Seismic Attenuation of the Nepal Himalaya using Coda Q tomography

We use 663 local waveforms, to study the seismic attenuation of the Nepal Himalaya, through estimation of temporal decay of coda wave amplitude. We assume a single back scattering model for the coda waves, which samples an elliptical area with the earthquake source and the receiver at the two focii. Our data samples the region from 23°N to 35°N and 76°E to 94°E. The frequency dependence of Q(f) has been computed for each coda waveform by estimating the Q at eight central frequencies between 1 and 14 Hz using 𝐴(𝑓,𝑡) = 𝑆 (𝑓) 𝑡-𝛽 ⅇ𝜋𝑓𝑡𝑄(𝑓), and the power law 𝑄=𝑄0 ( 𝑓 / 𝑓0 ) 𝜂. The coda lapse time was chosen at twice the S-wave arrival time (2ts) with a sampling window of 30 s. The estimated values of Q0 shows a large variation from 739 to 22 while 𝜂 varies from 1.02 to 1.20. 2-dimensional attenuation tomography was computed for the region using a grid size of 2° by 2°, to highlight the lateral variation in Q and η. Our results reveal two distinct regions, one with low Q0 in the Himalaya and the other with high Q in the adjacent Indian Shield, while 𝜂 shows large lateral variations. We perform a point spreading function resolution test to estimate the resolution of our tomography results. These reveal different degrees of amplitude recovery, while the shape recovery is satisfactory in the region of our interest.