NH4.4

Terre del Reno is a municipality in the Emilia-Romagna Region (Italy) that experienced relevant liquefaction events during the 2012 seismic crisis, which was characterised by two main shocks (ML 5.9 and 5.8).  Such events are mainly related to the complex geo-stratigraphic setting of the area. In this background, the present work is devoted to achieving two main objectives: i) define a new integrated methodology to assess liquefaction susceptibility in complex stratigraphic conditions through a multi-level approach; ii) perform a level 3 seismic microzonation study of Terre del Reno. To this purpose, more than one thousand geophysical and geotechnical measurements available from three different databases and some hundreds of new collected investigations were stored in a dedicated geodatabase. Data and metadata, that were spatially and statistically manipulated to guarantee their harmonization, standardization, and uniqueness, were explored to reconstruct a model for the Terre del Reno subsoil. Specifically, a geological model of the studied area (~ several hundreds of meters) was first reconstructed as well as the seismic bedrock geometry (the latter defines as the layer characterized by the stiffness requirement: Vs > 800 m/s). This model was obtained by integrating deep bore-hole data available from previous studies and geophysical and geotechnical investigations. Furthermore, a high-resolution geological reconstruction of the upper 30 m has also been performed through sedimentological and paleo morphological analysis to characterize the sedimentary units affected by liquefaction. This analysis may be used to compare both well-known and innovative geotechnical indicators for liquefaction susceptibility assessment. Thus, a set of acceleration time histories, that are spectrum-compatibles with the spectrum of reference input motion at outcropping bedrock of the site, were used as input in 1D and 2D site effect numerical modelling. The obtained results were synthetized and represented in a level 3 seismic microzonation study with the aim of providing operational indicators devoted to urban planning and for challenging problem related to liquefaction.

How to cite: Varone, C., Baris, A., Caciolli, M. C., Fabozzi, S., Fortunato, C., Gaudiosi, I., Giallini, S., Mancini, M., Martelli, L., Modoni, G., Moscatelli, M., Paolella, L., Simionato, M., Sirianni, P., Spacagna, R. L., Stigliano, F., Tentori, D., and Razzano, R.: An integrated approach for engineering - geological modelling in view of seismic microzonation, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8586, https://doi.org/10.5194/egusphere-egu22-8586, 2022.

How to cite: Hussain, Y., Cauchie, L., Mreyen, A.-S., and Havenith, H.-B.: The near-surface velocity structures of an incipient volcanic flank collapse revealed by geophysical studies (preliminary results), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8973, https://doi.org/10.5194/egusphere-egu22-8973, 2022.

Shallow sedimentary layers have a strong impact on seismic motion. These so-called site effects may be responsible for dramatic ground motion amplification and increase the duration of shaking when an earthquake occurs. The quantification of such amplification effects for specific sites might be challenging to carry out in low-to-moderate seismicity regions where moderate to large earthquakes have long return periods. Therefore, methods based on background ambient noise might be of great interest for these areas.

In this study, we investigate the potential of ambient noise in ground motion amplification assessment through SSRn (noise-based Standard Spectral Ratio) and SSRh (hybrid Standard Spectral Ratio, Perron et al., 2018) computation. We continuously recorded ambient noise from February to March 2020 on a 400-sensor seismic array covering an area of about 10 x 10 km in the Tricastin industrial region (French Rhone Valley) where critical facilities are located. This area is located on a very elongated valley, filled with Pliocene sediments (sands and clays), that was dug during the Messinian Salinity Crisis in Cretaceous sandstones and limestones. The strong lithological contrast between the sedimentary filling and the bedrock, as well as the valley's incised geometry, is prone to generate strong and complicated site effects.

Previous studies have shown that SSRn is not able to reproduce earthquake-based SSR amplification factor for frequencies higher than 1 Hz. This disagreement may be explained by the influence of local noise sources. Here, we introduce an approach to mitigate the influence of strong local sources in SSRn and SSRh. Our workflow relies on a clustering algorithm to select the Fourier Amplitude Spectrum (FAS) used in the SSRn and SSRh computation. By applying this method, we were able to remove strong anthropic transient signals at some sites and therefore improve the amplification assessment above 1Hz through the SSRn and SSRh. However, half part of the array is located nearby permanent anthropic sources that remain a major issue in quantifying the amplification at the scale of the valley. This study provides some insights into the conditions of applications of SSRn and SSRh in noisy industrialized environments.

How to cite: Gisselbrecht, L., Froment, B., and Boué, P.: Noise-based estimation of local seismic amplification in an industrialized area of the French Rhone Valley, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9511, https://doi.org/10.5194/egusphere-egu22-9511, 2022.

Amplification of seismic energy in steep topography is widespread and plays an important role affecting the locations of earthquake-induced damage and the distribution of earthquake-triggered landslides. Mountains, and especially the large freestanding massifs of the European Alps, represent extreme topography and may thus exhibit larger topographic amplification than features with less relief. However, suitable broadband seismic data from these locations are rare, in part due to difficult and often dangerous site access. Here we present ambient seismic data collected on two mountains in the Swiss Alps (the Matterhorn and Grosser Mythen), similar in shape but different in scale. At the Matterhorn, comparing data from seismic stations on the summit and ridge to a nearby local reference showed elevated spectral power on the mountain between 0.4 and 1 Hz, and directional site-to-reference spectral amplitude ratios up to 14, which we attribute in part to topographic resonance. We used ambient vibration modal analysis and numerical eigenfrequency modeling to identify the fundamental mode of the Matterhorn at 0.42 Hz, as well as evidence for a second, mutually-perpendicular mode at a similar frequency. Our data further show high modal damping ratios of ∼20% for these modes, which we ascribe to radiative energy loss. A short campaign measurement at Grosser Mythen, showed similar modal properties with a higher fundamental frequency of 1.8 Hz and peak spectral ratios of 6. At the Matterhorn, we analyzed 13 months of continuous data, showing that spectral peaks are stable over time and that the fundamental frequency of the mountain does not measurably vary. Our results aid estimation of topographic amplification for other mountain features.

How to cite: Weber, S., Beutel, J., Häusler, M., Geimer, P. R., Fäh, D., and Moore, J. R.: Ambient seismic amplification in extreme topography: instrumental evidence from the Matterhorn, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6950, https://doi.org/10.5194/egusphere-egu22-6950, 2022.

In the past decade, studies to determine amplification effects due to the local geology have been conducted in the Maltese islands (Central Mediterranean) by means of ambient noise techniques. Particular areas of interest include the north and north-western areas of the islands which are characterised by clay, that can reach a thickness of 75 m, buried beneath limestone.  This introduces a velocity inversion in the stratigraphy and consistent, characteristic peaks in the H/V spectral ratios. With the expansion of the Malta Seismic Network (MSN) to these geological areas of concern, the possibility of confirming and further investigating the results using empirical data arises. Here we present results, mainly in terms of Standard Spectral Ratio (SSR) and earthquake H/V, using 3 years of earthquake data at three stations of the MSN. In particular we note that the amplifications obtained using the SSR technique are significantly higher than those obtained using both noise and earthquake H/V techniques. While the peaks observed in the H/V spectra are also reproduced in the SSR curves using earthquake data, the latter exhibit important additional peaks at frequencies below 1 Hz, whose amplitude may be as high as 30. By separating the earthquake data set on the basis of distance from the islands, we show that the amplification is source-dependent, and that the high amplification values originate from larger, more distant earthquakes in the Hellenic arc. This has important implications for seismic hazard determination.

How to cite: Farrugia, D., Galea, P., and D'Amico, S.: Assessing local site response using earthquake data: The case of thick buried low-velocity layers, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9387, https://doi.org/10.5194/egusphere-egu22-9387, 2022.

Earthquake-induced landslides (EQILs) are severe secondary effects of a seismic event. These phenomena can produce direct and indirect damages to structure, infrastructures and to the society. EQILs can also trigger secondary hazard cascading effects and impact civil protection activities. For this reason, the scientific community over the past decades has dedicated an increasing attention to EQILs, with the publication of numerous scientific papers. In this work we describe a preliminary outcome of a comprehensive review of the main articles published on this topic from 1984 to 2021 in peer reviewed international journals. The selected articles, which have been identified after a systematic search on the Clarivate analytics’ Web of Science-Core Collection™ online platform, have been catalogued in a database, whose structure was designed to include the information preparatory for the analysis. Specifically, for each of the 798 articles we reported: a) the bibliometric information (i.e., article title, author(s), publication year, journal name and number of citations); b) the specific topic addressed by the article, which can be distinguished with respect to the scale of the analysis (i.e., regional or single slope) and the type of research (e.g., mapping, characterization/description, modelling); and c) the information related to the earthquake(s) considered in the article. In the database, we have identified 139 earthquakes whose main characteristics (e.g., date of occurrence, location, magnitude, focal mechanism) have been organized in a sub-section of the database. The analyses pointed out different commonalities between articles which allowed us to infer general aspects related to EQILs and, at the same time, to describe a comprehensive state of art on the topic.

How to cite: Schilirò, L., Rossi, M., Polpetta, F., Fiorucci, F., Fortunato, C., and Reichenbach, P.: A systematic review of scientific literature on earthquake-induced landslides, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7045, https://doi.org/10.5194/egusphere-egu22-7045, 2022.

First analyses of landslide distribution and triggering factors are presented for the region affected by the August, 14, Nippes, earthquake (Mw=7.2) in Haiti. Landslide mapping was mainly carried out by comparing pre- and post-event remote imagery (~0.5-1 m resolution) available on Google Earth® and with Sentinel 2 (10 m) satellite images. The first cover about 50% of the affected region (for post-event imagery), the latter were selected to cover the entire potentially affected zone. On the basis of the completed landslide inventory, comparisons are made with a catalogue compiled by the USGS for the January, 12, 2010 seismic event (Mw=7.0); additionally, we also analyzed the pre-2021 earthquake slope stability conditions. These comparisons show that the  total number of landslides mapped for the 2021 earthquake (=7091) is smaller than the one observed for the 2010 (=23567). However, these fewer landslides triggered in 2021 cover a much wider area of slopes (>80 km²) than those induced by the 2010 event (~25 km²). A simple statistical analysis indicates that the lower number of 2021-landslides can be explained by the ‘under-mapping’ of smallest landslides triggered in 2021, partly due to the lower resolution imagery available for most of the areas affected by the recent earthquake; this is also confirmed by an inventory completeness analysis based on size-frequency statistics. The much larger total area of landslides triggered in 2021, compared to the 2010 earthquake, can be related to different physical reasons: a) the larger earthquake magnitude in 2021; b) the more central location of the fault segment that ruptured in 2021 with respect to coastal zones; c) and possible climatic pre-conditioning of slope stability in the 2021-affected area. These observations are supported by (1) a new pre-2021 earthquake landslide map, (2) rainfall distribution maps presented for different periods (including October 2016 - when Hurricane Matthew had crossed the western part of Haiti), covering a region including both 2010- and 2021 affected zones, as well as (3) the shaking intensity prediction and related simplified Newmark Displacement maps.

How to cite: Havenith, H.-B., Guerrier, K., Schlögel, R., Mreyen, A.-S., Ulysse, S., Braun, A., Victor, K.-H., Saint-Fleur, N., Cauchie, L., Boisson, D., and Prépetit, C.: First analysis of landslides triggered by the August 14, 2021, Nippes (Haiti) earthquake, compared with the 2010 event, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1980, https://doi.org/10.5194/egusphere-egu22-1980, 2022.

The Apulian region (south-east of Italy) is extensively characterized by the presence of old underground cavities used in the past for the extraction of calcarenite rock, a very soft carbonate rock, and then abandoned over time. The assessment of the stability conditions for such caves is a very challenging problem, especially where the interaction of cavities with overlying structures or infrastructures is possible. The role of environmental factors in the triggering of cave failure processes has been widely studied in the literature, for instance by Parise & Lollino (2011), Castellanza et al. (2018), Perrotti et al. (2018) to mention a few. However, the instability processes related to dynamic loads are often underestimated. In fact, very few literature works exist in the specific field: Genis & Gercek (2003) have firstly demonstrated the role of dynamic waves in the enlargement of the yield zone around the cave; Genis & Aydan (2007; 2008) have carried out some studies applied to real cases, specifically focusing on the pillar behaviour. The effects of the interaction between adjacent cavities has been also investigated by Gercek (2005) and Landolfi (2013), all highlighting that the presence of cavities at short distance induces larger risk conditions under dynamic conditions.

This work is aimed at investigating the effects of dynamic loads, in accordance with regional seismicity, on the evolution of plasticity within man-made underground cavities in soft calcarenite. Both the seismic behaviour of single ideal caves and that of twin adjacent caves have been analyzed. In order to investigate the evolution of the stress-strain state of the cavity under dynamic loading and the corresponding equilibrium conditions, a parametric analysis was carried out. The parametric analysis was performed by varying both the geometrical features of the ideal cavity, in accordance with the typical values observed for the Apulian underground cavities system, and the seismic input characteristics. An elastic-perfectly plastic Mohr-Coulomb model, integrated by viscous damping according to the frequency-dependent Rayleigh formulation, has been adopted.

The numerical results highlight appreciable widening of the rock zones at yielding caused by the dynamic input, especially in the case of wide cavities. Also, the overburden roof thickness plays a significant influence, since a clear increment of the difference between the static and dynamic behaviour of the rock mass is observed when the roof thickness increases. The numerical results also indicate that the dynamic cavity stability depends on the energy content of the dynamic input.

In addition, the numerical model implementing the interaction between twin cavities under dynamic conditions shows the tendency to plastic failure in the septum, which is enhanced in the dynamic phase compared to the static one, and again dependent on the width of the cavity, the thickness of the roof and the energy content of the dynamic load.  

Lastly, the research has also proposed a methodology to calculate the factor of safety with respect to the occurrence of a general failure of underground cavities under dynamic conditions, which allows to quantify the change of the stability conditions from static to dynamic conditions.

How to cite: Lollino, P., de Lucia, D., and Fazio, N. L.: Failure susceptibility assessment under dynamic conditions of man-made underground caves in soft rocks, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7586, https://doi.org/10.5194/egusphere-egu22-7586, 2022.

Landslides are among the most frequent secondary effects related to seismic events. The prediction of the expected displacements of landslides activated by seismic shaking (1^st-time failures and reactivated landslides) is therefore a substantial feature for the hazard assessment in high seismicity regions. Several databases collecting geological and geometrical information on worldwide landslides events are available in literature. This study presents the result of statistical analyses on morphometric, topographic and geotechnical parameters extracted from existing landslide databases (Domej et al., 2020; Martino et al., 2019; Tanyas et al., 2019). The aim is to define a procedure to generate 2D step-like-slope landslide models representative of the most common landslides in terms of failure mechanism (divided into two main categories: purely rotational and translational landslides), volume, and geotechnical properties. Rock falls and toppling, flow-like landslides and deep-seated landslides were excluded from the initial dataset, because they are associated with peculiar physical processes during the failure and the propagation. The performed statistical analysis allowed to identify the most frequent values of depth/length ratios, volume and slope angle, from which other geometrical measurements were analytically derived. In addition, various landslide locations along the slope were considered to cover most of the real cases. This resulted in 36 different landslides/slopes shapes. Landslides dynamic/geotechnical parameters (shear wave velocity, density, strength) were selected to be consistent with those inferred for rocks, cohesive soils and granular soils by statistical analysis. The representativeness of the inferred models is assessed by comparing the theoretical geometries with the real ones detected in the Campotosto Basin (Central Apennines, Italy), a high seismicity area very close to the Amatrice village, which was strongly hit by the recent 2016-2017 (M 6.5) Central Italy seismic sequence as well as by the 2009 (M 5.9) L’Aquila earthquake. The simplified landslide models represent the first part of a major study on the prediction of seismically induced landslide displacements. The aim is to improve the existing Newmark’s approach–based PARSIFAL (Probabilistic Approach to pRovide Scenarios of earthquake-Induced slope FAiLures) method (Martino et al., 2019) to assess the earthquake-induced displacements at a regional scale, by introducing corrective factors derived from parametric dynamic numerical simulations on the simplified geometries; such factors should incorporate some aspects of the complex seismic waves/landslides slopes interaction. Such a procedure will allow to overcome the Newmark’s method limitations and to extend the advantages of the numerical analyses, the use of which is generally limited to studies at slope scale, over larger areas.

How to cite: Mita, M., Di Renzo, M. E., Bourdeau, C., Fiorucci, M., Marmoni, G., Antonielli, B., Esposito, C., Lenti, L., and Martino, S.: 2D simplified landslide models inferred by statistical analyses on existing landslide databases for multi-hazard analysis: an application to the Campotosto Lake basin (Central Apennines, Italy), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11746, https://doi.org/10.5194/egusphere-egu22-11746, 2022.

The 2005 Mw 7.6 Kashmir earthquake is the most devastating earthquake to occur along the Himalayan arc, resulting in 87,000 fatalities, 69,000 injuries, and 2.8 million people left homeless. The rupture occurred on a 30° NE dipping thrust fault and generated a ~70 km long surface rupture that concentrated much of the damage. Along with the primary hazard caused by the seismic shaking, many secondary hazards, including nearly 3,000 coseismic landslides, were initiated due to the shaking from this event. In the absence of seismic data recorded near the source of this earthquake, we attempt to understand the relationships between ground shaking and coseismic landslides by using numerical techniques to model the ground motions and topographic amplification from the Kashmir earthquake. We use the spectral element method implemented in SPECFEM3D to model kinematic rupture scenarios for the Kashmir earthquake in both high resolution and flat topography, obtaining a topographic amplification factor by comparing these simulations. We generate a range of seismic source models using the rupture generator FakeQuakes, starting with a mean slip model from the earthquake and adding stochastic variations to both static and kinematic rupture properties to produce variable rupture scenarios. The advantages of this technique, compared to calculating ground motions from one finite fault model, is that by adding stochastic variations, the source model has higher, more realistic, frequencies, and that it enables the investigation of how varying rupture properties affect topographic amplification. We calculate both peak ground velocity (PGV) and topographic amplification for each scenario and compare the average and standard deviations to locations of landslide initiation. Preliminary results from five earthquake sources shows that with changing source parameters, PGV and topographic amplification patterns remain relatively constant and that positive amplifications are concentrated at the peaks of ridges and negative amplifications are concentrated in valleys. There are no obvious relationships between the patterns of amplification and landsliding, possibly due to limitations in landslide mapping. Other causes of landslides--such as variations in lithology, distance to anthropogenic features (roads, construction), distance to faults, and distance to ridges, and rivers--will be investigated further to understand the relationships between topographic amplification and other triggers. Future work includes combining these results with similar studies for earthquakes with different source properties and in different topographic settings to further understand the controlling factors of topographic amplification as a trigger for coseismic landslides.

How to cite: Dunham, A., Kiser, E., Kargel, J., Haritashya, U., Watson, S., and Shugar, D.: Ground motion simulations relating topographic amplification and landslide initiation during the Mw7.6 2005 Kashmir Earthquake., EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10750, https://doi.org/10.5194/egusphere-egu22-10750, 2022.

Earthquake-induced ground effects are strongly related to the presence and distance of active and capable faults, and they play an extremely important role in the mitigation of seismic risk. The Italian Seismic Microzonation Guidelines subdivide the active and capable faults in ‘certain and defined’ and ‘uncertain’, attributing to them microzones with defined landscape uses: ‘Respect’ and ‘Susceptibility’ zones respectively. In this work, we present the methodology used to map and analyze the Montereale basin’s faults, located in the highly seismic region of the central Apennines of Italy. The Montereale faults (MFS) pertain to two fault systems with an en echélon array, namely the San Giovanni and Capitignano fault systems. Yet the great scientific attention in this region, these faults still lack clear evidence of relationships with the major active and capable structures in the neighboring area that are considered responsible for the seismic events that affected central Italy in recent decades.

The San Giovanni fault cuts in heterogeneous deposits consisting of calcareous lithotypes, which expose well defined fault planes and easily recognizable fault scarps. Instead, the Capitignano fault occurs on softer arenaceous-pelitic deposits, which make hard to identify tectonic discontinuities.

The approach, by which we have mapped the Capitignano fault and defined Susceptibility and Respect microzones for the MFS, is divided into the following phases: 1. Identification of morphotectonic elements by the analysis of digital terrain models (DTM 10 m and LiDAR 1 m), morphological elements (linear slopes, non-degraded triangular facets, anomalies in the drainage network, linear valleys, saddles, alignments of slope breaks) represent the most evident expression of active tectonics. 2. Geological and geomorphological survey for the interpretation of the elements recognized by remote sensing data. 3. Geophysical surveys (tomography electrical resistivity and seismic reflection), planned based on the morphotectonic features, identified in the previous stages. 4. Paleoseismological trenches, located where geophysical investigations have confirmed the presence of subsoil’s discontinuities. 5. Dating of faulted soils.

Following this method, the recognition of active and capable faults was possible, even where their morphological expression was not evident or completely absent. Moreover, the study outcomes provided new pieces of evidence for a comparison with the neighboring and well-studied fault systems allowing to propose eventual structural relationships.  Finally, we believe that the proposed approach can be a powerful tool in regions densely affected by earthquakes. In fact, a deep knowledge of fault network and their mutual interactions allows to limit damage to people and inhabited centers and to plan reconstruction works in areas affected by seismic events.

How to cite: Sciortino, A., Chiarini, E., Nirta, G., Spadi, M., Tallini, M., Ferri, F., Puzzilli, L. M., Sapia, V., and Materni, V.: Mapping Active and Capable faults in structural complex settings. A case study from central Apennines (Italy)., EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6017, https://doi.org/10.5194/egusphere-egu22-6017, 2022.

The Seismic Microzonation, as practiced in Italy, consists in defining microzones of the territory affected by homogeneous response to seismic ground shaking, defined as stable zones, vulnerable seismic amplification zones, and unstable zones. In detail, the unstable zones are affected by landslides, soil liquefaction, ground subsidence and surface faulting. In this framework, we conducted a study in the administrative district of L’Aquila (central Italy) aimed at the construction of a new school building, in an area indicated as prone to surface faulting (an active and capable fault was hypothesised in the area) and liquefaction.

We dug two trenches (named as A and B) perpendicular to the presumed active fault trace. The excavation walls exposed several different continental units mainly characterized by colluvial, organic-rich and “cultural” sediments, as well as paleosols. In trench A, some units, made of sandy-gravelly colluvial deposits, contained abundant pottery fragments, being intensely reworked by very recent human activity. These units were mainly composed of silt and sand sparse with carbonate clasts and directly overlying Middle Pleistocene alluvial deposits. Trench B only exposed units containing pottery fragments and hence pertaining to historical times. Several radiocarbon dating made on charcoal found within these units confirmed the recent age of the deposits, spanning from 25000 to 1800 years before the present. The analysis of the trench walls, the analysis of two boreholes, and field geological investigations revealed the absence of any surface faulting events affecting the stratigraphic sequence of the area, at least since the Middle Pleistocene, likely since the Early Pleistocene. Furthermore, trench B exposed several sedimentary dikes reaching up close to the ground surface, crossing the historical colluvial units, as well as other deformation features typical of liquefaction phenomena. The radiocarbon age determination and the sedimentological characteristics of the units indicate that the most recent liquefaction event occurred after 180 A.D.

Ultimately, this work represents a “best-practice” case study to investigate the occurrence of geological surface criticalities (such as surface faulting and liquefaction) at specific sites of interest.

How to cite: Spadi, M., Maceroni, D., Dixit Dominus, G., Tallini, M., Falcucci, E., Galadini, F., Gori, S., Moro, M., and Saroli, M.: Surface faulting and liquefaction hazard assessment in the central Apennines for land use practices: a case study from the L’Aquila urban area (central Italy), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6312, https://doi.org/10.5194/egusphere-egu22-6312, 2022.

The huge impact caused by liquefaction during past earthquakes stimulates the interest of researchers in investigating the factors ruling the susceptibility of subsoil and the triggering conditions. The concern of stakeholders raises the need for risk assessment methods applicable at the large scale. A crucial aspect for liquefaction risk assessment consists in the subsoil characterization, with the  stratigraphic classification into homogeneous soil layers and the identification of the susceptible volumes, with the aim of constructing 2D and 3D geo-mechanical models. In the current practice, the CPT-based soil behavior type (SBT) and the soil behavior type index (Ic), are widely used to identify soil boundaries discontinuities (Robertson, 2016). Sometimes, the interpretation of subsoil profile is not immediate and unique, due to the lack of evident boundary changes. In these cases, the need is felt for sound, widely applicable tools that provide univocal identification of subsoil strata. Statistical procedure, developed over the years, provides a less subjective interpretation of the subsoil and, in conjunction with artificial intelligence, can lead to improve the current methodology obtaining an objective and extensive site characterization. This work exposes a data-driven analysis for the subsoil stratigraphic recognition combining geostatistical tools and AI genetic algorithms. The presented procedure is calibrated and validated on the case study of Terre del Reno (Italy), severely struck by liquefaction during the 2012 Mw 6.1 earthquake and characterized by complex geo-stratigraphic conditions. The selected area, homogeneously covered by about 1700 geognostic surveys, is investigated within the "PERL" research project, carried out by the Emilia Romagna Region (RER), CNR-IGAG and UniCas-DiCeM, aiming to provide a reliable procedure for liquefaction risk assessment and a seismic microzonation. From the RER geodatabase, 102 pairs of complementary CPT and boreholes were extracted to calibrate the method, defined as the couples of surveys located at a relative distance less than 30m, considered for this purpose as spatially correlated. Starting from the information available from the boreholes, a geologic-sedimentologic study has been carried out to define the main stratigraphic units. In parallel, CPT profiles are processed with a statistical method based on the spatial variability analysis of the measured parameters, identifying statistically homogeneous layers and associating to each of them the correspondent stratigraphic unit reported in the complementary borehole. At this stage, an artificial intelligence algorithm has been calibrated merging the outcomes derived from couples of CPTs and boreholes. Subsequently, the procedure has been applied to the remaining CPTs, combining the geological and geotechnical knowledge of the subsoil in an efficient and automatic way to enable a large-scale reconstruction of the subsoil stratigraphy.

How to cite: Baris, A., Caciolli, M. C., Fabozzi, S., Gaudiosi, I., Mancini, M., Martelli, L., Modoni, G., Moscatelli, M., Paolella, L., Razzano, R., Simionato, M., Spacagna, R. L., Stigliano, F., Tentori, D., and Varone, C.: Automatized CPT-based soil profile characterization for liquefaction susceptibility assessment, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9779, https://doi.org/10.5194/egusphere-egu22-9779, 2022.