Intrinsic and scattering attenuation in western India from aftershocks of the 26 January, 2001 Kachchh earthquake

176 vertical-component, short period observations from aftershocks of the M_w 7.7, 26 January, 2001 Kachchh earthquake are used to estimate seismic wave attenuation in western India using uniform and two layer models. The magnitudes (M_w) of the earthquakes are less than 4.5, with depths less than 46 km and hypocentral distances up to 110 km. The studied frequencies are between 1 and 30 Hz. Two seismic wave attenuation factors, intrinsic absorption (Q_i^− 1) and scattering attenuation (Q_s^− 1) are estimated using the Multiple Lapse Time Window method which compares time integrated seismic wave energies with synthetic coda wave envelopes for a multiple isotropic scattering model. We first assume spatial uniformity of Q_i^− 1, Q_s^− 1 and S wave velocity (β). A second approach extends the multiple scattering hypothesis to media consisting of several layers characterized by vertically varying scattering coefficient (g), intrinsic absorption strength (h), density of the media (ρ) and shear wave velocity structure. The predicted coda envelopes are computed using Monte Carlo simulation. Results show that, under the assumption of spatial uniformity, scattering attenuation is greater than intrinsic absorption only for the lowest frequency band (1 to 2 Hz), whereas intrinsic absorption is predominant in the attenuation process at higher frequencies (2 to 30 Hz). The values of Q obtained range from Q_t = 118, Q_i = 246 and Q_s = 227 at 1.5 Hz to Q_t ≈ 4000, Q_i ≈ 4600 and Q_s ≈ 33,300 at 28 Hz center frequencies, being Q_t^− 1 a measure of total attenuation. Results also show that Q_i^− 1, Q_s^− 1 and Q_t^− 1 decrease proportional to f^−ν. Two rates of decay are clearly observed for the low (1 to 6 Hz) and high (6 to 30 Hz) frequency ranges. Values of ν are estimated as 2.07 ± 0.05 and 0.44 ± 0.09 for total attenuation, 1.52 ± 0.21 and 0.48 ± 0.09 for intrinsic absorption and 3.63 ± 0.07 and 0.06 ± 0.08 for scattering attenuation for the low and high frequency ranges, respectively. Despite the lower resolution in deriving the attenuation parameters for a two layered crust, we find that scattering attenuation is comparable to or smaller than the intrinsic absorption in the crust whereas intrinsic absorption dominates in the mantle. Also, for a crustal layer of thickness 42 km, intrinsic absorption and scattering estimates in the crust are lower and greater than those of the mantle, respectively.