SEISMIC TOMOGRAPHY METHODS AND SUBSURFACE STUDIES OF THE KUMAON HIMALAYA

Abstract

Seismic tomography evolved as one of the most important tools for both quantitative and qualitative determination and investigation of the internal structure of Earth including subducted slabs, sources of hotspots, convection pattern in the mantle and detailed subsurface velocity structure. This tool has been extensively used by researchers and engineers for setting up parameters required in making different structures such as dams, reservoirs and buildings as well as, for search of ground water. Mathematically tomography is an inverse technique which combines the idea of forward computation as well as inverse computation. Therefore the basis of tomography can be defined in terms of basics of inverse theory such as parameter representation, forward problem, inverse problem and analysis of robustness of solution. One of the most challenging problems that are exclusively related to tomography is accurate and fast solution of forward problem. The problem arises because of nonlinear relationship between velocity and ray path geometry. Many methods have been given in past to solve out different problems in this field. One of the major problems in the forward problem is computation of first arrivals as these phases are very crucial particularly for determination accurate velocity models and for resolving thin layers of a structural feature. Fast Marching Method (FMM) is a method that gives mathematical guarantee to solve this problem. Similar to other methods, the numerical problems are associated with this method. The source neighborhood errors cause a great problem in accuracy. Many authors in past have given several methods. Multi-Stencils Fast Marching (MSFM) Method is a method that mathematically address the problem more closely. However, this method solves the basic problems to some limited extent. In this thesis a new method called Multistencils Pseudoanisotropic Fast Marching Method (MPFMM), by extending the concept of MSFM method, has been developed and discussed. The developed method improves significantly the numerical errors associated with FMM and MSFM method. Conventional methods of ray path computations such as ray tracing are still used as these methods do not suffer from large numerical errors and provide better accuracy in computations. These are largely used in determination of preliminary 1D velocity models as well as 2D velocity models. One of the major problems associated with ray tracing is that this often fails to converge to true source-receiver ray paths and consumes time. This problem has been addressed in this thesis. A method has been developed for ray tracing which is parallel to shooting method of ray tracing. The method has been tied up with two techniques of ray path adjustments which have been named as Spiral Path Search Method iv and Gradient Path Search method. Both these method works in series and computes ray path very efficiently. Simultaneous inversion for the determination of a set of related parameters is always better than the inversion which considers only one parameter. There are many methods for computation of velocity and hypocenters with origin time. Determination of layered structure of earth is an important aspect of research because these are related with the earth’s feature called discontinuities in velocity or quality factor. In that context inversion scheme and algorithms have been developed which simultaneously determine body wave velocities, hypocenters with origin times as well as layer interfaces. In that inversion scheme both P-and S-wave data are considered to increase the amount of information from data. This has been done because tomography problems often face one practical problem that is the problem of scarcity of data. To remove this to some extent both P-and S-wave are inverted simultaneously in the scheme. The developed method has been given the name multiparameter inversion method. The multiparameter inversion method has been used for determination of 1D velocity structures, hypocenter parameters beneath the Kumaon Himalaya. The proposed MPFMM has been used to develop 2D and 3D traveltime tomography methods which have been applied successfully for the data of Kumaon. To study the attenuation structure of the same region, a methodology of 3D attenuation tomography using grid type of parameterization has been developed and discussed. Data of the present study have been taken from strong motion network deployed in the Kumaon part of Himalaya. A total of 870 first arrival P- and S-phases from a total of 116 earthquakes have been used in the study. The observation times of first arrival phases have been used as data for traveltime inversions. A total of 373 phases of first arrival P-waves have been used for determination of 2D P-wave velocity structure while a total of 497 first arrival S-phases have been used for determination of 2D S-wave velocity structure. In determination of 3D shear wave velocity structure, a total of 405 first arrival S-phases from a total of 98 earthquakes have been used. To study the 3D attenuation tomography a total of 344 S-phase spectra recorded by 17 stations from a total of 82 events have been considered. The located earthquakes form a shape that is oriented along Himalayan belt. Most of the epicenters form a group that lies between MCT and NAT in the areas of Baluakot, Dharchula and Joljibi. The number of earthquakes along MCT is fewer than that along MT. Most of the earthquakes are found to occur at shallow depths. Some of the earthquakes occurs at sub moho depth which support the idea that upper mantle deforms by brittle processes. v One dimensional velocity model shows the Indian moho beneath the Kumaon Himalayan region. The depth to the moho is ~50 km from mean sea level. This when tied up with other studies gives moho plane which has strike and dip respectively, N27E and 4.6° northeastern direction. At depth ~10 km P-wave velocity changes from ~5.7 km/s to ~6.1 km/s and S-wave velocity change from ~3.1 km/s to ~3.3 km/s. At depth ~50 km P-wave velocity changes from ~6.9 km/s to ~8.3 km/s and S-wave velocity change from ~4.1 km/s to ~3.7 km/s. 2D vertical body wave velocity images have been obtained from Lohaghat to northwestern side of Sobla, along a plane that strikes N26E. 2D Velocity models for body waves show the extensive disturbances of crust due to underthrusting of Indian plate beneath the Eurasian plate. The outcrops of crystalline complexes in Lesser Himalaya near Thal and Dharchula respectively are well resolved in 2D velocity sections especially in P-wave velocity section. The Conrad discontinuity lies around 12 km depth below mean sea level (msl). The P-wave velocity of upper crust in most of the portion in the section, vary from 4.2 km/s to 5.6 km/s and the same in lower crust vary from 5.6 km/s to slightly more than 7.8 km/s. The S-wave velocity in most of the upper crust in the section, vary from 2.9 km/s to 3.3 km/s and the same in lower crust vary from 3.3 km/s to slightly more than 4.4 km/s. A 3D shear wave velocity structure up-to a depth of 33 km beneath the Kumaon region has been obtained using the data and developed methodology of tomography using MPFMM. The obtained velocity structure clearly resolves the outcrops of crystalline complexes present in the study region. Alternating zones of low and high velocities have been observed with increasing depth in the upper crust. This may be related to the overturning nature of the crustal layers beneath the Kumaon Himalaya. Site amplification for all the station locations have been obtained and found some important characteristics. Askot and Didihat are the two areas which show gradual increase in site amplifications with the increases of frequencies, and Bageshwar and Muwani are two areas which show gradual decrease of site amplifications with the increase of frequency. Berinag and Kamedi Devi are the two areas which show almost frequency independent site amplification characteristics. 3D Attenuation tomography is performed up-to a depth of 33 km using the developed tomography method in the Kumaon Himalaya. The quality factor shows variation from near about 0 to 2300 beneath the study area. The crystalline complexes near Didihat and Dharchula are resolved and show quality factor (~650-700) against the background values of (~0-200). The variation of quality factor and wave velocity beneath the Kumaon Himalaya vi favours that subsurface layers beneath this region are suffered from overturning and probably rotation.