The Near-Surface Attenuation for Regional Seismic Network of Perm Region

Skorkina A.

Perm State National Research University, Perm, Russia

[e-mail]

To determine seismic moment and moment magnitude using spectral analysis the spectra must be corrected for attenuation. That is particularly important for getting the correct corner frequency. For local crust studies, we must at least separate the attenuation into two terms (near surface and the rest) as in (1) with a constant Q in each layer.

(1),

where A ₀ is the initial amplitude, A (f, t) the amplitude after the waves have traveled the time t, f is the frequency, k – near-surface attenuation and Q (f) is the general frequency dependent Q. The near surface attenuation is almost frequency independent and limited to the very near surface layers (similar travel time). A typical k -value is 0.05.

Taking the natural logarithm of (1) gives

(2).

If the signal is generated by an earthquake and we only use the part of the spectrum where f ₀ (corner frequency) > f, the spectral level will only depend on attenuation and (2) shows that plotting vs. f, will give a straight line with slop if Q is frequency independent. If , the slope will become . If is small, the slope will also be directly proportional to k. Thus to reliably determine k by plotting the spectral level, for , without knowing Q, short hypocentral distances should be used [4].

We analyzed 191 records of ground velocity registered by six stations of Perm region seismic network (PR0, PR1R, PR2, PR3, PR4, SOKR) to estimate near surface attenuation k. We selected records of explosions with high SNR and clear P- and S-wave arrivals, covering a range of distances between 20 and 70 km. We can therefore safely assume that the decay of the spectrum is due to near surface attenuation only. Small events recorded at short distances tend to have shorter time-window lengths, so we did not use events covering a range of distances between 0 and 20 km. To calculate spectra we used about the same window length that includes the maximum amplitude of the waves. The window lengths are 40 s and 90 s for P- and S-waves respectively. To estimate k we calculated the linear least-squares fit of the observed spectra in a log-linear domain using equation (2). Fitting the spectral level was performed within a frequency range between 0.5 and 4.5-5 Hz, where the shape of instrument response is linear and appropriate SNR is present. Since there are results which indicate that k is independent of event size within the magnitude range (M < 3.5) [2], we do no correction for event sizes and k -value is assumed to depend only on geologic conditions (rock type) near the recording station.

Based on the location of the stations within the superficial geology maps, we classified the stations in three groups: I, station on potassium salt rocks (SOKR); II, stations on gypsum rocks (PR3); and III, stations on sort of clay soils (the rest stations). In this case, k is considered to be the same for all the stations in the group.

The obtained k -values for Perm region are similar to results obtained in other regions. For instance, Castro [2] found that the spectral decay parameter k is smaller for stations located on hard rocks compared with stations located on sites with less competent geology. We compare k -values calculated in different regions. It varies from 0.0003-0.0035 (Anza, California, [3]) to 0.049 (Cajon Pass Scientific drill hole, [1]) or 0.05 (Central Europe, [5]). There are also values of 0.027-0.047 (Imperial fault, California, [7]), 0.03 (Northern Iran, [6]) even about 0.0842 (Northeastern Sonora [2]).

Estimations of near-surface attenuation k show that the station on potassium salt rocks (SOKR) has the smallest value. The k -value is 0.0253 for this group. The station on gypsum rocks (PR3) has a slightly bigger k than previous one and has value of 0.0301. We calculated the average value of near-surface attenuation for the III group of stations because it locates about on the same rock type. The k -value is 0.0341 for this group. And we found an average value of k of 0.03 sec for Perm region seismic network that will be used in further studies.

References:

1.Abercrombie, R. E. Near-surface attenuation and site effects from comparison of surface and deep borehole recordings. Bulletin of the Seismological Society of America, Vol. 87, pp. 731–744, 1997.

2.Fernandez A.L., Castro R.R., Carlos I. Huerta. The spectral decay parameter kappa in Northeastern Sonora, Mexico. – Bulletin of the Seismological Society of America, Vol. 100, No. 1, pp. 196-206, February 2010.

3.Hough, S. E., J. G. Anderson, J. Brune, F. Vernon, III, J. Berger, J. Fletcher, L. Haar, T. Hanks and L. Baker. Attenuation near Anza, California. Bulletin of the Seismological Society of America Vol. 78, pp. 672–691, 1988.

4.Havskov J., Ottemoller L.. Routine data processing in earthquake seismology with sample data, exercises and software. Springer Science+Business Media B.V. 2010.

5.Malagnini, L., R. B. Herrmann and K. Koch. Regional ground motion scaling in Central Europe. Bulletin of the Seismological Society of America, Vol. 90, pp.1052–1061, 2000.

6.Motazedian, D. Region-specific key seismic parameters for earthquakes in Northern Iran. Bulletin of the Seismological Society of America, Vol. 96, pp. 1383–1395, 2006.

7.Singh. S. K., R. J. Aspel, J. Fried and J. N. Brune. Spectral attenuation of SH waves along the Imperial fault. Bulletin of the Seismological Society of America, Vol. 72, pp. 2003–2016, 1982.

Дата добавления: 2015-10-29