Scenario-based tsunami hazard assessment at Kolumbo volcano

Scenario-based tsunami hazard assessment at Kolumbo volcano | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Scenario-based tsunami hazard assessment at Kolumbo volcano Alessandro Tadini, Matteo Cerminara, Raphaël Paris, Augusto Neri, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5700315/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Jun, 2025 Read the published version in Bulletin of Volcanology → Version 1 posted 6 You are reading this latest preprint version Abstract Volcanic-induced tsunamis have a potentially devastating impact, especially in densely populated and/or touristic coastal areas. Kolumbo submarine volcano (Greece) experienced in 1650 CE an explosive eruption with eyewitnesses’ accounts of major tsunamis along the coasts of Santorini (Thera) and other islands. We present a scenario-based tsunami hazard assessment at this volcano based on existing simulations from literature and new simulations of tsunamis triggered by a less investigated but important mechanism, i.e. submarine landslides on the volcano flanks or within its crater. Simulations results show that the remobilization of a landslide volume of 150–300 Mm 3 inside the crater can produce tsunami waves larger than 10 m high along the NE coast of Thera and of the order of 5 m along the E and SE coasts. The expected tsunami arrival time ranges from 2–3 minutes along the NE coast of Thera up to 8–10 min on its SE coast. Such scenarios produce inundation areas consistent with those reconstructed for the 1650 CE event, and tsunami waves propagating inland at velocities from 2 to 12 m/s. Simulation results also suggest that, given the landslide parameters assumed, it is unlikely to mobilize a landslide with a large volume from the SW-facing Kolumbo crater slopes, given the relatively gentle topo-bathymetry of this area. The study findings are relevant based on the outcomes of the expert elicitation exercise carried out in parallel, which indicate that chances of having waves larger than 1 m high on the NE coast of Thera have median probabilities of 50–60%. Kolumbo volcano (Greece) 1650 CE eruption tsunami hazard numerical simulation Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Introduction Despite not being a frequent event, volcanogenic tsunamis represent one of the deadliest phenomena potentially associated with volcanic eruptions (Auker et al. 2013; Paris et al. 2019; Schindelé et al. 2024), during which eruptive, volcano-tectonic, and gravitational phenomena can lead to the generation of tsunamis (Paris 2015). Tsunami hazard is indeed potentially high for submarine volcanoes in shallow-water conditions (Day 2015; Syamsidik et al. 2020; Terry et al. 2022), where tsunamis can be triggered by explosions, the gravitational collapse of the subaerial eruptive column, pyroclastic flows penetrating the sea, caldera collapse, or submarine flank instability (landslides). Moreover, violent explosive eruptions in shallow-water may also trigger meteo-tsunamis such as that occurred during the 2022 Hunga Tonga Hunga Ha'apai eruption (Shen et al. 2024). A shallow-water condition is met for Kolumbo, an active submarine volcano located ~7 km NE of Santorini (Thera in Greek language) island within the Aegean Sea (Fig. 1) which last erupted in 1650 CE (Vougioukalakis et al. 1996; Cantner et al. 2014; Vougioukalakis et al. this volume). Particularly, this explosive eruption has been characterized by several hazardous phenomena, the main ones being tephra fallout (Fuller et al. 2018), gas poisoning (Konstantinou 2020), and tsunami (Nomikou et al. 2014; Ulvrova et al. 2016; Katsigera et al. 2024). Building on the heavy consequences that this eruption had on population and on the proximity of Kolumbo volcano to the highly touristic Thera island, a comprehensive hazard assessment project has been carried out for Kolumbo volcano under the auspices of the Greek civil protection authorities (Sparks et al. this volume). A core aspect of this project has been the setup of an expert elicitation exercise (Bevilacqua et al. this volume) to quantify the major uncertainties associated with specific aspects of the reconstruction of the 1650 CE eruption but also with potential hazards, including tsunami, associated to future explosive eruptions of Kolumbo volcano. The aim of this paper is to present a scenario-based tsunami hazard assessment for Kolumbo volcano based on a review of existing modelling studies and new numerical simulations that were carried out to inform the expert elicitation target questions on tsunami hazard. The previous literature using numerical simulations includes a study by Ulrova et al. (2016), investigating a variety of possible mechanisms of the tsunamis in 1650 CE (including pyroclastic flows, underwater explosions, and caldera collapse) and a study by Karstens et al. (2023) on combining landslide and explosion mechanisms to reconstruct this event. These studies together with the outcomes of our new simulations can be used to investigate the tsunami hazard in future eruptions. In particular, our new simulations aimed to investigate the dynamics of tsunamis generated by the, so far, less studied mechanism, that is the formation of a submarine landslide on the volcano flanks or within the crater. The new simulations can describe the coupled dynamics of landslide and tsunamis thus illustrating the effect of the source process on the tsunami propagation. Simulation results also provide new insights on the potential of submarine landslides, in particular of those occurring inside the crater, to generate tsunamis on the coast of Thera, as well as on the interpretation of the genesis of the 1650 CE tsunamis. Modelling results are also compared to and discussed in relation to the tsunami-related outcomes of the expert elicitation procedure developed in the project (see Bevilacqua et al. this volume, target questions of block Q14). This paper is organized as follows: we begin with a review of i) the Kolumbo volcano and the 1650 CE eruption with particular focus on tsunami observations (Section “The Kolumbo volcano and the 1650 CE eruption”) and ii) the previous tsunami simulations (Section “Previous tsunami simulations”). Then, we illustrate our simulation strategy (Section “Materials and Methods”), present the results of the new simulations of submarine landslides and associated tsunamis also in relation to the outcomes of the expert elicitation questions (Section “Results”), and discuss the implications of the simulations results (both the new ones and those from literature) for the tsunami hazard assessment (Section “Discussion”). Finally, we summarize the main findings and some future perspectives in the “Conclusions” section. Background The Kolumbo volcano and its 1650 CE eruption The Kolumbo volcano is part of the Kolumbo volcanic field, which develops ~ 10 km NE of Thera island (Fig. 1, Vougioukalakis et al. this volume). The polygenetic and active Kolumbo central volcano is the southernmost edifice of this field, and its most recent eruption occurred from September to December 1650 CE with a major explosive eruption on 29 September. This powerful eruption left a 1700 m large, 500 m deep crater at the center of the volcano (Fig. 2; Nomikou et al. 2012). A detailed description of the available evidence and chronicles of the 1650 CE eruption of Kolumbo is reported in Vougioukalakis et al. (this volume) and Mastroianni et al. (this volume). In the following, we will briefly recall the main facts, with specific reference to the tsunamis. Most of the observations of the 1650 CE eruption and its related phenomena were compiled by Fouqué (1879) from earlier reports and descriptions of the phenomena. Between January 1649 CE and March 1650 CE, precursory earthquakes (with an intensity of VII on Mercalli scale) were felt on Thera Island. In September (14-26), subterranean roaring and green sea water attested the beginning of submarine volcanic activity, building up a submarine pumice tuff cone. Then, the intensity of the earthquakes increased, with several events reaching an intensity of VIII (Mercalli). Earthquakes persisted throughout the eruption, but numerical simulations suggest that they could not be the source of the largest tsunami that was observed during the paroxysmal stage of the eruption (Ulvrova et al. 2016). The islet emerged from seawater on 26 September, although its nature is not yet identified (Vougioukalakis et al. this volume). During the following days, phreatomagmatic explosions frequently broke the sea surface with intermittent jets of gas and ash. The sea was progressively covered with pumice. The following shallow-water phreatomagmatic phase (27-29 September) did not generate major tsunamis, but on 27 September a wave pushed a boat into the sea and then brought it back to the shore. The paroxysmal phase started in the early morning of 29 September and lasted up to a maximum of 2 days. The major explosions were heard as far away as 400 km in the Dardanelles. Fine ash was deposited up to western Turkey (150-200 km). Gas affected Thera’s inhabitants (20-50 deaths) and animals. Earthquakes accompanying the paroxysmal phase were felt up to Crete (120-140 km away), and one of the earthquakes was strong and long enough to cause damages to buildings in Thera. The largest tsunami occurred at some time on 29 September during this paroxysmal phase. Available testimonies did not report the timing and number of tsunamis thus it is difficult to associate the tsunami with a particular source mechanism. Waves were observed on the coast of Thera, where about 2 km² of land was eroded (revealing Hellenistic and Byzantine ruins at Kamari and Perissa), many trees uprooted, and five churches were destroyed (Dominey-Howes et al. 2000). The spatial distribution of the tsunami deposits allows estimating a minimum wave runup in the order of 15-20 m a.s.l. on the eastern coast of Thera (>630 m inundation distance near Monolithos settlement), and >3.5 m a.s.l. on the southern coast (>360 m inundation near Perissa) (Ulvrova et al. 2016). The coast inside the Thera caldera was not affected by the tsunami, consistent with numerical simulations (Ulvrova et al. 2016). The timing of the tsunami in the framework of the eruption chronology remains unclear, even if historical sources place it around nighttime. The tsunami also impacted nearby islands, especially Ios and Sikinos, located more than 20 km northwest of the volcano. On Ios, a local wave runup of 14-20 m a.s.l. associated with pumice deposition was observed on a rocky shore, but the precise location remains unknown. On Sikinos, the tsunami penetrated 240 m inland (Fouqué 1879: “350 pas”, erroneously translated as feet in many publications). Sea agitation and damage to ships are mentioned in Kea Island (150 km northwest of Kolumbo) and Crete (towns of Dia and Chania, 120 and 165 km to the south, respectively). Between 1 October and 20 December, the activity was characterized by periodic explosions (waning phase), gas release (e.g., 20 deaths on 4 November), and small tsunamis whose origin is unknown (possibly following a large explosion on 4 November, and renewed activity on 6 December). On 2 October, nine men approaching the eruption site by boat were found swollen and burnt. Previous tsunamis simulations All numerical simulations published so far aimed at understanding the possible source mechanism of the main tsunami observed at some time on 29 September during the paroxysmal phase of the 1650 CE eruption (Ulvrova et al. 2016; Karstens et al. 2023). The full list of parameters of the simulations presented in these two papers is provided in Table 1. Ulvrova et al. (2016) tested three different source mechanisms of tsunami: (a) Submarine explosions with energies ranging from 3x10 14 to 5.4x10 16 joules (plausible range of energies for shallow-water volcanic explosions). The depth of the explosion was set at 150 m, which corresponds to the approximate present-day rim of the crater. (b) Caldera collapses of different geometries (full collapse of the central crater, collapse of the upper half only, or deepening/collapse of the lower half), and different durations (from 1 minute to 1 hour). Recent examples of caldera collapse (e.g. Pinatubo 1991, Hunga Tonga 2022) suggest that this source mechanism takes at least 30 minutes (Schott et al. 1996; Gupta et al. 2022). (c) Pyroclastic density currents resulting from the gravitational collapse of the subaerial eruptive column, with different flow densities (from 1100 to 1500 kg/m 3 ), different flow velocities (from 5 to 30 m/s), different volumes (from 5 x 10 6 to 100 x 10 6 m³), and different volume flux (from 10 4 to 10 7 m 3 /s). Karstens et al. (2023) proposed a hybrid scenario, with a submarine landslide of the north-western flank of the volcano (thus accounting for a possible initial retreat of the sea on the eastern coast of Thera), followed by a powerful submarine explosion. They tested different flow densities (from 1250 to 1750 kg/m 3 ) and yield strength (from 5 to 10 kPa) for the landslide, but only one volume (1.2 km³). Explosion energies range from 3x10 14 to 2.2x10 16 Joules, following the methodology proposed by Ulvrova et al. (2016). They proposed that the slow movement of the landslide on the north-western flank of the volcano (of about 500-1000 m occurring in about 4 minutes) caused a major depressurization of the magmatic system and the associated major explosion. Ulvrova et al. (2016) also traced the limits of sedimentary deposits from the 1650 CE tsunami, giving an idea of the extent of the inundation. The waves thus reached minimum altitudes ranging between 3.5 m a.s.l. (Perissa, southern coast) and 20 m a.s.l. (Monolithos, eastern coast), corresponding to a minimum inundation of 360 and 630 m, respectively. Materials and Methods The new simulations of submarine landslides and associated tsunamis have been carried out by using the Multilayer-HySEA model (Fernández-Nieto et al. 2018 ; Macías et al. 2020a , b ; Esposti Ongaro et al. 2021 ). They investigated in detail three scenarios: the case of a “small-scale” landslide occurring along the outer SW flank of the Kolumbo volcano described by Simulation 1 with volume of about 1.8 Mm 3 , and two “large-scale” slope failures inside its central crater described by Simulations 2 and 3 with volumes of about 147 and 300 Mm 3 , respectively. Numerical model The Multilayer-HySEA code is a multilayer, non-hydrostatic model in which the three-dimensional model equations are depth-averaged across a number of vertical layers. The governing equations correspond to a semi-discretization for the vertical variables of the Euler equations. The total pressure is decomposed into a sum of hydrostatic and non-hydrostatic pressures. In this process, the horizontal and vertical velocities are assumed to have a constant vertical profile in each layer. The proposed model admits an exact energy balance and, when the number of layers increases, the linear dispersion relation of the linear model converges to the same of Airy's theory (Fernández-Nieto et al. 2018 ). Moreover, the Multilayer-HySEA model can simulate the two-way interaction of the tsunamis with a landslide: in this case the motion of the bottom surface is represented by an additional layer described by the shallow-water equations of granular material. In this way, the Multilayer-HySEA code incorporates the possibility of simulating the generation of tsunami produced by subaerial or submarine granular landslides. The motion of the landslide is described by a granular landslide model (Fernández-Nieto et al. 2008 ), in which the Pouliquen and Forterre ( 2002 ) friction law is implemented. This law characterizes the dependence of static and dynamic friction coefficients on landslide velocity and thickness through three parameters (three friction angles) δ 1 , δ 2, δ 3 (Macías et al., 2020b ; Esposti Ongaro et al. 2021 ). The granular landslide model is in turn weakly coupled with the non-hydrostatic multilayer model through the boundary conditions. For more details on the model description and validation tests we refer the reader to Macías et al. ( 2020a , b ). The Multilayer-HySEA numerical code is designed to run on Graphic Processing Unit (GPU) accelerated High-Performance Computing (HPC) architectures (Escalante et al. 2018 ; 2019 ). The model in this configuration has been already applied to tsunamis generated in volcanic islands such as Stromboli (Italy) as described in Esposti Ongaro et al. ( 2021 ). Topo-bathymetry For the topo-bathymetry of the study area, we considered a 23 x 24 km domain including the Kolumbo crater and the whole Thera Island (Fig. 1 ). The employed topo-bathymetry is the result of a merging between two different datasets: 1) the bathymetry of the surroundings of Kolumbo volcano and Thera Island and 2) the topography of Thera Island. The bathymetry was collected during several multibeam echosounder surveys by R/V Aegeo using Seabeam 2120 (20 kHz) echosounder during 2001, 2006, and 2017 cruises (Alexandri et al. 2003 ; Sigurdsson et al. 2006 ; Sakellariou et al. 2010 ; Nomikou et al. 2014 ; Freundt et al. 2017). The pixel size of the original file is 0.03125 arc minutes, (X Y coordinate system WGS 84/ EPSG:4326), converted to GGRS87’ (pixel size 50 m), to create the bathymetry file. The original file was provided by the Hellenic Center for Marine Research, Hellenic National Oceanographic Data Center. The topography instead derives from a dataset of 5-m pixel size orthophoto maps, created from color countrywide air photos acquired during the “Large Scale Orthophotos” (LSO) Project (2007–2009). The merging of the two source datasets was performed in two steps: i) by homogenizing the cell sizes with a resolution which would allow affordable computational times and ii) by performing a smoothing of the junction zones between the two original data to avoid sharp changes in the topo-bathymetry that could cause instabilities in the numerical code. These steps were performed within the ArcGIS10© platform and allowed to obtain a final topo-bathymetry with a 10-m cell size. We note that the employed topo-bathymetry is obviously different from that in 1650 CE prior to the tsunami-generating event. The simulations described in the following sections are not specifically aimed at reproducing the 1650 CE tsunami, although they can provide useful insights for the interpretation of the tsunami source mechanism of this event. Simulation source conditions To deal with a realistic source condition and to get an accurate dispersion modeling of the waves, the Multilayer-HySEA model was used by adopting a 3 vertical layers configuration: the lower granular layer, representing the tsunamis-generating landslide, is two-way coupled with water, represented by the remaining two layers. A crucial aspect of the simulations conducted involves delineating the volume involved in triggering the tsunamis through the generation of the landslide. This delineation was achieved by intersecting the current bathymetric data of the volcano with an ellipsoidal-shaped volume centered at a specified spatial coordinate and defined by its three semi-axes. This process entails carving out a region from the bathymetric surface representing the potentially unstable granular layer that could lead to landslide formation. Following the delineation of the ellipsoid-shaped region, the simulation further determines the actual mobilized volume primarily on the basis of the friction angle used to model the rheological properties of the granular material constituting the landslide. Importantly, both the amount of mobilized material and its dynamics are not prescribed a priori but are evaluated based on the shallow water equations governing the granular layer and its interaction with the water layers. The proposed methodology allows the investigation of dynamics induced by instabilities within the actual bathymetry, facilitating a deeper understanding of potential tsunami generation mechanisms. While the present set of simulations does not explicitly explore the addition and flow of juvenile material resulting from a volcanic eruption onto the bathymetry, we believe that some useful insights on the potential impact of such a situation can be inferred from the outcomes of these new simulations. The ellipsoid is horizontally oriented with semi-axis a , b , and c (Fig. 3 ). Its position, angle of rotation and size are defined by (see Table 2): i) the (x c ,y c ) horizontal coordinates of the center C; ii) the (x a ,y a ) coordinates of the point A at the end of horizontal semi-axis a; iii) the length of horizontal semi-axis b. In addition, the vertical coordinate of points C and A (z a ) is defined by the elevation of point A on topo-bathymetry plus 20 m of vertical offset to avoid a vertical cut in correspondence of point A. Finally, the vertical semi-axis c is prescribed by imposing that the lowest point L (which has the same horizontal coordinates of C) is on the topo-bathymetry: L(x c ,y c ,z l ). Points A and L should be selected so that the former is shallower than the latter. A sketch summarizing the above-mentioned geometrical elements is provided in Fig. 2 . The resulting ellipsoid uniquely intersects the topo-bathymetry, and the intersection space represents the involved volume that could potentially slide, which however does not necessarily coincide with the landslide real volume mobilized. Such a mobilized volume of the landslide is dependent on the chosen rheology, depending in turn on the landslide density and the assumed friction angle (see next section for more details on the volumes involved). By varying the above-mentioned parameters (Table 2), we carried out three simulations: Simulations 1, 2, and 3. As already mentioned and as we will see in details in the next subsection, Simulation 1 corresponds to a “small-scale” (about 1.8 Mm 3 in volume) landslide occurring along the outer SW flank of Kolumbo, whereas Simulations 2 and 3 to “large-scale” (of 147 and 300 Mm 3 in volume, respectively) landslides generated by slope failures of the internal walls of the Kolumbo crater. We remark that these three simulations were chosen in order to investigate two potentially dangerous generation mechanism of tsunamis at Kolumbo, such as: 1) a small-scale submarine landslide on the SW external slopes directly facing the island of Thera (Simualtion 1), and 2) a large scale submarine landslide produced by the internal failure of the crater slopes which could produce a remarkable impact on Thera and surrounding islands (Simulations 2 and 3). In addition to these values, for all the three simulations, a landslide density of 2,000 kg/m 3 and friction angles δ 1 = 10°, δ 2 = 15° and δ 3 = 10° were considered. The chosen friction angles are assumed equal to those used in similar studies carried out a Stromboli volcano (Italy) (Esposti Ongaro et al. 2021 ). The density of the porous material underwater should be considered larger than that measured in air because the pores of the lapilli contained by the landslide are partially filled by water. The present study does not investigate the effects of density and friction angle variations on tsunami wave dynamics. However, initial findings on simulations at Stromboli volcano suggest that altering landslide density from 2,500 kg/m 3 to 1,700 kg/m 3 leads to a variability in tsunami wave heights that is considerably less significant when compared to the substantial uncertainty associated with parameters such as instability volume and position (Cerminara et al. 2024 , Trolese et al. 2024 ). Simulations were carried out at INGV-Pisa on the “hpc-gpu” server (equipped with 20 Intel-Xeon CPU cores and 3 NVIDIA P100 GPUs). The computational time to simulate 1,000 s of dynamics (corresponding to 16.7 min) is 13 h (for Simulation 1), 20 h (for Simulation 2), and 29 h (for Simulation 3). Results Here we present the main outcomes of the three simulations. Some of them refer to the landslide generating the tsunami whereas others refer to the propagation of the tsunami itself. In Table 3 we summarize, for each simulation, the maximum landslide volumes as defined by intersecting the bathymetry with the source ellipsoid, the actual mobilized volume and the time of movement of the submarine landslide (i.e. the time necessary for 99% of the mobilized volume to stop). The variability of mobilized volumes is associated with the shape and geometry of the source landslide volume as well as with the slope of the bathymetry in the source area and with the rheology of the landslide. Vice versa, the movement times results are quite similar across the three simulations, despite the very different volume scales and geometries. Such a scale-independent property is mostly related to the adopted rheology of the landslide and represents an interesting topic for future research. Figure 4 shows the evolution in time of the ratio between the integral of the mobilized volume at a certain time and the total mobilized volume at the end of the simulation. In the following subsections, we present, for each simulation, the main results describing the source geometry and landslide dynamics (Figs. 5, 7, and 9 for Simulations 1, 2, and 3, respectively) and the tsunami propagation and impact (Figs. 6, 8 and 10, again for Simulation 1, 2 and 3, respectively). In particular, Figs. 5, 7, and 9 describe different views of the initial geometry of the landslide volume as well as the initial and final bathymetry. Vice versa Figs. 6, 8, and 10 illustrate the tsunami propagation over the whole computational domain by showing: a) the 1-cm wave arrival times, b) the maximum free surface heights, and c) the inundation areas (only for Simulations 2 and 3) of the Santorini archipelago in terms of water depth values (meters a.g.l.). Finally, in Fig. 10 we illustrate a comparison between the maximum free surface height and the 1-cm arrival times of the three simulations along the NE to SE coast of Thera whereas in Fig. 12 and Table 4 we show the same comparison for the maximum free surface height at several control points and related profiles perpendicular to the coastlines. Simulation 1 As shown in Fig. 5a, Simulation 1 describes a submarine landslide scenario from the outer SW flank of Kolumbo volcano, which is directly facing the NE coast of Thera Island (see Fig. 1). A more direct description of the dynamics of the landslide and associated tsunamis wave is provided by a video animation attached as Supporting Information 1. From Fig. 5a, it is evident how, despite the area enclosing the maximum landslide volume is quite large, the propagation of the landslide is much more limited due to the low slope angle of the bathymetry and the consequent relatively small volume mobilized (see also Fig. 1). As a result, the bathymetry variation due to the propagation of the landslide is very limited and restricted to two separated small areas, suggesting the occurrence of two small independent landslides, one moving from NE to SW along the Kolumbo external crater walls (Landslide 1 - Fig. 5a), and another one moving from NW to SE along a relief on the topo-bathymetry (Landslide 2 - Fig. 5a). This dynamic is clearly visible in the video of Supporting Information 1. The simulation outputs of Fig. 6 reflect the small size of the submarine landslide, as the maximum free surface elevation is ~ 1.5 m near the outer rim of the Kolumbo crater (Fig. 6b) and along the NE coast of Thera (Fig. 11b). As a consequence, the resulting inundation for Thera is in practice negligible and, for this reason, has not been illustrated. Arrival times for the 1-cm waves range from ~ 180 s along the NE coast of Thera up to ~ 540 s near Perissa, along the SE coast of Thera (Fig. 6a). Simulation 2 Simulation 2 presents a potential slope failure inside the central crater of Kolumbo, as the bathymetry variation is all contained within this crater (Fig. 7 and video included as Supporting Information 2). Such a scenario involves a much larger mobilized volume (about 147 Mm 3 ) and therefore a larger impact on the Santorini archipelago compared to Simulation 1. In particular, while the 1-cm wave arrival times do not differ significantly from those computed for Simulation 1 (Fig. 8a), the maximum free surface height can now reach values up to ~ 23 m near the Kolumbo crater (Fig. 8b) and up to 17 m near Koloumbo’s settlement on Thera (see also Fig. 11b). The resulting inundation could affect several areas on Thera Island, such as the NE coast but also large parts of the E coast, especially around the settlement of Monolithos (Fig. 8c). Simulation 3 Simulation 3 (Fig. 9) presents a similar scenario to Simulation 2, although at a larger scale (mobilized volume now of about 300 Mm 3 , i.e. about double that of Simulation 2). As shown in Fig. 9a, the area enclosing the maximum landslide volume is, in this case, large enough to also include some parts of the outer Kolumbo crater walls although the collapse is all internal at the crater. The video animation of this simulation is also provided as Supporting Information 3. The impact of such a scenario on the Santorini archipelago is the largest among the three simulated. Arrival times of the 1-cm wave do not differ significantly with respect to the other simulations, especially for the NE coast of Thera (Fig. 10a), but maximum free surface heights can now reach up to > 70 meters near the Kolumbo crater (Fig. 10b) and up to about 36 m along the NE coast of Thera (near Koloumbos settlement, see also Fig. 11b), resulting in large inundation areas from NE Thera (wide areas on both sides of Koloumbos settlement), to E Thera (from Monolithos settlement to Kamari) and even to SE Thera (Perissa and southern coast, see Fig. 10c). Hazard variables along the E coast of Thera For a more quantitative analysis of the hazard associated with the simulation outcomes, we also present two additional figures illustrating some key variables along the E coast of Thera: 1) The distributions, along the NE to SE coast of Thera (roughly from Koloumbos settlement to Perissa), of the arrival times of the 1-cm wave (assumed representative of the arrival time of the tsunami) and of the maximum free surface height of the tsunami (Fig. 11); 2) a comparison between the maximum surface height of the three tsunami simulations at eight control points and along their associated profiles perpendicular to the NE-SE coastline of Thera and the NW coastline of Thirasia (Fig. 12). Regarding the first figure (Fig. 11), we have chosen to investigate the distribution of arrival time and maximum surface height along this profile as it represents the closest area to Kolumbo volcano and the portion of Thera Island (especially from Monolithos toward S) with plain shores and relatively large settlements and critical infrastructures. For the two variables (i.e. arrival time and max free surface height) and the whole E coastline, we also show the minimum of the 1-cm arrival times (triangles in Fig. 11a) and the maximum of the free surface (diamonds in Fig. 11b). These outcomes have relevant applications from a civil protection point of view to identify areas with highest expected waves and shortest time delays between the triggering of the phenomenon at source and the associated effects along the coast. The delay between the triggering landslide and the wave arrival at the E coast is between about 3 minutes, at Koloumbus settlement, to about 10 minutes, at Perissa. The coastline of Thera Island inside the Santorini caldera is mostly characterized by high cliffs and it is also significantly more protected from large tsunami waves generated by the Kolumbo volcano. However, two important infrastructures are located along this coast, that are (see Fig. 2) the touristic ports of Thera and Athinios. In addition, another touristic port is located in front of the town of Thirasia in Thirasia Island (see Fig. 2). For these three localities, our largest simulation (Simulation 3) indicates: 1-cm wave arrival times of ~ 7 min, ~ 8 min, and ~ 6 min at Thera port, Athinios port and Thirasia port, respectively; maximum free surface heights of ~ 0.75 m, ~ 1.2 m, and ~ 0.8 m at Thera port, Athinios port, and Thirasia port, respectively; The islands inside the Santorini caldera (Nea Kameni, Palea Kameni, and Aspronisi) are not populated and therefore not considered in this analysis, although the density of people can be relatively high in Nea Kameni during tourist visits. Vice versa, based on the results of Ulvrova et al. (2016), the tsunami travel times (i.e., the arrival time of the first wave generated at the Kolumbo volcano) for the surrounding islands (see Fig. 1) are of the order of ~ 6 minutes for the S coasts of the island of Ios, 8 minutes for the island of Anafi and ~ 10 minutes for the island of Amorgos. The second figure (Fig. 12) analyzes the maximum surface height computed in the three simulations in selected locations along the outside Thera coast. These control points selected are similar to those used by Ulvrova et al. (2016). In particular, we drew a profile line passing through each point and perpendicular to the coastline and we show the topo-bathymetry (from − 30 to 10 m a.s.l.) and the maximum free surface height along it for the three simulations (Fig. 12). For such control points, we also show in Table 4 the corresponding depth from our topo-bathymetry, the 1-cm wave arrival time and the maximum free heigh. To have a better appreciation of the potential hazardous actions associated with tsunami waves generated by Kolumbo volcano, some semi-quantitative comparisons between available data of the 1650 CE event and some selected results of our simulations were performed. For instance, Fig. 13 shows a comparison between the inundation computed by Simulation 3 and the field evidence of the 1650 CE tsunami. These latter data have been documented by Ulvrova et al. (2016) on several outcrops along the NE and SE Thera coasts where tsunami deposits have been recognized and linked to the 1650 CE event. The figure clearly shows that simulation results are semi-quantitatively consistent with the historical inundation of the 1650 CE event as reconstructed from field observations. Similarly, Fig. 14 provides an estimation of the potential inundation flow velocities of Simulation 3. To get the flow velocity ( u ), we used the empirical correlation given by Matsutomi and Okamoto (2010), namely $$\:u\:=\:Fr\sqrt{{gh}_{f}}$$ 1 where \(\:{h}_{f}\) is the water depth measured at the front of an impacted element (e.g. building, hill), Fr the Froude number (assumed equal to 0.66) and g the gravitational acceleration. Values obtained with this relation, based on the largest potential fluid force exerted by the flow on the impacted element, range from 0.2 to 12 m/s and therefore can produce major devastation of the coastal areas. Outcomes of the expert elicitation on tsunami hazard To have a better appreciation of the values associated with the new tsunami simulation above described, in this section, we briefly report some of the outcomes of the expert elicitation procedure carried out during the project (Bevilacqua et al. this volume) with specific reference to the block Q14 (questions 14a-c, 14d1-14d3) about the tsunamis hazard at Kolumbo. The new tsunami simulations were carried out at the same time of the expert elicitation sessions and therefore the outcome of one informed and guided the other thus favoring a more robust and consistent description of the phenomenon (Sparks et al. this volume). For more details on the elicitation, we refer to Bevilacqua et al. (this volume). The elicitation outcomes, summarized in Table 5, clearly imply that tsunamis are a significant threat for Thera and the other islands around Kolumbo. Moreover, experts’ answers indicate that there is a medium to high probability of having a major tsunami in the circumstances of an explosive eruption occurring in the next 30 years. Tsunamis are more likely to occur from the main crater of Kolumbo, particularly during the paroxysmal phase of the eruption. A complementary result is that tsunamis are less likely to be generated from the other edifices of the volcanic field although the uncertainty of these estimates is large. In terms of tsunami wave height, median probabilities (for the [5th -50th -95th ] percentile values see Table 5) of having a wave higher than 1 m along the NE coast of Thera are, for the CM (Classical Model)-EW (Equal Weight), 52–60%, whereas they decrease to about 19–27% for waves higher than 5 m, and to 7–12% for waves higher than 10 m. All these probabilities refer to potential events occurring in a future period of 30 years. The three new simulations presented above were able to generate maximum wave heights of about 1.5, 17, and 36 m (for Simulation 1, 2, and 3, respectively - see Fig. 11b) along the NE coast of Thera and values of the order of 1, 5 and 10 m along approximately the central part of the NE coast of Thera (i.e the area between Koloumbus and Monolithos). As a consequence, the three simulations can be considered approximately representative of the three different scales associated with the elicitation questions (questions 14d1-14d3) and, in turn, to the probabilities of occurrence in case of an eruption in the next 30 years (Table 5). Discussion New simulated scenarios and associated hazard The new simulated scenarios allow us to discuss the dynamics of tsunamis triggered by submarine landslides at Kolumbo volcano and their potential in terms of hazard. This mechanism of tsunami generation has received less attention in the literature with respect to other mechanisms (Karstens et al. 2023 ). The choice to investigate the effect of this generation mechanism is largely dependent on the current geological and bathymetric conditions of the Kolumbo volcanic system which lacks a subaerial part. In particular, the pumice cone formed in the 1650 CE eruption becomes unstable in the case of a future explosive eruption or a regional earthquake such as the 1956 Amorgos earthquake (see e.g. Papazachos and Kkallas this volume). Particularly, a potential internal failure of the crater rim, as well as a landslide on the external rim of the crater, could be likely generated during a future explosive eruption (Katsigera et al. 2024 ) or large magnitude regional earthquake. For the assumed rheological properties of the landslide above described, results of Simulation 1 suggest that the SW-facing slope of Kolumbo volcano has a low potential to generate submarine landslides resulting in > 1 m tsunami waves on NE coast of Thera Island. This is interpreted as due to the low slope angles along the SW-side of the Kolumbo volcano, which are generally well below (about 10°) the value of friction angles considered in our simulations (see Fig. 1 ). However, the definition of realistic values of the friction angles of such landslides is still subject of debate (see e.g. Poulain et al. 2023 ). For the Kolumbo case, this task is furthermore complicated due to the complex stratigraphy of its outer slopes, where fine beds of pumice clasts could act as weak failure surfaces (especially if shaken by a regional earthquake) resulting in an increase in pore pressure and reduced friction angles (see e.g. d’Acremont et al. 2022; Tran et al. 2024 ). Therefore, further investigation should be undertaken to evaluate the effect of lower friction angles on the remobilized volume of the landslide and, thus, on its tsunamigenic potential. Vice versa, Simulations 2 and 3, as representative of the medium and largest scenarios here considered, describe the propagation of a tsunami triggered by a landslide occurring on the inner slopes of the Kolumbo crater with volumes of about 150 and 300 Mm 3 , respectively, and have the potential to represent a major hazard for the E coast of Thera (see Figs. 8 and 10 ). We note in fact that the first wave arrival times at the control points (see Table 4) are ~ 2–3 minutes on the NE Thera (e.g. Profiles 2–3) and ~ 8–10 minutes near Perissa on the SE coast of Thera (Profile 7). Moreover, if we consider the whole E coast of Thera, we observe that both the shortest arrival times (Fig. 11 a) and the highest waves (Fig. 11 b) of our simulated scenarios are expected near the settlement of Koloumbos, but significantly short times (~ 3–4 minutes) and high waves (~ 1–35 m) are also computed along the whole E coast. Such timings and potential wave heights should be properly taken into consideration when preparing emergency/evacuation plans and designing alert systems linked to the detection of the triggering phenomenon. Finally, although the aim of our simulations was not to reproduce the tsunamis occurred during the 1650 CE eruption, we note that the internal crater failure scenarios investigated (e.g. Simulations 2 and especially Simulation 3, see Figs. 10 ), produce an inundation of the Monolithos/Perissa areas (especially with Simulation 3, see Fig. 10 ) which is semi-quantitatively consistent with the historical accounts (see Section 1) and the field evidence of the 1650 CE tsunami (see Fig. 13 ). In addition, we observe that Simulation 3 produces an inundation along the E coast of Thera Island comparable with that produced by the simulations of Karstens et al. ( 2023 ) but with about 20% of their mobilized volume, a simpler triggering mechanism and a shorter emplacement time of the mass movement (4 minutes from Karstens et al. ( 2023 ) compared to the 48–62 s of the present simulations, see Table 3 and Fig. 4 ). With these observations, we are not claiming that the 1650 CE tsunami was certainly generated by an internal collapse of the Kolumbo crater, but this triggering mechanism is certainly plausible and compatible with some of its evidence. Further ad-hoc simulations should be done to better reconstruct the 1650 CE tsunami event. Simulated scenarios from literature and hazard assessment The considerable number of simulations done in previous work (Ulvrova et al. 2016 ; Karstens et al. 2023 ) and in the current study can provide a first comprehensive evaluation of tsunami hazard at Kolumbo. To this aim, we developed a simplified matrix diagram shown in Fig. 15 that links, from one side, the tsunami triggering mechanism such as submarine explosion, pyroclastic flow, caldera collapse, and submarine landslide, and, on the other side, the probabilities of having major tsunamis with different wave heights (i.e. 1, 5 and 10 m) along the NE coast of Thera in case of an eruption in the next 30 years as estimated by expert elicitation (Bevilacqua et al. this volume). For each of the source mechanisms of tsunami mentioned in Fig. 15 , we identified scenarios for which waves on NE Thera exceed the values used in the expert elicitation (i.e. 1, 5 and 10 m). These scenarios are based on the sensitivity of the simulation outcomes to key source parameters that have the highest influence on wave amplitude: explosion energy, pyroclastic flow volume flux, caldera collapse duration, and submarine landslide volume (Paris 2015 ; Schindelé et al. 2024 ). In the matrix diagram of Fig. 15 all the simulations reported are related to a source location inside the current Kolumbo crater. Moreover, even for the new simulations done assuming such a source, there are still some scenarios missing with respect to some values of the elicited wave height along NE Thera (i.e., crater internal failure generating > 1 m wave height and landslide along the crater external slopes generating > 10 m wave height), which should be considered in future developments. Finally, the values of wave heights obtained from simulations are given along the coast. During the inland inundation phase, tsunami flows typically have Froude numbers between 0.66 and 2 (Matsutomi and Okamoto 2010 ). As an example, a tsunami with initial wave heights of 1 m, 5 m, and 10 m at the coast will propagate inland (using Eq. 1 above) at velocities higher than 2 m/s, 4.5 m/s, and 6.5 m/s, respectively. For a 1650-like event, inundation flow velocity in Perissa and Monolithos will likely exceed, respectively, 4 m/s and 7 m/s. Such estimates are instead more difficult to obtain for the Cape Koloumbos area because of the steep topography in that part of the island. Conclusions Based on the findings of the expert elicitation carried out (Bevilacqua et al. this volume; Sparks et al. this volume), a future eruption in the Kolumbo volcanic field has a good chance to generate tsunamis. In particular, an eruption from the Main Cone meets conditions conducive to tsunami formation which includes the existence of a large unstable tephra cone and a high probability that the next eruption will be predominantly explosive. Elicitation outcomes indicate a medium to high likelihood of a major tsunami associated with a 30-year future eruption. In particular, chances of having a > 1 m wave on the NE coast of Thera have medians in the range 50–60%, although there is a large uncertainty associated with these estimates. Likelihood decreases to median values of about 7–12% of having > 10 m tsunami wave heights along the NE coast of Thera, although 95%ile values still reach 30–60%, depending on the elicitation decision maker adopted. In this paper, we reviewed the results of the tsunami simulations carried out in previous works (Ulvrova et al. 2016 ; Karstens et al. 2023 ) and new simulations produced during this project. This analysis allows us to outline a first comprehensive description of the tsunami scenario-based hazard at Kolumbo. Moreover, the new simulations carried out in this study allowed us to describe the dynamic evolution of two new triggering mechanisms involving the Kolumbo crater area: a landslide involving the outer, SW-facing crater slopes, and an internal failure of the crater rim (simulated by assuming two different volumes). The simulation outputs allowed us to obtain some important insights regarding tsunami hazard triggered by submarine landslides. A scenario involving a significant (i.e. volume of the order of 150–300 Mm 3 ) internal failure of the Kolumbo crater (not necessarily related to an eruption) is capable of producing tsunami waves up to > 10 m high along the NE coast of Thera. Such a scenario is made more likely by the steep slopes of the inner crater wall in agreement with Katsigera et al. ( 2024 ). These volume values are also associated with wave heights > 5 m along the E coast of Thera, and ~ 5 m along the SE coast of Thera. Tsunami risk associated with these scenarios is relatively low along the NE coast of Thera due to the low density of settlements, whereas it is potentially much larger along the E and SE coasts, given the presence of main infrastructures (i.e. Thera airport near Monolithos on the E coast) and highly touristic areas (near Perissa and Kamari on the E and SE coasts). The expected arrival times of the first waves are also remarkably short ranging from 2–3 minutes in the most exposed coasts up to 8–10 min on the south-eastern coast of Thera. Such tsunami waves would propagate inland at velocities typically ranging from 2 to 12 m/s. We refer to Sparks et al. (this volume) for a first quantitative analysis of the individual and societal risk associated with Kolumbo tsunamis. We have also shown how simulation results suggest that, given the specific set of landslide rheological parameters assumed, it is unlikely to mobilize a landslide with a large volume from the SW-facing Kolumbo crater slopes, given the relatively gentle topo-bathymetry of this area. Nevertheless more work is needed to investigate the effect of landslide rheological properties on this outcome. On the other hand, a scenario involving an internal failure of the Kolumbo crater could produce an inundation on Thera Island semi-quantitatively consistent with that observed during the paroxysm of the 1650 CE eruption, although we acknowledge that during such a complex eruption multiple mechanisms could have generated the reconstructed tsunamis. Finally, while we underline that the present study can contribute to the improvement of tsunami hazard assessment at Kolumbo, we believe that more research is still needed to fill some important knowledge gaps. In particular, additional simulations and analyses are required to investigate tsunami events triggered from i) different areas along the outer slopes of Kolumbo crater (e.g. to the NW), ii) smaller volumes due to internal crater failure, iii) different areas from the Kolumbo volcanic field, and iv) different physical and rheological properties of the collapsing material (in particular frictional parameters). In addition, a larger computational domain including the nearby islands (Ios, Anafi and Amorgos) is required to better characterize the volcanic tsunami hazard assessment in this region. Declarations Acknowledgments Large Scale Orthophotos” (LSO) Project was implemented by the Hellenic Cadastre and co-funded by the European Union within the framework of the Operational Program "Information Society" of the 3rd Community Support Framework and NSRF 2007 - 2013. Funding This research has been supported by the project “Hazard and risk assessment for Kolumbo Volcano, Greece”, Hellenic Survey of Geology & Mineral Exploration (HSGME). We thank Hellenic Survey of Geology & Mineral Exploration (HSGME) for hosting first project workshop on Santorini, and Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Pisa, for hosting second project workshop. Conflicts of interest/Competing interests The authors declare no conflict of interest/competing interests. Availability of data and material The online version contains supplementary material available as Supporting Information 1, Supporting Information 2 and Supporting Information 3. References Alexandri M, Papanikolaou D, Nomikou P (2003) Santorini Volcanic Field - New Insights Based On Swath Bathymetry. Abstracts Iugg, 30 June-11 july 2003, Sapporo, Japan. Auker MR, Sparks RSJ, Siebert L, Crosweller HS, Ewert J (2013) A statistical analysis of the global historical volcanic fatalities record. J Appl Volcanol 2:1-24. https://doi.org/10.1186/2191-5040-2-2. Bevilacqua A, de’Michieli Vitturi M, Tadini A, Neri A, Aspinall WP, Sparks RSJ (this volume) Kolumbo Volcano, Greece: expert elicitation findings supporting volcanic hazard and risk assessment. Cantner K, Carey S, Nomikou P (2014) Integrated volcanologic and petrologic analysis of the 1650 AD eruption of Kolumbo submarine volcano, Greece. J Volcanol Geotherm Res 269:28-43. https://doi.org/10.1016/j.jvolgeores.2013.10.004. Cerminara M, Esposti Ongaro T, De’ Michieli Vitturi M, Tadini A, Trolese M, Fornaciai A, Nannipieri L, Rodriguez Galvez JF (2024). Simulated scenarios of volcanic mass movements and associated tsunamis at Stromboli (Aeolian archipelago, Tyrrhenian sea, Italy). version 1 [Data set]. Istituto Nazionale di Geofisica e Vulcanologia (INGV). https://doi.org/10.13127/stromboli/sciara_del_fuoco_tsunami d'Acremont E, Lafuerza S, Rabaute A, Lafosse M, Castelot MJ, Gorini C, Alonso B, Ercilla G, Vazquez JT, Vandorpe T, Juan C, Migeon S, Ceramicola S, Lopez-Gonzalez N, Rodriguez M, El Moumni B, Benmarha O, Ammar A (2022). Distribution and origin of submarine landslides in the active margin of the southern Alboran Sea (Western Mediterranean Sea). Mar Geol 445, 106739. https://doi.org/10.1016/j.margeo.2022.106739. Day SJ (2015) Volcanic tsunamis. In: Sigurdsson H (Ed) The encyclopedia of volcanoes, 2nd ed. Academic Press, pp 993-1009. https://doi.org/10.1016/B978-0-12-385938-9.00058-4. Dominey-Howes DTM, Papadopoulos GA, Dawson AG (2000) Geological and historical investigation of the 1650 Mt. Columbo (Thera Island) eruption and tsunami, Aegean Sea, Greece. Nat Haz 21:83–96. https://doi.org/10.1023/A:1008178100633. Escalante C, de Luna TM, Castro MJ (2018) Non-hydrostatic pressure shallow flows: GPU implementation using finite volume and finite difference scheme. Appl Math Comput 338:631–659. https://doi.org/10.1016/j.amc.2018.06.035. Escalante C, Fernández-Nieto ED, de Luna TM, Castro MJ (2019) An efficient two-layer non-hydrostatic approach for dispersive water waves. J Sci Comput 79:273–320. https://doi.org/10.1007/s10915-018-0849-9. Esposti Ongaro T, de’ Michieli Vitturi M, Cerminara M, Fornaciai A, Nannipieri L, Favalli M, Calusi B, Macìas J, Castro MJ, Ortega S, González-Vida JM, Escalante C (2021) Modeling Tsunamis Generated by Submarine Landslides at Stromboli Volcano (Aeolian Islands, Italy): A Numerical Benchmark Study. Front Earth Sci 9:628652. http://doi.org/10.3389/feart.2021.628652 . Fernández-Nieto ED, Bouchut F, Bresch D, Castro Díaz MJ, Mangeney A (2008) A new Savage–Hutter type model for submarine avalanches and generated tsunami. J Comput Phys 227:7720–7754. https://doi.org/10.1016/j.jcp.2008.04.039. Fernández-Nieto ED, Garres-Díaz J, Mangeney A, Narbona-Reina G (2018) 2D granular flows with the μ (I) rheology and side walls friction: A well-balanced multilayer discretization. J Comput Phys 356:192-219. https://doi.org/10.1016/j.jcp.2017.11.038. Fouqué F (1879) Santorin et ses éruptions. Masson. Freundt A (2017) POS 513 cruise report (p. 18). GEOMAR. https://doi.org/10.3289/cr_pos513 Fuller S, Carey S, Nomikou P (2018) Distribution of fine-grained tephra from the 1650 CE submarine eruption of Kolumbo volcano, Greece. J Volcanol Geotherm Res 352:10-25. https://doi.org/10.1016/j.jvolgeores.2018.01.004. Gupta AK, Bennartz R, Fauria KE, Mittal T (2022) Eruption chronology of the December 2021 to January 2022 Hunga Tonga-Hunga Ha’apai eruption sequence. Commun Earth Environ 3:314. https://doi.org/10.1038/s43247-022-00606-3. Karstens J, Crutchley GJ, Hansteen TH, Preine J, Carey S, Elger J, Kühn M, Nomikou P, Schmid F, Dalla Valle G, Kelfoun K, Berndt, C. (2023) Cascading events during the 1650 tsunamigenic eruption of Kolumbo volcano. Nat Commun 14(1) 6606. https://doi.org/10.1038/s41467-023-42261-y. Katsigera A, Nomikou P, Pavlopoulos K (2024) A Preliminary Hazard Assessment of Kolumbo Volcano (Santorini, Greece). GeoHazards 5(3):816-832. https://doi.org/10.3390/geohazards5030041. Konstantinou KI (2020) Magma chamber evolution during the 1650 AD Kolumbo eruption provides clues about past and future volcanic activity. Sci Rep 10(1) 15423. https://doi.org/10.1038/s41598-020-71991-y. Macías J, Escalante C, Castro MJ (2020a) Multilayer-HySEA model validation for landslide generated tsunamis. Part I rigid slides. Nat Hazards Earth Syst Sci 1–32. https://doi.org/10.5194/nhess-2020-171. Macías J, Escalante C, Castro MJ (2020b) Multilayer-HySEA model validation for landslide generated tsunamis. Part II Granular slides. Nat Hazards Earth Syst Sci 1–34. https://doi.org/10.5194/nhess-2020-172. Mastroianni F, Vougioukalakis G, Petrone C, Braschi E, Fantozzi I, Francalanci L (this volume) Multiple reservoirs and different magma interaction processes in the feeding system of the 1650 CE eruption of Kolumbo submarine volcano, Greece. Matsutomi H, Okamoto K (2010) Inundation flow velocity of tsunami on land. Isl Arc 19:443-457. https://doi.org/10.1111/j.1440-1738.2010.00725.x. Nomikou P, Carey S, Papanikolaou D, Bell KC, Sakellariou D, Alexandri M, Bejelou K (2012) Submarine volcanoes of the Kolumbo volcanic zone NE of Santorini Caldera, Greece. Global Planet Change 90 : 135-151. https://doi.org/10.1016/j.gloplacha.2012.01.001. Nomikou P, Carey S, Bell K, Papanikolaou D, Bejelou K, Cantner K, Sakellariou D, Perros I (2014) Tsunami hazard risk of a future volcanic eruption of Kolumbo submarine volcano, NE of Santorini Caldera, Greece. Nat Haz 72:1375-1390. https://doi.org/10.1007/s11069-012-0405-0. Papazachos C, Kkallas C (this volume) Deterministic seismic hazard assessment for the Kolumbo area. Paris R (2015) Source mechanisms of volcanic tsunamis. Philos Trans R Soc Lond A 373 20140380. https://doi.org/10.1098/rsta.2014.0380. Paris R, Ulvrova M, Selva J, Brizuela B, Costa A, Grezio A, Lorito S, Tonini R (2019) Probabilistic hazard analysis for tsunamis generated by subaqueous volcanic explosions in the Campi Flegrei caldera, Italy. J Volcanol Geotherm Res 379:106-116. https://doi.org/10.1016/j.jvolgeores.2019.05.010. Poulain P, Le Friant A, Mangeney A, Viroulet S, Fernandez-Nieto E, Castro Diaz M, Peruzzetto M, Grandjean G, Bouchut F, Pedreros R, Komorowski JC (2023). Performance and limits of a shallow-water model for landslide-generated tsunamis: from laboratory experiments to simulations of flank collapses at Montagne Pelée (Martinique). Geophys J Int 233(2):796-825. https://doi.org/10.1093/gji/ggac482 Pouliquen O, Forterre Y (2002). Friction law for dense granular flows: application to the motion of a mass down a rough inclined plane. J Fluid Mech 453:1–19. https://doi.org/10.1017/S0022112001006796. Sakellariou D, Sigurdsson H, Alexandri M, Carey S, Rousakis G, Nomikou P, Georgiou P, Ballas D (2010) Active tectonics in the Hellenic Volcanic Arc: the Kolumbo submarine volcanic zone. Bull Geol Soc of Greece XLIII/2:1056–1063. Schindelé F, Kong L, Lane EM, Paris R, Ripepe M, Titov V, Bailey R (2024) A Review of Tsunamis Generated by Volcanoes (TGV) Source Mechanism, Modelling, Monitoring and Warning Systems. Pure Appl Geophys 181:1745–1792. https://doi.org/10.1007/s00024-024-03515-y. Scott WE, Hoblitt RP, Torres RC, Self S, Martinez ML, Nillos TJ (1996) Pyroclastic flows of the June 15, 1991, climatic eruption of Mount Pinatubo. In: Newhall CG, Punongbayan S (eds) Fire and Mud: Eruptions of Mount Pinatubo, Philippines. Philipp. Inst Volcanol and Seismol, Quezon City, pp. 545–570. Shen Y, Whittaker CN, Lane EM, White JD, Power W (2024) Physical experiments of waves generated by submerged steam eruptions with applications to volcanic tsunamis. Sci Adv 10(45) eadk6208. https://doi.org/10.1126/sciadv.adk6208. Sigurdsson, H., Carey, S., Alexandri, M., Vougioukalakis, G., Croff, K. L., Roman, C., Sakellariou, D., Anagnstou, C., Rousakis, G., Ioakim, C., Gogou, A., Ballas, D., Misaridis, T., and Nomikou, P. (2006). Marine investigations of Greece's Santorini volcanic field. Eos 87(34):337–342. https://doi.org/10.1029/2006EO340001. Sparks RSJ, Vougioukalakis G, Aspinall WP, Neri A, Antonakos A, Baxter PJ, Bevilacqua A, Cerminara M, de’ Michieli Vitturi M, Francalanci L, Koutroulli A, Maeno F, Mastroianni F, Papazachos K, Pardini F, Paris R, Tadini A, Vaselli O (this volume). Future eruptions of the Kolumbo volcanic field: prognosis with hazard and risk assessment. Syamsidik B, Luthfi M, Suppasri A, Comfort LK (2020) The 22 December 2018 Mount Anak Krakatau volcanogenic tsunami on Sunda Strait coasts, Indonesia: tsunami and damage characteristics. Nat Hazards Earth Syst Sci 20:549–565. https://doi.org/10.5194/nhess-20-549-2020. Terry JP, Goff J, Winspear N, Bongolan VP, Fisher S (2022) Tonga volcanic eruption and tsunami, January 2022: Globally the most significant opportunity to observe an explosive and tsunamigenic submarine eruption since AD 1883 Krakatau. Geosci. Lett., 9(1):24. https://doi.org/10.1186/s40562-022-00232-z. Tran QA, Sørlie E, Grimstad G, Eiksund G, Takahashi H, and Sassa S (2024). Influence of sediment permeability in seismic-induced submarine landslide mechanism: CFD-MPM validation with centrifuge tests and analysis. Comput Geotech, 174, 106588. https://doi.org/10.1016/j.compgeo.2024.106588. Trolese M, Cerminara M, Esposti Ongaro T, de' Michieli Vitturi M, Tadini A (2024). Modeling tsunami generation and propagation: Insights from sensitivity analysis of landslide parameters at Stromboli, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19260, https://doi.org/10.5194/egusphere-egu24-19260, 2024. Ulvrova M, Paris R, Nomikou P, Kelfoun K, Leibrandt S, Tappin DR, McCoy FW (2016) Source of the tsunami generated by the 1650 AD eruption of Kolumbo submarine volcano (Aegean Sea, Greece). J Volcanol Geotherm Res 321:125-139. https://doi.org/10.1016/j.jvolgeores.2016.04.034. Vougioukalakis G, Francalanci L, Mitropoulos D, Perissoratis K (1996) The 1649–1650 eruption of the Kolumbo submarine volcanic center, Santorini. In Second Workshop on European Laboratory Volcanoes, Thira, Santorini, Greece. Vougioukalakis G, Koutroulli A, Sparks RSJ, Francalanci L, Mastroianni F, Laurenzi MA, Schaen G, Papazachos C, Aspinall WP, Kanellopoulos C, Baxter PJ, Neri A, Tadini A, Bevilacqua A (this volume) The Kolumbo volcanic field. Tables Tables 1 to 5 are available in the Supplementary Files section. Supplementary Files SupportingInformation1.mp4 SupportingInformation2.mp4 SupportingInformation3.mp4 Table1.xlsx Table 1 Simulations performed by Ulvrova et al. (2016) and Karstens et al. (2023) to reproduce the 29 September 1650 CE tsunami event. Table2.xlsx Table 2 Parameters defining the ellipsoid used to excavate the topo-bathymetry for the three simulations performed in the study. Table3.xlsx Table 3 Maximum landslide volume (as defined by intersecting the bathymetry with the source ellipsoid), mobilized volume, and landslide movement time (i.e. the time necessary for 99% of the mobilized volume to stop) for the three simulations. Table4.xlsx Table 4 Maximum free surface heights and arrival times at the 8 control points of Figure 12 for the three new simulations carried out in this study. Table5.xlsx Table 5 Summary of key elicitation outcomes for tsunami-related questions. CM (Classical Model) and EW (Equal Weight) decision-maker solutions have been rounded to the first digit with respect to the values provided in Bevilacqua et al. (this volume). Cite Share Download PDF Status: Published Journal Publication published 07 Jun, 2025 Read the published version in Bulletin of Volcanology → Version 1 posted Editorial decision: Moderate revision (possibly re-reviewed) 01 Apr, 2025 Reviewers agreed at journal 13 Feb, 2025 Reviewers invited by journal 03 Feb, 2025 Editor invited by journal 13 Jan, 2025 Editor assigned by journal 25 Dec, 2024 First submitted to journal 23 Dec, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5700315","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":410610492,"identity":"66e43739-b6e0-4953-bcf5-700f8220e945","order_by":0,"name":"Alessandro Tadini","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIiWNgGAWjYDACdgjF2ACm2CQY2NgbwWwenFqYEVqAGKSF5yBEC049aFqATIkEiBAuLfzNzIc//qi5I7u9vYH9wY8yizw+ycdtD7+2McjY49AicZgtTZrn2DPjOWcOMDb2nJMoZpNObDeWbcPjsMM8ZswMbIcTZ0gkMDbwtkkktkkDkcQZ3FrkD/N//vjjH0RL41+QFsmD+LUYHOZhkOBtg2hpBtsiwdgm+aECtxbDw2xm0rx9h41n8BxsnC1zDqiFB+gwhgoJHp4D2LXIHW9+/PHHt8OyM9ibD3x8U1aXOL/9+DPJHwY29uwNuPwPB4wIJcxA1xJUj6b7B4kaRsEoGAWjYFgDADb5VQwjE+p1AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-3603-0853","institution":"Istituto Nazionale di Geofisica e Vulcanologia Sezione di Pisa","correspondingAuthor":true,"prefix":"","firstName":"Alessandro","middleName":"","lastName":"Tadini","suffix":""},{"id":410610493,"identity":"f9961745-3ee7-487f-af56-e6475b24751c","order_by":1,"name":"Matteo Cerminara","email":"","orcid":"","institution":"Istituto Nazionale di Geofisica e Vulcanologia Sezione di Pisa","correspondingAuthor":false,"prefix":"","firstName":"Matteo","middleName":"","lastName":"Cerminara","suffix":""},{"id":410610494,"identity":"360c75c0-7a80-49a4-8859-11e0dd0e8357","order_by":2,"name":"Raphaël Paris","email":"","orcid":"","institution":"Laboratoire Magmas et Volcans","correspondingAuthor":false,"prefix":"","firstName":"Raphaël","middleName":"","lastName":"Paris","suffix":""},{"id":410610495,"identity":"a87703e7-cbfa-4ad2-8c14-0714789877b3","order_by":3,"name":"Augusto Neri","email":"","orcid":"","institution":"Istituto Nazionale di Geofisica e Vulcanologia Sezione di Pisa","correspondingAuthor":false,"prefix":"","firstName":"Augusto","middleName":"","lastName":"Neri","suffix":""},{"id":410610496,"identity":"45f4f231-3909-4ef6-97bf-020612a732e1","order_by":4,"name":"Stephen R. J. Sparks","email":"","orcid":"","institution":"University of Bristol School of Earth Sciences","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"R. J.","lastName":"Sparks","suffix":""},{"id":410610497,"identity":"c0ecffa8-fe1a-4995-8622-1b12ce50d9bc","order_by":5,"name":"Georges Vougioukalakis","email":"","orcid":"","institution":"Hellenic Survey of Geology and Mineral Exploration","correspondingAuthor":false,"prefix":"","firstName":"Georges","middleName":"","lastName":"Vougioukalakis","suffix":""},{"id":410610498,"identity":"2af26a40-cd1f-4f06-843f-8a74b3ab54d7","order_by":6,"name":"Anna Koutroulli","email":"","orcid":"","institution":"Hellenic Survey of Geology and Mineral Exploration","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Koutroulli","suffix":""},{"id":410610499,"identity":"22ba649a-aae6-4788-b2ff-c1ec9890bdfe","order_by":7,"name":"Benedetta Calusi","email":"","orcid":"","institution":"Università degli studi di Firenze, Dipartimento di matematica e informatica Ulisse Dini","correspondingAuthor":false,"prefix":"","firstName":"Benedetta","middleName":"","lastName":"Calusi","suffix":""}],"badges":[],"createdAt":"2024-12-23 14:13:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5700315/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5700315/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00445-025-01837-w","type":"published","date":"2025-06-07T15:57:29+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":75550100,"identity":"3937bd7f-ebcc-4569-b590-41a8e7be5256","added_by":"auto","created_at":"2025-02-05 18:18:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3347516,"visible":true,"origin":"","legend":"\u003cp\u003eLocation of the Santorini archipelago (Thera, Thirasia, Nea Kameni, Palea Kameni and Aspronisi) and the Kolumbo volcano. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection. The inset in the upper left angle shows the location of the Santorini archipelago and the Kolumbo volcano with respect to mainland Greece and the nearby islands (Ios, Amorgos, and Anafi).\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/fd7ee86098485d404af64679.png"},{"id":75550106,"identity":"f8ed73c8-7d49-4269-997f-5dbce14f8d92","added_by":"auto","created_at":"2025-02-05 18:18:53","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2916827,"visible":true,"origin":"","legend":"\u003cp\u003eSlope distribution of the topography and bathymetry of Santorini with indication of its main villages or settlements (yellow dots), ports, and the coastline (blue line). The Kolumbo submarine volcano is shown on the NE side of Thera, about 7 km away from its coast. Slope angle values are derived from the topo-bathymetry described in Section “Topo-bathymetry”. The computational domain used in the numerical simulations is evidenced by the red box. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/f27ce0259f9e46461977ab9f.png"},{"id":75550108,"identity":"30d996ae-992e-4909-9d0c-c846460c19f6","added_by":"auto","created_at":"2025-02-05 18:18:54","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":190195,"visible":true,"origin":"","legend":"\u003cp\u003ea) Plain and b) vertical views of the ellipsoid (dotted lines) used to excavate the bathymetry. The black arrows indicate the line of sight from the opposite panel. The black solid line in panel b) represents a schematic topo-bathymetry.\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/5d78ae1e265671d731af3c29.png"},{"id":75550123,"identity":"bd4d8e30-87cc-4b13-bf49-0325f6838c09","added_by":"auto","created_at":"2025-02-05 18:18:56","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":243417,"visible":true,"origin":"","legend":"\u003cp\u003eGraph showing the evolution in time of the ratio between the integral of mobilized volume at a given time and the total volume remobilized for the three simulations carried out. The asterisks indicate, for each simulation the time necessary for 99% of the mobilized volume to stop.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/acf0dc55839650e8a7a3b343.png"},{"id":75550132,"identity":"df2e88a8-c13d-4dc7-add8-ba31465beea9","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":972100,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 1: a) Plain view of the ellipsoid (in blue) with the bathymetry variation (at the end of the landslide propagation) in color scale; b) vertical section (red line in panel a) of the bathymetry of the volcano along the main horizontal axis of the ellipsoid excavated on the topo-bathymetry (note that the vertical axis of the upper panel of b) has a vertical exaggeration; the true scale is provided in the lower panel). The light blue area is a section of the maximum volume of the landslide, whereas the continuous and dashed lines correspond to the profile of the bathymetry at the beginning and end of the simulation, respectively. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/880308c492e46c21b7195a86.png"},{"id":75550154,"identity":"e5231460-eb0e-48af-a2d7-632895781e9f","added_by":"auto","created_at":"2025-02-05 18:18:58","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3147892,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 1: a) 1-cm wave first arrival times (isolines values are in seconds). The black asterisk is the point along the NE to SE coast of Thera with the shortest value of arrival time (the green triangle in Fig. 11a); b) maximum free surface (m a.s.l.), the area with the maximum free surface height \u0026gt; 1.5 m is located on the outer rim of the Kolumbo crater and is evidenced by the black polygon. The green asterisk is the point along the NE to SE coast of Thera with the highest value of maximum free surface height (the green diamond in Fig. 11b). Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection. The computational domain is evidenced by the red box.\u003c/p\u003e","description":"","filename":"Fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/e31d3e49c96b3ad9e949ec43.png"},{"id":75550888,"identity":"d3165f62-bef4-4f56-b9ba-f464253bb637","added_by":"auto","created_at":"2025-02-05 18:26:54","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2506430,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 2: a) Plain view of the ellipsoid (in blue) with the bathymetry variation (at the end of the simulation) in color scale; b) vertical section (red line in panels a, c and d) of the bathymetry of the volcano along the main horizontal axis of the ellipsoid excavated on the topo-bathymetry (note that the vertical axis of the upper panel of b) has a vertical exaggeration; the true scale is provided in the lower panel). The light blue area is a section of the maximum volume of the landslide, whereas the continuous and dashed lines are the profile of the bathymetry at the beginning and end of the simulation, respectively; plain view of the topo-bathymetry c) before and d) after the landslide occurrence. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig7.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/01fbad44bcaa21652af670f5.png"},{"id":75550111,"identity":"e5e38fbc-91c2-4745-a440-bb41aabcfdcd","added_by":"auto","created_at":"2025-02-05 18:18:55","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":6581280,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 2: a) 1-cm wave first arrival times (isolines values are in seconds). The black asterisk is the point along the NE to SE coast of Thera with the smallest value of arrival time (the violet triangle in Fig. 11a); b) maximum free surface (m a.g.l.), the areas with the maximum free surface height \u0026gt; 20 m are shown by the black polygons. The green asterisk is the point along the NE to SE coast of Thera with the highest value of maximum free surface height (reported as violet diamond in Fig. 11b); c) inundation area with water depth values. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection. The computational domain is evidenced by the red box.\u003c/p\u003e","description":"","filename":"Fig8.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/8fd9df8d3fc4c35147d21057.png"},{"id":75550116,"identity":"25f15887-94f2-4b7e-95a6-f8f65fc1925b","added_by":"auto","created_at":"2025-02-05 18:18:55","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":2475917,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 3: a) Plain view of the ellipsoid (in blue) with the bathymetry variation (at the end of the simulation) in color scale; b) vertical section (red line in a, c and d) of the bathymetry of the volcano along the main horizontal axis of the ellipsoid excavated on the topo-bathymetry (note that the horizontal and vertical axes of the upper panel has vertical exaggeration; the true scale is provided in the lower panel). The light blue area is a section of the maximum volume of the landslide, whereas the continuous and dashed lines are the profile of the bathymetry at the beginning and end of the simulation respectively; plain view of the topo-bathymetry c) before and d) after the occurrence of the landslide. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig9.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/7993dd38a17ab3f8fc3c38ce.png"},{"id":75550133,"identity":"e7e81986-322e-4725-aaba-b7bd6922b630","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":6689466,"visible":true,"origin":"","legend":"\u003cp\u003eSimulation 3: a) 1-cm wave first arrival times (isolines values are in seconds). The black asterisk is the point along the NE to SE coast of Thera with the smallest value of arrival time (the light blue triangle in Fig. 11a); b) maximum free surface (m a.s.l.), the area with the maximum free surface height \u0026gt; 70 m is evidenced by the black polygon. The green asterisk is the point along the NE to SE coast of Thera with the highest value of maximum free surface height (the light blue diamond in Fig. 11b); c) inundation area with water depth values. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection. The computational domain is evidenced by the red box.\u003c/p\u003e","description":"","filename":"Fig10.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/c96415824b8bf1d9d9643c44.png"},{"id":75550109,"identity":"e173f4b9-30cd-4aa4-9c58-ca5b2f433ac8","added_by":"auto","created_at":"2025-02-05 18:18:54","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":1720479,"visible":true,"origin":"","legend":"\u003cp\u003ea) Arrival times (s) of the 1-cm wave and b) maximum free surface height (m a.s.l.) along the whole E coast of Thera (highlighted in blue in the small insets). Main towns/settlements along the coast are also indicated with their total inhabitants (2011 census) and some main infrastructures or attractions. The two plots also show the values of the minimum (for arrival times) and maximum (for free surface height) for the three simulations along the coast (triangles and diamonds, respectively). These latter are also reported, as respectively black and green asterisks, in Figs. 6a-b, 8a-b and 10a-b.\u003c/p\u003e","description":"","filename":"Fig11.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/7b4d09a49d19d0feb385f9da.png"},{"id":75550105,"identity":"0a9d0bab-b0b5-410e-9343-895755f293c4","added_by":"auto","created_at":"2025-02-05 18:18:53","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":3209578,"visible":true,"origin":"","legend":"\u003cp\u003eMap of the location of the control points (pale blue circles) and associated free surface elevation profiles (purple segments) perpendicular to the Thera coastline showing the maximum free surface heights as computed for Simulations 1, 2, and 3. Vertical red dashed lines on profile graphs indicate the location of the control points along each profile. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig12.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/8fd5b869d6aba59eeb641b1f.png"},{"id":75550173,"identity":"bdf14829-7032-44b8-9259-c4faffb9d758","added_by":"auto","created_at":"2025-02-05 18:19:18","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":7605696,"visible":true,"origin":"","legend":"\u003cp\u003eComparison between the inundation area (color legend refers to water depth in m a.g.l.) of Simulation 3 and locations with evidence of 1650 CE tsunami deposits (white squares, from Ulvrova et al. 2016). A, B, and C are enlargements of the main panel. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig13.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/19bd0723ce674518a4d1a749.png"},{"id":75550139,"identity":"f306cabf-4df5-42ca-ae2c-c346d24f1f8b","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"png","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":7498691,"visible":true,"origin":"","legend":"\u003cp\u003eComparison between the calculated flow speed (from Equation 1) of Simulation 3 and locations with evidence of 1650 CE tsunami deposits (white squares, from Ulvrova et al. 2016). A, B, and C are enlargements of the main panel. Coordinates expressed in the GGRS ’87/Greek Grid – EPSG:2100 UTM projection.\u003c/p\u003e","description":"","filename":"Fig14.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/163069f6d51676da99f346ab.png"},{"id":75550122,"identity":"0f6a162a-0417-4158-9633-4eb6572b2f42","added_by":"auto","created_at":"2025-02-05 18:18:56","extension":"png","order_by":15,"title":"Figure 15","display":"","copyAsset":false,"role":"figure","size":1016579,"visible":true,"origin":"","legend":"\u003cp\u003eMatrix diagram for volcanic tsunami hazard at Kolumbo volcano. Values in square brackets correspond to the [5\u003csup\u003eth\u003c/sup\u003e-\u003cem\u003e\u003cstrong\u003e50\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cstrong\u003eth\u003c/strong\u003e\u003c/em\u003e\u003c/sup\u003e-95\u003csup\u003eth\u003c/sup\u003e] percentiles of the CM decision-maker solutions (rounded to the first decimal digit) of the corresponding questions of the expert elicitation (see Table 5 and Bevilacqua et al. this volume). Colored areas indicate different scenarios produced by different tsunami-triggering mechanisms and report the value of the most important parameters required to generate waves of a given height on the NE coast of Thera, based on numerical simulations. We remark that each colored area refers just to one specific triggering mechanism which is here considered independent from all the others. “n.a.” indicates that a simulation describing such a specific scenario is currently missing.\u003c/p\u003e","description":"","filename":"Fig15.png","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/770772b8e6332f1d231a5f07.png"},{"id":84242626,"identity":"3c82de97-8991-42b7-b55e-761c305c16da","added_by":"auto","created_at":"2025-06-09 16:10:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":56170067,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/75be6c71-cf69-4394-b960-54cdefca9ea3.pdf"},{"id":75550101,"identity":"788c7a9e-7238-44f3-ac3d-e2909a3197dd","added_by":"auto","created_at":"2025-02-05 18:18:53","extension":"mp4","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":977783,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingInformation1.mp4","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/6693d7ac1f12166d0877e015.mp4"},{"id":75550172,"identity":"054a9216-8908-4252-b26a-b146912064a9","added_by":"auto","created_at":"2025-02-05 18:19:17","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1562038,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingInformation2.mp4","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/4e17b729a408adf8fcff91ca.mp4"},{"id":75550137,"identity":"3cb12e92-6f9f-417b-9d64-c8658e2b5ecc","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"mp4","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1614414,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingInformation3.mp4","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/d18d0bdba3fe0b68c6abcebc.mp4"},{"id":75550893,"identity":"472a2bbf-004e-4be2-918c-83e56636f507","added_by":"auto","created_at":"2025-02-05 18:26:57","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":12838,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e Simulations performed by Ulvrova et al. (2016) and Karstens et al. (2023) to reproduce the 29 September 1650 CE tsunami event.\u003c/p\u003e","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/0dc5b1acaa95ab872314f675.xlsx"},{"id":75550125,"identity":"f59f6962-bb39-420e-87f5-66bca2455fb6","added_by":"auto","created_at":"2025-02-05 18:18:56","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":9185,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 2\u003c/strong\u003e Parameters defining the ellipsoid used to excavate the topo-bathymetry for the three simulations performed in the study.\u003c/p\u003e","description":"","filename":"Table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/85f5dd4f8892561e4214a710.xlsx"},{"id":75550160,"identity":"db0fa69b-5b14-45e9-8d3e-a27e9f4ab718","added_by":"auto","created_at":"2025-02-05 18:18:58","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":9177,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e Maximum landslide volume (as defined by intersecting the bathymetry with the source ellipsoid), mobilized volume, and landslide movement time (i.e. the time necessary for 99% of the mobilized volume to stop) for the three simulations.\u003c/p\u003e","description":"","filename":"Table3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/abf8d772ea135610c0b1417f.xlsx"},{"id":75550141,"identity":"38782814-8a31-43d2-b06d-06669598e31e","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":9883,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e Maximum free surface heights and arrival times at the 8 control points of Figure 12 for the three new simulations carried out in this study.\u003c/p\u003e","description":"","filename":"Table4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/589bf3dc1c7aab63889f5572.xlsx"},{"id":75550140,"identity":"e1d26cc6-ca02-4406-87a9-7a7e7c93718b","added_by":"auto","created_at":"2025-02-05 18:18:57","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":10540,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 5\u003c/strong\u003e Summary of key elicitation outcomes for tsunami-related questions. CM (Classical Model) and EW (Equal Weight) decision-maker solutions have been rounded to the first digit with respect to the values provided in Bevilacqua et al. (this volume).\u003c/p\u003e","description":"","filename":"Table5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5700315/v1/6738a7e33abad7179f778d13.xlsx"}],"financialInterests":"","formattedTitle":"Scenario-based tsunami hazard assessment at Kolumbo volcano","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDespite not being a frequent event, volcanogenic tsunamis represent one of the deadliest phenomena potentially associated with volcanic eruptions (Auker et al. 2013; Paris et al. 2019; Schindelé et al. 2024), during which eruptive, volcano-tectonic, and gravitational phenomena can lead to the generation of tsunamis (Paris 2015). Tsunami hazard is indeed potentially high for submarine volcanoes in shallow-water conditions (Day 2015; Syamsidik et al. 2020; Terry et al. 2022), where tsunamis can be triggered by explosions, the gravitational collapse of the subaerial eruptive column, pyroclastic flows penetrating the sea, caldera collapse, or submarine flank instability (landslides). Moreover, violent explosive eruptions in shallow-water may also trigger meteo-tsunamis such as that occurred during the 2022 Hunga Tonga Hunga Ha'apai eruption (Shen et al. 2024).\u003c/p\u003e\n\u003cp\u003eA shallow-water condition is met for Kolumbo, an active submarine volcano located ~7 km NE of Santorini (Thera in Greek language) island within the Aegean Sea (Fig. 1) which last erupted in 1650 CE (Vougioukalakis et al. 1996; Cantner et al. 2014; Vougioukalakis et al. this volume). Particularly, this explosive eruption has been characterized by several hazardous phenomena, the main ones being tephra fallout (Fuller et al. 2018), gas poisoning (Konstantinou 2020), and tsunami (Nomikou et al. 2014; Ulvrova et al. 2016; Katsigera et al. 2024). Building on the heavy consequences that this eruption had on population and on the proximity of Kolumbo volcano to the highly touristic Thera island, a comprehensive hazard assessment project has been carried out for Kolumbo volcano under the auspices of the Greek civil protection authorities (Sparks et al. this volume). A core aspect of this project has been the setup of an expert elicitation exercise (Bevilacqua et al. this volume) to quantify the major uncertainties associated with specific aspects of the reconstruction of the 1650 CE eruption but also with potential hazards, including tsunami, associated to future explosive eruptions of Kolumbo volcano.\u003c/p\u003e\n\u003cp\u003eThe aim of this paper is to present a scenario-based tsunami hazard assessment for Kolumbo volcano based on a review of existing modelling studies and new numerical simulations that were carried out to inform the expert elicitation target questions on tsunami hazard. The previous literature using numerical simulations includes a study by Ulrova et al. (2016), investigating a variety of possible mechanisms of the tsunamis in 1650 CE (including pyroclastic flows, underwater explosions, and caldera collapse) and a study by Karstens et al. (2023) on combining landslide and explosion mechanisms to reconstruct this event. These studies together with the outcomes of our new simulations can be used to investigate the tsunami hazard in future eruptions.\u003c/p\u003e\n\u003cp\u003eIn particular, our new simulations aimed to investigate the dynamics of tsunamis generated by the, so far, less studied mechanism, that is the formation of a submarine landslide on the volcano flanks or within the crater. The new simulations can describe the coupled dynamics of landslide and tsunamis thus illustrating the effect of the source process on the tsunami propagation. Simulation results also provide new insights on the potential of submarine landslides, in particular of those occurring inside the crater, to generate tsunamis on the coast of Thera, as well as on the interpretation of the genesis of the 1650 CE tsunamis. Modelling results are also compared to and discussed in relation to the tsunami-related outcomes of the expert elicitation procedure developed in the project (see Bevilacqua et al. this volume, target questions of block Q14).\u003c/p\u003e\n\u003cp\u003eThis paper is organized as follows: we begin with a review of i) the Kolumbo volcano and the 1650 CE eruption with particular focus on tsunami observations (Section “The Kolumbo volcano and the 1650 CE eruption”) and ii) the previous tsunami simulations (Section “Previous tsunami simulations”). Then, we illustrate our simulation strategy (Section “Materials and Methods”), present the results of the new simulations of submarine landslides and associated tsunamis also in relation to the outcomes of the expert elicitation questions (Section \u0026nbsp;“Results”), and discuss the implications of the simulations results (both the new ones and those from literature) for the tsunami hazard assessment (Section “Discussion”). Finally, we summarize the main findings and some future perspectives in the “Conclusions” section.\u003c/p\u003e\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n"},{"header":"Background","content":"\u003cp\u003e\u003cstrong\u003eThe Kolumbo volcano and its 1650 CE eruption\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe Kolumbo volcano is part of the Kolumbo volcanic field, which develops ~ 10 km NE of Thera island (Fig. 1, Vougioukalakis et al. this volume). The polygenetic and active Kolumbo central volcano is the southernmost edifice of this field, and its most recent eruption occurred from September to December 1650 CE with a major explosive eruption on 29 September. This powerful eruption left a 1700 m large, 500 m deep crater at the center of the volcano (Fig. 2; Nomikou et al. 2012).\u003c/p\u003e\u003cp\u003eA detailed description of the available evidence and chronicles of the 1650 CE eruption of Kolumbo is reported in Vougioukalakis et al. (this volume) and Mastroianni et al. (this volume). In the following, we will briefly recall the main facts, with specific reference to the tsunamis.\u003c/p\u003e\u003cp\u003eMost of the observations of the 1650 CE eruption and its related phenomena were compiled by Fouqué (1879) from earlier reports and descriptions of the phenomena. Between January 1649 CE and March 1650 CE, precursory earthquakes (with an intensity of VII on Mercalli scale) were felt on Thera Island. In September (14-26), subterranean roaring and green sea water attested the beginning of submarine volcanic activity, building up a submarine pumice tuff cone. Then, the intensity of the earthquakes increased, with several events reaching an intensity of VIII (Mercalli). Earthquakes persisted throughout the eruption, but numerical simulations suggest that they could not be the source of the largest tsunami that was observed during the paroxysmal stage of the eruption (Ulvrova et al. 2016). The islet emerged from seawater on 26 September, although its nature is not yet identified (Vougioukalakis et al. this volume). During the following days, phreatomagmatic explosions frequently broke the sea surface with intermittent jets of gas and ash. The sea was progressively covered with pumice. The following shallow-water phreatomagmatic phase (27-29 September) did not generate major tsunamis, but on 27 September a wave pushed a boat into the sea and then brought it back to the shore.\u003c/p\u003e\u003cp\u003eThe paroxysmal phase started in the early morning of 29 September and lasted up to a maximum of 2 days. The major explosions were heard as far away as 400 km in the Dardanelles. Fine ash was deposited up to western Turkey (150-200 km). Gas affected Thera’s inhabitants (20-50 deaths) and animals. Earthquakes accompanying the paroxysmal phase were felt up to Crete (120-140 km away), and one of the earthquakes was strong and long enough to cause damages to buildings in Thera. The largest tsunami occurred at some time on 29 September during this paroxysmal phase. Available testimonies did not report the timing and number of tsunamis thus it is difficult to associate the tsunami with a particular source mechanism. Waves were observed on the coast of Thera, where about 2 km² of land was eroded (revealing Hellenistic and Byzantine ruins at Kamari and Perissa), many trees uprooted, and five churches were destroyed (Dominey-Howes et al. 2000). The spatial distribution of the tsunami deposits allows estimating a minimum wave runup in the order of 15-20 m a.s.l. on the eastern coast of Thera (\u0026gt;630 m inundation distance near Monolithos settlement), and \u0026gt;3.5 m a.s.l. on the southern coast (\u0026gt;360 m inundation near Perissa) (Ulvrova et al. 2016). The coast inside the Thera caldera was not affected by the tsunami, consistent with numerical simulations (Ulvrova et al. 2016). The timing of the tsunami in the framework of the eruption chronology remains unclear, even if historical sources place it around nighttime.\u003c/p\u003e\u003cp\u003eThe tsunami also impacted nearby islands, especially Ios and Sikinos, located more than 20 km northwest of the volcano. On Ios, a local wave runup of 14-20 m a.s.l. associated with pumice deposition was observed on a rocky shore, but the precise location remains unknown. On Sikinos, the tsunami penetrated 240 m inland (Fouqué 1879: “350 pas”, erroneously translated as feet in many publications). Sea agitation and damage to ships are mentioned in Kea Island (150 km northwest of Kolumbo) and Crete (towns of Dia and Chania, 120 and 165 km to the south, respectively).\u003c/p\u003e\u003cp\u003eBetween 1 October and 20 December, the activity was characterized by periodic explosions (waning phase), gas release (e.g., 20 deaths on 4 November), and small tsunamis whose origin is unknown (possibly following a large explosion on 4 November, and renewed activity on 6 December). On 2 October, nine men approaching the eruption site by boat were found swollen and burnt.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003ePrevious tsunamis simulations\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eAll numerical simulations published so far aimed at understanding the possible source mechanism of the main tsunami observed at some time on 29 September during the paroxysmal phase of the 1650 CE eruption (Ulvrova et al. 2016; Karstens et al. 2023). The full list of parameters of the simulations presented in these two papers is provided in Table 1. Ulvrova et al. (2016) tested three different source mechanisms of tsunami:\u003c/p\u003e\u003cp\u003e(a) Submarine explosions with energies ranging from 3x10\u003csup\u003e14\u003c/sup\u003e to 5.4x10\u003csup\u003e16\u003c/sup\u003e joules (plausible range of energies for shallow-water volcanic explosions). The depth of the explosion was set at 150 m, which corresponds to the approximate present-day rim of the crater.\u003c/p\u003e\u003cp\u003e(b) Caldera collapses of different geometries (full collapse of the central crater, collapse of the upper half only, or deepening/collapse of the lower half), and different durations (from 1 minute to 1 hour). Recent examples of caldera collapse (e.g. Pinatubo 1991, Hunga Tonga 2022) suggest that this source mechanism takes at least 30 minutes (Schott et al. 1996; Gupta et al. 2022).\u003c/p\u003e\u003cp\u003e(c) Pyroclastic density currents resulting from the gravitational collapse of the subaerial eruptive column, with different flow densities (from 1100 to 1500 kg/m\u003csup\u003e3\u003c/sup\u003e), different flow velocities (from 5 to 30 m/s), different volumes (from 5 x 10\u003csup\u003e6\u003c/sup\u003e to 100 x 10\u003csup\u003e6\u003c/sup\u003e m³), and different volume flux (from 10\u003csup\u003e4\u003c/sup\u003e to 10\u003csup\u003e7\u003c/sup\u003e m\u003csup\u003e3\u003c/sup\u003e/s).\u003c/p\u003e\u003cp\u003eKarstens et al. (2023) proposed a hybrid scenario, with a submarine landslide of the north-western flank of the volcano (thus accounting for a possible initial retreat of the sea on the eastern coast of Thera), followed by a powerful submarine explosion. They tested different flow densities (from 1250 to 1750 kg/m\u003csup\u003e3\u003c/sup\u003e) and yield strength (from 5 to 10 kPa) for the landslide, but only one volume (1.2 km³). Explosion energies range from 3x10\u003csup\u003e14\u003c/sup\u003e to 2.2x10\u003csup\u003e16\u003c/sup\u003e Joules, following the methodology proposed by Ulvrova et al. (2016). They proposed that the slow movement of the landslide on the north-western flank of the volcano (of about 500-1000 m occurring in about 4 minutes) caused a major depressurization of the magmatic system and the associated major explosion.\u003c/p\u003e\u003cp\u003eUlvrova et al. (2016) also traced the limits of sedimentary deposits from the 1650 CE tsunami, giving an idea of the extent of the inundation. The waves thus reached minimum altitudes ranging between 3.5 m a.s.l. (Perissa, southern coast) and 20 m a.s.l. (Monolithos, eastern coast), corresponding to a minimum inundation of 360 and 630 m, respectively.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003eThe new simulations of submarine landslides and associated tsunamis have been carried out by using the Multilayer-HySEA model (Fern\u0026aacute;ndez-Nieto et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Mac\u0026iacute;as et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003eb\u003c/span\u003e; Esposti Ongaro et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). They investigated in detail three scenarios: the case of a \u0026ldquo;small-scale\u0026rdquo; landslide occurring along the outer SW flank of the Kolumbo volcano described by Simulation 1 with volume of about 1.8 Mm\u003csup\u003e3\u003c/sup\u003e, and two \u0026ldquo;large-scale\u0026rdquo; slope failures inside its central crater described by Simulations 2 and 3 with volumes of about 147 and 300 Mm\u003csup\u003e3\u003c/sup\u003e, respectively.\u003c/p\u003e\n\u003ch3\u003eNumerical model\u003c/h3\u003e\n\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe Multilayer-HySEA code is a multilayer, non-hydrostatic model in which the three-dimensional model equations are depth-averaged across a number of vertical layers. The governing equations correspond to a semi-discretization for the vertical variables of the Euler equations. The total pressure is decomposed into a sum of hydrostatic and non-hydrostatic pressures. In this process, the horizontal and vertical velocities are assumed to have a constant vertical profile in each layer. The proposed model admits an exact energy balance and, when the number of layers increases, the linear dispersion relation of the linear model converges to the same of Airy's theory (Fern\u0026aacute;ndez-Nieto et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Moreover, the Multilayer-HySEA model can simulate the two-way interaction of the tsunamis with a landslide: in this case the motion of the bottom surface is represented by an additional layer described by the shallow-water equations of granular material. In this way, the Multilayer-HySEA code incorporates the possibility of simulating the generation of tsunami produced by subaerial or submarine granular landslides. The motion of the landslide is described by a granular landslide model (Fern\u0026aacute;ndez-Nieto et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2008\u003c/span\u003e), in which the Pouliquen and Forterre (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) friction law is implemented. This law characterizes the dependence of static and dynamic friction coefficients on landslide velocity and thickness through three parameters (three friction angles) \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e1\u003c/sub\u003e, \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e2,\u003c/sub\u003e \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e3\u003c/sub\u003e (Mac\u0026iacute;as et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2020b\u003c/span\u003e; Esposti Ongaro et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The granular landslide model is in turn weakly coupled with the non-hydrostatic multilayer model through the boundary conditions. For more details on the model description and validation tests we refer the reader to Mac\u0026iacute;as et al. (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003eb\u003c/span\u003e). The Multilayer-HySEA numerical code is designed to run on Graphic Processing Unit (GPU) accelerated High-Performance Computing (HPC) architectures (Escalante et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The model in this configuration has been already applied to tsunamis generated in volcanic islands such as Stromboli (Italy) as described in Esposti Ongaro et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eTopo-bathymetry\u003c/h3\u003e\n\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eFor the topo-bathymetry of the study area, we considered a 23 x 24 km domain including the Kolumbo crater and the whole Thera Island (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The employed topo-bathymetry is the result of a merging between two different datasets: 1) the bathymetry of the surroundings of Kolumbo volcano and Thera Island and 2) the topography of Thera Island. The bathymetry was collected during several multibeam echosounder surveys by R/V Aegeo using Seabeam 2120 (20 kHz) echosounder during 2001, 2006, and 2017 cruises (Alexandri et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Sigurdsson et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Sakellariou et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Nomikou et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Freundt et al. 2017). The pixel size of the original file is 0.03125 arc minutes, (X Y coordinate system WGS 84/ EPSG:4326), converted to GGRS87\u0026rsquo; (pixel size 50 m), to create the bathymetry file. The original file was provided by the Hellenic Center for Marine Research, Hellenic National Oceanographic Data Center. The topography instead derives from a dataset of 5-m pixel size orthophoto maps, created from color countrywide air photos acquired during the \u0026ldquo;Large Scale Orthophotos\u0026rdquo; (LSO) Project (2007\u0026ndash;2009).\u003c/p\u003e \u003cp\u003eThe merging of the two source datasets was performed in two steps: i) by homogenizing the cell sizes with a resolution which would allow affordable computational times and ii) by performing a smoothing of the junction zones between the two original data to avoid sharp changes in the topo-bathymetry that could cause instabilities in the numerical code. These steps were performed within the ArcGIS10\u0026copy; platform and allowed to obtain a final topo-bathymetry with a 10-m cell size.\u003c/p\u003e \u003cp\u003eWe note that the employed topo-bathymetry is obviously different from that in 1650 CE prior to the tsunami-generating event. The simulations described in the following sections are not specifically aimed at reproducing the 1650 CE tsunami, although they can provide useful insights for the interpretation of the tsunami source mechanism of this event.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eSimulation source conditions\u003c/h3\u003e\n\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eTo deal with a realistic source condition and to get an accurate dispersion modeling of the waves, the Multilayer-HySEA model was used by adopting a 3 vertical layers configuration: the lower granular layer, representing the tsunamis-generating landslide, is two-way coupled with water, represented by the remaining two layers.\u003c/p\u003e \u003cp\u003eA crucial aspect of the simulations conducted involves delineating the volume involved in triggering the tsunamis through the generation of the landslide. This delineation was achieved by intersecting the current bathymetric data of the volcano with an ellipsoidal-shaped volume centered at a specified spatial coordinate and defined by its three semi-axes. This process entails carving out a region from the bathymetric surface representing the potentially unstable granular layer that could lead to landslide formation. Following the delineation of the ellipsoid-shaped region, the simulation further determines the actual mobilized volume primarily on the basis of the friction angle used to model the rheological properties of the granular material constituting the landslide. Importantly, both the amount of mobilized material and its dynamics are not prescribed a priori but are evaluated based on the shallow water equations governing the granular layer and its interaction with the water layers. The proposed methodology allows the investigation of dynamics induced by instabilities within the actual bathymetry, facilitating a deeper understanding of potential tsunami generation mechanisms. While the present set of simulations does not explicitly explore the addition and flow of juvenile material resulting from a volcanic eruption onto the bathymetry, we believe that some useful insights on the potential impact of such a situation can be inferred from the outcomes of these new simulations.\u003c/p\u003e \u003cp\u003eThe ellipsoid is horizontally oriented with semi-axis \u003cem\u003ea\u003c/em\u003e, \u003cem\u003eb\u003c/em\u003e, and \u003cem\u003ec\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Its position, angle of rotation and size are defined by (see Table\u0026nbsp;2): i) the (x\u003csub\u003ec\u003c/sub\u003e,y\u003csub\u003ec\u003c/sub\u003e) horizontal coordinates of the center C; ii) the (x\u003csub\u003ea\u003c/sub\u003e,y\u003csub\u003ea\u003c/sub\u003e) coordinates of the point A at the end of horizontal semi-axis a; iii) the length of horizontal semi-axis b. In addition, the vertical coordinate of points C and A (z\u003csub\u003ea\u003c/sub\u003e) is defined by the elevation of point A on topo-bathymetry plus 20 m of vertical offset to avoid a vertical cut in correspondence of point A. Finally, the vertical semi-axis c is prescribed by imposing that the lowest point L (which has the same horizontal coordinates of C) is on the topo-bathymetry: L(x\u003csub\u003ec\u003c/sub\u003e,y\u003csub\u003ec\u003c/sub\u003e,z\u003csub\u003el\u003c/sub\u003e). Points A and L should be selected so that the former is shallower than the latter. A sketch summarizing the above-mentioned geometrical elements is provided in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe resulting ellipsoid uniquely intersects the topo-bathymetry, and the intersection space represents the involved volume that could potentially slide, which however does not necessarily coincide with the landslide real volume mobilized. Such a mobilized volume of the landslide is dependent on the chosen rheology, depending in turn on the landslide density and the assumed friction angle (see next section for more details on the volumes involved).\u003c/p\u003e \u003cp\u003eBy varying the above-mentioned parameters (Table\u0026nbsp;2), we carried out three simulations: Simulations 1, 2, and 3. As already mentioned and as we will see in details in the next subsection, Simulation 1 corresponds to a \u0026ldquo;small-scale\u0026rdquo; (about 1.8 Mm\u003csup\u003e3\u003c/sup\u003e in volume) landslide occurring along the outer SW flank of Kolumbo, whereas Simulations 2 and 3 to \u0026ldquo;large-scale\u0026rdquo; (of 147 and 300 Mm\u003csup\u003e3\u003c/sup\u003e in volume, respectively) landslides generated by slope failures of the internal walls of the Kolumbo crater.\u003c/p\u003e \u003cp\u003eWe remark that these three simulations were chosen in order to investigate two potentially dangerous generation mechanism of tsunamis at Kolumbo, such as: 1) a small-scale submarine landslide on the SW external slopes directly facing the island of Thera (Simualtion 1), and 2) a large scale submarine landslide produced by the internal failure of the crater slopes which could produce a remarkable impact on Thera and surrounding islands (Simulations 2 and 3).\u003c/p\u003e \u003cp\u003eIn addition to these values, for all the three simulations, a landslide density of 2,000 kg/m\u003csup\u003e3\u003c/sup\u003e and friction angles \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e1\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;10\u0026deg;, \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e2\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;15\u0026deg; and \u003cem\u003eδ\u003c/em\u003e\u003csub\u003e3\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;10\u0026deg; were considered. The chosen friction angles are assumed equal to those used in similar studies carried out a Stromboli volcano (Italy) (Esposti Ongaro et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The density of the porous material underwater should be considered larger than that measured in air because the pores of the lapilli contained by the landslide are partially filled by water. The present study does not investigate the effects of density and friction angle variations on tsunami wave dynamics. However, initial findings on simulations at Stromboli volcano suggest that altering landslide density from 2,500 kg/m\u003csup\u003e3\u003c/sup\u003e to 1,700 kg/m\u003csup\u003e3\u003c/sup\u003e leads to a variability in tsunami wave heights that is considerably less significant when compared to the substantial uncertainty associated with parameters such as instability volume and position (Cerminara et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2024\u003c/span\u003e, Trolese et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSimulations were carried out at INGV-Pisa on the \u0026ldquo;hpc-gpu\u0026rdquo; server (equipped with 20 Intel-Xeon CPU cores and 3 NVIDIA P100 GPUs). The computational time to simulate 1,000 s of dynamics (corresponding to 16.7 min) is 13 h (for Simulation 1), 20 h (for Simulation 2), and 29 h (for Simulation 3).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eHere we present the main outcomes of the three simulations. Some of them refer to the landslide generating the tsunami whereas others refer to the propagation of the tsunami itself.\u003c/p\u003e\n\u003cp\u003eIn Table\u0026nbsp;3 we summarize, for each simulation, the maximum landslide volumes as defined by intersecting the bathymetry with the source ellipsoid, the actual mobilized volume and the time of movement of the submarine landslide (i.e. the time necessary for 99% of the mobilized volume to stop). The variability of mobilized volumes is associated with the shape and geometry of the source landslide volume as well as with the slope of the bathymetry in the source area and with the rheology of the landslide. Vice versa, the movement times results are quite similar across the three simulations, despite the very different volume scales and geometries. Such a scale-independent property is mostly related to the adopted rheology of the landslide and represents an interesting topic for future research. Figure\u0026nbsp;4 shows the evolution in time of the ratio between the integral of the mobilized volume at a certain time and the total mobilized volume at the end of the simulation.\u003c/p\u003e\n\u003cp\u003eIn the following subsections, we present, for each simulation, the main results describing the source geometry and landslide dynamics (Figs.\u0026nbsp;5, 7, and 9 for Simulations 1, 2, and 3, respectively) and the tsunami propagation and impact (Figs.\u0026nbsp;6, 8 and 10, again for Simulation 1, 2 and 3, respectively).\u003c/p\u003e\n\u003cp\u003eIn particular, Figs.\u0026nbsp;5, 7, and 9 describe different views of the initial geometry of the landslide volume as well as the initial and final bathymetry. Vice versa Figs.\u0026nbsp;6, 8, and 10 illustrate the tsunami propagation over the whole computational domain by showing: a) the 1-cm wave arrival times, b) the maximum free surface heights, and c) the inundation areas (only for Simulations 2 and 3) of the Santorini archipelago in terms of water depth values (meters a.g.l.). Finally, in Fig.\u0026nbsp;10 we illustrate a comparison between the maximum free surface height and the 1-cm arrival times of the three simulations along the NE to SE coast of Thera whereas in Fig.\u0026nbsp;12 and Table\u0026nbsp;4 we show the same comparison for the maximum free surface height at several control points and related profiles perpendicular to the coastlines.\u003c/p\u003e\n\u003ch3\u003eSimulation 1\u003c/h3\u003e\n\u003cdiv\u003e\n \u003cp\u003eAs shown in Fig.\u0026nbsp;5a, Simulation 1 describes a submarine landslide scenario from the outer SW flank of Kolumbo volcano, which is directly facing the NE coast of Thera Island (see Fig.\u0026nbsp;1). A more direct description of the dynamics of the landslide and associated tsunamis wave is provided by a video animation attached as Supporting Information 1. From Fig.\u0026nbsp;5a, it is evident how, despite the area enclosing the maximum landslide volume is quite large, the propagation of the landslide is much more limited due to the low slope angle of the bathymetry and the consequent relatively small volume mobilized (see also Fig.\u0026nbsp;1). As a result, the bathymetry variation due to the propagation of the landslide is very limited and restricted to two separated small areas, suggesting the occurrence of two small independent landslides, one moving from NE to SW along the Kolumbo external crater walls (Landslide 1 - Fig.\u0026nbsp;5a), and another one moving from NW to SE along a relief on the topo-bathymetry (Landslide 2 - Fig.\u0026nbsp;5a). This dynamic is clearly visible in the video of Supporting Information 1.\u003c/p\u003e\n \u003cp\u003eThe simulation outputs of Fig.\u0026nbsp;6 reflect the small size of the submarine landslide, as the maximum free surface elevation is ~ 1.5 m near the outer rim of the Kolumbo crater (Fig.\u0026nbsp;6b) and along the NE coast of Thera (Fig.\u0026nbsp;11b). As a consequence, the resulting inundation for Thera is in practice negligible and, for this reason, has not been illustrated. Arrival times for the 1-cm waves range from ~ 180 s along the NE coast of Thera up to ~ 540 s near Perissa, along the SE coast of Thera (Fig.\u0026nbsp;6a).\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eSimulation 2\u003c/h3\u003e\n\u003cp\u003eSimulation 2 presents a potential slope failure inside the central crater of Kolumbo, as the bathymetry variation is all contained within this crater (Fig.\u0026nbsp;7 and video included as Supporting Information 2).\u003c/p\u003e\n\u003cp\u003eSuch a scenario involves a much larger mobilized volume (about 147 Mm\u003csup\u003e3\u003c/sup\u003e) and therefore a larger impact on the Santorini archipelago compared to Simulation 1. In particular, while the 1-cm wave arrival times do not differ significantly from those computed for Simulation 1 (Fig.\u0026nbsp;8a), the maximum free surface height can now reach values up to ~ 23 m near the Kolumbo crater (Fig.\u0026nbsp;8b) and up to 17 m near Koloumbo’s settlement on Thera (see also Fig.\u0026nbsp;11b). The resulting inundation could affect several areas on Thera Island, such as the NE coast but also large parts of the E coast, especially around the settlement of Monolithos (Fig.\u0026nbsp;8c).\u003c/p\u003e\n\u003cdiv id=\"Sec11\"\u003e\n \u003ch2\u003eSimulation 3\u003c/h2\u003e\n \u003cp\u003eSimulation 3 (Fig.\u0026nbsp;9) presents a similar scenario to Simulation 2, although at a larger scale (mobilized volume now of about 300 Mm\u003csup\u003e3\u003c/sup\u003e, i.e. about double that of Simulation 2). As shown in Fig.\u0026nbsp;9a, the area enclosing the maximum landslide volume is, in this case, large enough to also include some parts of the outer Kolumbo crater walls although the collapse is all internal at the crater. The video animation of this simulation is also provided as Supporting Information 3.\u003c/p\u003e\n \u003cdiv\u003e\n \u003cp\u003eThe impact of such a scenario on the Santorini archipelago is the largest among the three simulated. Arrival times of the 1-cm wave do not differ significantly with respect to the other simulations, especially for the NE coast of Thera (Fig.\u0026nbsp;10a), but maximum free surface heights can now reach up to \u0026gt; 70 meters near the Kolumbo crater (Fig.\u0026nbsp;10b) and up to about 36 m along the NE coast of Thera (near Koloumbos settlement, see also Fig.\u0026nbsp;11b), resulting in large inundation areas from NE Thera (wide areas on both sides of Koloumbos settlement), to E Thera (from Monolithos settlement to Kamari) and even to SE Thera (Perissa and southern coast, see Fig.\u0026nbsp;10c).\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\"\u003e\n \u003ch2\u003eHazard variables along the E coast of Thera\u003c/h2\u003e\n \u003cp\u003eFor a more quantitative analysis of the hazard associated with the simulation outcomes, we also present two additional figures illustrating some key variables along the E coast of Thera:\u003c/p\u003e\n \u003cp\u003e1) The distributions, along the NE to SE coast of Thera (roughly from Koloumbos settlement to Perissa), of the arrival times of the 1-cm wave (assumed representative of the arrival time of the tsunami) and of the maximum free surface height of the tsunami (Fig.\u0026nbsp;11);\u003c/p\u003e\n \u003cp\u003e2) a comparison between the maximum surface height of the three tsunami simulations at eight control points and along their associated profiles perpendicular to the NE-SE coastline of Thera and the NW coastline of Thirasia (Fig.\u0026nbsp;12).\u003c/p\u003e\n \u003cp\u003eRegarding the first figure (Fig.\u0026nbsp;11), we have chosen to investigate the distribution of arrival time and maximum surface height along this profile as it represents the closest area to Kolumbo volcano and the portion of Thera Island (especially from Monolithos toward S) with plain shores and relatively large settlements and critical infrastructures. For the two variables (i.e. arrival time and max free surface height) and the whole E coastline, we also show the minimum of the 1-cm arrival times (triangles in Fig.\u0026nbsp;11a) and the maximum of the free surface (diamonds in Fig.\u0026nbsp;11b). These outcomes have relevant applications from a civil protection point of view to identify areas with highest expected waves and shortest time delays between the triggering of the phenomenon at source and the associated effects along the coast.\u003c/p\u003e\n \u003cp\u003eThe delay between the triggering landslide and the wave arrival at the E coast is between about 3 minutes, at Koloumbus settlement, to about 10 minutes, at Perissa. The coastline of Thera Island inside the Santorini caldera is mostly characterized by high cliffs and it is also significantly more protected from large tsunami waves generated by the Kolumbo volcano. However, two important infrastructures are located along this coast, that are (see Fig.\u0026nbsp;2) the touristic ports of Thera and Athinios. In addition, another touristic port is located in front of the town of Thirasia in Thirasia Island (see Fig.\u0026nbsp;2). For these three localities, our largest simulation (Simulation 3) indicates:\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e1-cm wave arrival times of ~ 7 min, ~ 8 min, and ~ 6 min at Thera port, Athinios port and Thirasia port, respectively;\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003emaximum free surface heights of ~ 0.75 m, ~ 1.2 m, and ~ 0.8 m at Thera port, Athinios port, and Thirasia port, respectively;\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cp\u003eThe islands inside the Santorini caldera (Nea Kameni, Palea Kameni, and Aspronisi) are not populated and therefore not considered in this analysis, although the density of people can be relatively high in Nea Kameni during tourist visits. Vice versa, based on the results of Ulvrova et al. (2016), the tsunami travel times (i.e., the arrival time of the first wave generated at the Kolumbo volcano) for the surrounding islands (see Fig.\u0026nbsp;1) are of the order of ~ 6 minutes for the S coasts of the island of Ios, 8 minutes for the island of Anafi and ~ 10 minutes for the island of Amorgos.\u003c/p\u003e\n \u003cp\u003eThe second figure (Fig.\u0026nbsp;12) analyzes the maximum surface height computed in the three simulations in selected locations along the outside Thera coast. These control points selected are similar to those used by Ulvrova et al. (2016). In particular, we drew a profile line passing through each point and perpendicular to the coastline and we show the topo-bathymetry (from − 30 to 10 m a.s.l.) and the maximum free surface height along it for the three simulations (Fig.\u0026nbsp;12). For such control points, we also show in Table\u0026nbsp;4 the corresponding depth from our topo-bathymetry, the 1-cm wave arrival time and the maximum free heigh.\u003c/p\u003e\n \u003cp\u003eTo have a better appreciation of the potential hazardous actions associated with tsunami waves generated by Kolumbo volcano, some semi-quantitative comparisons between available data of the 1650 CE event and some selected results of our simulations were performed. For instance, Fig.\u0026nbsp;13 shows a comparison between the inundation computed by Simulation 3 and the field evidence of the 1650 CE tsunami. These latter data have been documented by Ulvrova et al. (2016) on several outcrops along the NE and SE Thera coasts where tsunami deposits have been recognized and linked to the 1650 CE event. The figure clearly shows that simulation results are semi-quantitatively consistent with the historical inundation of the 1650 CE event as reconstructed from field observations.\u003c/p\u003e\n \u003cp\u003eSimilarly, Fig.\u0026nbsp;14 provides an estimation of the potential inundation flow velocities of Simulation 3. To get the flow velocity (\u003cem\u003eu\u003c/em\u003e), we used the empirical correlation given by Matsutomi and Okamoto (2010), namely\u003c/p\u003e\n \u003cdiv id=\"Equ1\"\u003e\n \u003cdiv format=\"TEX\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e $$\\:u\\:=\\:Fr\\sqrt{{gh}_{f}}$$\u003c/div\u003e\n \u003cdiv\u003e1\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003ewhere \\(\\:{h}_{f}\\) is the water depth measured at the front of an impacted element (e.g. building, hill), \u003cem\u003eFr\u003c/em\u003e the Froude number (assumed equal to 0.66) and \u003cem\u003eg\u003c/em\u003e the gravitational acceleration.\u003c/p\u003e\n \u003cp\u003eValues obtained with this relation, based on the largest potential fluid force exerted by the flow on the impacted element, range from 0.2 to 12 m/s and therefore can produce major devastation of the coastal areas.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\"\u003e\n \u003ch2\u003eOutcomes of the expert elicitation on tsunami hazard\u003c/h2\u003e\n \u003cp\u003eTo have a better appreciation of the values associated with the new tsunami simulation above described, in this section, we briefly report some of the outcomes of the expert elicitation procedure carried out during the project (Bevilacqua et al. this volume) with specific reference to the block Q14 (questions 14a-c, 14d1-14d3) about the tsunamis hazard at Kolumbo. The new tsunami simulations were carried out at the same time of the expert elicitation sessions and therefore the outcome of one informed and guided the other thus favoring a more robust and consistent description of the phenomenon (Sparks et al. this volume). For more details on the elicitation, we refer to Bevilacqua et al. (this volume).\u003c/p\u003e\n \u003cp\u003eThe elicitation outcomes, summarized in Table\u0026nbsp;5, clearly imply that tsunamis are a significant threat for Thera and the other islands around Kolumbo. Moreover, experts’ answers indicate that there is a medium to high probability of having a major tsunami in the circumstances of an explosive eruption occurring in the next 30 years. Tsunamis are more likely to occur from the main crater of Kolumbo, particularly during the paroxysmal phase of the eruption. A complementary result is that tsunamis are less likely to be generated from the other edifices of the volcanic field although the uncertainty of these estimates is large. In terms of tsunami wave height, median probabilities (for the [5th -50th -95th ] percentile values see Table\u0026nbsp;5) of having a wave higher than 1 m along the NE coast of Thera are, for the CM (Classical Model)-EW (Equal Weight), 52–60%, whereas they decrease to about 19–27% for waves higher than 5 m, and to 7–12% for waves higher than 10 m. All these probabilities refer to potential events occurring in a future period of 30 years.\u003c/p\u003e\n \u003cdiv\u003e\n \u003cp\u003eThe three new simulations presented above were able to generate maximum wave heights of about 1.5, 17, and 36 m (for Simulation 1, 2, and 3, respectively - see Fig.\u0026nbsp;11b) along the NE coast of Thera and values of the order of 1, 5 and 10 m along approximately the central part of the NE coast of Thera (i.e the area between Koloumbus and Monolithos). As a consequence, the three simulations can be considered approximately representative of the three different scales associated with the elicitation questions (questions 14d1-14d3) and, in turn, to the probabilities of occurrence in case of an eruption in the next 30 years (Table\u0026nbsp;5).\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eNew simulated scenarios and associated hazard\u003c/h2\u003e \u003cp\u003eThe new simulated scenarios allow us to discuss the dynamics of tsunamis triggered by submarine landslides at Kolumbo volcano and their potential in terms of hazard. This mechanism of tsunami generation has received less attention in the literature with respect to other mechanisms (Karstens et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The choice to investigate the effect of this generation mechanism is largely dependent on the current geological and bathymetric conditions of the Kolumbo volcanic system which lacks a subaerial part.\u003c/p\u003e \u003cp\u003eIn particular, the pumice cone formed in the 1650 CE eruption becomes unstable in the case of a future explosive eruption or a regional earthquake such as the 1956 Amorgos earthquake (see e.g. Papazachos and Kkallas this volume). Particularly, a potential internal failure of the crater rim, as well as a landslide on the external rim of the crater, could be likely generated during a future explosive eruption (Katsigera et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) or large magnitude regional earthquake.\u003c/p\u003e \u003cp\u003eFor the assumed rheological properties of the landslide above described, results of Simulation 1 suggest that the SW-facing slope of Kolumbo volcano has a low potential to generate submarine landslides resulting in \u0026gt;\u0026thinsp;1 m tsunami waves on NE coast of Thera Island. This is interpreted as due to the low slope angles along the SW-side of the Kolumbo volcano, which are generally well below (about 10\u0026deg;) the value of friction angles considered in our simulations (see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). However, the definition of realistic values of the friction angles of such landslides is still subject of debate (see e.g. Poulain et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). For the Kolumbo case, this task is furthermore complicated due to the complex stratigraphy of its outer slopes, where fine beds of pumice clasts could act as weak failure surfaces (especially if shaken by a regional earthquake) resulting in an increase in pore pressure and reduced friction angles (see e.g. d\u0026rsquo;Acremont et al. 2022; Tran et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, further investigation should be undertaken to evaluate the effect of lower friction angles on the remobilized volume of the landslide and, thus, on its tsunamigenic potential.\u003c/p\u003e \u003cp\u003eVice versa, Simulations 2 and 3, as representative of the medium and largest scenarios here considered, describe the propagation of a tsunami triggered by a landslide occurring on the inner slopes of the Kolumbo crater with volumes of about 150 and 300 Mm\u003csup\u003e3\u003c/sup\u003e, respectively, and have the potential to represent a major hazard for the E coast of Thera (see Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e and \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e). We note in fact that the first wave arrival times at the control points (see Table\u0026nbsp;4) are ~\u0026thinsp;2\u0026ndash;3 minutes on the NE Thera (e.g. Profiles 2\u0026ndash;3) and ~\u0026thinsp;8\u0026ndash;10 minutes near Perissa on the SE coast of Thera (Profile 7). Moreover, if we consider the whole E coast of Thera, we observe that both the shortest arrival times (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e11\u003c/span\u003ea) and the highest waves (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e11\u003c/span\u003eb) of our simulated scenarios are expected near the settlement of Koloumbos, but significantly short times (~\u0026thinsp;3\u0026ndash;4 minutes) and high waves (~\u0026thinsp;1\u0026ndash;35 m) are also computed along the whole E coast. Such timings and potential wave heights should be properly taken into consideration when preparing emergency/evacuation plans and designing alert systems linked to the detection of the triggering phenomenon.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eFinally, although the aim of our simulations was not to reproduce the tsunamis occurred during the 1650 CE eruption, we note that the internal crater failure scenarios investigated (e.g. Simulations 2 and especially Simulation 3, see Figs.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e), produce an inundation of the Monolithos/Perissa areas (especially with Simulation 3, see Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e) which is semi-quantitatively consistent with the historical accounts (see Section 1) and the field evidence of the 1650 CE tsunami (see Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e). In addition, we observe that Simulation 3 produces an inundation along the E coast of Thera Island comparable with that produced by the simulations of Karstens et al. (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) but with about 20% of their mobilized volume, a simpler triggering mechanism and a shorter emplacement time of the mass movement (4 minutes from Karstens et al. (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) compared to the 48\u0026ndash;62 s of the present simulations, see Table\u0026nbsp;3 and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). With these observations, we are not claiming that the 1650 CE tsunami was certainly generated by an internal collapse of the Kolumbo crater, but this triggering mechanism is certainly plausible and compatible with some of its evidence. Further ad-hoc simulations should be done to better reconstruct the 1650 CE tsunami event.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eSimulated scenarios from literature and hazard assessment\u003c/h2\u003e \u003cp\u003eThe considerable number of simulations done in previous work (Ulvrova et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Karstens et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and in the current study can provide a first comprehensive evaluation of tsunami hazard at Kolumbo. To this aim, we developed a simplified matrix diagram shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e15\u003c/span\u003e that links, from one side, the tsunami triggering mechanism such as submarine explosion, pyroclastic flow, caldera collapse, and submarine landslide, and, on the other side, the probabilities of having major tsunamis with different wave heights (i.e. 1, 5 and 10 m) along the NE coast of Thera in case of an eruption in the next 30 years as estimated by expert elicitation (Bevilacqua et al. this volume).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFor each of the source mechanisms of tsunami mentioned in Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e15\u003c/span\u003e, we identified scenarios for which waves on NE Thera exceed the values used in the expert elicitation (i.e. 1, 5 and 10 m). These scenarios are based on the sensitivity of the simulation outcomes to key source parameters that have the highest influence on wave amplitude: explosion energy, pyroclastic flow volume flux, caldera collapse duration, and submarine landslide volume (Paris \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Schindel\u0026eacute; et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In the matrix diagram of Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e15\u003c/span\u003e all the simulations reported are related to a source location inside the current Kolumbo crater. Moreover, even for the new simulations done assuming such a source, there are still some scenarios missing with respect to some values of the elicited wave height along NE Thera (i.e., crater internal failure generating\u0026thinsp;\u0026gt;\u0026thinsp;1 m wave height and landslide along the crater external slopes generating\u0026thinsp;\u0026gt;\u0026thinsp;10 m wave height), which should be considered in future developments.\u003c/p\u003e \u003cp\u003eFinally, the values of wave heights obtained from simulations are given along the coast. During the inland inundation phase, tsunami flows typically have Froude numbers between 0.66 and 2 (Matsutomi and Okamoto \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). As an example, a tsunami with initial wave heights of 1 m, 5 m, and 10 m at the coast will propagate inland (using Eq.\u0026nbsp;\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e above) at velocities higher than 2 m/s, 4.5 m/s, and 6.5 m/s, respectively. For a 1650-like event, inundation flow velocity in Perissa and Monolithos will likely exceed, respectively, 4 m/s and 7 m/s. Such estimates are instead more difficult to obtain for the Cape Koloumbos area because of the steep topography in that part of the island.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eBased on the findings of the expert elicitation carried out (Bevilacqua et al. this volume; Sparks et al. this volume), a future eruption in the Kolumbo volcanic field has a good chance to generate tsunamis. In particular, an eruption from the Main Cone meets conditions conducive to tsunami formation which includes the existence of a large unstable tephra cone and a high probability that the next eruption will be predominantly explosive. Elicitation outcomes indicate a medium to high likelihood of a major tsunami associated with a 30-year future eruption. In particular, chances of having a\u0026thinsp;\u0026gt;\u0026thinsp;1 m wave on the NE coast of Thera have medians in the range 50\u0026ndash;60%, although there is a large uncertainty associated with these estimates. Likelihood decreases to median values of about 7\u0026ndash;12% of having\u0026thinsp;\u0026gt;\u0026thinsp;10 m tsunami wave heights along the NE coast of Thera, although 95%ile values still reach 30\u0026ndash;60%, depending on the elicitation decision maker adopted.\u003c/p\u003e \u003cp\u003eIn this paper, we reviewed the results of the tsunami simulations carried out in previous works (Ulvrova et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Karstens et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and new simulations produced during this project. This analysis allows us to outline a first comprehensive description of the tsunami scenario-based hazard at Kolumbo. Moreover, the new simulations carried out in this study allowed us to describe the dynamic evolution of two new triggering mechanisms involving the Kolumbo crater area: a landslide involving the outer, SW-facing crater slopes, and an internal failure of the crater rim (simulated by assuming two different volumes).\u003c/p\u003e \u003cp\u003eThe simulation outputs allowed us to obtain some important insights regarding tsunami hazard triggered by submarine landslides. A scenario involving a significant (i.e. volume of the order of 150\u0026ndash;300 Mm\u003csup\u003e3\u003c/sup\u003e) internal failure of the Kolumbo crater (not necessarily related to an eruption) is capable of producing tsunami waves up to \u0026gt;\u0026thinsp;10 m high along the NE coast of Thera. Such a scenario is made more likely by the steep slopes of the inner crater wall in agreement with Katsigera et al. (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). These volume values are also associated with wave heights\u0026thinsp;\u0026gt;\u0026thinsp;5 m along the E coast of Thera, and ~\u0026thinsp;5 m along the SE coast of Thera. Tsunami risk associated with these scenarios is relatively low along the NE coast of Thera due to the low density of settlements, whereas it is potentially much larger along the E and SE coasts, given the presence of main infrastructures (i.e. Thera airport near Monolithos on the E coast) and highly touristic areas (near Perissa and Kamari on the E and SE coasts). The expected arrival times of the first waves are also remarkably short ranging from 2\u0026ndash;3 minutes in the most exposed coasts up to 8\u0026ndash;10 min on the south-eastern coast of Thera. Such tsunami waves would propagate inland at velocities typically ranging from 2 to 12 m/s. We refer to Sparks et al. (this volume) for a first quantitative analysis of the individual and societal risk associated with Kolumbo tsunamis.\u003c/p\u003e \u003cp\u003eWe have also shown how simulation results suggest that, given the specific set of landslide rheological parameters assumed, it is unlikely to mobilize a landslide with a large volume from the SW-facing Kolumbo crater slopes, given the relatively gentle topo-bathymetry of this area. Nevertheless more work is needed to investigate the effect of landslide rheological properties on this outcome.\u003c/p\u003e \u003cp\u003eOn the other hand, a scenario involving an internal failure of the Kolumbo crater could produce an inundation on Thera Island semi-quantitatively consistent with that observed during the paroxysm of the 1650 CE eruption, although we acknowledge that during such a complex eruption multiple mechanisms could have generated the reconstructed tsunamis.\u003c/p\u003e \u003cp\u003eFinally, while we underline that the present study can contribute to the improvement of tsunami hazard assessment at Kolumbo, we believe that more research is still needed to fill some important knowledge gaps. In particular, additional simulations and analyses are required to investigate tsunami events triggered from i) different areas along the outer slopes of Kolumbo crater (e.g. to the NW), ii) smaller volumes due to internal crater failure, iii) different areas from the Kolumbo volcanic field, and iv) different physical and rheological properties of the collapsing material (in particular frictional parameters). In addition, a larger computational domain including the nearby islands (Ios, Anafi and Amorgos) is required to better characterize the volcanic tsunami hazard assessment in this region.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLarge Scale Orthophotos\u0026rdquo; (LSO) Project was implemented by the Hellenic Cadastre and co-funded by the European Union within the framework of the Operational Program \u0026quot;Information Society\u0026quot; of the 3rd Community Support Framework and NSRF 2007 - 2013.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research has been supported by the project \u0026ldquo;Hazard and risk assessment for Kolumbo Volcano, Greece\u0026rdquo;, Hellenic Survey of Geology \u0026amp; Mineral Exploration (HSGME). We thank Hellenic Survey of Geology \u0026amp; Mineral Exploration (HSGME) for hosting first project workshop on Santorini, and Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Pisa, for hosting second project workshop.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest/Competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest/competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe online version contains supplementary material available as Supporting Information 1, Supporting Information 2 and Supporting Information 3.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAlexandri M, Papanikolaou D, Nomikou P (2003) Santorini Volcanic Field - New Insights Based On Swath Bathymetry. Abstracts Iugg, 30 June-11 july 2003, Sapporo, Japan.\u003c/li\u003e\n\u003cli\u003eAuker MR, Sparks RSJ, Siebert L, Crosweller HS, Ewert J (2013) A statistical analysis of the global historical volcanic fatalities record. J Appl Volcanol 2:1-24. https://doi.org/10.1186/2191-5040-2-2.\u003c/li\u003e\n\u003cli\u003eBevilacqua A, de\u0026rsquo;Michieli Vitturi M, Tadini A, Neri A, Aspinall WP, Sparks RSJ (this volume) Kolumbo Volcano, Greece: expert elicitation findings supporting volcanic hazard and risk assessment.\u003c/li\u003e\n\u003cli\u003eCantner K, Carey S, Nomikou P (2014) Integrated volcanologic and petrologic analysis of the 1650 AD eruption of Kolumbo submarine volcano, Greece. J Volcanol Geotherm Res 269:28-43. https://doi.org/10.1016/j.jvolgeores.2013.10.004.\u003c/li\u003e\n\u003cli\u003eCerminara M, Esposti Ongaro T, De\u0026rsquo; Michieli Vitturi M, Tadini A, Trolese M, Fornaciai A, Nannipieri L, Rodriguez Galvez JF (2024). Simulated scenarios of volcanic mass movements and associated tsunamis at Stromboli (Aeolian archipelago, Tyrrhenian sea, Italy). version 1 [Data set]. Istituto Nazionale di Geofisica e Vulcanologia (INGV). https://doi.org/10.13127/stromboli/sciara_del_fuoco_tsunami\u003c/li\u003e\n\u003cli\u003ed\u0026apos;Acremont E, Lafuerza S, Rabaute A, Lafosse M, Castelot MJ, Gorini C, Alonso B, Ercilla G, Vazquez JT, Vandorpe T, Juan C, Migeon S, Ceramicola S, Lopez-Gonzalez N, Rodriguez M, El Moumni B, Benmarha O, Ammar A (2022). Distribution and origin of submarine landslides in the active margin of the southern Alboran Sea (Western Mediterranean Sea). Mar Geol 445, 106739. https://doi.org/10.1016/j.margeo.2022.106739.\u003c/li\u003e\n\u003cli\u003eDay SJ (2015) Volcanic tsunamis. In: Sigurdsson H (Ed) The encyclopedia of volcanoes, 2nd ed. Academic Press, pp 993-1009. https://doi.org/10.1016/B978-0-12-385938-9.00058-4.\u003c/li\u003e\n\u003cli\u003eDominey-Howes DTM, Papadopoulos GA, Dawson AG (2000) Geological and historical investigation of the 1650 Mt. Columbo (Thera Island) eruption and tsunami, Aegean Sea, Greece. Nat Haz 21:83\u0026ndash;96. https://doi.org/10.1023/A:1008178100633.\u003c/li\u003e\n\u003cli\u003eEscalante C, de Luna TM, Castro MJ (2018) Non-hydrostatic pressure shallow flows: GPU implementation using finite volume and finite difference scheme. Appl Math Comput 338:631\u0026ndash;659. https://doi.org/10.1016/j.amc.2018.06.035.\u003c/li\u003e\n\u003cli\u003eEscalante C, Fern\u0026aacute;ndez-Nieto ED, de Luna TM, Castro MJ (2019) An efficient two-layer non-hydrostatic approach for dispersive water waves. J Sci Comput 79:273\u0026ndash;320. https://doi.org/10.1007/s10915-018-0849-9.\u003c/li\u003e\n\u003cli\u003eEsposti Ongaro T, de\u0026rsquo; Michieli Vitturi M, Cerminara M, Fornaciai A, Nannipieri L, Favalli M, Calusi B, Mac\u0026igrave;as J, Castro MJ, Ortega S, Gonz\u0026aacute;lez-Vida JM, Escalante C (2021) Modeling Tsunamis Generated by Submarine Landslides at Stromboli Volcano (Aeolian Islands, Italy): A Numerical Benchmark Study. Front Earth Sci 9:628652. \u003cu\u003ehttp://doi.org/10.3389/feart.2021.628652\u003c/u\u003e.\u003c/li\u003e\n\u003cli\u003eFern\u0026aacute;ndez-Nieto ED, Bouchut F, Bresch D, Castro D\u0026iacute;az MJ, Mangeney A (2008) A new Savage\u0026ndash;Hutter type model for submarine avalanches and generated tsunami. J Comput Phys 227:7720\u0026ndash;7754. https://doi.org/10.1016/j.jcp.2008.04.039.\u003c/li\u003e\n\u003cli\u003eFern\u0026aacute;ndez-Nieto ED, Garres-D\u0026iacute;az J, Mangeney A, Narbona-Reina G (2018) 2D granular flows with the \u0026mu; (I) rheology and side walls friction: A well-balanced multilayer discretization. J Comput Phys 356:192-219. https://doi.org/10.1016/j.jcp.2017.11.038.\u003c/li\u003e\n\u003cli\u003eFouqu\u0026eacute; F (1879) Santorin et ses \u0026eacute;ruptions. Masson.\u003c/li\u003e\n\u003cli\u003eFreundt A (2017) POS 513 cruise report (p. 18). GEOMAR. https://doi.org/10.3289/cr_pos513\u003c/li\u003e\n\u003cli\u003eFuller S, Carey S, Nomikou P (2018) Distribution of fine-grained tephra from the 1650 CE submarine eruption of Kolumbo volcano, Greece. J Volcanol Geotherm Res 352:10-25. https://doi.org/10.1016/j.jvolgeores.2018.01.004.\u003c/li\u003e\n\u003cli\u003eGupta AK, Bennartz R, Fauria KE, Mittal T (2022) Eruption chronology of the December 2021 to January 2022 Hunga Tonga-Hunga Ha\u0026rsquo;apai eruption sequence. Commun Earth Environ 3:314. https://doi.org/10.1038/s43247-022-00606-3.\u003c/li\u003e\n\u003cli\u003eKarstens J, Crutchley GJ, Hansteen TH, Preine J, Carey S, Elger J, K\u0026uuml;hn M, Nomikou P, Schmid F, Dalla Valle G, Kelfoun K, Berndt, C. (2023) Cascading events during the 1650 tsunamigenic eruption of Kolumbo volcano. Nat Commun 14(1) 6606. https://doi.org/10.1038/s41467-023-42261-y.\u003c/li\u003e\n\u003cli\u003eKatsigera A, Nomikou P, Pavlopoulos K (2024) A Preliminary Hazard Assessment of Kolumbo Volcano (Santorini, Greece). GeoHazards 5(3):816-832. https://doi.org/10.3390/geohazards5030041.\u003c/li\u003e\n\u003cli\u003eKonstantinou KI (2020) Magma chamber evolution during the 1650 AD Kolumbo eruption provides clues about past and future volcanic activity. Sci Rep 10(1) 15423. https://doi.org/10.1038/s41598-020-71991-y.\u003c/li\u003e\n\u003cli\u003eMac\u0026iacute;as J, Escalante C, Castro MJ (2020a) Multilayer-HySEA model validation for landslide generated tsunamis. Part I rigid slides. Nat Hazards Earth Syst Sci 1\u0026ndash;32. https://doi.org/10.5194/nhess-2020-171.\u003c/li\u003e\n\u003cli\u003eMac\u0026iacute;as J, Escalante C, Castro MJ (2020b) Multilayer-HySEA model validation for landslide generated tsunamis. Part II Granular slides. Nat Hazards Earth Syst Sci 1\u0026ndash;34. https://doi.org/10.5194/nhess-2020-172.\u003c/li\u003e\n\u003cli\u003eMastroianni F, Vougioukalakis G, Petrone C, Braschi E, Fantozzi I, Francalanci L (this volume) Multiple reservoirs and different magma interaction processes in the feeding system of the 1650 CE eruption of Kolumbo submarine volcano, Greece.\u003c/li\u003e\n\u003cli\u003eMatsutomi H, Okamoto K (2010) Inundation flow velocity of tsunami on land. Isl Arc 19:443-457. https://doi.org/10.1111/j.1440-1738.2010.00725.x.\u003c/li\u003e\n\u003cli\u003eNomikou P, Carey S, Papanikolaou D, Bell KC, Sakellariou D, Alexandri M, Bejelou K (2012) Submarine volcanoes of the Kolumbo volcanic zone NE of Santorini Caldera, Greece. Global Planet Change 90\u003cem\u003e:\u003c/em\u003e135-151. https://doi.org/10.1016/j.gloplacha.2012.01.001.\u003c/li\u003e\n\u003cli\u003eNomikou P, Carey S, Bell K, Papanikolaou D, Bejelou K, Cantner K, Sakellariou D, Perros I (2014) Tsunami hazard risk of a future volcanic eruption of Kolumbo submarine volcano, NE of Santorini Caldera, Greece. Nat Haz 72:1375-1390. https://doi.org/10.1007/s11069-012-0405-0.\u003c/li\u003e\n\u003cli\u003ePapazachos C, Kkallas C (this volume) Deterministic seismic hazard assessment for the Kolumbo area.\u003c/li\u003e\n\u003cli\u003eParis R (2015) Source mechanisms of volcanic tsunamis. Philos Trans R Soc Lond A 373 20140380. https://doi.org/10.1098/rsta.2014.0380.\u003c/li\u003e\n\u003cli\u003eParis R, Ulvrova M, Selva J, Brizuela B, Costa A, Grezio A, Lorito S, Tonini R (2019) Probabilistic hazard analysis for tsunamis generated by subaqueous volcanic explosions in the Campi Flegrei caldera, Italy. J Volcanol Geotherm Res 379:106-116. https://doi.org/10.1016/j.jvolgeores.2019.05.010.\u003c/li\u003e\n\u003cli\u003ePoulain P, Le Friant A, Mangeney A, Viroulet S, Fernandez-Nieto E, Castro Diaz M, Peruzzetto M, Grandjean G, Bouchut F, Pedreros R, Komorowski JC (2023). Performance and limits of a shallow-water model for landslide-generated tsunamis: from laboratory experiments to simulations of flank collapses at Montagne Pel\u0026eacute;e (Martinique). Geophys J Int 233(2):796-825. https://doi.org/10.1093/gji/ggac482\u003c/li\u003e\n\u003cli\u003ePouliquen O, Forterre Y (2002). Friction law for dense granular flows: application to the motion of a mass down a rough inclined plane. J Fluid Mech 453:1\u0026ndash;19. https://doi.org/10.1017/S0022112001006796.\u003c/li\u003e\n\u003cli\u003eSakellariou D, Sigurdsson H, Alexandri M, Carey S, Rousakis G, Nomikou P, Georgiou P, Ballas D (2010) Active tectonics in the Hellenic Volcanic Arc: the Kolumbo submarine volcanic zone. Bull Geol Soc of Greece XLIII/2:1056\u0026ndash;1063.\u003c/li\u003e\n\u003cli\u003eSchindel\u0026eacute; F, Kong L, Lane EM, Paris R, Ripepe M, Titov V, Bailey R (2024) A Review of Tsunamis Generated by Volcanoes (TGV) Source Mechanism, Modelling, Monitoring and Warning Systems. Pure Appl Geophys 181:1745\u0026ndash;1792. https://doi.org/10.1007/s00024-024-03515-y.\u003c/li\u003e\n\u003cli\u003eScott WE, Hoblitt RP, Torres RC, Self S, Martinez ML, Nillos TJ (1996) Pyroclastic flows of the June 15, 1991, climatic eruption of Mount Pinatubo. In: Newhall CG, Punongbayan S (eds) Fire and Mud: Eruptions of Mount Pinatubo, Philippines. Philipp. Inst Volcanol and Seismol, Quezon City, pp. 545\u0026ndash;570.\u003c/li\u003e\n\u003cli\u003eShen Y, Whittaker CN, Lane EM, White JD, Power W (2024) Physical experiments of waves generated by submerged steam eruptions with applications to volcanic tsunamis. Sci Adv 10(45) eadk6208. https://doi.org/10.1126/sciadv.adk6208.\u003c/li\u003e\n\u003cli\u003eSigurdsson, H., Carey, S., Alexandri, M., Vougioukalakis, G., Croff, K. L., Roman, C., Sakellariou, D., Anagnstou, C., Rousakis, G., Ioakim, C., Gogou, A., Ballas, D., Misaridis, T., and Nomikou, P. (2006). Marine investigations of Greece\u0026apos;s Santorini volcanic field. Eos 87(34):337\u0026ndash;342. https://doi.org/10.1029/2006EO340001.\u003c/li\u003e\n\u003cli\u003eSparks RSJ, Vougioukalakis G, Aspinall WP, Neri A, Antonakos A, Baxter PJ, Bevilacqua A, Cerminara M, de\u0026rsquo; Michieli Vitturi M, Francalanci L, Koutroulli A, Maeno F, Mastroianni F, Papazachos K, Pardini F, Paris R, Tadini A, Vaselli O (this volume). Future eruptions of the Kolumbo volcanic field: prognosis with hazard and risk assessment.\u003c/li\u003e\n\u003cli\u003eSyamsidik B, Luthfi M, Suppasri A, Comfort LK (2020) The 22 December 2018 Mount Anak Krakatau volcanogenic tsunami on Sunda Strait coasts, Indonesia: tsunami and damage characteristics. Nat Hazards Earth Syst Sci 20:549\u0026ndash;565. https://doi.org/10.5194/nhess-20-549-2020.\u003c/li\u003e\n\u003cli\u003eTerry JP, Goff J, Winspear N, Bongolan VP, Fisher S (2022) Tonga volcanic eruption and tsunami, January 2022: Globally the most significant opportunity to observe an explosive and tsunamigenic submarine eruption since AD 1883 Krakatau. Geosci. Lett., 9(1):24. https://doi.org/10.1186/s40562-022-00232-z.\u003c/li\u003e\n\u003cli\u003eTran QA, S\u0026oslash;rlie E, Grimstad G, Eiksund G, Takahashi H, and Sassa S (2024). Influence of sediment permeability in seismic-induced submarine landslide mechanism: CFD-MPM validation with centrifuge tests and analysis. Comput Geotech, 174, 106588. https://doi.org/10.1016/j.compgeo.2024.106588.\u003c/li\u003e\n\u003cli\u003eTrolese M, Cerminara M, Esposti Ongaro T, de\u0026apos; Michieli Vitturi M, Tadini A (2024). Modeling tsunami generation and propagation: Insights from sensitivity analysis of landslide parameters at Stromboli, EGU General Assembly 2024, Vienna, Austria, 14\u0026ndash;19 Apr 2024, EGU24-19260, https://doi.org/10.5194/egusphere-egu24-19260, 2024.\u003c/li\u003e\n\u003cli\u003eUlvrova M, Paris R, Nomikou P, Kelfoun K, Leibrandt S, Tappin DR, McCoy FW (2016) Source of the tsunami generated by the 1650 AD eruption of Kolumbo submarine volcano (Aegean Sea, Greece). J Volcanol Geotherm Res 321:125-139. https://doi.org/10.1016/j.jvolgeores.2016.04.034.\u003c/li\u003e\n\u003cli\u003eVougioukalakis G, Francalanci L, Mitropoulos D, Perissoratis K (1996) The 1649\u0026ndash;1650 eruption of the Kolumbo submarine volcanic center, Santorini. In Second Workshop on European Laboratory Volcanoes, Thira, Santorini, Greece.\u003c/li\u003e\n\u003cli\u003eVougioukalakis G, Koutroulli A, Sparks RSJ, Francalanci L, Mastroianni F, Laurenzi MA, Schaen G, Papazachos C, Aspinall WP, Kanellopoulos C, Baxter PJ, Neri A, Tadini A, Bevilacqua A (this volume) The Kolumbo volcanic field.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 5 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bulletin-of-volcanology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buvo","sideBox":"Learn more about [Bulletin of Volcanology](http://link.springer.com/journal/445)","snPcode":"445","submissionUrl":"https://www.editorialmanager.com/buvo/default2.aspx","title":"Bulletin of Volcanology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Kolumbo volcano (Greece), 1650 CE eruption, tsunami hazard, numerical simulation","lastPublishedDoi":"10.21203/rs.3.rs-5700315/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5700315/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eVolcanic-induced tsunamis have a potentially devastating impact, especially in densely populated and/or touristic coastal areas. Kolumbo submarine volcano (Greece) experienced in 1650 CE an explosive eruption with eyewitnesses\u0026rsquo; accounts of major tsunamis along the coasts of Santorini (Thera) and other islands. We present a scenario-based tsunami hazard assessment at this volcano based on existing simulations from literature and new simulations of tsunamis triggered by a less investigated but important mechanism, i.e. submarine landslides on the volcano flanks or within its crater. Simulations results show that the remobilization of a landslide volume of 150\u0026ndash;300 Mm\u003csup\u003e3\u003c/sup\u003e inside the crater can produce tsunami waves larger than 10 m high along the NE coast of Thera and of the order of 5 m along the E and SE coasts. The expected tsunami arrival time ranges from 2\u0026ndash;3 minutes along the NE coast of Thera up to 8\u0026ndash;10 min on its SE coast. Such scenarios produce inundation areas consistent with those reconstructed for the 1650 CE event, and tsunami waves propagating inland at velocities from 2 to 12 m/s. Simulation results also suggest that, given the landslide parameters assumed, it is unlikely to mobilize a landslide with a large volume from the SW-facing Kolumbo crater slopes, given the relatively gentle topo-bathymetry of this area. The study findings are relevant based on the outcomes of the expert elicitation exercise carried out in parallel, which indicate that chances of having waves larger than 1 m high on the NE coast of Thera have median probabilities of 50\u0026ndash;60%.\u003c/p\u003e","manuscriptTitle":"Scenario-based tsunami hazard assessment at Kolumbo volcano","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-02-05 18:18:43","doi":"10.21203/rs.3.rs-5700315/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Moderate revision (possibly re-reviewed)","date":"2025-04-01T13:00:56+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-02-13T09:34:38+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-02-03T15:06:37+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Bulletin of Volcanology","date":"2025-01-13T12:04:27+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-12-26T02:17:46+00:00","index":"","fulltext":""},{"type":"submitted","content":"Bulletin of Volcanology","date":"2024-12-23T09:09:29+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bulletin-of-volcanology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buvo","sideBox":"Learn more about [Bulletin of Volcanology](http://link.springer.com/journal/445)","snPcode":"445","submissionUrl":"https://www.editorialmanager.com/buvo/default2.aspx","title":"Bulletin of Volcanology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"2d066a3e-1021-4e05-abf9-a0a9c5822a9d","owner":[],"postedDate":"February 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-06-09T16:02:02+00:00","versionOfRecord":{"articleIdentity":"rs-5700315","link":"https://doi.org/10.1007/s00445-025-01837-w","journal":{"identity":"bulletin-of-volcanology","isVorOnly":false,"title":"Bulletin of Volcanology"},"publishedOn":"2025-06-07 15:57:29","publishedOnDateReadable":"June 7th, 2025"},"versionCreatedAt":"2025-02-05 18:18:43","video":"","vorDoi":"10.1007/s00445-025-01837-w","vorDoiUrl":"https://doi.org/10.1007/s00445-025-01837-w","workflowStages":[]},"version":"v1","identity":"rs-5700315","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5700315","identity":"rs-5700315","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}