ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 1 Site effect studies following the 2016 Mw 6.0 Amatrice Earthquake (Italy): the Emersito Task Force activities GIOVANNA CULTRERA1*, EZIO D’ALEMA2, SARA AMOROSO3, BARBARA ANGIONI1, PAOLA BORDONI1, LUCIANA CANTORE3, FABRIZIO CARA1, ARRIGO CASERTA4, ROCCO COGLIANO5, MARIA D’AMICO 2, GIUSEPPE DI GIULIO 3, DEBORAH DI NACCIO3, DANIELA FAMIANI1, CHIARA FELICETTA2, ANTONIO FODARELLA5, SARA LOVATI2, LUCIA LUZI2, CLAUDIA MASCANDOLA2, MARCO MASSA2, ALESSIA MERCURI1, GIULIANO MILANA1, FRANCESCA PACOR2, MARTA PISCHIUTTA1, STEFANIA PUCILLO5, RODOLFO PUGLIA2, GAETANO RICCIO5, GABRIELE TARABUSI1, MAURIZIO VASSALLO3 1Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italy 2Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano, Milano, Italy 3Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, L'Aquila, Italy 4Istituto Nazionale di Geofisica e Vulcanologia, Sezione CNT, Roma, Italy 5Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Grottaminarda (AV), Italy * giovanna.cultrera@ingv.it Abstract On August 24, 2016, at 01:36 UTC a MW 6.0 earthquake struck an extensive area of the Central Apennines (Italy) be- tween the towns of Norcia and Amatrice. Due to the mainshock magnitude and the widespread damaging level of build- ings in the epicentral area, the Emersito task force has been mobilized by the Istituto Nazionale di Geofisica e Vulcanologia (INGV). The aim of Emersito is to carry out and coordinate the monitoring of local site effects, caused by geological and geomorphological settings. During the first days of the seismic emergency, Emersito installed a tempo- rary seismic network for site effect studies at 4 municipalities close to the epicentral area (Amandola, Civitella del Tronto, Montereale and Capitignano), using 22 stations equipped with both velocimetric and accelerometric sensors. The selection of the sites where stations have been installed was mainly driven by the proximity to the epicentral area (without interfere with the rescue operations) and by peculiar geologic and geomorphologic settings (topographic irregu- larities, fault zones, alluvial plains). Preliminary analyses performed on ambient noise and aftershocks signals show that directional amplification effects may have occurred at stations installed on the top of topographic irregularities. We also observed the lengthening and amplification of the seismograms and a variability of the peaked frequency across the sedi- mentary basin between Montereale and Capitignano, probably related to a different thickness of the deposits. Further analyses are necessary to assess the correlation with surface geology. ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 2 I. INTRODUCTION he earthquake ground-motion depends on the source process, on the propagation of the incoming wave-field through the crustal medium, and on the transmission of the seismic waves from bedrock to the Earth sur- face (free-field motion). The last contribution is known as site effect, and substantially modifies the amplitude, the frequency content and du- ration of the seismic waveforms. Three main types of site effects can be identi- fied: impedance contrast between bedrock and softer soil, the lateral heterogeneities in the elastic and/or geometrical properties of the soil and the topographic settings that can pro- duce 2D or 3D effects. Site effects due to soft soil layers (e.g. alluvial or sedimentary cover) overlapping hard rock produce seismic wave amplification whose ex- tent depends on the impedance contrast be- tween different layers. These effects are usual- ly modeled by 1D wave propagation through the soil column. When lateral heterogeneities of the subsoil me- chanical properties are not negligible (e.g. sed- iment filled valleys), 2D or 3D geometries are more appropriate to describe the complexity of the ground motion (Kawase, 1996; Bindi et al., 2009; Chaljub et al., 2010; Di Giulio et al., 2016). In the last decade particular attention has been devoted to the ground motion directional am- plification in fault damage zones, related to fracture networks and the resulting stiffness anisotropy of the rock mass (Pischiutta et al., 2012). Concerning the topographic effects, a correla- tion between pattern of damage and topo- graphic elevation has been recently investigat- ed (Paolucci, 2002; Massa et al., 2014), although there are fewer evidences compared to the previous two types of site effects. Usually, these studies use seismic noise and background seismicity because stronger events are available for a target area during a seismic sequence only. Emersito is a task force of the Istituto Naziona- le di Geofisica e Vulcanologia (INGV) that in- volves several headquarters (Milano, Roma, L’Aquila, Grottaminarda; Italy). Its aim is to coordinate and carry out seismic monitoring to investigate possible site effects caused by seismic events of moderate–to-large magni- tude (M≥5) in the Italian territory. In case of relevant seismic sequences, a rapid installation of temporary seismic stations is fundamental to seismic monitoring, to the research in the field of seismic hazard and, in some cases, to support post-emergency activities (e.g. microzonation studies). Emersito operated for the first time during the Emilia 2012 sequence (Bordoni et al, 2012), and it was officially es- tablished in 2015 (DDG, 2015); mobilization procedures have been test during the INGV Emergency simulation performed on 2015 (Pondrelli et al, 2016). On August 24, 2016 (01:36 UTC) a MW6.0 earthquake struck an extensive area in the Central Apennines (Italy) between the munici- palities of Norcia and Amatrice. The epicenter was located by the INGV close to the Accumoli village. A large aftershock (MW 5.3, August 24 02:33 UTC) and over 9000 aftershocks have been located up to September 16, 2016, in an T ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 3 extended area of about 50 km in the Appennine direction (from the town of Ussita to the North, to the lake of Campotosto in the South), and about 15 km width (INGV work- ing group, 2016). In the first days of the seismic emergency (August 24th-25th, 2016), we have chosen several urban areas located at the mar- gin of the epicentral area and characterized by geological and structural features able to pro- duce local site effects (Figure 1). Several issues were considered for the selection of the areas: availability of geological maps and first level microzonation studies, location of recording stations of temporary and/or permanent seis- mic networks (Italian National Seismic Net- work operated by INGV and Italian Strong Motion Network operated by the Italian De- partment of Civil Protection), rescue activities in the epicentral area. The available infor- mation was centralized in a web Geographic Information System (Arcgis on-line) to facili- tate the data sharing to all Emersito members (Figure 1). During August 26th-31st, 22 tempo- rary seismic stations were installed in four municipalities: Amandola (FM), Civitella del Tronto (TE), Montereale (AQ), and Capitignano (AQ). Additional 5 recording sta- tions were set in Acquasanta Terme (AP) by the INGV National Earthquake Centre in co- operation with the Emersito task force. Figure 1. Thematic levels of Emersito ArcGis on-line with the 2016 Amatrice seismic sequence (from August 24th to September 28th) together with the selected sites and the temporary network. ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 4 II. INSTRUMENTAL SETTING The instrumental setting was composed of 22 digitizers (Reftek 130 and Quanterra), equipped with both Lennartz-5s velocimeter and Episensor accelerometer (Emersito report, 2016). All the stations recorded in continuous mode: five of them transferred data to a re- mote server in real-time, while the remaining stations were in a stand-alone configuration. Data will be distributed through the European Integrated Data Archive (EIDA, http://eida.rm.ingv.it/), by using the interna- tional XO code assigned on August, 29th to the temporary Emersito network by the Federation of Digital Seismograph Networks (www.fdsn.org/networks/detail/XO_2016). The XO network was named “Seismic Net- work for Site Effect Studies in Amatrice Area (Central Italy)” and the station codes were reg- istered at the International Seismological Cen- tre (ISC) on September 6th. Table 1 reports network and station code, municipality, site coordinate, digitizer, sensor, type of seismic data acquisition (real-time vs local), and instal- lation period. III. DESCRIPTION OF THE SELECTED SITES AND PRELIMINARY RESULTS Amandola The village of Amandola is located 30 km Northeast from the epicentral area (Figure 1). It extends along a relief, with SW-NE trend, constituted by clayey sandstones. Several buildings of the historical center, and also the hospital, suffered some damage after the mainshock. Figure 2 shows the location of the 5 stations (Table 1), which were installed along the ridge and at the base of the relief, with the aim of investigating possible topographic ef- fects. In particular, the hospital area was moni- tored by AM03 station. Figure 2. Stations installed at Amandola. All stations installed at Amandola sent data in real time to the remote INGV acquisition cen- ter and were soon available (i.e. in average 2 hours later each event origin time) and down- loadable from the real time INGV Strong Mo- tion Database (ISMD; http://ismd.mi.ingv.it). Due to the broadening of the sequence to- wards North (Gruppo di Lavoro INGV, 2016), AM05 was integrated to the Italian Seismic Network with the aim to improve the location operated by the INGV National Earthquake Centre. A preliminary analysis of the Horizontal to Vertical Spectral Ratio (HVSR) was performed considering at all stations 1 hour of ambient noise (running window length of 120 seconds, 4th order Butterworth filter in the frequency band 0.1-15 Hz, tapering 5%, Konno-Ohmachi logarithmic smoothing with b=20). Possible transient signals were deleted by visual inspec- tion. In general, all stations installed at the top of the ridge show HVSR with amplification ranging from 3 to 4 in the frequency range 3-4 ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 5 Hz. The maximum amplification was observed in the N-NW direction (Figure 2); AM05, in- stalled at the base of the ridge, can be consid- ered the reference site in the range of interest for topographic effect, showing amplification at frequency > 8 Hz due to a very thin layer of landfill. Further investigations are needed to ascribe the recordings behavior to a topo- graphical effect. Figure 3. HVSR on noise recorded at AM02, AM03 and AM05: (top) quadratic mean of the two horizontal com- ponents ± 1 standard deviation; (bottom) rotated spectral ratio at different azimuths. Montereale and Capitignano The villages of Montereale and Capitignano are located 20 km South of the mainshock and in correspondence with the southernmost epi- centres of the sequence (Figure 1). Both munic- ipalities suffered slight damage after the mainshock and the larger events. Figure 4 shows the 12 stand-alone stations installed to study topographic and basin effects, as well as to investigate possible fault effects close to CP04, CP05 and CP06 (Table 1). Figure 4. Stations installed at Montereale and Capitignano, superimposed to the Geological map of Italy at 1:100,000 scale. Displayed seismograms refer to the recorded NS component of the Ml 3.6 earthquake (01/09/2016 03:53, hypocentral distance R=16km); rose diagrams indicate the horizontal polarization averaged on 19 earthquakes with Ml≥ 3 recorded in the period 30/08- 1/09/2016. MN04, MN08 and MN06 are located on the top of a topographic irregularity, characterized by a maximum height of 140 m from the nearby level ground and maximum slope of 18°. They are on Miocene flysch units (Figure 4) and, as a first approximation, they can be considered stiff rock sites. Signal polarization analysis per- formed in the time domain (Jurkevics, 1988) indicates a predominant N-NE/S-SW direction of the horizontal motion at these sites (Figure 4). The directional amplification effect is con- firmed by the HVSR performed on noise and on earthquakes (MN04 in Figure 5). The stations on Quaternary alluvial deposits are affected by the lengthening and amplifica- tion of the seismograms (MN02 and MN03 in Figure 4). At those stations, signal polarization analysis shows an isotropic distribution of po- ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 6 larization azimuths, i.e. the site tends to pro- duce a similar amplitude level on the two hor- izontal components. Moreover, the HVSR on noise recorded at MN02 and CP07 show an amplification peak at 1.5 Hz and 6 Hz, respec- tively, that it is probably related to a different thickness of the basin deposits (Figure 5). Fur- ther investigations are needed to verify the persistence of the behavior and the correlation with the topography. Figure 5. Rotated HVSR on noise (left) and on earth- quakes (right) recorded at MN04, MN02 and CPT07 of Figure 4. The HVSR on earthquakes is averaged on 19 events with Ml=3.0÷3.9 recorded in the period 30/08- 1/09/2016; CP07 data were not available in this period. Civitella del Tronto Civitella del Tronto, a village located at about 30-40 km East of the epicentral area (Figure 1), did not suffer significant structural damages for the August 24th mainshock. The municipali- ty is located on a topographic relief, with a SW-NE trend, consisting of travertine overlay- ing sandstones. Five stand-alone stations were installed along the top and at the base of the relief to detect possible topographic effects (Figure 6). Data were not available at the time of this study because the stations were dis- mounted at the end of September 2016. Figure 6. Location of the stations installed at Civitella del Tronto. IV. CONCLUSIVE REMARKS The Emersito efforts during the 2016 Amatrice seismic emergency consisted in the deploy- ment of 22 stations equipped with both velocimeter and accelerometer within few days after the mainshock of August 24th 2016, with the aim of studying site effects for select- ed areas. Usually, these studies use back- ground seismicity while during a seismic se- quence stronger events are available and they can be recorded with a rapid installation of a temporary seismic network. Due to the morphology of the settlements stroke by the earthquake, we focused in the investigation of the site effects of topography, small intra-mountain basin and fault zone to seismic motion. Preliminary results suggest the presence of such effects, requiring further ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 7 analyses to assess the correlation with surface geology. The collected dataset will be available in EIDA node and could support other seismo- logical studies (e.g. source and directivity ef- fects, role of fluids, earthquake location). The remarks in this paper refer to the first phase of the Central Italy seismic sequence, up to the end of September 2016; at the time of publication, several large magnitude earth- quakes that contributed to modify the damag- ing level of the studied villages have hit the area. ACKNOWLEDGEMENTS We are grateful to the local authorities and the inhabitants of the investigated municipalities that supported us during the network man- agement. The study benefited from funding provided by the Italian Presidenza del Consiglio dei Ministri, Dipartimento della Protezione Civile (DPC); scientific papers funded by DPC do not represent its official opinion and policies. REFERENCES [Bindi et al., 2009] Bindi, D and Parolai, S and Cara, F and Di Giulio, G and Ferretti, G and Luzi, L and Monachesi, G and Pacor, F and Rovelli, A (2009). Site amplifications observed in the Gubbio Basin, Central Italy: hints for lat- eral propagation effects. Bulletin of the Sei- smological Society of America, 99(2A), 741-760. [Bordoni et al, 2012] Bordoni, P., Azzara, R.M., Cara, F., Cogliano, R., Cultrera, G., Di Giulio, G., Fodarella, A., Milana, G., Pucillo, G.,Riccio G., Rovelli, A., Augliera, P.,Luzi, L., Lovati, S.,Massa, M., Pacor, F., Puglia, R., Ameri, G. (2012). Preliminary results from EMERSITO, a rapid response network for site-effect studies. Annals of Geophysics, 55, 4, 2012; doi: 10.4401/ag-6153. [Chaljub et al, 2010] Chaljub, E., Moczo, P., Tsuno, S., Bard, P. Y., Kristek, J., Käser, M., Stupazzini, M., Kristekova, M. (2010). Quanti- tative comparison of four numerical predic- tions of 3D ground motion in the Grenoble val- ley, France. Bulletin of the Seismological Socie- ty of America, 100(4), 1427-1455. [DDG, 2015] Decreto Direttore Generale INGV n. 337 e Allegati, 2015; http://istituto.ingv.it/l- ingv/amministrazione-trasparente- 1/contenuti-at-vari/decreti/anno-2015- decreti-presidente-e-direttore-generale). [Di Giulio et al, 2016] Di Giulio, G., de Nardis, R., Boncio, P., Milana, G., Rosatelli, G., Stoppa, F., Lavecchia, G. (2016). Seismic response of a deep continental basin including velocity in- version: the Sulmona intramontane basin (Cen- tral Apennines, Italy). Geophysical Journal In- ternational, 204(1), 418-439. [Emersito report, 2016] EMERSITO working group (2016). Rapporto preliminare sulle atti- vità svolte dal gruppo operativo EMERSITO a seguito del terremoto di Amatrice Mw 6.0 (24 agosto 2016, Italia centrale). doi: 10.5281/zenodo.151655 [INGV working group, 2016] INGV working group on the Amatrice earthquake (2016). Se- cond summary report on the M6.0 Amatrice earthquake of August 24, 2016 (Central Italy), doi: 10.5281/zenodo.16624 Jurkevics, A. (1988), Polarization analysis of three component array data, Bull. Seismol. Soc. Am., 78, 1725–1743. [Kawase, 1996] Kawase, H. (1996). The cause of the damage belt in Kobe: the basin-edge effect, constructive interference of the direct S-wave with the basin-induced diffracted/Rayleigh waves. Seismol Res Lett, 67:25–34. ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 8 [Massa, 2014] Marco Massa, Simone Barani and Sara Lovati, Overview of topographic ef- fects based on experimental observations: meaning, causes and possible interpretations, Geophysical Journal International, doi: 10.1093/gji/ggt341. [Paolucci, 2002] Paolucci, R. (2002). Amplifica- tion of earthquake ground motion by steep topographic irregularities. Earthq. Eng. Struct. Dyn., 31:1831–1853. [Pischiutta et al., 2012] Pischiutta, M., Salvini, F., Fletcher, J., Rovelli, A., & Ben-Zion, Y. (2012). Horizontal polarization of ground mo- tion in the Hayward fault zone at Fremont, California: dominant fault-high-angle polariza- tion and fault-induced cracks. Geophysical Journal International, 188(3), 1255-1272. [Pondrelli et al., (2016)] Pondrelli, S., Amato, A., Massa, M., Montone, P., Crescimbene, M., La Longa, F., Emergenza sismica working group (2016). "Pianificazione e gestione di un’emergenza sismica: esercitazione INGV del 26 novembre 2015 effettuata nell’ambito della Linea di Attività T5 “Sorveglianza sismica e operatività post terremoto”. Quaderni di Geofisica, 317, novembre 2016. ISSN 1590-2595 [in Italian]. ANNALS OF GEOPHYSICS, 59, Fast Track 5, 2016; DOI: 10.4401/ag-7189 9 Table 1. Emersito seismic stations installed at the beginning of the 2016 Amatrice seismic emergency (network code XO) Station code Municipality Lat. [°] Lon. [°] Elev. [m] Digitizer Vel. Sensor Acc. Sensor Acquisition Installation period AM01 Amandola 42.980556 13.358708 549 REFTEK130 LE3D-5S EPISENSOR Real-time 29/08-10/10/2016 AM02 Amandola 42.979597 13.353573 516 REFTEK130 LE3D-5S EPISENSOR Real-time 30/08-10/10/2016 AM03 Amandola 42.981622 13.362768 511 REFTEK130 LE3D-5S EPISENSOR Real-time 29/08-10/10/2016 AM04 Amandola 42.983345 13.365007 455 REFTEK130 LE3D-5S EPISENSOR Real-time 30/08-10/10/2016 AM05 Amandola 42.977404 13.352786 464 REFTEK130 LE3D-5S EPISENSOR Real-time from 30/08/2016 CV01 Civitella del Tronto 42.772736 13.666229 642 Q330 LE3D-5S EPISENSOR Stand-alone 31/08-26/09/2016 CV02 Civitella del Tronto 42.773104 13.669496 600 Q330 LE3D-5S EPISENSOR Stand-alone 30/08-26/09/2016 CV03 Civitella del Tronto 42.772644 13.672969 540 Q330 LE3D-5S EPISENSOR Stand-alone 30/08-26/09/2016 CV04 Civitella del Tronto 42.772273 13.666690 585 Q330 LE3D-5S EPISENSOR Stand-alone 31/08-26/09/2016 CV05 Civitella del Tronto 42.771551 13.663721 605 Q330 LE3D-5S EPISENSOR Stand-alone 31/08-26/09/2016 CP01 Capitignano 42.52458 13.28758 885 REFTEK130 LE3D-5S EPISENSOR Stand-alone 29/08-17/09/2016 CP02 Capitignano 42.52389 13.29596 898 REFTEK130 LE3D-5S EPISENSOR Stand-alone 29/08-19/09/2016 CP04 Capitignano 42.52529 13.30163 957 REFTEK130 LE3D-5S --- Stand-alone 30/08-16/09/2016 CP05 Capitignano 42.52052 13.30267 925 REFTEK130 LE3D-5S EPISENSOR Stand-alone 29/08-16/09/2016 CP06 Capitignano 42.53498 13.28591 900 REFTEK130 LE3D-5S EPISENSOR Stand-alone 29/08-17/09/2016 CP07 Capitignano 42.50748 13.27508 842 REFTEK130 LE3D-5S EPISENSOR Stand-alone 30/08-16/09/2016 MN02 Montereale 42.51393 13.26704 827 REFTEK130 LE3D-5S EPISENSOR Stand-alone 26/08-16/09/2016 MN03 Montereale 42.51916 13.25532 823 REFTEK130 LE3D-5S EPISENSOR Stand-alone 26/08-17/09/2016 MN04 Montereale 42.52906 13.23519 977 REFTEK130 LE3D-5S EPISENSOR Stand-alone 27/08-16/09/2016 MN06 Montereale 42.52400 13.24480 923 REFTEK130 LE3D-5S EPISENSOR Stand-alone 26/08-16/09/2016 MN08 Montereale 42.52624 13.24125 916 REFTEK130 LE3D-5S EPISENSOR Stand-alone 27/08-16/09/2016 MN09 Montereale 42.52987 13.24546 827 REFTEK130 LE3D-5S EPISENSOR Stand-alone 27/08-16/09/2016