from the Antarctica and Australia during Cretaceous (e.g. Kayal, 2008). The g. ( g y ,. ) continental collision at the Himalayan arc, between India and Asia, is an.
BODY WAVES ATTENUATION OF KUMAUN HIMALAYA
Mahak Singh Chauhan, A.R.Bansal • University of Naples ‘Federico II’ Naples, Italy Institute • National Geophysical Research Institute, Hyderabad, India
I t d ti Introduction y The Himalayan belt is formed due to the collision of Indian and Eurasian
plates l t in i the th period i d 50-55 50 55 million illi years ago. y This belt is seismically active and earthquake of varying magnitude are being observed. Most of these earthquakes are associated with great loss of lif and life d destruction d t ti off property. t For F understanding d t di off the th nature t off earthquakes and for reliable assessment of seismic risk in the Himalayan belt, knowledge and understanding of seismicity and the attenuation of strong ground motion are essential. essential y Due to the vast natural resources in the Himalayan region, development of the region is being planned (Tehri dam, Tunnels projects). y People P l living li i in i this thi region i are concerned d about b t their th i survival i l due d to t the th active seismicity of the region.
y In the light of the higher seismicity, an appraisal of the relation of
earthquake occurrences with geology and tectonics of the region is very essential to make an assessment of the seismic potentialities, for survival of the lives and natural resources, and in designing of the major structures. y The designing of earthquake resistant structures is a major challenge to the Civil Engineers. This challenge can be met if we develop ability to predict ground motion due to future earthquakes. y The important structure such as nuclear power plants, plants dams, dams and high-rise buildings require estimate of ground motion for earthquake resistant designing. y In the present work, we have made effort to understand the Body wave attenuation in Kumaun Himalaya using strong ground motion data.
G l Geology off the th Kumaun K Hi l Himalaya y Himalaya is a large geodynamic laboratory of nature where orogeny is still
in youth to early mature phases of evolution. y The Kumaun region of the Himalaya lies near the center of the Himalayan fold-and-thrust belt and is situated between the Kali River in the east and Sutlej in the west, including a 320 km stretch of mountainous terrain. This part of Himalaya exposes all the four major litho-tectonic subdivisions of the Himalaya from South to North. They are Sub-Himalaya, Lesser Hi l Himalaya, G t Himalaya Great Hi l and d Tethys T th Himalaya. Hi l All the th litho-tectonic lith t t i zones are bound on either side by longitudinally continuous tectonic surfaces such as Main Boundary Thrust (MBT), Main Central Thrust (MCT) South Tibetan Detachment (STD) system and Indus Tsangpo Suture (MCT), Zone (ITSZ) (Valdiya, 1980). y A generalized tectonic sequence for the Lesser Kumaun Himalaya (Valdiya, 1978) is tabulated. tabulated
Seismicity of the Himalaya y The geophysical data in Indian Ocean floor suggest that India got separated
from the Antarctica and Australia during g Cretaceous ((e.g. g Kayal, y , 2008). ) The continental collision at the Himalayan arc, between India and Asia, is an example of crustal thickening and mountain building on a large scale. y The Himalaya y is the g greatest mountain chain in the world,, 2500 km long g with loftiest peaks from west to east, viz, Naga Parbat (8,125 m), Everest (8,848 m) and Namcha Barwa (7,755 m). The width of the belt varies from 250 km to 350 km (Kayal, 2008). The uplift process is still going on at about one centimetre per year along with continued erosion and denudation, with different rates from time to time and place to place. y Seismicity of the Himalayas suggests that at present the Indian Plate underthrusts the Eurasian plate. Due to active movement of Indian plate, stress is developing and releasing which is main cause of the earthquakes in this region.
h recent years, ddamaging i strong earthquakes h k occurred d in i the h Himalaya. i l y In the e.g - Bihar-Nepal earthquake (MS 6.6) of August 20, 1988 in the Himalaya fore deep, Uttarkashi earthquake (mb 6.6) of September 30, 1991, Chamoli earthquake h k (mb ( b 6.3)) off March h 28, 1999 in i Garhwal h l Himalaya, i l Kashmir h i earthquake of October 8,2005 (M 7.6) and the Sikkim earthquake of September 18, 2011 (M 6.9). y The h main i Himalaya i l seismic i i belt b l (MHSB) ( S ) and d its i fore f d deep region i in i the h Indian Himalaya, between the western syntaxis and the eastern syntaxis, may be divided into three segment: (i) northwestern h and d western Himalayas, Hi l b between l i d 72◦-80◦E, longitude 72 80 E which hi h include Jammu & Kashmir (J & K) Himalaya and Himachal Pradesh in the north-western part, Garhwal and Kumaun Himalaya in the western part, (ii) the h centrall and d eastern Himalayas Hi l which hi h include i l d Nepal N l Himalaya Hi l i the in h central part and Sikkim and Darjeeling Himalaya in the eastern part, between longitude 80◦-90◦E, and (iii) the h northeastern h Hi l Himalaya which hi h include i l d Bhutan Bh and d Arunachal A h l Pradesh P d h between longitude 90◦-98◦E (Kayal, 2008).
METHODOLGY : Coda normalization method The coda normalization method is based on the idea that at lapse time, the seismic energy gy is uniformly y distributed in some volume surrounding g the source (Sato & Fehler, 1998). In this method spectral amplitude of the earthquake source is normalized by coda waves at a fixed lapse time. It is based on the idea that coda waves consist of scattered S waves from random heterogeneities in the Earth (Aki 1969; Aki & Chouet, 1975; Sato 1977). Roughly lapse time is taken twice of the direct S-wave travel time and spectrall amplitude li d off coda d at a lapse l time i tC , AC(f,t (f C), ) is i independent i d d off hypocentral distance r in the regional distance range, and can be described (Aki 1980)
the spectral p amplitude p of the direct S-wave, As(f (f,r)) , can be expressed p as
To normalize the spectral amplitude of S wave, eq. (2) is divided by (1):
Taking ki the h logarithm l i h off (3): (3)
Eq. (5) was first proposed by Aki(1980) to measure QS
Yoshimoto et al. (1993) extended this method for the measurement of QP by assuming that earthquakes within a small range of magnitude have the same spectral ratio of P- to S-wave radiation within a narrow frequency range f ± ∆f for different spectral shapes of P and S waves (Molnar, Tucker & Brune 1973; Rautian et al. 1978).
Where SP(f) is the source spectral amplitude of P waves. In this case AC(f tC) will also be proportional to SP(f) as AC(f,tC)
For P waves eq. 5 can be written as
Error analysis Suppose we have a linear equation as following:
Where m is the slope and c is the intercept. In case of least square fit of a straight line in observed data the standard error in slope (SEs) and standard error in intercept (SEi) can be estimated as :
Where
and
Where n is the number of points. points The misfit of the data can also be studied by estimating the degree of scatter or coefficient of determination (R2) which can be defined as fraction of the total variation in y explained by variation in x:
Where
ANALYSIS Data
A network t k off eight i ht strong t motion ti accelerographs l h off Kinemetrics, Ki t i USA, USA have been installed in the Kumaun, Himalaya under the major research project sponsored by the Department of Science and Technology/MOES, Government of India, India in March 2006. 2006
(a) Study area, (b) Location Map
The parameter of events recorded by the network.
Projection of earthquakes on different stations
A unprocessed accelerogram recorded at Dharchula.
Picked phase of P and S waves in a accelerograms recorded at Dharchula.
A 5 sec P wave window with and without zoom in top and below Figures respectively.
A 5 sec S wave window with and without zoom in top and below Figures respectively.
A 5 sec coda wave window with and without zoom in top and below Figures respectively.
Filtering Each seismogram was band-pass filtered by using a Butterworth filter with five pass-bands
On the filtered seismograms, we measured the root mean square amplitude (Ab b ki (Abubakirov & Gusev G 1990 Yoshimoto, 1990; Y hi t 2012, 2012 personall communication) i ti ) off direct P and S waves and coda spectral amplitude AC (f, tc) in a 5 sec time window for each frequency band. When the coda amplitude at 40s was lower than the noise level then we used value for earlier lapse time tC = 30 or 20 sec.
Filtered accelerograms with band pass of frequency 1‐2 Hz.
Filtered accelerograms with band pass of frequency 2‐4 Hz.
Filtered accelerograms with band pass of frequency 4‐8 Hz.
Filtered accelerograms with band pass of frequency 8‐16 Hz.
Filtered accelerograms with band pass of frequency 16‐32 Hz.
Estimation of QP and QS •We substituted the observed amplitudes and hypocentral distances values into equations (5) and (8), to estimate QP and QS. •The average velocities of P and S waves in the lithosphere are used as VP = 6.5 km/s (Sharma, 2008) and VS = 3.5 km/s (Joshi, 2006). • In equation (5) and (8) ∆r is also a crucial parameter. We used different values of ∆r varies between 5 to 20 km and found almost same value of QP andd QS. The Th only l difference diff i that is th t degree d off scattering tt i R2 improves i with ith higher values of ∆r. •So we fixed ∆r =20 km for further analysis. • We also assumed that geometrical spreading is proportional to r -1.
RESULT AND DISCUSSION QP and d QS Average value of QP and QS at different central frequency.
Plots of QP and QS values frequency for whole Kumaun region.
C l i Conclusion y The strong motion data of digital network in Kumaun Himalaya is analyzed
y y y
y
y
from 2006 to 2008 in this study. study Nine earthquakes are used from the data set which are recorded more than three stations with large S/N and less location errors. QP and QS in the Kumaun Himalaya region are found to be strongly frequency dependent and increase with frequency. The QS/QP ≥ 1 is found for all frequency range. Th low The l values l off QP and d QS correspond d to t seismically i i ll active ti areas with ith tectonic complexity due to the ongoing convergence between Indian and Eurasian plate. It is i found f d that th t the th attenuation tt ti is i stronger t f P wave than for th S waves for f the th entire frequency range and this probably reflects the high degree of heterogeneity presence in the crust of Kumaun Himalaya. O results Our lt are well ll comparable bl to t the th other th tectonically t t i ll active ti regions i characterized by high degree of heterogeneity reported globally.
