Irpinia is one of the Italian regions with the highest seismic hazard. The last destructive event was the Irpinia earthquake of 1980 with magnitude 6.9 and caused 3,000 fatalities. The Irpinia earthquake was a pure normal fault event that activated at least three different fault segments, which are displayed in the figure together with the area of the aftershocks. Currently the area is characterized by microseismicity which broadly nucleates within the same volume activated during the Irpinia earthquake. The area is monitored by the ISNet seismic network, that is composed by 33 seismic stations covering an area of about 70 by 100 Km. For this work, we have focused on the microseismicity nucleated in the time interval 2008-2020 within the volume ruptured during the first sub-event of the Irpinia earthquake, which hosted the hypocenter and activated the fault segments of Cervialto and Marzano - they are sketched in the figure as yellow rectangles. Our dataset was composed of about 500 events, on which we have looked for repeating earthquakes by clustering events with similar waveforms. Initial step was the definition of the cross-correlation matrix calculated on the P phases for each station. A Hierarchical Agglomerative Clustering was then applied to the multi-channel cross-correlation matrix in a way that earthquakes belonging to the same cluster shared similar waveforms. Only clusters with high similarity were selected thresholding the mean cross-correlation values. The selected clusters form the clustered seismicity. This set of earthquakes is further separated into clusters of repeaters and of sequences depending on whether the clusters were also clustered in time or not. The rest of the seismicity is defined as background. In the figure an example of repeaters is displayed. The clustered seismicity was localized with a double-difference approach. The results are displayed in this figure, where squares show location of sequences, and dots those of the repeaters. As you can see, most of the clustered seismicity is localized on the Cervialto segment. Moreover, sequences and repeaters are spatially disjoined along strike. Along dip hypocenters of clustered seismicity seem to skirt when moving from Cervialto segment southeastwards along strike. In this figure we compare the stress drop averaged over distance intervals along strike of background and clustered seismicity. We note that the largest stress drops appear close to the southern boundary of Cervialto segment. Moreover, repeaters seem to have larger stress drop than the rest of seismicity. In this figure the clustered seismicity is localized with respect to the b-value map of the area (found in a previous paper). As you can see, clustered seismicity mostly nucleated in areas of low b-values. In particular, repeaters nucleate in a narrow fault segment with very low b values edging with a segment with high b-value. The fault segments where repeaters and sequences nucleate appear to be characterized also by different values of Vp/Vs ratio, here inferred by tomography. Sequences are localized within areas with Vp/Vs ratios of 1.8 or more, while repeaters originate where Vp/Vs ratio lower than 1.75. These results are confirmed also by local estimations of Vp/Vs ratios obtained using a modified Wadati approach, even though the spatial resolution is lower. This picture somehow summarizes the results. Together with the know separation of Cervialto and Marzano segments, where most of the coseismic slip of 1980 earthquake occurred, and the Sele gap, where instead no slip occurred, we postulate a further segmentation of the Cervialto fault with potentially different fault couplings. We also find a sharp contrast of behaviours at the southern boundary of Cervialto segment, which could be caused by an extensional transfer zone. In conclusions, the three main outcomes of this work are: 1. Existence of repeating earthquakes in the Irpinia Fault; 2. Indications of small-scale segmentations on the Cervialto fault; 3. A sharp contrast of properties south of the Cervialto segment to be explained.
Earthquake repeaters and small-scale Vp/Vs ratio in the Irpinia region (Southern Italy): spatiotemporal patterns and relationship with seismic source properties / Palo, Mauro; Picozzi, Matteo; De Landro, Grazia; Zollo, Aldo. - (2022). (Intervento presentato al convegno Third European Conference on Earthquake Engineering and Seismology (3ECEES) tenutosi a Bucharest, Romania nel September 4-9 2022).
Earthquake repeaters and small-scale Vp/Vs ratio in the Irpinia region (Southern Italy): spatiotemporal patterns and relationship with seismic source properties
Mauro PaloPrimo
;Matteo Picozzi;Grazia De Landro;Aldo Zollo
2022
Abstract
Irpinia is one of the Italian regions with the highest seismic hazard. The last destructive event was the Irpinia earthquake of 1980 with magnitude 6.9 and caused 3,000 fatalities. The Irpinia earthquake was a pure normal fault event that activated at least three different fault segments, which are displayed in the figure together with the area of the aftershocks. Currently the area is characterized by microseismicity which broadly nucleates within the same volume activated during the Irpinia earthquake. The area is monitored by the ISNet seismic network, that is composed by 33 seismic stations covering an area of about 70 by 100 Km. For this work, we have focused on the microseismicity nucleated in the time interval 2008-2020 within the volume ruptured during the first sub-event of the Irpinia earthquake, which hosted the hypocenter and activated the fault segments of Cervialto and Marzano - they are sketched in the figure as yellow rectangles. Our dataset was composed of about 500 events, on which we have looked for repeating earthquakes by clustering events with similar waveforms. Initial step was the definition of the cross-correlation matrix calculated on the P phases for each station. A Hierarchical Agglomerative Clustering was then applied to the multi-channel cross-correlation matrix in a way that earthquakes belonging to the same cluster shared similar waveforms. Only clusters with high similarity were selected thresholding the mean cross-correlation values. The selected clusters form the clustered seismicity. This set of earthquakes is further separated into clusters of repeaters and of sequences depending on whether the clusters were also clustered in time or not. The rest of the seismicity is defined as background. In the figure an example of repeaters is displayed. The clustered seismicity was localized with a double-difference approach. The results are displayed in this figure, where squares show location of sequences, and dots those of the repeaters. As you can see, most of the clustered seismicity is localized on the Cervialto segment. Moreover, sequences and repeaters are spatially disjoined along strike. Along dip hypocenters of clustered seismicity seem to skirt when moving from Cervialto segment southeastwards along strike. In this figure we compare the stress drop averaged over distance intervals along strike of background and clustered seismicity. We note that the largest stress drops appear close to the southern boundary of Cervialto segment. Moreover, repeaters seem to have larger stress drop than the rest of seismicity. In this figure the clustered seismicity is localized with respect to the b-value map of the area (found in a previous paper). As you can see, clustered seismicity mostly nucleated in areas of low b-values. In particular, repeaters nucleate in a narrow fault segment with very low b values edging with a segment with high b-value. The fault segments where repeaters and sequences nucleate appear to be characterized also by different values of Vp/Vs ratio, here inferred by tomography. Sequences are localized within areas with Vp/Vs ratios of 1.8 or more, while repeaters originate where Vp/Vs ratio lower than 1.75. These results are confirmed also by local estimations of Vp/Vs ratios obtained using a modified Wadati approach, even though the spatial resolution is lower. This picture somehow summarizes the results. Together with the know separation of Cervialto and Marzano segments, where most of the coseismic slip of 1980 earthquake occurred, and the Sele gap, where instead no slip occurred, we postulate a further segmentation of the Cervialto fault with potentially different fault couplings. We also find a sharp contrast of behaviours at the southern boundary of Cervialto segment, which could be caused by an extensional transfer zone. In conclusions, the three main outcomes of this work are: 1. Existence of repeating earthquakes in the Irpinia Fault; 2. Indications of small-scale segmentations on the Cervialto fault; 3. A sharp contrast of properties south of the Cervialto segment to be explained.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


