RISK ASSESSMENT OF IDENTIFIED SEISMO-TECTONICALLY SUSCEPTIBLE AREAS IN WESTERN HIMALAYA

Abstract

This research work was carried out with the following five objectives, viz. identification of the most hazardous seismogenic source zones, identification of seismo-tectonically susceptible areas, segmentation of longer tectonic units, viz., Main Boundary Thrust and Main Central Thrust, estimation of a spatio-temporal regime for predictive earthquakes in terms of probable location, probable time of occurrence and probable range of magnitude using the above three objectives and the assessment of seismic risk to property and life associated with the largest predictive earthquake. These five studies were carried out in a 12o by 12o study area with coordinates between longitude: 72oE to 84oE and latitude: 25oN to 37oN, in the western Himalaya. Seismicity and tectonic data were prepared through a rigorous exercise for the study area. A comprehensive earthquake catalogue was created by combining earthquake catalogues from the Indian Meteorological Department (IMD), the International Seismological Center (ISC), and the United States Geological Survey (USGS), homogenizing the various magnitude scales to the same scale, i.e., moment magnitude, and declustering the merged, homogenized earthquake catalogue to remove dependent seismic events from the independent seismic events. This resulted in Merged, Homogenized and Declustered (MHD) earthquake catalogue. The Merged, Homogenized and Declustered (MHD) earthquake catalogue was divided into two different catalogues for the same study area and was used for separate purposes. Merged, Homogenized and Declustered- 1 (MHD-1) earthquake catalogue and Merged, Homogenized and Declustered-2 (MHD-2) earthquake catalogue. MHD-1 consists of 2376 earthquakes within a magnitude range of 3.5 ≤ Mw ≤ 8.0, for the time-period 1501-2010 and was used for various analyses such as hazard assessment, identification of seismo-tectonically susceptible areas, and segmentation of longer lineaments. MHD-2 consists of 809 earthquakes within a magnitude range of 3.5 ≤ Mw ≤ 6.9, for the time-period 2011-2020 and was used for the validation of results. Tectonic data for the study area was used as per Narula et al. 2000 and Taylor et. al. 2003. The digitized tectonic map has 65 named tectonic units and 366 unnamed tectonic units. The seismotectonic map was prepared by superimposing seismicity data on tectonic data for the study area. x In the first study, study area was divided into twelve seismogenic source zones (SSZ). Seismicity data and tectonic data were used for dividing the study area into twelve seismogenic source zones (SSZ) and named them are as Kangra seismogenic source zone, Uttarakhand seismogenic source zone, Western Syntaxes seismogenic source zone, Kaurik seismogenic source zone, Kashmir seismogenic source zone, Western Tibet seismogenic source zone, Karakoram seismogenic source zone, Jhelum seismogenic source zone, Indo- Gangetic seismogenic source zone, Gonda Bijori seismogenic source zone, Bhatinda Amritsar seismogenic source zone, and Beng Co seismogenic source zone. Several hazard parameters required for the assessment of seismic hazard were computed. These parameters included Gutenberg- Richter recurrence law parameters, (‘a’, ‘b’ values); mean annual rate of exceedance, (m); Magnitude of completeness, (Mc); maximum observed magnitude, (Mmax, -obs); maximum calculated magnitude, (Mmax, -cal) and return periods for various magnitudes. The great Kangra earthquake occurred in 1905, and the next 8.0 magnitude earthquake is expected to occur between the years 2022 and 2150, based on hazard parameter computation. Similar computations were conducted for other seismogenic source zones. It was observed that the Kangra seismogenic source zone and the Uttarakhand seismogenic source zone are within the seismic gap defined by the Kangra earthquake of 1905 and the Bihar – Nepal earthquake of 1934. Therefore, an earthquake of high magnitude cannot be ruled out within these seismogenic source zones. The highest PGA was observed in the Kangra seismogenic source zone. This leads to the conclusion that the Kangra seismogenic source zone is the most vulnerable seismogenic source zone, followed by the Uttarakhand seismogenic source zone in the study area. In the second study, six major steps of the pattern recognition techniques were explained with its application in the field of identification of seismo-tectonically susceptible areas and the segmentation of longer tectonic units. These steps include identification of features, extraction of identified and selected features, categorizing the dataset based on several classification criteria, discriminant analysis, training exercise based on the dataset and decision-making exercise based on the results of training exercise. Initially, 24 features were identified from seismicity, tectonic and river data. Out of these 24 features, twelve features were finally used in different studies. Five classification criteria were used for the identification of seismo-tectonically susceptible areas and two classification criteria were used for the segmentation of longer tectonic units, i.e., the Main Boundary Thrust and the Main Central Thrust. xi Discriminant analysis was conducted for these seven classification criteria. These analyses were named as Pattern Recognition model -1 (PR model- 1) to Pattern Recognition model -7 (PR model- 7). Discriminant function obtained from Pattern Recognition model -1 to Pattern Recognition model -5 were used to identify the Pattern Recognition model with maximum percentage of classification of Class A and Class B. Pattern Recognition model -3 showed the maximum percentage of classification. Hence it was finally selected for training exercise and decision-making exercise. This resulted into the classification of study area into three types of seismo-tectonically susceptible areas viz. most susceptible area, moderately susceptible area, and the least susceptible area. These three types of seismo-tectonically susceptible areas were validated by using Merged, Homogenized and Declustered -2 earthquake catalogue. These earthquakes occurred after the completion of Merged, Homogenized and Declustered -1 earthquake catalogue and were not part of the training exercise. The Merged, Homogenized and Declustered-2 (MHD-2) earthquake catalogue, was considered for validation of results. Out of the 809 earthquakes, 513 earthquakes occurred in the most vulnerable area, while 283 occurred in the moderately susceptible area. It is worth noting that all four earthquakes of Mw ≥ 6.0 occurred in the most susceptible and moderately susceptible areas. The presence of smaller magnitude earthquakes alongside the larger magnitude earthquakes in the most susceptible area indicates that this area is seismically more active in the other areas. After identifying the three types of seismo-tectonically susceptible areas, Pattern Recognition technique was used for segmentation of longer tectonic units. Since long tectonic units do not rupture all at once rather, they are active in smaller segments, hence an attempt has been made for identifying the most active segments of the seismically active tectonic units. Crossover earthquakes were identified in the training exercise, and when these crossover epicentres were superimposed on the seismotectonic map it leads to a decision. Decision making exercise resulted in seven segments of Main Boundary Thrust and seven segments of Main Central Thrust. The seven segments identified for Main Boundary Thrust are: Poonch segment, Udhampur segment, Kangra segment, Solan segment, Dehradun segment, Nainital segment and Gonda Segment, and seven segments for Main Central Thrust are: Mashko segment, Chenab segment, Kinnaur segment, Uttarkashi segment, Bageshwar segment, Pithoragarh Segment and Pokhara Segment. These segments of Main Boundary Thrust and Main Central Thrust were further used to identify the location of the predictive earthquake. xii In central portion of the study area, both Main Boundary Thrust and Main Central Thrust seem to be seismically more active. As the Kangra and Solan segments of Main Boundary Thrust are within the truncated area of Kangra and Uttarakhand seismogenic source zones, these were retained for further narrowing down of the possible area where the most severe case of the predictive earthquake may be placed. The length of Kangra segment of Main Boundary Thrust in truncated area of Kangra seismogenic source zone is 114 km. Similarly, the length of Solan segment of Main Boundary Thrust in the truncated Kangra seismogenic source zone is 132 km, it is the eastward continuation of the Kangra segment upto the Solan district. This cumulative length of 246 km of Main Boundary Thrust is hereafter referred to as the Kangra - Solan segment of Main Boundary Thrust. The intersection of Kangra and Solan segments is estimated by the coordinates 76.92°E, 31.86°N on the surface and is used to locate a predictive earthquake. A predictive earthquake zone map was also prepared for the study area to assess seismic risk. Seismic risk was assessed for the predictive earthquake zones in terms of number of houses which will need either retrofitting or reconstruction and casualties expected in damaged houses. Populations living in predictive earthquake zones were at high seismic risk on account of high seismic hazard coupled with the vulnerable housing data. States within the predictive earthquake zones are, in alphabetical order: Chandigarh, Haryana, Himachal Pradesh, Jammu & Kashmir, Punjab, Uttarakhand and Uttar Pradesh. Ambala, Bilaspur, Chamba, Chandigarh, Hamirpur, Kangra, Kullu, Mandi, Panchkula, Rupnagar, Sirmaur, Solan, Sahibzada Ajit Singh Nagar and Shahid Bhagat Singh Nagar districts lie completely within the predictive earthquake zones. And Dehradun, Gurdaspur, Hoshiarpur, Fatehgarh Sahib, Jalandhar, Jammu & Kashmir, Kapurthala, Kinnaur, Kurukshetra, Lahul & Spiti, Ludhiana, Patiala, Shimla, Saharanpur, Tehri Garwal, Una, Uttarkashi and Yamunanagar districts lie partially within the predictive earthquake zones. When the maximum hazard level in the predictive earthquake zone VIPE was considered equivalent to the intensity IX as per European Macroseismic scale-98 (EMS 98), then total number of houses that need reconstruction were 1157371 and when a higher hazard in the predictive earthquake zone VIPE equivalent to intensity X as per European Macroseismic scale-98 (EMS 98) was considered then the number of houses that need reconstruction were increased to 1306977. An increase of approximately 13% was observed in the number of houses that will need special attention in future. The methodology of risk assessment is elaborated for Mandi xiii district because of the 32 districts of the predictive earthquake zones, Mandi district has observed the highest number of property and life loss. 87.02% of all existing houses in Mandi district are expected to suffer G5 and G4 grade of damage, and consequently will require either reconstruction or retrofitting. Moreover, if average population per house is 2.39, then 10767 casualties are expected in the district. And if occupancy was increased to 5, then life loss increased to 22525 casualties. The outcome of this study indicates that several disastrous earthquakes of various magnitudes are overdue in the study area. Hence, the present study, dealing with assessing seismic risk for predictive earthquakes, can be the basis and starting point of detailed planning for earthquake disaster mitigation.