1 Introduction

The coastal areas of the Adriatic Sea are exposed to a range of natural hazards, including floods, windstorms, droughts, earthquakes, and tsunamis. Although strong tsunamigenic earthquakes are not frequent in the Adriatic Sea, their impacts could be considerable given the coastal exposure and vulnerability. Indeed, even a moderate tsunami may have a relevant effect on some areas due to the presence of sites of high historical/cultural/touristic interest (e.g. the city of Venice) and fragile ecosystems (e.g. lagoons and river deltas). Defining hazard scenarios for a wide range of possible tsunamigenic earthquakes is essential to reduce their potential socioeconomic impacts on coastal communities and set up sustainable development plans for inland and offshore areas. The adequate estimation of expected hazards, along with the characterization of exposed assets and their vulnerability, provides the basis for tsunami risk assessment.

Historical records indicate that the Adriatic region experienced considerable impacts from earthquakes and tsunamis in the past (Papadopoulos et al. 2014; Maramai et al. 2014; 2019; 2021). For instance, the 30 July 1627 Gargano earthquake (Mw6.7) and subsequent tsunami caused widespread destruction along the Eastern Adriatic coast (Patacca and Scandone 2004). More recently, the 15 April 1979 Montenegro earthquake (Mw6.9) generated a local tsunami that affected the coastal areas in Montenegro and Albania. Overall, Pasarić et al. (2012) identified 27 tsunamigenic events that occurred during the last 600 years: 9 of them are located on the west side of the Adriatic Sea, while 18 along its eastern side (6 are classified as reliable and 10 as unreliable, while 2 are categorized as events of meteorological origin). In 2019, an updated tsunami catalogue for the Mediterranean region was published, namely, the Euro-Mediterranean Tsunami Catalogue (EMTC), which includes a detailed survey of tsunamis of seismic origin in the Adriatic Sea (Maramai et al. 2019). This study identified the potential tsunamigenic events, including a level of uncertainty in the location and magnitude of seismic events.

Significant tsunamis are rare in the Northern Adriatic, where only a few historical events have been reported in the literature, with sources mostly located along the central and southern parts of the Adriatic coasts (Maramai et al. 2007; Tiberti et al. 2008). Based on historical information, recent review studies support the relevant impacts of past tsunamis in the Northeast Adriatic region in 1348 and 1511 (Maramai et al. 2021). Despite the evidence of sporadic but possibly damaging events, a detailed investigation of the potential impact of tsunamis of seismic origin on coastal activities and communities along the Northern Adriatic coasts is still unavailable. Therefore, this work aims to provide different scenario-based tsunami hazard metrics along the shoreline for the area of interest (e.g. maximum wave amplitude, estimated time of arrivals) given the available source, bathymetry and topography datasets. The results are essential for a comprehensive understanding and assessment of exposure, vulnerability and risk, thus for increasing the resilience of coastal communities in the Northeastern Adriatic region. Moreover, these results are also needed to develop emergency plans, as well as to suitably inform residents and visitors about the potential tsunami risk in the region and to convey adequate instructions in the event of a tsunami warning.

Recently, a tsunami early warning system has been established for the entire Northeast Atlantic and the Mediterranean region, referred to as NEAMTWS (Tsunami Warning and Mitigation System in the Northeastern Atlantic, the Mediterranean and connected seas). In this framework, the CAT-INGV system (Centro Allerta Tsunami, https://cat.ingv.it/en/), managed by INGV (National Institute of Geophysics and Volcanology), operates as a tsunami warning centre for the entire Mediterranean, including the Adriatic Sea (Amato et al. 2021). At the national level, CAT-INGV operates within the formal framework of the National Alert System for Earthquake-Generated Tsunamis (SiAM—Sistema Allertamento Maremoti), established in 2017 (Directive of the President of the Council of Ministers, 17 February 2017) to disseminate tsunami alert messages to the territory, including local authorities. A debate is ongoing about the possibility of delivering last-mile messages to alert the population (Papadopoulos et al. 2020). In case a potential tsunamigenic event occurs, the system provides alert messages expressed by the following three levels, in order of increasing severity: "information" (green), "advisory" (orange) and "watch" (red), based on the specific decision matrix (Fig. 1) reported in Amato et al. (2021). According to this matrix, the distance between a site of interest and the epicentre is divided into three possible categories: local (distance ≤ 100 km), regional (100 km < distance ≤ 400 km) and basin-wide (distance > 400 km). The "watch" level is defined when seismic information indicates that the coast may be hit by a tsunami with a wave amplitude greater than 0.5 m offshore (along 50 m isobaths) and/or when the tsunami run-up at the shoreline is expected to be greater than 1 m. The "advisory" level is indicated, instead, when the run-up is expected to be less than 1 m. Since 2017, several tsunamis have been located and assigned a watch alert level by the CAT-INGV system, including the Aegean Sea earthquake in 2017 (Mw = 6.6) and the more recent events in Crete (2nd May 2020), Aegean Sea (30th October 2020), Northern Algeria (18th March 2021) and the Turkey earthquakes (6th February 2023).

Fig. 1
figure 1

CAT-INGV tsunami decision matrix (https://cat.ingv.it/en/tsunami-alert/alert-procedures/decision-matrix, last accessed May 2023)

Full size image

A still open challenge is to go beyond the alerts issued by the early warning system and define detailed impact scenarios that allow evacuation plans and mitigation strategies to be developed with local authorities (e.g. civil protection and municipalities). Such strategies require knowledge of the local context and a detailed definition of hazard/risk scenarios compatible with the potential tsunamis that may affect the region under consideration. Recent studies have been conducted for the Heraklion City, Crete (Triantafyllou et al. 2019) as well as for Southern Italy, both on expected inundation assessment (Tonini et al. 2021) and tsunami risk perception (Cerase et al. 2019). The study conducted by Tonini et al. (2021) focused on two locations along the Eastern coast of Sicily, namely Catania and Siracusa. Their research specifically examined the local tsunami hazard in these areas. To assess the hazard, they used an enhanced digital terrain model (DTM) with a spatial resolution of 10 m for the tsunami simulations. This DTM was created by combining and resampling various raster datasets with different horizontal cell resolutions.

Additionally, for the land portion of the study area, they incorporated a raster dataset with a 1-m cell resolution obtained from the Geology and Geotechnologies Laboratory at INGV, specifically near the coastal zone. However, similar research has not yet been carried out for the Northern Adriatic, where the bathymetric and topographic settings are substantially different. The available tsunami hazard estimates (e.g. NEAMTHM18 maps, Basili et al. 2019), in fact, are provided for sites located along the 50 m depth isobaths; given the very shallow bathymetry of the Northern Adriatic, these sites are very far from the areas of interest (Fig. 2), at a distance of more than 100 km from the coastline (e.g. Amato et al. 2021 and references therein). Usually, to calculate the run-up to the coastline (i.e. the maximum topographic height, compared to the mean sea level, reached by the tsunami wave during its ingression), an empirical coefficient (i.e. a multiplication factor) is applied to the maximum amplitude of the tsunami wave estimated along the 50 m isobaths (Tonini et al. 2021). However, in the Northern Adriatic, the distance between these isobaths and the coastline is very large, and such an empirical relationship may prove inadequate. In addition, in the alert messages, a minimum wave amplitude (or a run-up) level is given for the area of interest, but an upper bound is not provided. According to the mentioned SiAM directive (see Attachment 1 of the directive), the accuracy and reliability of inundation data can be significantly improved at a local scale (e.g. municipal level and port areas), especially for those areas particularly critical due to the presence of strategic infrastructure, high population density, etc., where it is appropriate to integrate its results with detailed studies through physical–numerical modelling. Accordingly, this study aims to provide a more complete and detailed description of the tsunami hazard, following a multi-scenario approach, for specific locations along the northeastern coast of the Adriatic Sea, namely, the cities of Trieste, Monfalcone, Grado and Lignano (Fig. 2), hereinafter referred to as areas of interest (AOI).

Fig. 2
figure 2

Regional and zoomed maps of the study region: (top) map of Northeast Adriatic Sea and photos of the areas of interest, AOI; (bottom) map of tsunami hazard estimates for the NEAM macro-area "Northeastern Atlantic, the Mediterranean and connected seas" (NEAMTHM18; http://ai2lab.org/tsumapsneam/interactive-hazard-curve-tool/, last accessed on March 2022) and map showing the distance between the sites where hazard estimates are provided (50 m isobaths) and the areas of interest

Full size image

The urban centres selected as AOI for this study (Fig. 2) represent important economic, historical and cultural centres. Trieste, the largest city in the Friuli Venezia Giulia Region (FVG), is of fundamental economic importance due to the presence of ports and other infrastructure, making it a strategic site for trade and tourism and its notable historical relevance. Monfalcone is an industrial site with several shipyards and a textile and chemical production centre. The cities of Grado and Lignano are among the leading Italian seaside tourist centres. Lignano, in particular, is a primary tourist site, reaching a population of several hundred thousand visitors during the summer period; the considerable variability of the population throughout the year makes it an area of particular interest for future tsunami risk assessment. Moreover, between Grado and Lignano there is a large lagoon, characterized by a peculiar and fragile ecosystem, with a high degree of vulnerability, due to its direct connection with the Adriatic Sea basin.

Modelling the hazard associated with possible tsunamigenic earthquakes occurring in the Adriatic Sea is therefore carried out to contribute to tsunami risk assessment and emergency management for the selected urban areas located along the Northeastern Adriatic coasts. This research builds on previous studies (Peresan and Hassan 2022) and aims to update the existing tsunami hazard models for the Northeastern Adriatic coastal cities of Lignano, Grado, Monfalcone and Trieste. Recent results based on the computation of an extensive set of tsunami scenarios are presented, considering the possible sources defined in the most updated database of seismogenic sources; parametric tests are performed to account for seismic source variability. Specifically, this study is accomplished by accounting for the recently updated DISS database of seismogenic sources in the Adriatic Sea, considering different potential tsunamigenic sources of tectonic origin located in three distance ranges from the areas of interest (namely at Adriatic-wide, regional and local scales). Selected basin-wide sources in the Mediterranean Sea are considered as well to assess the possible relevance to the AOI of tsunamis occurring outside the Adriatic Sea.

Tsunami modelling is performed by the NAMI DANCE software (Dogan et al. 2021a; Yalciner and Zaytsev 2017; Yalciner et al. 2017; Lynett et al. 2017; Velioglu et al. 2016), which allows accounting for seismic source properties, variable bathymetry and nonlinear effects in waves generation, propagation and inundation. The software has been successfully tested for recent tsunamis in the Mediterranean area and other regions (Celikbas et al. 2023; Dogan et al. 2021a, b, c; Dogan et al. 2023; Velioglu et al. 2016; Aytore et al. 2016). Three nested grids of bathymetry and topography are used in the computations, i.e. a coarse grid of about 400 m (Adriatic and Mediterranean scale) extracted from GEBCO 2020 data (https://download.gebco.net/), a medium grid of about 111 m (North Adriatic) produced from EMODnet data and a fine grid of 20 m (area of interest) obtained from the work of Trobec et al. (2018). Accordingly, a set of maps of computed tsunami maximum amplitude are developed for sources located at different distances from the Northeastern Adriatic region, considering, in particular, the distance ranges adopted in the decision matrix (Fig. 1), and therefore relevant towards emergency management and planning mitigation actions at the selected sites (namely, Lignano, Grado, Monfalcone and Trieste).

This work consists of two main parts: in the first, we evaluate the hazard from the potential tsunamigenic sources located within the Adriatic Sea, which are categorized in this study into Adriatic-wide, regional and local sources, based upon their distance from the site of interest and given the tsunami decision matrix currently in force (Fig. 1). We provide maps of maximum tsunami wave amplitude for the whole grid till the shoreline of the area of interest, and nearshore wave amplitude at point of interests (POIs) (shown in supplementary material) for local, regional and Adriatic scales. Also, we provide the same maps for an aggregated scenario extracted from all possible scenarios within the Adriatic Sea at whatever distance. In the second part, we compute the tsunami scenarios for possible very distant sources located outside the Adriatic Sea, hereinafter called basin-wide, from five main tsunamigenic earthquake zones (West Greece, Calabrian, West Hellenic, East Hellenic and Cyprian Arcs). The tsunami scenarios of the maximum expected earthquake (based on historical seismicity or geological surveys) within each source were computed, and an aggregated scenario was extracted. Moreover, maps of maximum wave amplitude for aggregated scenarios at basin and area of interest scales are given for each tsunamigenic earthquake zone.

2 Tsunami modelling

Tsunami modelling usually relies on a specific model, namely a mathematical formulation that describes the physical characteristics of the tsunami, to compute the generation and propagation of tsunami waves and their coastal impact. The tsunami modelling process comprises three phases: (a) source modelling, in which the initial conditions of the tsunami source are computed (i.e. initial sea surface displacements and velocity fields); (b) propagation modelling, to be used to describe the propagation of tsunamis in the open sea; (c) inundation modelling, in which coastal modification of tsunami waves and run-up, in shallow regions and on land, is treated. The last phase is possibly the most important in tsunami modelling, as suitable tools and data are essential for accurate quantification.

Numerical models (i.e. computer-derived simulation packages) generally use a grid system for the area of interest that includes information about the source, bathymetry, topography and surface roughness or land use. Therefore, a numerical model can incorporate complicated geographic variations in bathymetry, topography and land uses, and can simulate different stages of tsunamis, including wave amplitude, current speed and inundation depth variations. The recent tsunami models simulate tsunami generation, propagation and inundation stages, overcoming the challenge of abrupt changes in conditions at the shore, which is the most dynamic and complex stage of tsunami modelling. Among the available models, COMCOT (Cornell University, USA; GNS Science, New Zealand), TUNAMI-N1/N2 (Tohoku University, Japan) and MOST (National Center for Tsunami Research, USA) have demonstrated their capacity in investigating the three stages of tsunami evolution. In this study, we perform the tsunami numerical modelling by the code NAMI DANCE (Middle East Technical University-METU, Turkey), which represents an extension and update of the model TUNAMI-N1/N2 and is based on the solution of the nonlinear form of the long wave equations with respect to related initial and boundary conditions. In general, the most common and proper solution upon which the codes are developed is the numerical solution in nonlinear shallow water approximation (Imamura 1988), which neglects the vertical variation of velocity over water depth because tsunami wavelengths are usually much larger than ocean depths.

The tsunami modelling computation carried out in this study is based on a nonlinear shallow water equation. The role of roughness is considered by attributing a friction coefficient of 0.015, taken from Chow (1960). As far as the source model is concerned, the NAMI DANCE code follows the Okada (1985) relations, which provide the numerical result of the elastic seabed displacement of the co-seismic deformation, to compute the initial condition for the shallow water wave equation (Dogan et al. 2021a; Yalciner and Zaytsev 2017; Yalciner et al. 2017; Lynett et al. 2017; Velioglu et al. 2016). For modelling purposes, a finite rectangular source model is defined by the following parameters: epicentral coordinates, strike, dip, rake, depth, amount of slip, length L and width W. The Okada (1985) model for fault slip is linear; hence, if necessary, it is possible to model a complex rupture as a group of source segments, each with a different amount of slip. However, Okada's model neglects the spatiotemporal variations in slip, which are essential for assessing tsunami impacts in regions close to the earthquake source. Geist and Dmowska (1999) showed that local tsunami run-up can vary over a factor of 3, depending on the slip distribution. Therefore, spatiotemporal variation in fault rupture is considered when treating the local tsunamigenic sources in the Gulf of Trieste by segmenting the source into sub-sources with variable rupture slip and velocity.

3 Bathymetry and topographic dataset

Three nested grids of bathymetry and topography have been used in the computations, namely: a coarse grid of about 400 m spatial resolution at the Adriatic and Mediterranean scale, a medium grid of about 111 m resolution for North Adriatic and a fine grid of 20 m resolution for the area of interest.

The bathymetry and topography data were extracted, inspected, processed and then converted into the format required by the tsunami modelling code (NAMI DANCE). The coarse-domain bathymetry and topography grids of about 400 m spatial resolution (parent) for the Eastern Mediterranean and Adriatic seas were extracted from GEBCO-2020 (GEBCO Compilation Group 2020). Also, medium-domain resolution grids of about 111 m were obtained and processed from the EMODnet portal (https://portal.emodnet-bathymetry.eu/). Finally, a fine bathymetry and topography grid of about 10 m resolution for the Northeastern Adriatic, where the municipalities of interest (AOI) are located, was provided by Trobec et al. (2018), who used various geophysical datasets to develop a detailed bathymetry model of the Gulf of Trieste (Fig. 3). Topographic or land data were extracted from Copernicus DEM- Global and European Digital Elevation Model (COP-DEM) of 10 m cell size. It is worth noting that in this study, the grid of 10 m resolution was de-sampled into a 20 m resolution grid due to current computational limits; the lower resolution grid, however, might not allow to capture some crucial features of the wetlands within the area of interest (e.g. Marano Lagoon, between Lignano and Grado). However, while the 20 m resolution grid might not be sufficient for a detailed inundation modelling, it could still be adequate for estimating the maximum tsunami amplitude and the estimated time of arrivals (ETA) for the area of interest, namely to meet the requirements of the civil protection of the FVG region. A careful examination of the considered bathymetry dataset revealed some artefacts (e.g. closed openings) that could affect the propagation of tsunami waves behind the shoreline. For instance, the closed openings to lagoons in the de-sampled bathymetry dataset may prevent tsunamis from progressively propagating within the wetlands. Therefore, the bathymetry data for these areas require careful analysis in the preprocessing stage and should be properly modified to accurately reflect the real situation, to allow computing the tsunami inundation. A focused and detailed mapping is foreseen in these areas, particularly the Marano Lagoon, and it must be developed for effective exposure, vulnerability and risk analysis at some sites of interest.

Fig. 3
figure 3

Bathymetry and topography of the study area and the area of interest. The large domain, with coarse bathymetry, is taken from GEBCO-2020. The smaller domain is taken from EMODnet (a) and Trobec et al. (2018) (b)

Full size image

The maps of topography and bathymetry of the study region (Fig. 3) indicate that the deeper part is located in the southern part of the Adriatic Sea, while the northern part is very shallow, with a water depth of less than 50 m. Noteworthy, due to the shallow bathymetry, the tsunami waves propagate with very low velocity in the Northern Adriatic Sea. This is a significant advantage, especially for distant tsunamigenic sources, as more time is available to alert people and to take emergency actions.

4 Computation of tsunamigenic earthquake scenarios in the Adriatic Sea

One of the goals of this study is to provide physically consistent tsunami hazard estimates based on the modelling of tsunami wave propagation from an extensive set of possible sources to be considered in case an alert is issued by the CAT-INGV tsunami warning system. Hence, we first defined the possible tsunamigenic earthquake sources located in the Adriatic Sea based on the recently updated DISS database (DISS-3.3 http://diss.rm.ingv.it/diss/) and following the tsunami decision matrix developed by CAT-INGV (Fig. 1). According to the decision matrix, a seismogenic source is considered tsunamigenic, of potential weak local tsunami (distance ≤ 100 km), if it is capable of producing shallow offshore earthquakes of magnitude M ≥ 6. Therefore, we compiled a database (Table 1 and Fig. 4) of local tsunamigenic earthquake sources of Mmax ≥ 6 relative to our area of interest from the DISS-3.3 database (Fig. 4) and Mmax ≥ 6.5 for regional (100 km < distance ≤ 400) and basin-wide (distance > 400 km) tsunamis, to evaluate the potential tsunami hazard at different distances and to serve as pre-computed scenarios in case a future tsunami alert is activated. This way, the best-matched pre-computed scenario (in location and magnitude) results can be identified, extracted and used for a quick response in case of tsunamigenic earthquake occurrence. Notably, the maximum magnitude of most seismogenic sources within the DISS-3.3 was upgraded (i.e. generally increased) in a 0–1.7 magnitude unit (Table 1) relative to the previous DISS-3.2.1 version. Therefore, according to the updated DISS-3.3, a more substantial number of seismogenic sources prove capable of generating tsunamis, compared to the former DISS version used by Peresan and Hassan (2022), because the magnitude increased for several sources.

Table 1 Parameters of offshore tsunamigenic sources in the Adriatic Sea and the representative tsunami scenarios
Full size table
Fig. 4
figure 4

Seismogenic sources from DISS-3.3 (DISS Working Group 2021), including a all sources and b the selected tsunamigenic sources in the Adriatic Sea located at different locations in the area of interest. Orange and yellow lines indicate the 100 km and 400 km limits, respectively

Full size image

Based on parametric studies, for each scenario, the parameters composing the worst-case scenario in terms of strike, dip, rake and focal depth were extracted from the given range in the DISS database and then adopted (geometry and kinematic parameters). In addition, the fault length, width and slip were estimated from the source scaling relationship of Wells and Coppersmith (1994) and recent studies, as reported in Table 1. More than a hundred individual scenarios were simulated during this work; maps of maximum tsunami wave amplitude at each grid and POIs nearshore (listed in supplementary material, Table S1) were then developed for local, regional and Adriatic-wide tsunamis (the computed scenarios and the related maps are available in digital format).

4.1 Adriatic-wide tsunamigenic earthquake sources

Following the described procedure, 13 seismogenic sources with Mmax ≥ 6.5, located at distances greater than 400 km from our area of interest, were selected to compute Adriatic-wide tsunami scenarios. These sources are: ALCS002, ALCS004, ALCS018, ALCS020, HRCS001, HRCS002, HRCS004, HRCS007, HRCS010, ITCS059, ITCS074, MECS001 and MECS007 and their Mmax range between 6.9 and 8; of note, most of the powerful tsunamigenic sources are located along the Eastern coast of the Adriatic Sea. The scenarios for the selected sources (Fig. 5) are simulated individually, and then an aggregated scenario is developed (Fig. 6). The computed maximum tsunami wave amplitude at each grid point till shoreline for our AOI is up to 0.5 m, as shown in Fig. 6a. The maximum wave amplitude is provided at given POIs as well (Fig. 6b) and indicates maximum wave amplitude of 0.5, 0.4, 0.2 and 0.2, for Lignano, Grado, Monfalcone and Trieste, respectively. The estimated arrival time (ETA) varies from source to source, with a minimum arrival time of 272–280 min for these cities.

Fig. 5
figure 5

Tsunamigenic earthquake sources at Adriatic-wide scale (distance from the AOI is > 400 km). The yellow line defines the sources at more than 400 km, while red rectangles are the tsunamigenic sources

Full size image
Fig. 6
figure 6

a Map of the maximum tsunami wave amplitude computed for the Adriatic-wide sources (listed in Table 1) located at distances greater than 400 km from the area of interest (AOI). b The maximum tsunami wave amplitude computed for the Adriatic-wide sources (listed in Table 1) at the selected POIs

Full size image

4.2 Regional tsunamigenic earthquake sources

From now on, following the decision matrix, the seismogenic sources with Mmax ≥ 6.5 located at a distance between 100 and 400 km from our area of interest were considered. Accordingly, more than 15 seismogenic sources from DISS were defined as potentially tsunamigenic (Fig. 7). They were simulated individually, and then an aggregated scenario was developed (Fig. 8). The computed maximum tsunami wave amplitude for our AOI was up to 1.0 m at some sites (Fig. 8a). The regional tsunami sources were found to impose a higher hazard to the AOI, compared to the Adriatic-wide tsunami sources. The maximum wave amplitude along the shoreline is provided at given POIs as well (Fig. 8b). The simulated scenarios indicate a maximum wave amplitude of 0.9, 1.0, 0.4 and 0.4 m for Lignano, Grado, Monfalcone and Trieste, respectively. Moreover, the minimum arrival time ETA ranged between 107 and 115 min for these cities.

Fig. 7
figure 7

Tsunamigenic earthquake sources at regional scale (distance from the AOI is > 100 and ≤ 400 km). The orange line defines the 100 km epicentral distance, while the yellow line defines the 400 km epicentral distance; red rectangles represent tsunamigenic sources

Full size image
Fig. 8
figure 8

a Map of the maximum tsunami wave amplitude computed for the regional tsunamigenic sources (listed in Table 1). b The maximum tsunami wave amplitude computed for the regional tsunamigenic sources at POIs (listed in Table S1)

Full size image

4.3 Local tsunamigenic earthquake sources

Two local offshore seismogenic sources were defined in the Northeast Adriatic region according to the DISS-3.3 database (Table 1 and Fig. 4; for further details, see DISS Working Group 2021). These earthquake sources were named the Southern and the Northern Trieste Gulf, as indicated in Table 1, and assigned magnitudes Mmax 6.5 and 6.9, respectively. These two sources are considered tsunamigenic because of their assigned location, magnitude and focal mechanism, and they might impact the AOI with strong ground motion and tsunami waves if they occur.

In the current study, the local seismogenic sources were considered in developing the tsunami hazard estimation. The spatiotemporal variation in fault rupture was considered here by segmenting the sources into sub-sources with rupture slip and velocity that varied along the scenario fault. The offshore source was segmented into 7-km-long sub-sources, with slip and rupture velocities increasing in a unilateral NW direction. This source shall be studied in detail in future developments, considering different rupture styles (i.e. unilateral, bilateral) and directivity angles, representing different potential scenarios in future ruptures. The computed maximum tsunami wave amplitude at POIs was up to 2.0 m (Fig. 9). Specifically, a maximum wave amplitude of 2.0 m is estimated for Trieste and Grado, while Monfalcone and Lignano have 1.0 and 0.4 m, respectively.

Fig. 9
figure 9

a The maximum tsunami wave amplitudes for local sources (i.e. ITCS100 and ITCS101) located at distances less than 100 km from the area of interest. b The maximum tsunami wave amplitude computed for the local tsunamigenic sources (i.e. ITCS100 and ITCS101) at POIs

Full size image

5 Basin-wide tsunami sources (out of the Adriatic Sea)

Based on the CAT-INGV’s tsunami matrix (Fig. 1), an offshore or inland, shallow earthquake (i.e. depth < 100 km) of magnitude larger than 7.5 has the potential to generate a destructive tsunami affecting the whole Mediterranean basin at any distance. Accordingly, tsunami scenarios were modelled for potential seismogenic zones of magnitude higher than 7.5 outside the Adriatic Sea, namely: West Greece, West Hellenic, Calabrian Arc, East Hellenic and Cyprian Arc zones. GEBCO 2020 bathymetry and topography data of about 400 m resolution were used for the entire basin (Fig. 10). The maximum wave amplitude was computed for the area of interest in the Northern Adriatic (Fig. 11), considering the seismic source parameters shown in Table 2. Fault slip parameters for tsunami scenario computation were collected from DISS-3.3 and individual studies (England et al. 2015; Hassan et al. 2023). For each seismogenic zone separately, several scenarios accounting for source uncertainties were computed, and tsunami wave amplitude maps developed as aggregated scenarios of maximum estimated values.

Fig. 10
figure 10

Basin-wide tsunami sources (red rectangles) within which individual scenarios are modelled, and the maps of maximum wave amplitude are computed and presented for the AOI (blue rectangle)

Full size image
Fig. 11
figure 11

Aggregated scenario from the individual scenarios (listed in Table 2) computed for the West Greece zone at basin-wide scale (top) and a zoom on the area of interest (bottom)

Full size image
Table 2 Parameters of basin-wide tsunamigenic sources in the Mediterranean Sea adopted in the computation of maps of the maximum tsunami wave amplitude
Full size table

The computed scenarios can be used, jointly with real-time seismic and sea-level tide measurements, to provide a quantitative basis for the decision-making process by emergency managers and first responders in the FVG region in case of future large tsunamigenic earthquake occurrence in one of the considered zones.

5.1 West Greece zone

The western coast of Greece lies along the Ionian Sea, which is relatively less prone to tsunamis compared to other areas of the Mediterranean Sea, such as the eastern coast of the Mediterranean. However, it is important to note that tsunamis can still occur in the Ionian Sea due to various factors such as earthquakes, underwater landslides and volcanic eruptions. The most significant event to hit the western coast of Greece in recent history was associated with the 1953 Ionian earthquake, which triggered a tsunami that caused substantial damage and loss of life in the region (Mavroulis and Lekkas 2021).

Five individual tsunami scenarios (Table 2) were modelled and an aggregated scenario of maximum wave amplitude was extracted (Fig. 11). The resulting maximum wave amplitude at the AOI ranges between 0.10 and 0.44 m as shown in the bottom panel of Fig. 11.

5.2 Calabrian Arc zone

The Calabrian Arc is considered a high-risk area for tsunamis due to seismic activity and underwater landslides. The area has experienced several significant tsunamis in the past, including the 1908 Messina earthquake and tsunami (e.g. Schambach et al. 2020), which caused widespread devastation and loss of life in the region.

Eight individual tsunami scenarios (Table 2) were modelled within the Calabrian Arc zone (Fig. 10), and an aggregated scenario was then produced (Fig. 12). The computed maximum nearshore wave amplitude, computed at the AOI, is about 0.2 m (Fig. 12, bottom panel).

Fig. 12
figure 12

Aggregated scenario from the individual scenarios (listed in Table 2) computed for the Calabrian Arc zone at basin-wide scale (top) and a zoom on the area of interest (bottom)

Full size image

5.3 West Hellenic Arc zone

The West Hellenic Arc is a seismically active area located in the Eastern Mediterranean Sea (Fig. 10). This arc is part of the larger Hellenic Arc, which is a curving chain of tectonic plate boundaries that extends from the western coast of Greece to the southern coast of Turkey. The Hellenic Arc results from the subduction of the African Plate beneath the Eurasian Plate and is associated with frequent seismic and volcanic activity. The most famous tsunami event was associated with the Mw8.5 Crete earthquake, an extremely powerful earthquake that occurred on July 21, 365 CE, near the island of Crete in the Eastern Mediterranean. This earthquake and the related tsunami were among the most devastating events in the past, causing widespread destruction across the Mediterranean region. The tsunami had a far-reaching impact, affecting coastal areas in Africa, the Adriatic, Greece and Sicily, as shown by Shaw et al. (2008) and Samaras et al. (2015).

Six individual tsunami scenarios adopted from England et al. (2015) and Hassan et al. (2023) (Table 2) were modelled to evaluate the impact on AOI of possible tsunamis originating within this source zone, which is capable of generating basin-wide tsunamis (Fig. 13). The computed nearshore maximum tsunami wave amplitude at AOI is quite large. It ranges between 0.15 and 0.5 m, which makes this source the most powerful compared to the other sources outside the Adriatic Sea (Fig. 13, bottom panel).

Fig. 13
figure 13

Aggregated scenario from the individual scenarios (listed in Table 2) computed for the West Hellenic zone at basin-wide scale (top) and a zoom on the area of interest (bottom)

Full size image

5.4 East Hellenic Arc zone

The East Hellenic Arc is another subduction zone in the Aegean Sea between Greece and Turkey. Like the West Hellenic Arc, this area is also seismically active and has experienced several significant earthquakes and tsunamis. While the impact of a tsunami from the East Hellenic Arc on AOI is not as high as from other areas, it is still fundamental for authorities and residents in the region to be aware of the potential related risks and take appropriate measures to mitigate them. To this end, an aggregated scenario was computed from three individual scenarios (Table 2) in the East Hellenic Zone (Fig. 14). The maximum computed wave amplitude at the AOI was 0.1–0.16 m (Fig. 14, bottom panel).

Fig. 14
figure 14

Aggregated scenario from the individual scenarios (listed in Table 2) computed for the East Hellenic Zone at basin-wide scale (top) and a zoom on the area of interest (bottom)

Full size image

5.5 Cyprian Arc zone

The Cyprian Arc is a subduction zone in the Eastern Mediterranean Sea, between Cyprus and the Turkish coast (Fig. 10). This area is also seismically active and has experienced several significant earthquakes and tsunamis.

While the Adriatic Sea is relatively far from the Cyprian Arc, pre-computed scenarios may still provide useful information for emergency managers in AOI. Overall, the possible impact of a tsunami with origin in the Cyprian Arc on the coast of the AOI proves very low and almost negligible (Fig. 15). The results in Fig. 15 indicate that the tsunami associated with the largest earthquake within this zone (Table 2) might be insignificant (i.e. 0.03 m) in terms of hazard to the AOI.

Fig. 15
figure 15

Aggregated scenario from the individual scenarios (listed in Table 2) computed for the Cyprian Arc Zone at a basin-wide scale (top) and a zoom on the area of interest (bottom)

Full size image

6 Conclusions and future developments

This work modelled a set of tsunami scenarios, considering all potential seismogenic sources capable of generating tsunamis according to the CAT-INGV’s decision matrix. To compute tsunami hazard maps (local, regional and basin-wide), tsunamigenic sources available in the recently updated DISS-3.3 were modelled, indicating a significant increase in the Mmax of most seismogenic sources. Tsunami maximum wave amplitude, which may enable developing inundation scenarios, was computed over a grid and for the selected POIs, including Trieste, Monfalcone, Lignano and Grado.

Adriatic-wide and regional tsunamis can produce tsunami waves with amplitudes as large as 0.7 and 1.1 m, respectively. Local tsunamis are capable of producing tsunami wave amplitude between 2.0 and 2.2 m; their highest impact could be on the coastal zone of Trieste and Grado, while it could reach 1.2 m for Monfalcone and about 0.7 m for Lignano, as indicated in Fig. 9 and Tables 3 and 4. The reader should bear in mind that the tables presented here provide just a summary representation of the modelling results. For further details and a comprehensive description of the results, it is necessary to consider the full set of output files (in kml, grid, and dat formats), which contain complete numerical information from the simulations. Detailed analysis of these local sources, as well as the development of more detailed scenarios for the selected AOI, especially in lagoon areas, will be the main focus of future developments.

Table 3 Summary of estimated maximum tsunami wave amplitude (range) at the cities of interest due to tsunami earthquake sources located at the local, regional and Adriatic scales
Full size table
Table 4 List of estimated maximum tsunami wave amplitude and estimated arrival times (ETA) for the cities of interest associated with the DISS tsunamigenic sources located at the local, regional and Adriatic scales
Full size table

Basin-wide tsunamis from West Greece, Calabrian Arc and West Hellenic Arc should be taken into account because tsunamigenic earthquakes of Mmax > 7.5 prove capable of impacting the AOI at orange alert level (amplitude less than 0.5 m) and even at red alert level (up to 0.8 m for a magnitude M = 8.5 earthquake in the West Hellenic Arc), as inferred from the modelling. Instead, the scenarios computed for the East Hellenic and Cyprian Arcs provide amplitudes corresponding to the information level.

The modelling results obtained so far provide a primary physics-based description of the tsunami hazard along Northeastern Adriatic coasts, associated with possible tsunamis generated by offshore earthquakes at different epicentral distances. Though the obtained information supplies some quantitative basis in support of the decision-making process, which could be considered if a warning message is issued by the CAT-INGV tsunami alert system, several aspects still need further investigation. To achieve an improved (i.e. more detailed, complete and robust) tsunami hazard assessment, the envisaged future developments include:

  • In-depth revision of the high-resolution bathymetry dataset to improve the match with observations and to remove possible artefacts (e.g. closed openings and/or channels) which could impact the propagation of tsunami waves within lakes or lagoons behind the shoreline.

  • Development of inundation maps by modelling ingression of tsunami waves based upon very high-resolution bathymetry and topography data.

  • Sensitivity analysis with respect to Mmax (including the maximum credible earthquake) to account for the uncertainty on the maximum magnitude associated with the tsunamigenic sources, demonstrated by the recent revision of the DISS database.

  • Detailed investigation of local tsunamigenic sources, considering different earthquake rupture styles (i.e. unilateral, bilateral), variability and directivity angles, which represent different potential scenarios that may occur in future ruptures.

  • Assessment of the hazard from near-offshore tsunamigenic sources, according to guidelines in the decision matrix, taking advantage from earlier studies in the region (e.g. Paulatto et al. 2007) and theoretical considerations (Yanovskaya et al. 2003).