GNGTS 2017 SeSSione 1.1

�3

BaCK azImuth determInatIon of regIonal earthQuaKes usIng ColoCated measurements of ground rotatIons and translatIons durIng the 2016 Central ItalY seIsmIC seQuenCe A. Simonelli1,2, H. Igel1, J. Wassermann1, J. Belfi2, A. Di Virgilio2, N. Beverini2, G. De Luca3,G. Saccorotti3

1 Ludwig-Maximilians-Universitaet, Münik, Germany2 Istituto Nazionale di Fisica Nucleare, Pisa, Italy3 Istituto Nazionale di Geofisica e Vulcanologia, Italy

Introduction. On August 24, 2016, at 01:36:32 UTC a Mw=6.0 struck the central sector of the Apennines chain (Italy), causing almost 300 casualties and extensive destruction. During the following two months, both rate and energy of aftershocks decreased progressively. On October 26, 2016, the activity renewed with two energetic events (Mw=5.4 and Mw=5.9) until climaxing, four days later, with a Mw=6.5 shock. The colocated observation of ground translations and vertical rotations permits, with a single station approach, to estimate the Back azimuth (hereinafter BAZ) of the incoming wave-field generated by seismic events. The measurements that are object of this study are obtained my means of a large ring laser gyroscope named Gingerino and characterized in (Belfi et al., 2017). This instrument is located, as shown in Fig. 1, inside the underground laboratory of the INFN in Gran Sasso and on top of his granite frame is located a broadband seismometer. We selected 33 events with the criterion of the best s/n ratio and we studied systematically the relative BAZ with a novel method based on rotation-to-translation comparison.

Method. The classical seismological observations are based on the data recorded by seismometers, this sensors provide a measure of the ground displacement. During the transit of a seismic wave by the way the ground is not only translating but it also rotates; this means that the standard seismometer recordings of an earthquake are not a complete measure of the real ground motion. The experiment Gingerino permits to directly measure the vertical ground rotation rate. A comparison with the seismometer data, performed with the processing described below, allow an estimation of the direction of the incoming wavefield for S-waves horizontally polarized and Lg waves. The direct measure of the direction of the wavefield, based on physical principles, is of great importance itself as a property of the seismic event and can help to highlight geological/structural effects that can cause anomalous off-path propagation of the seismic waves. We assume the plane-wave propagation and we process the data set in order to get an experimental estimation of the events back azimuth. We compare this results to the theoretical

�4

GNGTS 2017 SeSSione 1.1

ones. The horizontal components of ground acceleration are rotated in steps ��� of one degree in the span��� of one degree in the span of one degree in the span [0, 2�]. From a system oriented to�]. From a system oriented to]. From a system oriented to geographical coordinates (N-E) we create a set of ray parameter oriented horizontal acceleration traces {a_R(��),��),), a_T(��)�, where �� is the unknown��)�, where �� is the unknown)�, where �� is the unknown�� is the unknown is the unknown Back azimuth. Provided the plane wave assumption and linear elasticity we know that vertical rotation and transverse acceleration (Aki and Richards, 2009; Cochard et al., 2006) should show in the seismograms as the same waveform scaled by the frequency dependent phase velocity C(f). Given the last assumption we use the Wavelet coherence tool (WTC) (Grinsted et al., 2004) to obtain time-frequency maps of correlation between the vertical rotation rate Ω_z and theΩ_z and the_z and the transverse accelerations set {a_T(��)�,��)�,)}, derived after the previous step of traces rotation. The result of this processing is an array of correlation values C(��,t,f) that are function of the��,t,f) that are function of the,t,f) that are function of the time, the frequency and the rotation step ��� of the seismometer horizontal components. This��� of the seismometer horizontal components. This of the seismometer horizontal components. This representation will allow us to have a time-frequency estimation of the back azimuth. This

Fig. 1 - The experiment location inside the “Laboratori Nazionali del Gran sasso” (top right) and the picture of the Gingerino Ring Laser Gyroscope (bottom right), in the center of the granite frame is located the broadband seismometer. In the large background picture we see the events used (see Tab. 1) for this study, the epicenters, and for the first ten events of Tab. 1 we provide the relative focal mechanisms.

Fig. 2 - The back azimuth analysis in different frequency bands for the Visso M 5.9 main shock, In this plot the red color means that the correlation is equal to one i.e. the waveforms are identical and in phase. On top of the figure the superposition of rotation rate (red) and transverse acceleration (black).

GNGTS 2017 SeSSione 1.1

��

analysis is shown in Fig. 2 in the case of the the Visso MW 5.9 earthquake. The solid line in Fig. 2 represents the theoretical back azimuth. We can see that for this event the surface Love waves are very clear also in the seismogram and that, at periods longer than 3 seconds, the BAZ estimation is in good accord with the theoretical one. In the band represented by a central frequency of 2 Hz we can see that this method is very effective in identifying a region of high coherence corresponding to what we identify as the SH arrival with the correct BAZ.

The analysis shown in Fig. 2 basically allow us to understand from where the different frequencies are coming from in the different seismic phases, we remark that this kind of analysis is only possible by exploiting seismometer array measures. We also remark that this method is

Tab. 1 - Parameters of the main events of the sequence.

StartTime Mag Dist[km] BAZ[deg] Depth[km] EventN.

2�-Oct-201�1�:1�:0� �.� �2.3 324.� �.� 1

2�-Oct-201�1�:10:3� �.4 ��.� 322.� �.� 2

01-Nov-201�0�:��:3� 4.� ��.� 331.0 �.� 3

03-Nov-201�00:3�:00 4.� ��.0 32�.4 �.4 4

30-Oct-201�13:34:�4 4.� �1.2 31�.� �.2 �

30-Oct-201�12:0�:�� 4.� ��.4 31�.2 �.� �

2�-Oct-201�21:41:�� 4.� ��.1 321.4 �.� �

2�-Oct-201�0�:21:4� 4.3 �0.� 320.� �.4 �

31-Oct-201�0�:0�:44 4.2 ��.4 320.1 10.0 �

30-Oct-201�10:1�:2� 4.1 �3.3 31�.1 10.� 10

2�-Oct-201�03:1�:2� 4.0 ��.� 321.� �.2 11

1�-Oct-201�0�:32:34 4.0 4�.1 31�.4 �.2 12

31-Oct-201�0�:1�:1� 3.� 4�.3 31�.� �.� 13

2�-Oct-201�1�:22:22 3.� ��.� 31�.1 �.0 14

0�-Oct-201�1�:11:0� 3.� 44.� 31�.1 �.� 1�

0�-Nov-201�1�:��:1� 3.� ��.4 324.� �.1 1�

2�-Oct-201�1�:��:�� 3.� �2.� 31�.2 �.� 1�

2�-Oct-201�1�:43:42 3.� �3.� 320.1 12.� 1�

0�-Nov-201�0�:13:0� 3.� 3�.� 30�.� 10.� 1�

30-Oct-201�12:32:�� 3.� 3�.� 31�.3 �.2 20

30-Oct-201�11:14:20 3.� 4�.� 321.3 �.4 21

2�-Oct-201�1�:��:31 3.� ��.� 323.� 13.2 22

2�-Oct-201�21:24:�1 3.� �1.3 31�.� 10.3 23

0�-Nov-201�1�:1�:1� 3.� �0.� 321.2 �.� 24

0�-Nov-201�0�:1�:3� 3.� 44.3 30�.2 11.1 2�

31-Oct-201�0�:34:1� 3.� �3.1 31�.� �.2 2�

30-Oct-201�23:��:1� 3.� ��.4 31�.� �.� 2�

30-Oct-201�10:2�:24 3.� ��.1 31�.2 10.� 2�

0�-Oct-201�04:42:42 3.� 4�.0 31�.3 11.� 2�

02-Nov-201�0�:41:12 3.� �0.� 31�.1 10.3 30

01-Nov-201�1�:��:12 3.� �3.1 31�.� 10.� 31

30-Oct-201�13:14:1� 3.� �4.4 30�.� �.� 32

2�-Oct-201�23:1�:0� 3.� �1.� 320.� 14.0 33

��

GNGTS 2017 SeSSione 1.1

very effective in identifying the regional SH onset, this is traditionally an hard task since this onset is often buried in the coda signal and is again based on a real ground motion measure. A more quantitative and statistically consistent analysis of the back azimuth for the entire event database is given in the next steps. The C(��,t,f) array is calculated for every event. We find the��,t,f) array is calculated for every event. We find the,t,f) array is calculated for every event. We find the maxima of correlation in a time window that goes from the beginning of the S-coda to the end of the surface waves phase. The obtained values are binned in histograms and the distribution is modeled with a gaussian function (KDE Gaussian). We apply this processing to all the events and we resume the analysis by plotting the estimated Back azimuth and the theoretical one for the entire set in the left panel of Fig. 3. In the right panel of Fig. 3 we represent also the polar histogram of the misfits and the relative gaussian kernel modeling of the distribution.

Conclusions. The theoretical BAZ is just an indication of the possible direction of the wave field. In a complex topography like the Gran Sasso region and specially in an underground environment at 1400 m from the free surface is not trivial whether the seismic waves should follow the theoretical BAZ. Nevertheless for teleseisms in (Simonelli et al., 2016) we measured a misfit of five degrees after the Love waves onset at periods longer than 10 seconds. From the analysis on this entire data set we can state that we observe a 10 degrees systematic misfit that can be compatible with the orientation error of the seismometer or can be due to a structural effect. A future measure with a triaxial fiber optic gyroscope, used as a gyroscopic compass will allow us to orient our instrument and measure the previous orientation with a precision lower than 0.1 deg. We tried a cluster analysis in order to check if the misfit could be dependent on the events parameters of Tab. 1 and on the S/N ratio but the result does not show any clear dependence. In conclusion an average misfit of ��misf. = 10 ± 18 deg. is observed.��misf. = 10 ± 18 deg. is observed.misf. = 10 ± 18 deg. is observed.ReferencesAki, K. and Richards, P. G. (2009). Quantitative Seismology. University Science Books. Belfi, J., Beverini, N., Bosi, F., Carelli, G., Cuccato, D., Luca, G. D., Virgilio, A. D., Gebauer, A., Maccioni, E., Ortolan,

A., Porzio, A., Saccorotti, G., Simonelli, A., and Terreni, G. (2017). Deep underground rotation measurements:Deep underground rotation measurements: Gingerino ring laser gyroscope in gran sasso. Review of Scientific Instruments, 88(3):034502.

Cochard, A., Igel, H., Schuberth, B., Suryanto, W., Velikoseltsev, A., Schreiber, U., Wassermann, J., Scherbaum, F., and Vollmer, D. (2006). Rotational motions in seismology: Theory, observation, simulation. In Teisseyre, R., Majewski, E., and Takeo, M., editors, Earthquake Source Asymmetry, Structural Media and Rotation Effects, pages 391–411. Springer, New York.

Grinsted, A., Moore, J. C., and Jevrejeva, S. (2004). Application of the cross wavelet transform and wavelet coherence to geophysical time series. Nonlinear Processes in Geophysics, 11(5/6):561–566.Nonlinear Processes in Geophysics, 11(5/6):561–566.

Simonelli, A., Belfi, J., Beverini, N., Carelli, G., Virgilio, A. D., Maccioni, E., Luca, G. D., and Saccorotti, G. (2016). First deep underground observation of rotational signals from an earthquake at teleseismic distance using a large ring laser gyroscope. Annals of Geophysics, 59(0).

Fig. 3 - The theoretical BAZ ad the measured one for all the events indexed as in Tab. 1 (left). Misfit distribution and the relative gaussian KDE modeling in solid red line (right).