation of seismic activity rate method developed by Woo (1996). In this method the activity rate at each grid location has been carried out using kernel iv estimation method by using all the un-declustered dataset. Employing this approach the estimated PGA values found to vary from 0.02g to 0.33g for return period of 475 years. It is noteworthy to mention that the zone free and moment slip approaches for seismic hazard assessment in probabilistic terms has been employed for the first time for Himalayan region in this research work. Under zoning approach three different geometry and dimensions for seismogenic zones were attempted and it is found that the smaller sized seismogenic zones resulted highest ground motion hazard zones at three different locations and for the intermediate and larger sized zones the highest ground motion hazard zones are two and one respectively with increasing dimensions. The annual rate of exceedance for the seismogenic zones plays important role in prediction of ground motion and it is generally higher if annual rate is higher. Estimated spectral acceleration levels are comparable between the results of intermediate sized zones in zoning method and zone free method. However, moment slip predicted higher level of ground motion. Zone free method predicted maximum ground motion in two areas falling in Kashmir valley and eastern part of Uttarakhand and westernmost part of Nepal. Ground motion ranges from 0.547g to 0.609g for 2475 return period. Anisotropic parameter of the kernel used in this study controls shape of predicted hazard zones as the hazard zone took a elongated shape following the Himalayan structural features and geometrical structure of seismicity. The adopted moment slip approach predicted higher order of ground motion ranging from 0.687g to 0.749g for 2475 years return period in the westernmost part of Nepal. Higher ground motion estimated because there are uncertainties in calculation of fault length which tend to increase the maximum magnitude. Highest predicted ground motion of about 0.6g for MCE and 0.36g for DBE conditions as predicted in this research work appears to be meaningful and justifiable in view of expected earthquake potential of the Himalayan tectonic belt. The strong ground motion estimated at the bed rock level can be used for the estimation of ground motion at the surface wherever required by incorporating the site amplification factor. The derived information will be useful in engineering design and microzonation work.