The Mediterranean region is highly exposed to tsunami risk, as witnessed by several recent events (Yalçıner et al., 2017; Cirella et al., 2020; Dogan et al., 2021; Kalligeris et al., 2022), basin-wide or local historical events (Solov'ev et al., 2000; Mauelshagen, 2007; Maramai et al., 2014; Papadopoulos et al., 2014; Maramai et al., 2021), and a recent assessment of seismically induced tsunami hazard (Sørensen et al., 2012; Basili et al., 2021a). However, the tsunami risk in the Mediterranean is thought to be underrated, due to the low frequency of events, as in other regions of the world (Dawson et al., 2004; Dogulu et al., 2014; UNESCO-IOC, 2017; Amato, 2020; Necmioğlu et al., 2021). It is therefore important to raise awareness among people, as well as of that local authorities responsible for civil protection measures and emergency management and national/regional authorities.

The Sendai Framework for Disaster Risk Reduction 2015–2030 gives high attention to early warning systems (EWSs), suggesting to “invest in, develop, maintain and strengthen people-centered multi-hazard, multisectoral forecasting and early warning systems …; develop such systems through a participatory process; [and] tailor them to the needs of users, including social and cultural requirements, in particular gender”. The strong emphasis on people means that any communication strategy should be based on a preliminary assessment of people's knowledge, awareness, and perception of the risk. Moreover, the “needs of users” must be studied and understood to reach people and communities in the right way using the best language and communication channels and to have an optimal response in the case of an impending inundation.

The Italian Tsunami Alert Centre (Centro Allerta Tsunami, CAT) of the Italian National Institute of Geophysics and Volcanology (INGV) is part of the national tsunami warning system called SiAM (Italian national warning system for tsunami of seismic origin), established in 2017 with a prime minister directive (G.U. n.128 del 05-06-2017). SiAM is coordinated by the national Civil Protection Department (DPC) and, besides CAT-INGV, includes the Italian Institute for Environmental Protection and Research (ISPRA, Istituto Superiore per la Protezione e la Ricerca Ambientale), which manages the national sea level network and contributes to mapping the tsunami inundation zones. CAT acts as the tsunami service provider (TSP) for UNESCO member states of the North-Eastern Atlantic, Mediterranean and connected seas (NEAM) region (Amato et al., 2021), as well as the national tsunami warning center (NTWC) and tsunami warning focal point (TWFP) for Italy. Among the tasks of CAT, besides the tsunami surveillance/warning and the hazard assessment, there are also scientific and risk communications activities on tsunamis. CAT manages a dedicated website (http://www.ingv.it/cat/en/, last access: 18 July 2022), where people can find information on tsunami hazard/risk, on the warning system, and on historical events, as well as news on projects, papers, campaigns, and drills.

From this perspective, the activities of CAT include assessing people's perception of tsunami risk. This research aims to improve risk analysis and decision-making, develop methods for eliciting opinions about risk, and provide a basis for understanding and anticipating possible public reactions to tsunami hazards, enhancing risk communication among lay people, technical experts, and policy makers (Slovic et al., 1982; Slovic, 1987; Wildavsky and Dake, 1990; Slovic, 2001; Rippl, 2002).

This paper presents data on tsunami risk perception collected in central and southern Italy between 2018 and 2021. The study involved the administration of a semi-structured questionnaire to a sample of 5842 people in 450 coastal municipalities.

Risk perception studies taking into account socio-cultural and psychological aspects assess people's response to natural hazards and the behaviors they would adopt in response to risks. People's perceptions – individually or collectively – of a natural hazard are also influenced by individual factors such as personality, age, beliefs, gender, education level, knowledge, and culture (Slovic et al., 1982; Slovic and Peters, 2006; McIvor et al., 2009; McNeill et al., 2013; Wachinger et al., 2013).

Human behavior is driven by perceptions (Slovic, 1987) rather than scientific knowledge about “facts” (Renn, 1990). Therefore, it becomes strategic for those involved in risk mitigation and communication to have in-depth studies on the process that influences our ability to assess the risk of a natural phenomenon (Slovic, 1982) like tsunamis. Tsunamis are known to be a phenomenon with a low probability of occurrence but high impact, able to produce devastating consequences that can affect large areas and have serious consequences for human lives (Behrens et al., 2021; Rafliana et al., 2022). Their low frequency of occurrence is one of the variables that directly affects risk perception. In some Mediterranean areas (such as Italian coastal regions), tsunamis seem to be out of the collective consciousness, leading authorities to underestimate the significance of tsunami risk perception among the population and not take effective action to reduce the risk.

However, tsunamis' low frequency of occurrence does not reduce their destructive potential. Moreover, how important it is to study people's perceptions of natural hazards (Lindell and Perry, 2000; Paton et al., 2010; Wachinger et al., 2013; Bonaiuto et al., 2016), particularly tsunami risk perceptions, has emerged in various studies conducted in countries that have been affected by tsunamis, such as the 2004 Indian Ocean tsunami or the 2011 Tōhoku, Japan, tsunami (Kurita et al., 2007; Sugimoto et al., 2010; Alam, 2016; Arias et al., 2017; Akbar et al., 2020).

A historical catalog of observed tsunami effects in the Mediterranean (Maramai et al., 2014) cites over 200 documented events for the whole area, 90 % of which were caused by earthquakes. More recently, in the Euro-Mediterranean area, 31 earthquakes above magnitude 5.5 occurred at sea or near the coast between 2017 and October 2022, triggering the activation of CAT. Among them, 10 earthquakes generated an alert level for possible sea level change, including 6 advisory (possible sea level change with estimated run-up less than 1 m) and 4 watch (possible sea level change with estimated run-up values above 1 m) alerts. Two of them caused damage in Greece and Turkey (Yalçıner et al., 2017; Dogan et al., 2019, 2021; Triantafyllou et al., 2021; Kalligeris et al., 2022).

In addition, variables to be considered from a comprehensive perspective include the large growth in population living along the Euro-Mediterranean coasts. This phenomenon, which has occurred since the Second World War and has intensified in recent decades, also includes the development of tourist facilities and large industrial complexes. These intensive forms of settlements require multi-risk analytical hazard approaches, where the socio-cultural and psycho-social aspect becomes prominent. In this framework, strengthening tsunami risk perception studies is needed, surveying the opinions of employees; daily commuters; seasonal workers; and tourists, whose presence dramatically increases the risk during holiday seasons. The need to assess tsunami risk perception has been highlighted by several authors as a key to improving emergency behaviors and minimizing population risk by limiting casualties and infrastructure damage (Ho et al., 2008; Martin et al., 2009; Ritchie and Roser, 2014; EMDAT, 2019).

Since 2018 CAT-INGV has been promoting tsunami risk perception studies to provide oriented support to civil protection activities and develop data-driven, context-appropriate risk communication strategies. We adopted CATI (computer-assisted telephone interviewing) questionnaire structuring and administration methodology since it proved to be an excellent tool to collect a large, standardized, retraceable, and cost-effective number of data (Dawson et al., 2004; Cerase et al., 2019). Furthermore, the questionnaire as a survey method to study tsunami risk perception is widely used in the international context (see for example Apatu et al., 2016; Sun et al., 2013; Lindell et al., 2015, 2016; Jon et al., 2016; Fraser et al., 2016; Wei et al., 2017; Buylova et al., 2020).

The principal results of the three surveys are presented in two subsections: “Tsunami risk perception” and “Tsunami knowledge”.

The map in Fig. 2 shows the geographical positions of the coastal areas where the survey was conducted (by region) and the seas surrounding Italy.

The three surveys on tsunami risk perception, conducted between 2018 and 2021, started from the necessity to study and understand the level of knowledge of tsunami risk and the awareness of Italian citizens living in or visiting coastal areas exposed to tsunami hazard.

To date, these surveys represent a relevant sample of the Italian population, both for the number of interviewees and for the adopted methodology (5842 CATI interviews carried out on over 6000 km of coastline with moderate to very high tsunami hazard, plus a nationwide “Telepanel” sample representative of the whole Italian population).

The main results show that the tsunami risk perception varies significantly according to the coastal region. In particular, regions in the Adriatic (Apulia, Molise) show very low levels of risk perception compared to Calabria, Sicily, and Campania, on both the Ionian and the Tyrrhenian seaside. Latium and Sardinia lie in the middle, with equal numbers of people thinking that a tsunami could hit their region.

The educational degree affects tsunami risk perception: the higher the educational degree, the higher the tsunami risk perception.

On the contrary, data analyses show that tsunami risk perception is not influenced by either gender or age. A slight difference is observed in the middle age group in which interviewees of 35–49 years appear to have a slightly higher perception. Elderly people show the lowest frequency percentage in the response modality “I don't know” associated with low tsunami risk perception.

These data are consistent with several studies on local cultures, showing how communities that have previously experienced disasters are more likely to develop better resilience to and preparedness (Dekens, 2007) for these calamitous events. Such knowledge, which differs from scientific knowledge (Flavier et al., 1995), is associated with the historical memory of past experiences learned and transmitted through rituals, traditions, narratives, and folk songs (e.g., smong songs on the island of Simeulue) (McAdoo et al., 2006; Rahman et al., 2017, 2018; Sutton et al., 2021) and defined in a different way to “local knowledge”, “traditional knowledge”, “indigenous technical knowledge”, “peasant knowledge”, “traditional environmental knowledge”, and “folk knowledge” (Sillitoe, 1998; Mercer et al., 2007, 2010).

Memories of previous disasters both inform people's knowledge of their environment and vulnerability and influence their risk interpretation and response to future disasters (Arunotai, 2008). Collective memory, relying on oral tradition shared by a specific group, most commonly the family, tends to disappear with the death of the last eyewitness to the event (the three-generation limit). Cultural memory, supported by documents (such as newspapers, archives, images) and memorials as tangible signs for the community, ensures that disaster meanings and interpretations are recorded and transmitted from generation to generation. These forms of cultural memory, considered by Mercer et al. (2010) an existing or acquired knowledge set by local communities, are born and maintained through the accumulation of experiences, social relations, and community practices and institutions and their transmission across generations.

The study of Garnier and Lahournat (2022) highlights how Japanese stone monuments, representing elements of both tangible and intangible culture for the population, demonstrate the existence of disaster memory and reflect a desire to commemorate and transmit significant past events to current and future generations. These findings highlight the importance of oral transmission between generations regarding tsunami risk and could be very useful for designing effective information and communications activities about tsunami risk reduction (Spahn et al., 2010; Løvholt et al., 2014; Oktari et al., 2018; Sutton et al., 2020).

In this work, we also carried out a comparative analysis on tsunami risk perception between metropolitan areas and their respective coasts (see Sect. 4.1.1), starting from the hypothesis that risk perception could be different among the population residing in a large city and the population of small municipalities distributed along the seaside. The territorial units analyzed were selected based on exposure (higher coastal population density or proximity of major urban settlements to the coast) and territorial vulnerability (high concentration of anthropogenic activities such as industries or intensive tourism activities) according to ISTAT (Italian National Institute of Statistics) data.

The goal of this comparison is both to highlight possible differences in perceptions associated with densely populated urban areas and to provide a solid basis for targeted risk mitigation actions in specific contexts (for example to improve and better address the UNESCO Tsunami Ready program; Valbonesi et al., 2019, 2021; Valbonesi, 2022).

The results indicate that in Reggio Calabria, Catania, and Naples there are indeed significant differences in tsunami risk perception compared to in the respective seaside areas. These data are most evident in Reggio Calabria (71 %) with a difference of over 25 percentage points compared to the Tyrrhenian (45 %) and Ionian (46 %) sides of the same region (Calabria).

The data continue to be significant, albeit less evidently, for Catania (50 %) and the Ionian seaside (46 %) and for the city of Naples (47 %) and the Tyrrhenian seaside (45 %). It is useful to underline that the metropolitan city of Bari, despite being on a stretch of coast considered to be at high risk, has significantly lower values of perception of the tsunami risk (34 %), in line with the data for the whole Adriatic coast (32 %). Tsunami risk perception in the metropolitan area of Rome seems to be lower than on the Tyrrhenian coast. This is probably because Rome cannot really be considered a coastal town because most of its over 3 million inhabitants live far from the sea. The presence within the metropolitan area of Rome of the town of Ostia, a very populous municipality (over 231 000 inhabitants) lying along the Tyrrhenian coast, is not enough to raise the overall risk perception of the capital. It would be interesting to deepen this aspect with specific analysis of the local residents, commuters, and visitors. Given the city's proximity to the coast (about 20 km), it is also worth considering the large daily (as well as seasonal) tourist flows of the summer period.

The Telepanel data (n = 1500), collected in the same time period as the third survey, deserve particular attention because they provide a representative sample of the population at the national level, which means people predominantly living far from the seaside. Comparing the tsunami risk perception of the Telepanel with the surveys carried out on the coastal population, it emerges that the coastal population on average has a perception of risk significantly higher than the national average of about 20 percentage points. These data suggest that it is necessary to pay particular attention to the coastal touristic areas where every year millions of people spend their holidays. Providing information panels along the beaches and in the harbors, indicating evacuation routes and meeting points and establishing redundant alarm systems are the main tools to be put in place to reduce the tsunami risk of both residents and tourists.

As for knowledge of tsunamis, we started by considering the different associations linked to the two terms that are commonly used in the Italian language: tsunami and maremoto.

Results indicate that most of the sample is more familiar with the term tsunami (57 %), particularly those with a higher level of education (66 %), and the term is less familiar to people over 65 years old.

However, there are some differences that appear to be linked to local characteristics and which will be the subject of further investigation.

It is interesting to notice that the two terms are associated with different phenomena. For instance, the term tsunami is mostly related to the occurrence of a great destructive wave, while the word maremoto is more often related to the occurrence of an earthquake.

As for knowledge of the causes that generate tsunamis, in general the sample correctly attributes the occurrence of a tsunami to earthquakes and secondarily to volcanic eruptions, in line with scientific knowledge. Furthermore, tsunami knowledge does not seem to be linked to the gender and the areas of residence of the interviewees but increases according to their educational degree and inversely with age (people with a higher education degree and who are younger than 50 years old appear to be more informed).

Our data indicate that in general people are aware of the possible effects of tsunamis on their regions. However, this knowledge appears to be closely influenced by the media representation of the great tsunamis that occurred in Tōhoku, Japan (2011), and Sumatra–Andaman (2004). This type of media representation could turn out to be misleading with respect to the more modest phenomena that can be generated by smaller tsunamis, which are more frequently expected in the Italian coastal territory but are also capable of causing serious damage and creating victims.

However, the probability of a tsunami of significant size – such as the one that occurred in 1908 in southern Italy – is not entirely negligible.

Moreover, our data indicate that television continues to play a central role in conveying information relating to tsunami risk, while the information role played by social media and the web appears to still be marginal. A particularly problematic aspect concerns the poor visibility of scientific institutions and civil protection and local authorities in being recognized by the interviewees as official sources of information on the tsunami risk.

Our results suggest that the loss of memory of past events affects the perception of risk by citizens and communities, making risk mitigation actions more difficult and even ineffective (see for example Kurita et al., 2007; Sugimoto et al., 2010; Arias et al., 2017; Wei et al., 2017). Memory and recollections (such as commemorations of past events) are relevant for the development of risk mitigation strategies and to increase population resilience. More specifically, awareness raising aimed at attributing a sense and a meaning to the memory is needed in order to reduce risk. We believe that the results of this study, although limited to central and southern Italy, can be used in other countries of the NEAM region and worldwide to orient communication strategies and risk reduction actions.

In the near future, given the large number of data collected in the three surveys, we will focus on the creation of synthetic indexes for the perception and knowledge of tsunami risk.

We are also working to deepen knowledge related to cultural differences in the perception of risk, which seem to be very influenced by the local culture of reference. We are confident that this will allow us to better explain the differences in perception and knowledge that the data show in the different sea areas (coastal sides) and metropolitan cities. Finally, our efforts will be focused on translating the results of risk perception analyses into effective communication strategies for tsunami risk reduction. Not least, the 2018 pilot survey results – published in a previous paper by Cerase et al. (2019) – have already been extensively used in developing and improving the CAT website content. Moreover, several relevant aspects of the study made it possible to better address approaches in risk awareness campaigns such as “Io Non Rischio” (a civil protection campaign) as well as to undertake dissemination campaigns aimed both at raising awareness and at surveying tsunami risk perception in schools (one of the Tsunami Ready program indicators).