GNGTS 2013 - Atti del 32° Convegno Nazionale

NW-SE axis. This feature does not ensure an adequate azimuthal coverage. The problem may be faced and overcome through a focused analysis on the noise source temporal variations. Emilia: i) The peculiarity in this case is not due to the seismicity (which is a classicmainshock- aftershock sequence), but to the geographical area: the alluvial Po plain, is characterized by impressive resonance effects due to surface sediments that cause abnormal amplifications under 1 Hz (~ 0.8 Hz and down to 0.2 Hz). This effect interferes and superimposes on the oceanic microseismicity, which usually is taken as seismic noise source in this type of analysis. It is therefore necessary to remove the local amplification from the cross-correlations, for example by using adaptive filters, or by changing the frequency range under consideration. This second option, however, requires a choice of frequencies compatible with the spatial configuration of the seismic stations, as well as a detailed study of the spatial variation of noise sources at different frequencies. The anisotropic parameters are shown using a stereographic projection (Fig.. 2A): each segment is oriented along the fast direction and its length is proportional to the delay time. We divided dataset into 6 periods: 23/11/2011 M=3.6, 28/05/2012 M=4.3, 14/09/2012 M=3.7, 01/10/2012 M=3.6, 25/10/2012 M=5.0. The center of stereographic projection represents the station MMN while the position of the bar represents the back-azimuth of the event and the distance from the center is a function of the angle of incidence geometry (the outer circle represent 45 degrees). Temporal trend of fast directions and delay time at MMN, averaged over time for the period 2010 – 2010 is shown in Fig. 2B. The vertical bars and colour changes represent 6 events with magnitude greater than 3.5 (24/11/2011; 28/05/2012M= 4.3; 19/08/2012M= 3.7; 14/09/2012M= 3.7; 01/10/2012 M = 3.6; 25/10/2012 M = 5.0). The gray circles are the individual measurements, the green lines represent averaged values over 50 measurements. We considered all parameters with cc greater than 0.7 and delay time greater than 0.02 s. The averaged trends are obtained using the running average algorithm and an overlap length of SM-1 points. Frequency plots of fast direction for each period are shown in different colours, the red bar is the average for the period. Although time series obtained have interesting fluctuations, these are not easily classified as anomalies or seismic precursors mainly because the magnitude of the strongest earthquake that has affected the area is only M = 5. However they represent one of the longer anisotropic time series ever obtained and therefore it is an important step to identify a proper case-study useful to understand whether this methodology can provide a glimpse in the identification of precursory phenomena of strong earthquakes. Finally, to have an insight on the meaning of the parameter fluctuations over time we compare the time series of WP1 and WP2 with the trend of Vp/Vs evaluated using the same seismicity in the period range - second half of 2012 - beginning of 2013 (Fig. 3). The seismic event of magnitude 5 is indicated by the last vertical line in the graphs. Analyzing the trends, an important oscillation in both the delay time and the Vp/Vs is visible and it also corresponds to a decrease in Δvs (this decrease had already started before the event). Fluctuations of anisotropic parameters are fairly common in the months before the magnitude 5 earthquake and seems to be correlated with the Vp/Vs trend. The results presented here can help in getting new insights in the identification of precursory phenomena of strong earthquakes; further investigations are needed to better understand their statistical significance and what is the physical phenomenon that produces them. Acknowledgements. The temporary stations have been installed by the INGV group of the Seismic Mobile Network of Rome and Grottaminarda in collaboration with the University of Calabria and the German Research Centre for Geoscience (GFZ). We thanks Aladino Govoni e Milena Moretti for the P –S arrival times lecture of the dataset of temporary stations and INGV “surfanelpick group” for the event locations. References Brenguier F., Campillo M., Hadziioannou C., Shapiro N.M., Nadau R.M., Larose E., (2008). Postseismic relaxation along the San Andreas fault at Parkfield from continuous seismological observations. Science 321, 1478-1481. 114 GNGTS 2013 S essione 2.1