STUDY OF INTRINSIC AND SCATTERING ATTENUATION IN THE GARHWAL HIMALAYA

Abstract

The seismic wave attenuation in high frequency range (from 1 Hz to 30 Hz) has emerged as one of the promising field for indepth understanding of the seismotectonic characteristics of a region. The knowledge of seismic wave attenuation has wide ranging applications. Primarily, it is required for the precise estimation of earthquake source parameters by applying the path correction. Secondly, it is one of the major requirements in the prediction of strong ground motion at a site due to probable earthquakes in the region. The temporal and spatial variation of seismic wave attenuation in a prescribed spatial window has also been studied as an earthquake precursor. The seismic wave attenuation characteristics of the region is determined by the estimation of non-dimensional parameter, Q, called the quality factor. It is measured from the decay of seismic wave amplitudes with distance or time of different wave groups such as, P-waves (Q), S-waves (Q), Coda waves (Q) and Lg waves (QLg). In the present study, the seismic wave attenuation characteristics has been measured of the Uttarkashi and Chamoli regions in the Garhwal Himalaya from the Coda waves and S-waves of local earthquakes - and separated the intrinsic and scattering effect from the total attenuation. Aki (1969), first time addressed the importance and characteristics of coda waves of local earthquakes. Subsequently, Aki and Chouet (1975) devised the single back-scattering model to describe the properties of coda waves of local earthquakes and estimated the quality factor of coda waves (Q) for the central California region in USA and western Japan. Since then a number of studies have been carried out world over to estimate Qc of tectonically active and stable regions and established its relation with the scismicity and tectonics of the region. S-waves part of seismograms of local earthquakes has also been used to estimate the quality factor of S-waves (Q) using the single station ratio method given by Aki (1980). The Garhwal Himalaya, which forms the part of Northwest Himalaya, is seismically very active region. Two moderate sized earthquakes, the Uttarkashi earthquake of 1991 (mb=6.6) and the Chamoli earthquake of 1999 (mb6.4) have occurred in the region in last two and half decades. In the present study, the seismic wave attenuation characteristics of high frequency seismic waves of two sub-regions of the Garhwal Himalaya namely, Uttarkashi and Chamoli has been conducted from the analysis of Coda waves and S-waves of local earthquakes. Following are the brief account of the various components of the present study. To estimate the quality factor of Coda waves, Q. of Uttarkashi and Chamoli regions of the Garhwal Himalaya and to develop frequency dependent attenuation relations for both the regions. To investigate the lapse time dependence of Q with a view to study the variation of Qc with depth. To estimate the quality factor of S-waves, Q, of Uttarkashi and Chamoli regions of the Garhwal Himalaya using local earthquakes and to develop frequency dependent attenuation relationships for both the regions. To compare the attenuation characteristics of both the regions. To decompose the total attenuation into the intrinsic and scattering attenuation for both the regions and to study its tectonic implication. The single back-scattering model given by Aki and Chouet (1975), has been - adopted to estimate the quality factor of coda waves, Q,, whereas, the single station ratio method, given by Aki (1980a), has been employed to estimate the quality factor of S-waves. Q. According to the single back scattering model, the coda waves are considered as superposition of backscattered shear waves generated from numerous inhomogenities present in the crust and upper part of the mantle. These waves arrive long after the arrival of all direct waves and hence provide the average attenuation characteristics of the given region. The single station ratio method makes use of decay of maximum amplitude of S-wave normalized by the coda waves observed at the single station with the hypocentral distance due to a number of local earthquakes falling along a same azimuth. Therefore, this method provides the attenuation characteristics of a single path between the recording station and the earthquake sources. Separation of intrinsic and scattering attenuation effect from the 11 measurement of total attenuation has been carried out based on the methodology given by Wennerberg (1993). The local earthquakes occurred around both the regions during 2008-2012 have been used to study the seismic wave attenuation characteristics. For coda waves analysis, 1499 seismograms from 159 local earthquakes recorded during 2008-2009 have been considered for estimation of Q. These events have been divided into two groups. The first group of 96 events occurred in the epicentral distance range from 5.0 to 94.0 km, focal depth range from 1.6 to 42.0 km, and coda magnitude range from 0.2 to 3.0 are located around Uttarkashi region whereas, second group of 63 events occurred in the epicentral distance range from 20 to 109 km, focal depth range from 1.4 to 41 km, and coda magnitude range of 1.0 to 3.0 are located around Chamoli region. For S-wave analysis, 127 local events occurred during 2008-2012 and recorded at Rajgadi (RAJ) station have been used for estimation of Q. The local magnitudes, focal depths and hypocentral distances of these 127 events lie in the range of 0.3ML3.2, 2.8 km to 41 km and 10 km to 210 - km respectively. These events are confined in the selected azimuthal window. This data-set is also divided into two groups. The first group of 91 events occurred around Uttarkashi and confined in the hypocentral distance range between 10 and 80 km whereas, the second group of 36 events around Chamoli occurred in hypocentral distance range between 80 and 210 km with respect to RAJ station. The quality factor of coda waves, Q. have been computed in the frequency range from 1.5 to 24 Hz in the selected coda window. The seismograms at different central frequencies were filtered adopting the Butterworth band-pass filter of eight-pole. Then RMS amplitudes are calculated for smoothening the envelope of given coda window. Logarithmic of geometrically corrected RMS amplitudes of filtered coda waves at each frequency are plotted as a function of lapse time and slope of the best fit linear equation provides the Coda Q (Q) at particular central frequency. In this way, the coda waves of 30 sec duration of 1499 seismograms in four to five - lapse time windows range viz., 10-40, 20-50, 30-60, 40-70 and 50-80 sec are analysed at seven frequency bands for the estimation of Q in Uttarkashi and 111 epicentral distance the level of heterogeneities seems to be more pronounced around the Uttarkashi region as compared to the Chamoli region. At higher depth and longer epicentral distance the trend is reversed. The degree of frequency dependence, r, that represents the level of tectonic activity shows that among both the studied regions, the Uttarkashi region seems to be tectonically more active as compared to the Chamoli region up to depth of 96 km as n-value in the Uttarkashi region decreases from 1.20 to 1.03 as compare to the Chamoli region where it remains stationary at 0.95. Further, the level of tectonic activity around Uttarkashi region seems to be high in the shallow crust that decreases with increasing depth whereas, around Chamoli region the level of tectonic activity remains stationary with increasing depth. This coincides with the high rate of occurrence of small earthquakes around the Uttarkashi region at very shallow depth (< 10 km) in the close proximity of Rajgadi and Giyanja stations. A comparison of Qc' obtained (only for LTW of 30 sec) in this study with that obtained for other tectonic regions in India brought out that the North-East region of India, which is considered - seismically most active, exhibited the medium of high attenuation whereas, the Indian shield region, which is considered seismically stable region, showed the medium of low attenuation as compared to the other Indian regions. Attenuations of other declare regions of India are found to fall in between both these regions. The analysis of S-waves for the Uttarkashi region (hypocentral distance upto 80 km) and for Chamoli region (hypocentral distance from 80 to 210 km) has brought out Qp estimates from 32 at 1.5 Hz to 630 at 24.0 Hz for Uttarkashi region and vary from 54 at 1.5 Hz to 1670 at 24.0 Hz for Chamoli region. This showed that Qp is a function of frequency at different hypocentral distances. For the Uttarkashi 1.05±0.29 region, the frequency dependence Q relation, Q (18+8.4)f , and for Chamoli region, Q (26110.64)f'20123, have been obtained. Both the relations represent S-wave attenuation properties of the shallow or upper part of the crust upto 20-25 km depth underneath the studied region connecting the path of S-wave from earthquake sources (epicenters) to the recording station. Thus, Qp estimates have brought out that the S-wave attenuation around Uttarkashi region is higher as V compared to Chamoli region. A comparison of Q estimates for both the regions showed Uttarkashi region is more attenuating as compared to Chamoli region. This finding validates the spatial variation of Q. The separation of intrinsic attenuation (Q") and scattering attenuation (Q1), - from the measurements of Qc and Qp for the Uttarkashi and Chamoli regions of the Garhwal Himalaya has been carried out employing Wermerberg (1993) methodology. The results showed that for the Uttarkashi region, Qi' estimates vary from 0.010 at 1.5 Hz to 0.0005 at 24.0 Hz and QS1 estimates vary from 0.021 at 1.5 Hz to 0.001 at 24.0 Hz (for LTW 30-60 sec) and for the Chamoli region. Q1 estimates vary from 0.009 at 1.5 Hz to 0.0005 at 24.0 Hz and QS1 estimates vary from 0.010 at 1.5 Hz to 0.0001 at 24.0 Hz (for LTW 30-60 sec). The seismic albedo (130), computed based on Q1 and QS', is found to be greater than 0.5 for 1.5 Hz to 24.0 Hz for Uttarkashi. This signifies that the scattering attenuation dominants over all frequencies in this area. However, for the Chamoli region, the seismic albedo is greater than 0.5 for frequency range from 1.5 Hz to 12.0 Hz and less than 0.5 for the frequency range from 12.0 Hz to 24.0 Hz. This signifies that for the Chamoli region scattering dominants over the frequency range from 1.5 Hz to 12.0 Hz while the intrinsic attenuation dominants in frequency range of greater than 12.0 Hz to 24.0 Hz.