5. Attenuation (G and Kappa) and source parameters for regional earthquakes in the
5.3.1 The main regression analysis
Results of the main regression are presented in Table 5.1. The far-source geometric attenuation rate b has resulted in negative values for the selected frequency range. Values are approximately between -0.1 and -0.6, and the average is close to -0.5.
Table 5.1 Results of the main regression [using equation (5.5) or (5.7)] for the selected frequency range (see end of the table for definitions of terms)
f (Hz) Regression coefficients for equation (5.7) σest R2 SS(reg) SS(resi) C1 σ C2 σ b σ CT σ
0.5 0.577 0.10 -0.196 0.18 -0.1 0.19 -2.474 0.56 0.22 0.72 18.81 7.17 0.6 0.529 0.09 -0.069 0.17 -0.2 0.17 -2.264 0.50 0.20 0.78 19.96 5.76 0.7 0.443 0.10 0.078 0.18 -0.1 0.19 -2.353 0.56 0.22 0.73 19.08 7.19 0.8 0.381 0.09 0.223 0.18 -0.2 0.18 -1.970 0.55 0.22 0.74 19.94 6.88 0.9 0.375 0.09 0.253 0.17 -0.3 0.17 -1.621 0.53 0.21 0.76 20.54 6.38 1 0.375 0.09 0.223 0.18 -0.3 0.18 -1.547 0.55 0.22 0.74 19.29 6.80 1.1 0.378 0.10 0.169 0.18 -0.2 0.18 -1.719 0.56 0.22 0.71 17.51 7.16 1.2 0.383 0.10 0.116 0.19 -0.2 0.19 -1.889 0.58 0.23 0.67 15.90 7.72 1.3 0.352 0.10 0.152 0.19 -0.2 0.19 -1.817 0.58 0.23 0.66 15.04 7.64 1.4 0.337 0.10 0.172 0.18 -0.2 0.19 -1.753 0.56 0.23 0.67 14.70 7.23 1.5 0.335 0.10 0.202 0.18 -0.3 0.18 -1.424 0.56 0.22 0.69 15.52 7.07 1.6 0.348 0.09 0.187 0.18 -0.3 0.18 -1.419 0.54 0.22 0.71 15.90 6.63 1.7 0.332 0.09 0.220 0.18 -0.3 0.18 -1.406 0.54 0.22 0.71 16.03 6.64 1.8 0.312 0.09 0.250 0.18 -0.3 0.18 -1.202 0.53 0.21 0.71 15.63 6.48
93
1.9 0.285 0.09 0.280 0.17 -0.3 0.17 -1.198 0.53 0.21 0.70 14.87 6.35 2 0.266 0.09 0.307 0.17 -0.4 0.17 -1.005 0.51 0.21 0.71 14.43 6.03 2.1 0.273 0.09 0.286 0.16 -0.4 0.16 -0.828 0.50 0.20 0.71 14.17 5.66 2.2 0.282 0.08 0.274 0.16 -0.4 0.16 -0.759 0.49 0.20 0.72 14.29 5.47 2.3 0.275 0.09 0.277 0.16 -0.5 0.16 -0.728 0.50 0.20 0.71 13.95 5.68 2.4 0.250 0.09 0.325 0.17 -0.5 0.17 -0.560 0.51 0.20 0.70 13.94 5.86 2.5 0.227 0.09 0.352 0.17 -0.5 0.17 -0.560 0.51 0.20 0.69 13.41 5.91 2.6 0.219 0.09 0.345 0.17 -0.5 0.17 -0.654 0.50 0.20 0.69 12.63 5.77 2.7 0.225 0.09 0.318 0.16 -0.5 0.16 -0.631 0.50 0.20 0.68 12.14 5.64 2.8 0.224 0.08 0.312 0.16 -0.5 0.16 -0.578 0.49 0.19 0.69 11.86 5.38 2.9 0.239 0.08 0.276 0.16 -0.4 0.16 -0.658 0.48 0.19 0.69 11.63 5.22 3 0.248 0.08 0.247 0.15 -0.4 0.15 -0.623 0.46 0.19 0.70 11.25 4.90 3.1 0.268 0.08 0.201 0.15 -0.4 0.15 -0.593 0.46 0.18 0.70 11.03 4.73 3.2 0.286 0.08 0.166 0.15 -0.4 0.15 -0.520 0.46 0.18 0.70 11.02 4.83 3.3 0.283 0.08 0.166 0.15 -0.5 0.15 -0.397 0.47 0.19 0.69 10.88 5.00 3.4 0.288 0.08 0.154 0.16 -0.5 0.16 -0.351 0.48 0.19 0.67 10.81 5.25 3.5 0.289 0.08 0.144 0.16 -0.5 0.16 -0.236 0.49 0.20 0.66 10.53 5.48 3.6 0.296 0.09 0.119 0.16 -0.5 0.16 -0.177 0.50 0.20 0.65 10.24 5.59 3.7 0.292 0.09 0.120 0.16 -0.5 0.17 -0.196 0.50 0.20 0.64 10.03 5.72 3.8 0.297 0.09 0.106 0.16 -0.5 0.16 -0.216 0.50 0.20 0.64 9.89 5.67 3.9 0.305 0.09 0.091 0.16 -0.5 0.17 -0.194 0.50 0.20 0.63 9.89 5.72 4 0.301 0.09 0.095 0.16 -0.5 0.17 -0.199 0.50 0.20 0.63 9.80 5.73 4.1 0.295 0.09 0.099 0.16 -0.5 0.16 -0.275 0.50 0.20 0.63 9.56 5.60 4.2 0.311 0.09 0.060 0.16 -0.5 0.17 -0.295 0.50 0.20 0.62 9.39 5.70 4.3 0.321 0.08 0.035 0.16 -0.5 0.16 -0.283 0.49 0.20 0.63 9.22 5.46 4.4 0.327 0.08 0.017 0.16 -0.5 0.16 -0.293 0.49 0.20 0.63 9.06 5.40 4.5 0.318 0.08 0.028 0.16 -0.4 0.16 -0.311 0.49 0.20 0.62 8.89 5.40 4.6 0.330 0.08 0.004 0.16 -0.5 0.16 -0.273 0.49 0.20 0.62 8.87 5.42 4.7 0.338 0.08 -0.009 0.16 -0.5 0.16 -0.173 0.49 0.20 0.62 8.95 5.43 4.8 0.343 0.08 -0.021 0.16 -0.5 0.16 -0.102 0.49 0.20 0.62 8.88 5.42 4.9 0.349 0.08 -0.041 0.16 -0.5 0.16 -0.093 0.49 0.20 0.61 8.70 5.48 5 0.353 0.08 -0.049 0.16 -0.5 0.16 -0.086 0.49 0.19 0.62 8.66 5.38 5.1 0.361 0.08 -0.062 0.16 -0.5 0.16 0.001 0.48 0.19 0.62 8.74 5.28 5.2 0.362 0.08 -0.061 0.16 -0.5 0.16 0.060 0.48 0.19 0.63 8.85 5.19 5.3 0.358 0.08 -0.053 0.16 -0.5 0.16 0.074 0.47 0.19 0.63 8.83 5.10 5.4 0.351 0.08 -0.042 0.16 -0.6 0.16 0.150 0.48 0.19 0.63 8.76 5.18 5.5 0.348 0.08 -0.040 0.16 -0.5 0.16 0.127 0.47 0.19 0.63 8.66 5.15 5.6 0.346 0.08 -0.036 0.15 -0.5 0.16 0.132 0.47 0.19 0.63 8.67 5.03 5.7 0.336 0.08 -0.015 0.15 -0.5 0.15 0.124 0.47 0.19 0.64 8.71 4.95 5.8 0.321 0.08 0.014 0.15 -0.5 0.15 0.101 0.46 0.18 0.64 8.67 4.82 5.9 0.316 0.08 0.021 0.15 -0.5 0.15 0.117 0.45 0.18 0.65 8.59 4.68 6 0.317 0.08 0.019 0.15 -0.5 0.15 0.118 0.45 0.18 0.65 8.59 4.59 6.1 0.306 0.08 0.037 0.15 -0.5 0.15 0.101 0.44 0.18 0.65 8.50 4.50 6.2 0.302 0.08 0.047 0.14 -0.5 0.14 0.131 0.44 0.18 0.66 8.57 4.37 6.3 0.295 0.07 0.060 0.14 -0.5 0.14 0.150 0.44 0.17 0.66 8.56 4.32 6.4 0.297 0.07 0.057 0.14 -0.5 0.14 0.166 0.43 0.17 0.67 8.55 4.31 6.5 0.293 0.08 0.062 0.14 -0.5 0.14 0.177 0.44 0.18 0.66 8.51 4.35 6.6 0.290 0.07 0.069 0.14 -0.5 0.14 0.188 0.43 0.17 0.66 8.52 4.30 6.7 0.290 0.07 0.071 0.14 -0.5 0.14 0.191 0.43 0.17 0.67 8.56 4.22 6.8 0.289 0.07 0.077 0.14 -0.6 0.14 0.259 0.42 0.17 0.68 8.70 4.11 6.9 0.292 0.07 0.072 0.14 -0.6 0.14 0.349 0.42 0.17 0.68 8.73 4.07 7 0.291 0.07 0.077 0.14 -0.6 0.14 0.377 0.42 0.17 0.69 8.78 4.03 7.1 0.297 0.07 0.065 0.14 -0.6 0.14 0.373 0.42 0.17 0.69 8.80 4.00
94
7.2 0.306 0.07 0.050 0.14 -0.6 0.14 0.381 0.41 0.17 0.69 8.84 3.92 7.3 0.312 0.07 0.041 0.14 -0.6 0.14 0.410 0.41 0.17 0.70 8.93 3.87 7.4 0.319 0.07 0.028 0.14 -0.6 0.14 0.404 0.41 0.16 0.70 8.97 3.85 7.5 0.325 0.07 0.016 0.13 -0.6 0.14 0.357 0.41 0.16 0.70 8.97 3.82 7.6 0.333 0.07 0.002 0.13 -0.6 0.13 0.383 0.40 0.16 0.71 9.02 3.69 7.7 0.336 0.07 -0.002 0.13 -0.6 0.13 0.354 0.40 0.16 0.71 9.04 3.62 7.8 0.346 0.07 -0.022 0.13 -0.6 0.13 0.337 0.39 0.16 0.72 9.04 3.51 7.9 0.350 0.07 -0.031 0.13 -0.6 0.13 0.333 0.38 0.15 0.73 9.03 3.38 8 0.356 0.06 -0.041 0.12 -0.6 0.12 0.329 0.38 0.15 0.74 9.05 3.24 8.1 0.357 0.06 -0.047 0.12 -0.6 0.12 0.255 0.37 0.15 0.75 8.99 3.07 8.2 0.362 0.06 -0.057 0.12 -0.5 0.12 0.239 0.36 0.14 0.75 8.97 2.94 8.3 0.368 0.06 -0.067 0.12 -0.6 0.12 0.237 0.36 0.14 0.76 8.98 2.88 8.4 0.373 0.06 -0.079 0.12 -0.5 0.12 0.197 0.36 0.14 0.76 8.96 2.88 8.5 0.377 0.06 -0.090 0.12 -0.5 0.12 0.142 0.36 0.14 0.75 8.88 2.96 σ- Standard error of the coefficient
σest- Standard error of the estimate R2- Coefficient of determination SS(reg)-Regression Sum of Squares SS(res)-Residual Sum of Squares
The variation of b with the wave frequency is graphically indicated in Figure 5.1, in which it shows slightly lower attenuation rates (less than R-0.5) scattered between R-0.1 and R-0.5 are given for low frequencies (below about 2 Hz). At high frequencies, attenuations follow the common form R-0.5 which has been previously established by many researchers based on the studies of regional seismic phases such as Lg (Herrmann and Kijko, 1983; Kennett, 1986; Shin and Herrmann, 1987). Although, the resulted average far-source attenuation rate coincides with previous studies, the explanatory reason may need to be placed in a different context. The attenuation rate R-0.5, found in the above studies, is generally based on earthquake data of continental crusts, whereas the data in the present study are originally from a thinner oceanic crust. Studies have shown that regionally dominant phases like Lg is not effective or totally absent to traverse through a thinner oceanic crust due to inconsistency in the wave field and crustal structure, which may be attributed with physical features like thickness and composition (presence of sediments) (Kennett, 1986; Knopoff et al, 1979). Hence, oceanic crusts such as in the northern Indian Ocean could be less effective in transmitting Lg waves through the medium, in which case a direct comparison of the far-source attenuation rate obtained in the present study with that in the aforementioned studies on continental crusts, is a little intricate. However, despite the difference in the regional wave field and physical attributes between the two crustal types, a similar far-source geometric attenuation rate as of a continental crust has resulted for the northern Indian Oceanic region. This would imply that the attenuation rate is more independent of both the type of the crust (whether continental or oceanic) and regional seismic wave field associated with the medium. Consequently, multiple reflections and refractions of seismic waves taking place at far-source distances, appear insensitive to the dominant wave type in the particular case (e.g., Lg
for continental crusts and may be Sn for oceanic crusts).
95
Figure 5.1 Variation of the far-source geometric attenuation rate, b, with wave frequency. Values are approximately within -0.1 and -0.6, while being the average close to -0.5. Slightly lower attenuation rates especially scattered at low frequencies up to about 2 Hz are noticeable. At high frequencies, attenuations show generally following -0.5 rate as indicated previously by many others based on studies of regional seismic phases such as Lg (Herrmann and Kijko, 1983; Kennett, 1986; Shin and Herrmann, 1987)
Figure 5.2 shows a comparison of recorded and predicted [from equation (5.7)] acceleration amplitudes for two sample frequencies 4.5 and 7.0 Hz. Two magnitude bands (mb 4.0-4.5 and 4.5- 5.0) are selected for the comparison. The predicted amplitudes are based on the median magnitude in each selected magnitude band, i.e., mb 4.25 and 4.75, respectively, for mb 4.0-4.5 and 4.5-5.0.
Recorded amplitudes show a good comparison with the predicted ones for the chosen magnitude- distance combinations.
A residual analysis is carried out to see whether any unforeseen tradeoff exists between the results and regression variables. This partly validates the selected independent variables and their coherence to the regression equation. Residuals are defined as the difference between log observed amplitudes and log predicted amplitudes. Figure 5.3 shows the variation of residuals with the earthquake magnitude (mb) and with the logarithmic value of the hypocentral distance (log R). The figures do not show any apparent trend in residuals with the selected variables, and thus ensure the reliability of the selected attenuation equation for defining the actual dataset.
Hypocentral depth dependency on the regression is not considered since all the selected events are shallow crustal with depths around 10 km maximum.
96
Figure 5.2 Comparison of recorded vertical component acceleration amplitudes with predicted ones from the applied attenuation relationship [equation (5.7)], at two selected frequencies (4.5 and 7.0 Hz), for two magnitude bands mb 4.0-4.5 and 4.5-5.0. The predicted amplitudes are based on the median magnitude in each selected magnitude band, i.e., mb 4.25 and 4.75, respectively, for mb 4.0- 4.5 and 4.5-5.0. Solid line represents the mean predicted amplitudes and diamonds represent recorded amplitudes. Dash lines denote standard error on mean. (a) For mb 4.0-4.5. (b) For mb 4.5- 5.0.
97
Figure 5.3 Log residual (= log observed amplitude-log predicted amplitude) variation for selected frequencies (4.5 and 7.0 Hz). (a) Log residual vs. magnitude (mb). (b) Log residual vs. Log hypocentral distance. Figures show no any apparent trend in residuals in respect to the selected variables.