A Logic-Tree Based Probabilistic Seismic Hazard Assessment for the Central Ionian Islands of Cephalonia and Ithaca (Western Greece)

A Logic-Tree Based Probabilistic Seismic Hazard Assessment for the Central Ionian Islands of Cephalonia and Ithaca (Western Greece) | 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 A Logic-Tree Based Probabilistic Seismic Hazard Assessment for the Central Ionian Islands of Cephalonia and Ithaca (Western Greece) George Kaviris, Angelos Zymvragakis, Vasilis Kapetanidis, Vasiliki Kouskouna, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3991269/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Sep, 2024 Read the published version in Journal of Seismology → Version 1 posted 7 You are reading this latest preprint version Abstract The Central Ionian Islands of Cephalonia and Ithaca belong to the most seismically active Greek region, mainly due to the presence of the dextral Cephalonia-Lefkada Transform Fault Zone. The study area has experienced strong earthquakes in the 20 th century, including the destructive 1953 sequence with maximum intensity 9.0. The Paliki peninsula, western Cephalonia, hosted two strong earthquakes (M w = 6.1 and 5.8) in 2014, with ground acceleration reaching ~560 cm/s 2 and 735 cm/s 2 , respectively. This study updates the seismic hazard evaluation in Cephalonia and Ithaca using new data and computational techniques to reduce epistemic uncertainties. The probabilistic approach of Cornell and McGuire was used, and the uncertainties are reduced through data variability of the source models, seismicity data, and Ground Motion Prediction Equations using a logic tree approach, sampled by implementing the Latin Hypercube Sampling method. The spatial distribution of Peak Ground Acceleration and Peak Ground Velocity for return periods of 475 and 950 years indicates low variation in the entire study area and that the Paliki peninsula possesses the highest level of seismic hazard. Additionally, site-specific analysis across the three main towns, Lixouri and Argostoli in Cephalonia and Vathi in Ithaca, reveals that Lixouri has the greatest level of seismic hazard, while Vathi the lowest. PSHA PGA PGV UHS Sa Logic tree Epistemic uncertainty Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction The Central Ionian Islands (Figure 1) stand out as the most seismically hazardous area in Greece (EAK, 2003). Cephalonia Island, in particular, has hosted two M w ≥6.7 earthquakes (Makropoulos et al., 2012) within a 30-year span. The heightened seismic activity in this area is mainly attributed to the presence of the Cephalonia-Lefkada Transform Fault Zone (CLTFZ), a dextral plate boundary trending SSW-NNE. This fault zone bounds the western coasts of Cephalonia and Lefkada Islands and intersects with a complex network of onshore active faults, many of which are delineated in Figure 1. The CLTFZ, roughly outlined by the mapped earthquake epicenters of Figure 1, serves as a link between the NW-SE-trending major features of the Hellenic Arc in the south, and the collision front of the Apulian platform with the Hellenic foreland in the north. This configuration creates a shear zone of intense deformation, leading to major earthquakes that have caused substantial damage and loss of life in the recent past, particularly on Cephalonia and Ithaca Islands. The CLTFZ exhibits a strike of ~N15°E on the latitude of Lefkada Island, while further to the south, its trend shifts to ~N30°E as it passes offshore Cephalonia (Figure 1). The two slightly offset segments of the CLTFZ interact within Myrtos Gulf at northern Cephalonia, likely forming a transfer zone of extensional step-overs (Karakostas et al., 2015). The regional crustal stress field promotes strike-slip faulting (Kapetanidis and Kassaras, 2019), but the convergence of the African plate with the Aegean microplate introduces compression in a WSW-ENE direction. Differential GPS measurements have revealed a clockwise rotation of Cephalonia Island, relative to a station located on Aenos mountain, with the largest values observed at the western (Paliki) and northern (Erissos) parts of the island (Lagios et al., 2007). The broader area of Cephalonia features east-dipping NW- to NNW-striking thrust structures (Stiros et al., 1994), which may sporadically undergo seismic reactivation, resulting in earthquakes with a significant reverse dip-slip component. The Ionian Thrust traverses the southeastern part of Cephalonia Island and likely extends offshore, passing between Cephalonia and Ithaca (Underhill, 1989). Additional thrust structures are observed at Aenos Mountain and Argostoli, extending northward and separating Paliki from the rest of Cephalonia. Cephalonia has a historical record of significant earthquakes. The SHEEC catalog (Stucchi et al., 2013) and the AHEAD database (Albini et al., 2013) contain 14 historical earthquakes occurring between 1469 and 1867 with macroseismic epicenters near Lixouri and Argostoli. In the instrumental era, a significant M w =6.1 earthquake occurred on 12 January 1912, south of Argostoli (Makropoulos et al., 2012; Figure 1), killing 8 people and injuring 40 in Poros (Papazachos and Papazachou, 2003). Its focal mechanism, estimated from the directivity of macroseismic data, indicates strike-slip faulting (Papazachos et al., 1999). Another major earthquake of M w =6.1 occurred on 27 January 1915 near Ithaca (Makropoulos et al., 2012; Figure 1). It also exhibited an estimated dextral SW-NE strike-slip focal mechanism (Papazachos et al., 1999) and was responsible for the collapse or extensive damage of many houses (Papazachos and Papazachou, 2003). On 9 and 11 August 1953, two earthquakes of M w =5.9 and 6.6, respectively (Makropoulos et al., 2012), occurred east of Cephalonia Island, the latter with an epicenter on Ithaca Island. They were followed on 12 August by an even larger earthquake of M w =7.0 (Makropoulos et al., 2012) at the southeastern part of Cephalonia Island, with a location error estimate of ~50 km (Anderson and Jackson, 1987). Its focal mechanism, determined from first motion polarities, indicates reverse faulting in a NNW-SSE direction (McKenzie, 1972), although estimates from macroseismic data suggest strike-slip faulting (Papazachos et al., 1999). In both instances, the P-axis aligns with the direction of maximum horizontal compression (Kapetanidis and Kassaras, 2019), as well as with the direction of shortening (N258°E) in the area of the Central Ionian Sea, maintaining an extension-to-shortening ratio of 1:3 (Ganas et al., 2013b). The maximum intensity value of the 1953 earthquakes at Cephalonia Island was I max =9/10, observed at five localities, among which Argostoli and Lixouri (Sakkas et al., 2010), however estimated even higher due to cumulative damage. The deformation resulting from the 12 August 1953 event is evident in the observed coastal uplift near Poros, at the southeastern part of Cephalonia Island, as manifested through notches on the eroded rocks. Mushroom-shaped formations in uplifted rocks about 50 m offshore also suggest that a similar paleoseismic event must have occurred (Stiros et al., 1994). The coastal uplift observations support a piston-like motion on two subparallel, east-dipping reverse faults (Stiros et al., 1994). The 1953 earthquake sequence also revealed the necessity for the implementation of a National Building Code, to mitigate the risk from seismic hazards in Greece. Another significant earthquake was an M w =6.7 event (Makropoulos et al., 2012) that occurred on 17 January 1983, approx. 15 km southwest of Paliki. Due to its offshore epicenter, it reportedly caused minor damage to Cephalonia Island (EMS intensity IV at Argostoli), whereas its strongest aftershock, on 23 March 1983, was more damaging (max intensity VII), as its epicenter was located further north, closer to the island (Papazachos and Papazachou, 2003). The mainshock has been interpreted as a strike-slip rupture on a southeast-dipping fault with a relatively low angle (Papadimitriou, 1988), likely with a thrust component (Scordilis et al., 1985). A similar faulting type, i.e. dextral strike-slip on an east-dipping low-angle fault, was determined for a recent major earthquake (M w =6.7) that occurred southwest of Zakynthos Island on 25 October 2018 (Papadimitriou et al., 2021). The latter is located in a transition zone between the southern end of the CLTFZ and the northwestern edge of the Hellenic Arc. Pure reverse faulting, consistent with SW-NE compression, occurs further south. The most recent major seismic activity on Cephalonia Island comprised of an earthquake “doublet” that occurred on 26 January and 3 February 2014, with M w =6.1 and 5.8, respectively, on Paliki (Papadimitriou et al., 2014; Karakostas et al., 2015; Karastathis et al., 2015; Sokos et al., 2015; Sakkas et al., 2022). Moment tensor inversions for both events support dextral strike-slip faulting on SSW-NNE-trending, steep-dipping faults, with most solutions indicating a slight tilt toward an east-dipping direction. Fault plane models, constructed using the observed co-seismic deformation, reveal that the two earthquakes occurred on two sub-parallel fault segments, in the southern and the northern parts of the Paliki peninsula, respectively (Sakkas and Lagios, 2015). It seems likely that the 2014 earthquake “doublet” may have accelerated the occurrence of a major earthquake (M w =6.3) on 17 November 2015 (Papadimitriou et al., 2017) on Lefkada (Figure 1); an area that was already stress-loaded after the M w =6.3 earthquake of 14 August 2003 (Papadimitriou et al., 2006). The 2014 Cephalonia earthquakes produced a rich aftershock sequence, with the seismicity rate remaining elevated for about 4 years (Sakkas et al., 2022). The seismic activity was further enhanced by the occurrence of the 2015 Lefkada earthquake, which led to a slight increase in the seismicity rate in Cephalonia. This was particularly observed in the area of Myrtos Gulf, where small structures, likely trending E-W, being antithetic and transverse to the CLTFZ, appear to be easily triggered by major earthquakes in the vicinity (Sakkas et al., 2022). The 1953 earthquake sequence on Cephalonia is among the most significant ones to have occurred in Greece during the instrumental period, leaving a profound societal imprint. The devastation was extended on the islands of Cephalonia, Ithaca, and Zakynthos, resulting in the destruction of ~83% of the building stock. The human toll was significant, with 455 fatalities, 21 individuals reported missing and 2412 people sustaining injuries (Papazachos and Papazachou, 2003). It was also the cause of a decline in the population of Cephalonia, Ithaca, and Zakynthos during the following decades, as people abandoned the islands and migrated mainly to urban centers of Greece or abroad (Mavroulis and Lekkas, 2021). Seismic hazard is dedicated to investigating the phenomena triggered by earthquakes, with ground motion being the most significant as it acts as catalyst for potential secondary catastrophic events like rockfalls and liquefaction (Wang, 2005). To assess seismic hazard, we quantify ground motion by estimating Intensity Measurement Types (IMT), such as Peak Ground Acceleration (PGA), Peak Ground Velocity (PGV) and Spectral acceleration (Sa), through statistical methods (Gumbel and Lieblein, 1954; Cornell, 1968). The most common approach to compute the aforesaid parameters is the classic probabilistic method, initially proposed by Cornell (1968) and later commercialized by McGuire (1976), that introduces the usage of a seismotectonic model. Given the high seismicity of Greece, many seismic hazard assessment studies have been conducted, dating as far back as 1985 (Makropoulos and Burton, 1985) and continuing to the present (Bonatis, 2020; Bonatis et al., 2021; Pavlou et al., 2021; Kaviris et al., 2022a, 2022b, 2023). The Ionian Islands are widely recognized as an area of high seismic activity and are among the most seismically active regions in Europe. Therefore, a constantly updated with recent data seismic hazard assessment is crucial for urban planning to mitigate seismic risk. Earthquakes in the region have produced high PGA values, for example, the two 2014 Cephalonia events produced ground acceleration of about 560 cm/s 2 and 735 cm/s 2 , respectively (Kassaras et al., 2017). These values well exceed the regulations proposed by the 2003 National Building Code (EAK, 2003), highlighting the need for an updated seismic hazard model. It is worth noting that previous studies have been conducted for Cephalonia and Ithaca, as evidenced by the work of cited researchers (Bonatis, 2020; Bonatis et al., 2021; Sakkas et al., 2022). Nevertheless, in line with Cornell's insights, we investigate the possibility of reducing the considerable uncertainties regarding seismogenic sources and seismic wave attenuation. In this work, we assess seismic hazard by estimating PGA and PGV for Cephalonia and Ithaca, and Sa exclusively for the most populated localities of these islands, i.e. Lixouri, Argostoli, and Vathi. To mitigate uncertainties, we have developed an extensive and “non–trivial” logic tree decision diagram for estimating IMT. In addition, we adopt a stochastic statistical sampling method to capture a substantial portion of the Probability Density Function (PDF) for each IMT. By executing these techniques, the goal of this study is to effectively reduce the prevailing uncertainties and establish an accurate seismic hazard model that will have the potential to be applied in more precise seismic risk assessments for this highly seismically active study area. 2. Methodology The PSHA results of the approach introduced by Cornell (1968) and McGuire (1976) are in the form of annual probabilities of surpassing specific random values of acceleration, velocity, or spectral acceleration. Subsequently, the final output consists of PGA, PGV, or Sa levels corresponding to the selected return period. A pioneering aspect of this methodology, at the time of its inception, was the incorporation of a source model to depict the seismotectonic attributes of each study region. These models are classified according to the knowledge of the existing active faults within a region. For instance, a model may include the actual fault sources of the study area and report the annual exceedance rates for each magnitude bin per fault. This requires a good awareness of the dynamic characteristics of each fault. On the other hand, a source area model takes the form of polygons where seismicity is distributed, following a normal distribution within each polygon, and all attributes are consistent. This model type is typically employed in regions where the occurrence rates of all potential active faults are not well documented or are unmapped. This is partly the situation in our study area, Cephalonia and Ithaca, where the majority of active faults are offshore, and their dynamic characteristics are not known in detail. Consequently, the source area type of model was adopted for this study. The use of a single model may introduce high epistemic uncertainties regarding the seismological attributes of the broader study area. A conventional strategy for addressing epistemic uncertainties in PSHA involves introducing variability in the input data (Bommer and Scherbaum, 2008; Atkinson et al., 2014; Marzocchi et al., 2015; Kaviris et al., 2023). To address this concern, we integrated three models into our computational framework, i.e., the European Seismic Hazard Model 2013 (ESHM13) by Woessner et al. (2015), its subsequent update, ESHM20, developed by Danciu et al. (2021), and the local Ionian–Island source area model formulated by Bonatis (2020), herein referred to as BON20. Visual representations of these models are available in Figures 2a–c. ESHM13, ESHM20, and BON20 were treated as equivalent in our analysis, as each one exhibits distinct advantages and disadvantages in comparison to the others, thus preventing in designating one as superior. This decision was substantiated by examining the surface area of each source zone within these models. Small polygons, like the ones of ESHM13 and BON20 in the Central Ionian Islands (Figures 2a and 2c), provide a more precise description of the seismotectonic features within the area, whereas larger polygons, as those of ESHM20 (Figure 2b), are less accurate in describing these features, but with a higher number of earthquakes for statistical analysis. The latter provides insight into comparing the seismicity parameters obtained for each source model. BON20 (Figure 2c) has smaller zones providing a more accurate description of the seismotectonics of the Ionian Islands, when compared to the other two models. However, the required seismicity data of the small source areas around Cephalonia and Ithaca will result in a complex and non-smooth spatial distribution of the computed PGA and PGV values, which may hinder drawing conclusions about the seismic hazard of the study area. Consequently, implementing both local and non-local models is the best option to address this problem. The selection of an earthquake catalogue is a crucial step, as it is the basis for extracting events in each source area and conducting subsequent statistical analyses to derive essential seismicity parameters. Therefore, it is imperative to rely on a trustworthy and validated data source. In Greece, the instrumental earthquake catalogue of Makropoulos et al. (2012) is specifically tailored for seismic hazard studies, as it has a threshold magnitude of M s = 4.0 and M w = 4.1. It covers the period from 1900 to 2009, with the significant limitation of not including earthquakes since 2010. This lack of data is noteworthy because, as mentioned in the introduction, events such as those that occurred in 2014 and 2015 would be omitted from the computational framework. To address this issue, we extended the earthquake catalogue to 2019 in a consistent manner, following the same methodology as Makropoulos et al. (2012), i.e. incorporating reviewed events from the Bulletin of the International Seismological Centre (ISC). A common practice in PSHA is to decluster earthquake catalogues in order to retain only the parent earthquakes, removing foreshocks and aftershocks. However, this practice is debatable, as there have been reports suggesting that aftershocks can generate significant ground motions and are often responsible for additional damage (Marzocchi and Taroni, 2014; Taroni and Akinci, 2021). In this study, we opted not to decluster the earthquake catalogue, as this would result in loss of valuable data, especially considering that Cephalonia and Ithaca are regions characterized by frequent earthquake sequences. The seismicity parameters of each source area required for PSHA encompass the magnitude of completeness (M c ), the annual rate of Mc exceedance (λM c ), the maximum expected magnitude (M u ), and the b–value of the Gutenberg–Richter Frequency–Magnitude Distribution (FMD) (Gutenberg and Richter, 1944). We herein employ two methods for calculating M c and the b–value. The first is the classical maximum curvature method (MAXC), which was originally introduced by Wiemer and Wyss (2000) and identifies M c by pinpointing the maximum curvature of the FMD curve through the calculation of its first derivative's highest value. Subsequently, the b–value is determined using the maximum likelihood method introduced by Aki (1965). Overall, MAXC is highly reliable and robust and has undergone validation in several PSHA studies (Zhou et al., 2018; Pavlou et al., 2021; Kaviris et al., 2023). The second method, proposed by Godano and Petrillo (2023), offers a rapid and straightforward estimation of M c . It relies on the observation that the Gutenberg–Richter distribution exhibits an exponential behavior for magnitudes exceeding M c and a linear behavior for the smaller ones. Consequently, the average magnitude value (M a ) exhibits linear increase as the threshold magnitude (M th ) increases and the deviations from this behavior allow for an accurate M c computation, while the linearity of M a versus M th facilitates the b–value estimation. In this study, we seize the opportunity to also explore this new method for M c and b–value estimation, aiming to comprehend its advantages and limitations for future assessments. Variability was also considered for the M u parameter, because of the absence of a definitive estimation technique. Specifically, three techniques are utilized to estimate M u : the first, yielding the lowest possible M u , is based on the maximum observed earthquake magnitude (M maxobs ) within each source area. An intermediate M u value is estimated through the Robson–Whitlock–Cooke (RWC) technique, as described by Robson and Whitlock (1964) and Cooke (1979). RWC introduces a small positive factor based on the second maximum earthquake magnitude (M max n–1 ). The highest potential M u is determined by adding the positive factor 0.5 to M maxobs . Finally, λM c is computed using the maximum likelihood estimator technique outlined in the works of Kijko and Sellevoll (1989) and Kijko and Smit (2012). Summarizing, a strong variability has been included regarding the adoption of source models and the calculation of seismicity parameters. The Ground Motion Prediction Equation (GMPE) plays a pivotal role in PSHA by providing estimations of peak ground motions that take into account the earthquake magnitude, distance from the site, focal mechanism type, and soil conditions. It is important to emphasize that the prediction of ground motions resulting from an earthquake may introduce errors, especially in the near–field, hence GMPEs are empirical relationships that possess notable epistemic uncertainties. In this context, variability is a critical aspect, allowing us to incorporate a range of PGA, PGV, and Sa values for the same independent variables, such as earthquake magnitude and distance. In our study, we employed GMPEs that have undergone rigorous testing, validation, and ranking in recent PSHA studies (Pavlou et al., 2021; Kaviris et al., 2022a, 2022b; Sakkas et al., 2022; Kaviris et al., 2023). For PGA, we utilized the GMPEs proposed by Skarlatoudis et al. (2003) [SKA03], Danciu and Tselentis (2007) [DAT07], Sakkas (2016) [SAK16] and Chousianitis et al. (2018) [CHO18]. For PGV, the GMPEs of Skarlatoudis et al. (2007) [SKA07], Danciu and Tselentis (2007) [DAT07] and Chousianitis et al. (2018) [CHO18] were selected and for Sa the GMPE of Danciu and Tselentis (2007) [DAT07] was chosen. For PSHA purposes, it is essential not to assign to the mentioned GMPEs only one type of focal mechanism (normal or non–normal) to each source area, which can lead to overestimation or underestimation of peak ground motions. To address this challenge, we utilized the focal mechanism catalogue proposed by Kapetanidis and Kassaras (2019) to determine the accurate percentages of normal and non–normal focal mechanisms for each source area. We then applied those percents to the selected GMPEs to ensure accurate proportions between the normal and non–normal versions of each GMPE. This approach allows to avoid the need for interpolating the type of focal mechanism for each source area. The percentages can be found in Table S1. The concept of logic tree diagrams, originally introduced by Kulkarni et al. (1984), has consistently proven to be a reliable method for mitigating uncertainties (Bommer and Scherbaum, 2008; Atkinson et al., 2014; Marzocchi et al., 2015). In PSHA, a logic tree comprises multiple branches, each representing a potential seismic hazard outcome, reflecting the associated uncertainties. These branches are created to account for different choices that the analyst deems feasible and, to express the level of confidence in each one, every branch is assigned a normalized weight. Within the context of this paper, the logic tree visually represents the steps discussed (Figure 3). The process commences with the source models, where three branches are established, one for each model, with equal weighting. Subsequently, for every source area of each source model, additional branches are introduced to accommodate variables, such as b–values, M c , and λ(M c ), determined using the MAXC technique and the newly proposed method by Godano and Petrillo (2023). Moreover, variations in M u are considered by incorporating three branches, representing low, intermediate, and high M u levels. Furthermore, the selected GMPEs are integrated into the analysis, each one associated with a minor logic tree that accounts for the percentages of normal and non-normal focal mechanism types within each source area. It is important to note that this complexity refers to a single source area only. Consequently, the total number of logic tree samples becomes extraordinarily large, akin to the situation in PSHA for Canada (Kolaj et al., 2020). In such cases, it is essential to implement statistical sampling methods, as emphasized by Pagani et al. (2014). For our specific needs, the Latin Hypercube Sampling (LHS) method, initially proposed by McKay et al. (1979), is deemed the most suitable. Unlike the random sampling of the Monte Carlo technique, LHS divides the input ranges into equal intervals, ensuring that only one value is selected from each interval. This approach offers a more systematic and representative way to sample the various branches of the logic tree, thereby capturing the full range of scenarios. The number of samples used for the sampling process was determined through a trial-and-error approach, yielding stable results for PGA, PGV and Sa. It was decided that 10,000 samples would be employed, consistent with the procedure followed in ESHM13 and ESHM20. 3. Results The results are presented in the form of PGA and PGV spatial distribution maps for Cephalonia and Ithaca (Figures 4 and 5, respectively), considering return periods of 475 (Figures 4a and 5a) and 950 years (Figures 4b and 5b). Additionally, PGA–hazard curves were generated for the three most densely populated towns in our study area: Lixouri, Argostoli, and Vathi (Figure 6a). Those were produced to demonstrate PGA values for a wide range of return periods. Moreover, for the same sites, Uniform Hazard Spectra (UHS) were constructed by utilizing Sa levels, which correspond to their natural periods (Figure 6b). Based on the spatial distribution of both PGA and PGV for return periods of 475 and 950 years, the lowest values are observed in the southeastern portion of the study area. As we move towards the north and west, the values increase, reaching their peak mainly in Paliki and a small area east of the Myrtos Gulf. The pattern observed may be attributed mainly to the small distance of Paliki peninsula from the Cephalonia segment of the CLTFZ, characterized by high seismicity, as well as to the onshore faulting system, where the 2014 Cephalonia earthquakes occurred. The peak near Myrtos Gulf, where intermediate magnitude events have occurred due to the activation of secondary smaller structures transverse to the axis of the CLTFZ (Sakkas et al., 2022), may be aleatoric due to the small contour area. The lowest PGA value for return period of 475 years is approximately 460 cm/s², and the highest around 580 cm/s² (Figure 4a). The PGA range is about 120 cm/s², with the most intense variation occurring along the southeastern edge of Cephalonia. This indicates that intermediate to high PGA values are prevalent throughout most of the study area. Similar observations can be made for the return period of 950 years (Figure 4b), where the difference between the highest (around 700 cm/s²) and the lowest (approximately 560 cm/s²) PGA is about 140 cm/s². Again, the high PGA variation mainly occurs along the south easternmost edge of the study area. Regarding PGV, for a return period of 475 years (Figure 5a), some differences can be noticed compared to PGA. In particular, the spatial distribution is smoother, with the lowest PGV value approximately 33 cm/s, while the highest is around 36 cm/s. This relative stability in PGV values can be attributed to the smaller number of GMPEs used for PGV computation, as SAK16 does not propose a PGV model. This implies that the logic tree has significantly fewer total branches, resulting in a lower complexity of the spatial distribution. Continuing to the results of PGV for the return period of 950 years (Figure 5b), we find a similar situation. The highest value is ~49 cm/s, and the lowest is about 45 cm/s, with a small deviation of approximately 4 cm/s. Continuing to the site–specific analysis and the PGA–hazard curves (Figure 6a), we can observe the PGA levels across a range of probabilities of exceedance in 50 years. Lixouri and Argostoli exhibit similar values, with Lixouri's curve being slightly higher than that of Argostoli by approximately 10 cm/s² for all probabilities of exceedance. This small difference can possibly be attributed to their distance from the CLTFZ, as Lixouri is closer to the fault than Argostoli. Vathi displays the lowest hazard curve, which aligns with the spatial distribution of PGA values for both return periods, as Vathi is situated in a region characterized by intermediate PGA values. It is worth noting that even for the highest presented return periods, ground motions do not exceed 1 g. Regarding the UHS for the same towns (Figure 6b), Lixouri exhibits the highest Sa levels compared to Argostoli and Vathi across the entire range of natural periods. This observation is in agreement with the previously mentioned results. Furthermore, the UHS provides information about the natural period of the single–degree freedom system that experiences the highest Sa value, which in our case is 0.25 s. However, it is essential to acknowledge a slightly lower peak at 0.45 s, which should be considered, especially for Lixouri which experiences Sa values that are nearly identical for these two periods. 4. Discussion The primary objective of this study is to conduct a reassessment of seismic hazard in Cephalonia and Ithaca by involving the integration of new statistical techniques aimed at reducing epistemic uncertainties related to source models, seismicity data and GMPEs. The first set of results pertains to the spatial distribution of PGA and PGV for return periods 475 and 950 years across Cephalonia and Ithaca and the second one focuses exclusively on the three most densely populated localities, namely Lixouri, Argostoli, and Vathi. For these towns, PGA hazard curves and UHS were developed to illustrate the variation in PGA values over a wide range of probabilities of exceedance and the Sa distribution across various natural periods, respectively. The study suggests that the area of higher seismic hazard is the Paliki peninsula in western Cephalonia. This is mainly due to the vicinity of Paliki to the Cephalonia segment of the CLTFZ, maybe the most seismically active structure in Greece. In addition, high PGA values in Paliki are also influenced by the onshore local faults related with the 2014 Cephalonia earthquakes (M w =6.1 and M w =5.8), which also caused local ground deformation (Sakkas et al., 2022). Therefore, future infrastructure or seismic retrofitting in Paliki require special attention. Moreover, the area east of Myrtos Gulf has a high level of seismic hazard, although this finding may contain a considerable level of uncertainty. In the work of Sakkas et al. (2022), it was demonstrated that the post-seismic activity of the 2014 earthquake sequence primarily migrated northward, with clusters also located within the Myrtos Gulf. The site-specific results show that Lixouri has the highest PGA hazard curve and UHS, while Argostoli has intermediate curves, similar with those of Lixouri. In contrast, Vathi has the lowest maximum expected ground motions among the towns. Lixouri's proximity to the CLTFZ, in comparison to Argostoli and Vathi, may be the reason for this distribution. The findings for both return periods offer valuable insights for structural design and engineering purposes. Specifically, the information obtained by the spatial distribution of PGA and PGV serves as a crucial reference for engineers, enabling them to design infrastructure capable of withstanding the maximum anticipated ground motions. Furthermore, for existing buildings, the results can guide seismic retrofitting efforts by developing a more accurate seismic risk assessment of the area, ensuring that they meet safety standards. Such practices are essential in regions characterized by high seismic activity, as they aid to ensure the resilience of structures against infrequent but potentially destructive earthquakes. The outcomes of the site–specific analysis for the three towns offer valuable insights into the anticipated maximum ground motions over a 50–year timeframe, taking into account varying probabilities of occurrence. Additionally, these findings help in identifying measures to prevent resonance phenomena linked to the prevailing soil period (which was at 0.25 s). As previously mentioned, there have been several studies conducted to assess seismic hazard in Cephalonia and Ithaca. The one of Bonatis (2020) employed the same PSHA methodology as the one outlined here. However, differences exist between the source models and GMPEs used in the herein proposed PSHA. Specifically, Bonatis (2020) utilized a single seismotectonic model, BON20, and calculated PGA for a 475–year return period using various GMPEs, which, however, were not combined through a logic tree technique. The PGA results ranged from 200 to 900 cm/s 2 , while the herein obtained values are in a much narrower range. Nonetheless, the spatial distribution of PGA remained consistent, particularly regarding the high variability observed at the southern edge of Cephalonia. In this case, the PGA values ranged from 50 to 450 cm/s 2 , and PGV from 0 to 25 cm/s. However, it is important to note that these results cannot be directly compared to those of the current study, given the fundamental differences in the methodologies employed. In their recent work, Sakkas et al. (2022) used the ESHM13 source model and the GMPE developed by Danciu and Tselentis (2007), only for non-normal focal mechanisms. The MAXC method was used to obtain their seismicity parameters. Therefore, due to differences in the preprocessing part, variations in the results are expected between their work and this study. The logic tree used by Sakkas et al. (2022) had significantly fewer branches compared to the one used in this PSHA, resulting in a more smoothed spatial distribution of PGA. In their computational grid, Sakkas et al. (2022) included all the Ionian Islands. Therefore, for visualization clarity, the values in Cephalonia and Ithaca are depicted as a single value of approximately 500 cm/s 2 for a return period of 475 years. Τhe present PSHA was performed aiming to reduce the epistemic uncertainties, lowering them when compared to the aforementioned studies. This was achieved via a sophisticated logic tree approach and the application of a reliable sampling technique. The logic tree contains a very large number of different seismic hazard outcomes, considering the sublogic trees employed for each source area. In the process of estimating the b-value and Mc, two methods were considered, one of which is the new technique proposed by Godano and Petrillo (2023). There is a limitation in this, as it tends to estimate higher Mc values compared to other techniques (for instance, the MAXC that was also utilized in this study). The generation of high Mc values could potentially result in a limited number of data points for the estimation of the b-value, thereby leading to higher uncertainty in the regression model. The lack of a reliable b-value is a considerable drawback, given its crucial role in characterizing the seismicity for each source area. Nevertheless, the technique of Godano and Petrillo (2023) can be chosen in regions of high seismicity, where an adequate number of earthquakes certainly exists. Consequently, this method can be selected for the region of Cephalonia and Ithaca. Given the high degree of uncertainty regarding the Mu, we chose to designate three levels for each source area (low, intermediate, and high) in order to capture a broad range of seismic hazard outcomes related to this aspect. Furthermore, by incorporating precise weights into both the normal and non-normal versions of each GMPE at each source zone, we were able to manage the uncertainty related to the extrapolation of the focal mechanism. This is a critical step in preventing the overestimation and underestimation of the maximum expected ground motions. It is worth noting that a reliable sampling technique was required for this extensive logic tree. The LHS was chosen due to its non-memoryless nature, which is particularly significant in this context, as it effectively samples a considerable portion of the entire distribution of logic tree branches. This PSHA has certain limitations, for instance, it does not account for soil conditions, which can influence the results through amplification or attenuation phenomena. In the absence of information on true soil conditions, the study of Allen and Wald (2009), which relies only on topographic data can be used as a proxy for seismic site conditions, but it was not chosen for this PSHA in order to avoid additional potential uncertainties. Another limitation is the absence of a GMPE ranking system in the analysis, as it could provide valuable insights into which one of the selected empirical models best matches the recorded strong motion values and their relative weights for inclusion in the logic tree approach. However, it is worth noting that a recent study by Kaviris et al. (2023) conducted a GMPE ranking system for the same GMPEs as those selected in the herein computational framework, and the results indicated that the relative weights were very close to each other. This suggests that the PGA, PGV and Sa obtained here may not significantly differ from those generated using a non-equal-weighted logic tree approach. Future research endeavors could involve conducting fieldwork to capture ambient noise in three spatial dimensions at various sites. This data collection would facilitate the determination of the fundamental resonance frequency of the ground across Cephalonia and Ithaca using a spatial grid. Understanding this parameter is of great significance in earthquake engineering, as it provides valuable insights into site–specific effects. Another avenue for exploration is the development of a new GMPE that specifically predicts PGA and PGV for the vertical component of ground motion. Such information would be particularly beneficial for the construction of bridges, especially in regions with high seismic activity. Furthermore, a seismic risk assessment could be undertaken, exploiting the herein proposed PSHA as an input to determine the maximum expected ground motions. 5. Conclusion The aim of this study is to re–evaluate the seismic hazard for Cephalonia and Ithaca, taking into account new data and statistical methodologies. It is recognized that seismic hazard assessment is a field characterized by significant epistemic uncertainty, especially in terms of source models, seismicity parameters, and GMPEs (Cornell, 1968). Therefore, in regions like the one of the present work, which are subject to high seismic hazard, as highlighted by the current National Building Code (EAK, 2003), it is essential to update the maximum expected ground motions in order to have a reliable input for future seismic risk studies (Cornell, 1968). A well stablished and dependable approach to addressing epistemic uncertainty involves the use of a comprehensive logic tree decision graph. This method generates various seismic hazard outcomes, all taken into account in the final results. The most prevalent sources of bias and uncertainty are the seismotectonic models, seismicity parameters and GMPEs. In the herein proposed PSHA, we incorporate three groups of source areas, i.e., two European and one local for Cephalonia and Ithaca. These models are considered with equal weights in the computational schemes due to the lack of specific criteria indicating one model's superiority over another. Significant attention is given to the computation of seismicity parameters, where two methods were implemented for calculating the pair of Mc and b–value, and three for estimating Mu. Due to the absence of adequate criteria, we do not assign different weights among these techniques; they are treated as equal in the logic tree. The most critical aspect is addressing the uncertainty regarding GMPEs, as these empirical relations predict the maximum expected ground motions (PGA, PGV, and Sa). We select several GMPEs that consider the epicentral type of distance, since we do not have faults, but source areas, and the type of focal mechanism. In PSHA, it is common to extrapolate the focal mechanism type for each source area, leading to the selection of GMPEs for one specific faulting type only. In this updated PSHA, we create a sub–logic tree for each GMPE that considers the relative percentages of normal and non–normal types of focal mechanisms. When considering all possible seismic hazard outcomes, it becomes clear that the final number of branches is large, making a complete enumeration almost impossible. Therefore, we opt for a sampling of the logic tree using reliable techniques to capture a significant portion of it. Consequently, we believe that this study's results have thus far achieved the lowest degree of uncertainty. The findings of this study are presented as spatial distributions of PGA and PGV for return periods of 475 and 950 years, along with site–specific results for Lixouri, Argostoli, and Vathi (PGA–hazard curves and UHS). The highest ground motions are observed in the western portions of Cephalonia and Ithaca, with significant variability in the southeastern edge of Cephalonia. Among the towns studied, Lixouri exhibits the highest level of seismic hazard, while Vathi the lowest. All three towns have a dominant frequency of 0.25 s. The PGA results for the first return period can be compared to the reference value proposed by EAK (2003), which divides Greece into three seismic hazard zones, each being attributed a specific PGA for bedrock conditions. For Cephalonia and Ithaca, this reference value is approximately 360 cm/s², the highest in the country. However, the maximum PGA from our proposed PSHA model is 580 cm/s², underscoring the need to update the reference value. The results for the second return period can aid the construction of critical structures, such as medical centers. The site–specific analysis results provide insights on the estimated fundamental resonance frequency and the maximum expected spectral acceleration. 6. References Aki, K., 1965. Maximum likelihood estimate of b in the formula log N = a-bM and its confidence limits. Bull. Earthq. Res. Inst. 43, 237–239. Albini, P., Locati, M., Rovida, A., Stucchi, M., 2013. European Archive of Historical EArthquake Data (AHEAD). Istituto Nazionale di Geofisica e Vulcanologia (INGV). https://doi.org/10.6092/ingv.it-ahead Allen, T.I., Wald, D.J., 2009. On the Use of High-Resolution Topographic Data as a Proxy for Seismic Site Conditions (VS30). Bulletin of the Seismological Society of America 99, 935–943. https://doi.org/10.1785/0120080255 Anderson, H., Jackson, J., 1987. Active tectonics of the Adriatic Region. Geophys. J. Int. 91, 937–983. https://doi.org/10.1111/j.1365-246X.1987.tb01675.x Atkinson, G.M., Bommer, J.J., Abrahamson, N.A., 2014. Alternative Approaches to Modeling Epistemic Uncertainty in Ground Motions in Probabilistic Seismic‐Hazard Analysis. Seismological Research Letters 85, 1141–1144. https://doi.org/10.1785/0220140120 Bommer, J.J., Scherbaum, F., 2008. The Use and Misuse of Logic Trees in Probabilistic Seismic Hazard Analysis. Earthquake Spectra 24, 997–1009. https://doi.org/10.1193/1.2977755 Bonatis, P., 2020. Strong ground motion simulation in the Central Ionian Islands using a hybrid (deterministic and stochastic) approach. Master Thesis, School of Geology, Aristotle University of Thessaloniki. 129. Bonatis, P., Akinci, A., Karakostas, V., Papadimitriou, E., Kaviris, G., 2021. Near-Fault Broadband Ground Motion Simulation Applications at the Central Ionian Islands, Greece. Pure Appl. Geophys. 178, 3505–3527. https://doi.org/10.1007/s00024-021-02825-9 Chousianitis, K., Del Gaudio, V., Pierri, P., Tselentis, G.-A., 2018. Regional ground-motion prediction equations for amplitude-, frequency response-, and duration-based parameters for Greece. Earthquake Engng Struct Dyn 47, 2252–2274. https://doi.org/10.1002/eqe.3067 Cooke, P., 1979. Statistical inference for bounds of random variables. Biometrika 66, 367–374. https://doi.org/10.1093/biomet/66.2.367 Cornell, C., 1968. Engineering Seismic Risk Analysis. Bulletin of the Seismological Society of America. 58, 1583–1606. Danciu, L., Nandan, S., Reyes, C., Basili, R., Weatherill, G., Beauval, C., Rovida, A., Vilanova, S., Sesetyan, K., Bard, P.-Y., Cotton, F., Wiemer, S., Giardini, D., 2021. ESHM20 - EFEHR Technical ReportThe 2020 update of the European Seismic Hazard Model - ESHM20: Model Overview. EFEHR European Facilities of Earthquake Hazard and Risk. https://doi.org/10.12686/A15 Danciu, L., Tselentis, G.-A., 2007. Engineering Ground-Motion Parameters Attenuation Relationships for Greece. Bulletin of the Seismological Society of America 97, 162–183. https://doi.org/10.1785/0120050087 EAK, 2003. Greek seismic code edited by: Earthquake planning and protection organization. Athens, Greece. Ganas, A., Oikonomou, I.A., Tsimi, C., 2013a. NOAfaults: a digital database for active faults in Greece. geosociety 47, 518–530. https://doi.org/10.12681/bgsg.11079 Ganas, A., Marinou, A., Anastasiou, D., Paradissis, D., Papazissi, K., Tzavaras, P., Drakatos, G., 2013b. GPS-derived estimates of crustal deformation in the central and north Ionian Sea, Greece: 3-yr results from NOANET continuous network data. J. Geodyn. 67, 62–71. https://doi.org/10.1016/j.jog.2012.05.010 Ganas, A., 2023. NOAFAULTS KMZ layer Version 5.0 (V5.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.8075517 Godano, C., Petrillo, G., 2023. Estimating the Completeness Magnitude mc and the b-Values in a Snap. Earth and Space Science 10, e2022EA002540. https://doi.org/10.1029/2022EA002540 Gumbel, E., Lieblein, J., 1954. Statistical Theory of Extreme Values and Some Practical Applications: A Series of Lectures. National Bureau of Standards, US Government Printing Office, Washington. 33. Gutenberg, B., Richter, C.F., 1944. Frequency of earthquakes in California. Bulletin of the Seismological society of America 34, 185–188. Kapetanidis, V., Kassaras, I., 2019. Contemporary crustal stress of the Greek region deduced from earthquake focal mechanisms. Journal of Geodynamics 123, 55–82. https://doi.org/10.1016/j.jog.2018.11.004 Karakostas, V., Papadimitriou, E., Mesimeri, M., Gkarlaouni, C., Paradisopoulou, P., 2015. The 2014 Kefalonia Doublet (MW6.1 and MW6.0), Central Ionian Islands, Greece: Seismotectonic Implications along the Kefalonia Transform Fault Zone. Acta Geophys. 63, 1–16. https://doi.org/10.2478/s11600-014-0227-4 Karastathis, V.K., Mouzakiotis, E., Ganas, A., Papadopoulos, G.A., 2015. High-precision relocation of seismic sequences above a dipping Moho: the case of the January–February 2014 seismic sequence on Cephalonia island (Greece). Solid Earth 6, 173–184. https://doi.org/10.5194/se-6-173-2015 Kassaras, I., Papadimitriou, P., Kapetanidis, V., Voulgaris, N., 2017. Seismic site characterization at the western Cephalonia Island in the aftermath of the 2014 earthquake series. Geo-Engineering 8, 7. https://doi.org/10.1186/s40703-017-0045-z Kaviris, G., Zymvragakis, A., Bonatis, P., Kapetanidis, V., Spingos, I., Mavroulis, S., Kotsi, E., Lekkas, E., Voulgaris, N., 2023. A Logic-Tree Approach for Probabilistic Seismic Hazard Assessment in the Administrative Region of Attica (Greece). Applied Sciences 13. https://doi.org/10.3390/app13137553 Kaviris, G., Zymvragakis, A., Bonatis, P., Kapetanidis, V., Voulgaris, N., 2022a. Probabilistic and Scenario-Based Seismic Hazard Assessment on the Western Gulf of Corinth (Central Greece). Applied Sciences 12. https://doi.org/10.3390/app122111152 Kaviris, G., Zymvragakis, A., Bonatis, P., Sakkas, G., Kouskouna, V., Voulgaris, N., 2022b. Probabilistic Seismic Hazard Assessment for the Broader Messinia (SW Greece) Region. Pure and Applied Geophysics 179, 551–567. https://doi.org/10.1007/s00024-022-02950-z Kijko, A., Sellevoll, M.A., 1989. Estimation of earthquake hazard parameters from incomplete data files. Part I. Utilization of extreme and complete catalogs with different threshold magnitudes. Bulletin of the Seismological Society of America 79, 645–654. https://doi.org/10.1785/BSSA0790030645 Kijko, A., Smit, A., 2012. Extension of the Aki‐Utsu b‐Value Estimator for Incomplete Catalogs. Bulletin of the Seismological Society of America 102, 1283–1287. https://doi.org/10.1785/0120110226 Kolaj, M., Adams, J., Halchuk, S., 2020. The 6th generation seismic hazard model of Canada, in: 17th World Conference on Earthquake Engineering. pp. 1–12. Kulkarni, R., Youngs, R., Coppersmith, K., 1984. Assessment of confidence intervals for results of seismic hazard analysis, in: Proceedings of the Eighth World Conference on Earthquake Engineering. San Francisco, pp. 263–270. Lagios, E., Sakkas, V., Papadimitriou, P., Parcharidis, I., Damiata, B.N., Chousianitis, K., Vassilopoulou, S., 2007. Crustal deformation in the Central Ionian Islands (Greece): Results from DGPS and DInSAR analyses (1995–2006). Tectonophysics 444, 119–145. https://doi.org/10.1016/j.tecto.2007.08.018 Lekkas, E.L., Mavroulis, S.D., 2016. Fault zones ruptured during the early 2014 Cephalonia Island (Ionian Sea, Western Greece) earthquakes (January 26 and February 3, Mw 6.0) based on the associated co-seismic surface ruptures. J. Seismol. 20, 63–78. https://doi.org/10.1007/s10950-015-9510-3 Makropoulos, K., Kaviris, G., Kouskouna, V., 2012. An updated and extended earthquake catalogue for Greece and adjacent areas since 1900. Natural Hazards and Earth System Sciences 12, 1425–1430. https://doi.org/10.5194/nhess-12-1425-2012 Makropoulos, K.C., Burton, P.W., 1985. Seismic hazard in Greece. II. Ground acceleration. Tectonophysics 117, 259–294. https://doi.org/10.1016/0040-1951(85)90274-4 Marzocchi, W., Taroni, M., 2014. Some Thoughts on Declustering in Probabilistic Seismic‐Hazard Analysis. Bulletin of the Seismological Society of America 104, 1838–1845. https://doi.org/10.1785/0120130300 Marzocchi, W., Taroni, M., Selva, J., 2015. Accounting for Epistemic Uncertainty in PSHA: Logic Tree and Ensemble Modeling. Bulletin of the Seismological Society of America 105, 2151–2159. https://doi.org/10.1785/0120140131 Mavroulis, S., Lekkas, E., 2021. Revisiting the Most Destructive Earthquake Sequence in the Recent History of Greece: Environmental Effects Induced by the 9, 11 and 12 August 1953 Ionian Sea Earthquakes. Appl. Sci. 11, 8429. https://doi.org/10.3390/app11188429 McGuire, R.K., 1976. FORTRAN computer program for seismic risk analysis (Report No. 76–67), Open-File Report. https://doi.org/10.3133/ofr7667 McKay, D., Beckman, R., Conover, W., 1979. A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code. Technometrics 21, 239–245. https://doi.org/10.2307/1268522 McKenzie, D., 1972. Active Tectonics of the Mediterranean Region. Geophys. J. R. Astron. Soc. 30, 109–185. https://doi.org/10.1111/j.1365-246X.1972.tb02351.x Pagani, M., Monelli, D., Weatherill, G., Danciu, L., Crowley, H., Silva, V., Henshaw, P., Butler, L., Nastasi, M., Panzeri, L., Simionato, M., Vigano, D., 2014. OpenQuake Engine: An Open Hazard (and Risk) Software for the Global Earthquake Model. Seismological Research Letters 85, 692–702. https://doi.org/10.1785/0220130087 Papadimitriou, E., Karakostas, V., Mesimeri, M., Chouliaras, G., Kourouklas, C., 2017. The Mw6.5 17 November 2015 Lefkada (Greece) Earthquake: Structural Interpretation by Means of the Aftershock Analysis. Pure Appl. Geophys. 174, 3869–3888. https://doi.org/10.1007/s00024-017-1601-3 Papadimitriou, P., 1988, Etude de la structure du manteau superieur de l’Europe et modelisation des ondes de volume engendrees par des seismes Egeens, PhD thesis. University of Paris, Paris. Papadimitriou, P., Kaviris, G., Makropoulos, K., 2006. The Mw = 6.3 2003 Lefkada Earthquake (Greece) and induced stress transfer changes. Tectonophysics 423, 73–82. https://doi.org/10.1016/j.tecto.2006.03.003 Papadimitriou, P., Voulgaris, N., Kouskouna, V., Kassaras, I., Kaviris, G., Pavlou, K., Karakonstantis, A., Bozionelos, G., Kapetanidis, V., 2014. The Kefallinia Island earthquake sequence January–February 2014. In Proceedings of the Second European Conference on Earthquake Engineering and Seismology (2ECEES), Istanbul, Turkey, 24–29 August 2014. Papadimitriou, P., Kapetanidis, V., Karakonstantis, A., Spingos, I., Pavlou, K., Kaviris, G., Kassaras, I., Sakkas, V., Voulgaris, N., 2021. The 25 October, 2018 Zakynthos (Greece) earthquake: seismic activity at the transition between a transform fault and a subduction zone. Geophys. J. Int. 225, 15–36. https://doi.org/10.1093/gji/ggaa575 Papazachos, B.C., Papazachou, C.B., 2003. The Earthquakes of Greece. Ziti Publications, Thessaloniki, 273 p. (In Greek). Pavlou, K., Κaviris, G., Kouskouna, V., Sakkas, G., Zymvragakis, A., Sakkas, V., Drakatos, G., 2021. Minor seismic hazard changes in the broader area of Pournari artificial lake after the first filling (W. Greece). Results in Geophysical Sciences 100025. https://doi.org/10.1016/j.ringps.2021.100025 Robson, D.S., Whitlock, J.H., 1964. Estimation of a truncation point. Biometrika 51, 33–39. https://doi.org/10.1093/biomet/51.1-2.33 Sakkas, G., 2016. Calculation and analysis of the seismic motion rotational components in Greece, Ph.D. Thesis. Geophysics-Geothermics Department, Faculty of Geology, University of Athens, Greece (in Greek). 278. http://dx.doi.org/10.12681/eadd/39773 Sakkas, G., Kouskouna V., Makropoulos, K., 2010. Seismic hazard analysis in the Ionian Islands using macroseismic intensities. Hell. J. Geosci. 45, 239–248 Sakkas, V., Lagios, E., 2015. Fault modelling of the early-2014 ~M6 Earthquakes in Cephalonia Island (W. Greece) based on GPS measurements. Tectonophysics 644–645, 184–196. https://doi.org/10.1016/j.tecto.2015.01.010 Sakkas, V., Kapetanidis, V., Kaviris, G., Spingos, I., Mavroulis, S., Diakakis, M., Alexopoulos, J.D., Kazantzidou-Firtinidou, D., Kassaras, I., Dilalos, S., Vassilakis, E., Kotsi, E., Tselentis, G., Lekkas, E., Voulgaris, N., 2022. Seismological and Ground Deformation Study of the Ionian Islands (W. Greece) during 2014–2018, a Period of Intense Seismic Activity. Applied Sciences 12. https://doi.org/10.3390/app12052331 Scordilis, E.M., Karakaisis, G.F., Karacostas, B.G., Panagiotopoulos, D.G., Comninakis, P.E., Papazachos, B.C., 1985. Evidence for transform faulting in the Ionian sea: The Cephalonia island earthquake sequence of 1983. Pure Appl. Geophys. PAGEOPH 123, 388–397. https://doi.org/10.1007/BF00880738 Skarlatoudis, A.A., Papazachos, C.B., Margaris, B.N., Theodulidis, N., Papaioannou, Ch., Kalogeras, I., Scordilis, E.M., Karakostas, V., 2007. Erratum to Empirical Peak Ground-Motion Predictive Relations for Shallow Earthquakes in Greece. Bulletin of the Seismological Society of America 97, 2219–2221. https://doi.org/10.1785/0120070176 Skarlatoudis, A.A., Papazachos, C.B., Margaris, B.N., Theodulidis, N., Papaioannou, Ch., Kalogeras, I., Scordilis, E.M., Karakostas, V., 2003. Empirical Peak Ground-Motion Predictive Relations for Shallow Earthquakes in Greece. Bulletin of the Seismological Society of America 93, 2591–2603. https://doi.org/10.1785/0120030016Stiros, S.C., Pirazzoli, P.A., Laborel, J., Laborel-Deguen, F., 1994. The 1953 earthquake in Cephalonia (Western Hellenic Arc): coastal uplift and halotectonic faulting. Geophys. J. Int. 117, 834–849. https://doi.org/10.1111/j.1365-246X.1994.tb02474.x Sokos, E., Kiratzi, A., Gallovič, F., Zahradník, J., Serpetsidaki, A., Plicka, V., Janský, J., Kostelecký, J., Tselentis, G.A., 2015. Rupture process of the 2014 Cephalonia, Greece, earthquake doublet (Mw6) as inferred from regional and local seismic data. Tectonophysics 656, 131–141. https://doi.org/10.1016/j.tecto.2015.06.013 Stucchi, M., Rovida, A., Gomez Capera, A.A., Alexandre, P., Camelbeeck, T., Demircioglu, M.B., Gasperini, P., Kouskouna, V., Musson, R.M.W., Radulian, M., Sesetyan, K., Vilanova, S., Baumont, D., Bungum, H., Fäh, D., Lenhardt, W., Makropoulos, K., Martinez Solares, J.M., Scotti, O., Živčić, M., Albini, P., Batllo, J., Papaioannou, C., Tatevossian, R., Locati, M., Meletti, C., Viganò, D., Giardini, D., 2013. The SHARE European Earthquake Catalogue (SHEEC) 1000-1899. J. Seismol. 17, 523–544. https://doi.org/10.1007/s10950-012-9335-2 Taroni, M., Akinci, A., 2021. Good practices in PSHA: declustering, b-value estimation, foreshocks and aftershocks inclusion; a case study in Italy. Geophysical Journal International 224, 1174–1187. https://doi.org/10.1093/gji/ggaa462 Underhill, J.R., 1989. Late Cenozoic deformation of the Hellenide foreland, western Greece. Geol. Soc. Am. Bull. 101, 613–634. https://doi.org/10.1130/0016-7606(1989)1012.3.CO;2 Wang, Z., 2005. A Clear Definition of Seismic Hazard and Risk: A Basis for Hazard and Risk Assessment, Communication, and Management. Presented at the AGU Fall Meeting Abstracts, pp. S53B-1110. Wiemer, S., Wyss, M., 2000. Minimum Magnitude of Completeness in Earthquake Catalogs: Examples from Alaska, the Western United States, and Japan. Bulletin of the Seismological Society of America 90, 859–869. https://doi.org/10.1785/0119990114 Woessner, J., Laurentiu, D., Giardini, D., Crowley, H., Cotton, F., Grünthal, G., Valensise, G., Arvidsson, R., Basili, R., Demircioglu, M.B., Hiemer, S., Meletti, C., Musson, R.W., Rovida, A.N., Sesetyan, K., Stucchi, M., The SHARE Consortium, 2015. The 2013 European Seismic Hazard Model: key components and results. Bulletin of Earthquake Engineering 13, 3553–3596. https://doi.org/10.1007/s10518-015-9795-1 Zhou, Y., Zhou, S., Zhuang, J., 2018. A test on methods for MC estimation based on earthquake catalog. Earth and Planetary Physics 2, 150–162. https://doi.org/10.26464/epp2018015 Additional Declarations No competing interests reported. Supplementary Files kavirisetal2024seismicitydata.xlsx Cite Share Download PDF Status: Published Journal Publication published 06 Sep, 2024 Read the published version in Journal of Seismology → Version 1 posted Editorial decision: Revision requested 05 Aug, 2024 Reviews received at journal 03 Aug, 2024 Reviewers agreed at journal 03 May, 2024 Reviewers invited by journal 04 Mar, 2024 Submission checks completed at journal 26 Feb, 2024 Editor assigned by journal 26 Feb, 2024 First submitted to journal 26 Feb, 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-3991269","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":275094345,"identity":"4b466119-19c8-4916-aaa4-e605056259d7","order_by":0,"name":"George Kaviris","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYDACHiBOYJBg4AdxHpCkRbIBzCBWCwgYHCBWi7zP4WcSD/dY2BufX3zwQQLDYTn+BvZrH/BpMTzbZiaR8EwicduNZ8kGQC3GEgd4imfg1dLPYGyQcEAiwezGGaBehsOJDQd4kvE6zLCf/TNIi73xjPPfwFrmE9Iiz9tj+ACohXEDfw8bWMuGA+yH8Wox4DlTCNKSOOMGG9CFBunGhod5mPHb0pO+4eCPA3X2/P2HHz74UGEtJ3e8/TF+Ww7AWEBXAblAzMxjgN+WBhiLH66Z/QFeLaNgFIyCUTDiAAB/Uks1z/HH7AAAAABJRU5ErkJggg==","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"George","middleName":"","lastName":"Kaviris","suffix":""},{"id":275094346,"identity":"7560550b-f8f1-490e-9eff-9bcf6ed5de9e","order_by":1,"name":"Angelos Zymvragakis","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Angelos","middleName":"","lastName":"Zymvragakis","suffix":""},{"id":275094347,"identity":"1d8629b8-39da-48b6-a500-70adffd2872a","order_by":2,"name":"Vasilis Kapetanidis","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Vasilis","middleName":"","lastName":"Kapetanidis","suffix":""},{"id":275094348,"identity":"b91625e1-67fd-4975-9202-258159bef0a2","order_by":3,"name":"Vasiliki Kouskouna","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Vasiliki","middleName":"","lastName":"Kouskouna","suffix":""},{"id":275094349,"identity":"54df04e0-d589-49f1-ba37-201d4af9dbd7","order_by":4,"name":"Ioannis Spingos","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ioannis","middleName":"","lastName":"Spingos","suffix":""},{"id":275094350,"identity":"4cf73814-e34a-482c-b9dc-dd468d708857","order_by":5,"name":"Nikolaos Sakellariou","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Nikolaos","middleName":"","lastName":"Sakellariou","suffix":""},{"id":275094351,"identity":"7ae4a5c9-4bcd-4693-b30a-8b3a81544c65","order_by":6,"name":"Nicholas Voulgaris","email":"","orcid":"","institution":"National and Kapodistrian University of Athens (NKUA)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Nicholas","middleName":"","lastName":"Voulgaris","suffix":""}],"badges":[],"createdAt":"2024-02-26 15:12:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3991269/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3991269/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10950-024-10242-3","type":"published","date":"2024-09-06T16:05:28+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":51826004,"identity":"5cb829b9-04e8-45dd-ab7b-902c69471733","added_by":"auto","created_at":"2024-02-29 17:03:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":10483319,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e(Left) Map of the Central Ionian Islands in Western Greece (red rectangle in the inset map of Greece), presenting the seismicity catalogue of Makropoulos et al. (2012), covering the period 1900-2009, extended up to 2019 (this study). (Right) Close-up of the Cephalonia and Ithaca Islands. Earthquakes with M\u003c/em\u003e\u003csub\u003e\u003cem\u003ew\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e≥5.8 are presented as stars. The dates and magnitudes of significant earthquakes of the period 1900-2019 are marked on the map. The fault lines are from the NOAFAULTs v5.0 database (Ganas et al., 2013a; Ganas, 2023). CLTFZ: Cephalonia-Lefkada Transform Fault Zone; NAF: North Anatolian Fault.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/a85d2a1301c7c9922e4434c9.png"},{"id":51826000,"identity":"2fd53bea-247c-46da-a857-d6e5dd30148f","added_by":"auto","created_at":"2024-02-29 17:03:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":15086327,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eThe three chosen source area models employed in this study: (a) ESHM13 (Woessner et al., 2015), (b) ESHM20 (Danciu et al., 2021), and (c) BON20 (Bonatis, 2020). To determine the source areas from ESHM13 and ESHM20 contributing to the ground motions in Cephalonia and Ithaca, a selection distance of 100 km was applied (Kaviris et al., 2022a, 2022b, 2023).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/3475aa96d24a7343a004ea91.png"},{"id":51825991,"identity":"4d1b311f-28f6-474a-a993-a858bdda4d2e","added_by":"auto","created_at":"2024-02-29 17:03:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2330870,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eA section of the logic tree diagram presented for the current PSHA computational method. The branches are provided in detail exclusively for the GRAS369 source area of the ESHM13 source model, specifically for the lowest Mu value. It is worth noting that the complete set of branches for GRAS369 would also include two additional branches for the intermediate and the highest Mu values, while the remainder of the logic tree remains consistent. The same decision–making process was applied uniformly across all source areas within each source model, with the only difference associated with the weights assigned to the final column, representing the percentages of Normal (N) and Non–Normal (NN) focal mechanisms. Seismicity data are presented in Table S1.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/8ac0f2248633820c8a10b48d.png"},{"id":51826003,"identity":"10936221-962d-4e23-8ae4-b81765d63625","added_by":"auto","created_at":"2024-02-29 17:03:19","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":4609802,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePGA results for return periods 475 and 950 years (a and b, respectively).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/2c983790894ff42550112a67.png"},{"id":51825997,"identity":"bd7a30f5-cbd4-483a-929a-d6d8804a77f1","added_by":"auto","created_at":"2024-02-29 17:03:19","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":4059038,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePGV results for return periods of 475 and 950 years (a and b, respectively).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/9d5fa300797022b2e69dd98b.png"},{"id":51825996,"identity":"2ad20773-9d84-490a-9b3f-8734a0d08b63","added_by":"auto","created_at":"2024-02-29 17:03:18","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1173087,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePGA–hazard curves (a) and UHS in terms of Sa (b) for Lixouri, Argostoli and Vathi.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/7d735e97036fbd4472075e62.png"},{"id":64185839,"identity":"cd56617e-8c6b-4094-bd35-ed72d3a44714","added_by":"auto","created_at":"2024-09-09 16:22:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":46587809,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/3208d43d-0e9d-428c-b4b2-9852389512b2.pdf"},{"id":51825992,"identity":"58ace0e4-92ec-4d7c-8616-476bba7d3d6d","added_by":"auto","created_at":"2024-02-29 17:03:18","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":110827,"visible":true,"origin":"","legend":"","description":"","filename":"kavirisetal2024seismicitydata.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-3991269/v1/6fb62a5a64713865d77721fe.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Logic-Tree Based Probabilistic Seismic Hazard Assessment for the Central Ionian Islands of Cephalonia and Ithaca (Western Greece)","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe Central Ionian Islands (Figure 1) stand out as the most seismically hazardous area in Greece (EAK, 2003). Cephalonia Island, in particular, has hosted two M\u003csub\u003ew\u003c/sub\u003e \u0026ge;6.7 earthquakes (Makropoulos et al., 2012) within a 30-year span. The heightened seismic activity in this area is mainly attributed to the presence of the Cephalonia-Lefkada Transform Fault Zone (CLTFZ), a dextral plate boundary trending SSW-NNE. This fault zone bounds the western coasts of Cephalonia and Lefkada Islands and intersects with a complex network of onshore active faults, many of which are delineated in Figure 1. The CLTFZ, roughly outlined by the mapped earthquake epicenters of Figure 1, serves as a link between the NW-SE-trending major features of the Hellenic Arc in the south, and the collision front of the Apulian platform with the Hellenic foreland in the north. This configuration creates a shear zone of intense deformation, leading to major earthquakes that have caused substantial damage and loss of life in the recent past, particularly on Cephalonia and Ithaca Islands. The CLTFZ exhibits a strike of ~N15\u0026deg;E on the latitude of Lefkada Island, while further to the south, its trend shifts to ~N30\u0026deg;E as it passes offshore Cephalonia (Figure 1). The two slightly offset segments of the CLTFZ interact within Myrtos Gulf at northern Cephalonia, likely forming a transfer zone of extensional step-overs (Karakostas et al., 2015). The regional crustal stress field promotes strike-slip faulting (Kapetanidis and Kassaras, 2019), but the convergence of the African plate with the Aegean microplate introduces compression in a WSW-ENE direction. Differential GPS measurements have revealed a clockwise rotation of Cephalonia Island, relative to a station located on Aenos mountain, with the largest values observed at the western (Paliki) and northern (Erissos) parts of the island (Lagios et al., 2007). The broader area of Cephalonia features east-dipping NW- to NNW-striking thrust structures (Stiros et al., 1994), which may sporadically undergo seismic reactivation, resulting in earthquakes with a significant reverse dip-slip component. The Ionian Thrust traverses the southeastern part of Cephalonia Island and likely extends offshore, passing between Cephalonia and Ithaca (Underhill, 1989). Additional thrust structures are observed at Aenos Mountain and Argostoli, extending northward and separating Paliki from the rest of Cephalonia.\u003c/p\u003e\n\u003cp\u003eCephalonia has a historical record of significant earthquakes. The SHEEC catalog (Stucchi et al., 2013) and the AHEAD database (Albini et al., 2013) contain 14 historical earthquakes occurring between 1469 and 1867 with macroseismic epicenters near Lixouri and Argostoli. In the instrumental era, a significant M\u003csub\u003ew\u003c/sub\u003e=6.1 earthquake occurred on 12 January 1912, south of Argostoli (Makropoulos et al., 2012; Figure 1), killing 8 people and injuring 40 in Poros (Papazachos and Papazachou, 2003). Its focal mechanism, estimated from the directivity of macroseismic data, indicates strike-slip faulting (Papazachos et al., 1999). Another major earthquake of M\u003csub\u003ew\u003c/sub\u003e=6.1 occurred on 27 January 1915 near Ithaca (Makropoulos et al., 2012; Figure 1). It also exhibited an estimated dextral SW-NE strike-slip focal mechanism (Papazachos et al., 1999) and was responsible for the collapse or extensive damage of many houses (Papazachos and Papazachou, 2003).\u003c/p\u003e\n\u003cp\u003eOn 9 and 11 August 1953, two earthquakes of M\u003csub\u003ew\u003c/sub\u003e=5.9 and 6.6, respectively (Makropoulos et al., 2012), occurred east of Cephalonia Island, the latter with an epicenter on Ithaca Island. They were followed on 12 August by an even larger earthquake of M\u003csub\u003ew\u003c/sub\u003e=7.0 (Makropoulos et al., 2012) at the southeastern part of Cephalonia Island, with a location error estimate of ~50 km (Anderson and Jackson, 1987). Its focal mechanism, determined from first motion polarities, indicates reverse faulting in a NNW-SSE direction (McKenzie, 1972), although estimates from macroseismic data suggest strike-slip faulting (Papazachos et al., 1999). In both instances, the P-axis aligns with the direction of maximum horizontal compression (Kapetanidis and Kassaras, 2019), as well as with the direction of shortening (N258\u0026deg;E) in the area of the Central Ionian Sea, maintaining an extension-to-shortening ratio of 1:3 (Ganas et al., 2013b). The maximum intensity value of the 1953 earthquakes at Cephalonia Island was I\u003csub\u003emax\u003c/sub\u003e=9/10, observed at five localities, among which Argostoli and Lixouri (Sakkas et al., 2010), however estimated even higher due to cumulative damage. The deformation resulting from the 12 August 1953 event is evident in the observed coastal uplift near Poros, at the southeastern part of Cephalonia Island, as manifested through notches on the eroded rocks. Mushroom-shaped formations in uplifted rocks about 50 m offshore also suggest that a similar paleoseismic event must have occurred (Stiros et al., 1994). The coastal uplift observations support a piston-like motion on two subparallel, east-dipping reverse faults (Stiros et al., 1994). The 1953 earthquake sequence also revealed the necessity for the implementation of a National Building Code, to mitigate the risk from seismic hazards in Greece.\u003c/p\u003e\n\u003cp\u003eAnother significant earthquake was an M\u003csub\u003ew\u003c/sub\u003e=6.7 event (Makropoulos et al., 2012) that occurred on 17 January 1983, approx. 15 km southwest of Paliki. Due to its offshore epicenter, it reportedly caused minor damage to Cephalonia Island (EMS intensity IV at Argostoli), whereas its strongest aftershock, on 23 March 1983, was more damaging (max intensity VII), as its epicenter was located further north, closer to the island (Papazachos and Papazachou, 2003). The mainshock has been interpreted as a strike-slip rupture on a southeast-dipping fault with a relatively low angle (Papadimitriou, 1988), likely with a thrust component (Scordilis et al., 1985). A similar faulting type, i.e. dextral strike-slip on an east-dipping low-angle fault, was determined for a recent major earthquake (M\u003csub\u003ew\u003c/sub\u003e=6.7) that occurred southwest of Zakynthos Island on 25 October 2018 (Papadimitriou et al., 2021). The latter is located in a transition zone between the southern end of the CLTFZ and the northwestern edge of the Hellenic Arc. Pure reverse faulting, consistent with SW-NE compression, occurs further south.\u003c/p\u003e\n\u003cp\u003eThe most recent major seismic activity on Cephalonia Island comprised of an earthquake \u0026ldquo;doublet\u0026rdquo; that occurred on 26 January and 3 February 2014, with M\u003csub\u003ew\u003c/sub\u003e=6.1 and 5.8,\u0026nbsp;respectively, on Paliki\u0026nbsp;(Papadimitriou et al., 2014; Karakostas et al., 2015; Karastathis et al., 2015; Sokos et al., 2015; Sakkas et al., 2022). Moment tensor inversions for both events support dextral strike-slip faulting on SSW-NNE-trending, steep-dipping faults, with most solutions indicating a slight tilt toward an east-dipping direction. Fault plane models, constructed using the observed co-seismic deformation, reveal that the two earthquakes occurred on two sub-parallel fault segments, in the southern and the northern parts of the Paliki peninsula, respectively (Sakkas and Lagios, 2015). It seems likely that the 2014 earthquake \u0026ldquo;doublet\u0026rdquo; may have accelerated the occurrence of a major earthquake (M\u003csub\u003ew\u003c/sub\u003e=6.3) on 17 November 2015 (Papadimitriou et al., 2017) on Lefkada (Figure 1); an area that was already stress-loaded after the M\u003csub\u003ew\u003c/sub\u003e=6.3 earthquake of 14 August 2003 (Papadimitriou et al., 2006). The 2014 Cephalonia earthquakes produced a rich aftershock sequence, with the seismicity rate remaining elevated for about 4 years (Sakkas et al., 2022). The seismic activity was further enhanced by the occurrence of the 2015 Lefkada earthquake, which led to a slight increase in the seismicity rate in Cephalonia. This was particularly observed in the area of Myrtos Gulf, where small structures, likely trending E-W, being antithetic and transverse to the CLTFZ, appear to be easily triggered by major earthquakes in the vicinity (Sakkas et al., 2022).\u003c/p\u003e\n\u003cp\u003eThe 1953 earthquake sequence on Cephalonia is among the most significant ones to have occurred in Greece during the instrumental period, leaving a profound societal imprint. The devastation was extended on the islands of Cephalonia, Ithaca, and Zakynthos, resulting in the destruction of ~83% of the building stock. The human toll was significant, with 455 fatalities, 21 individuals reported missing and 2412 people sustaining injuries (Papazachos and Papazachou, 2003). It was also the cause of a decline in the population of Cephalonia, Ithaca, and Zakynthos during the following decades, as people abandoned the islands and migrated mainly to urban centers of Greece or abroad (Mavroulis and Lekkas, 2021).\u003c/p\u003e\n\u003cp\u003eSeismic hazard is dedicated to investigating the phenomena triggered by earthquakes, with ground motion being the most significant as it acts as catalyst for potential secondary catastrophic events like rockfalls and liquefaction (Wang, 2005). To assess seismic hazard, we quantify ground motion by estimating Intensity Measurement Types (IMT), such as Peak Ground Acceleration (PGA), Peak Ground Velocity (PGV) and Spectral acceleration (Sa), through statistical methods (Gumbel and Lieblein, 1954; Cornell, 1968). The most common approach to compute the aforesaid parameters is the classic probabilistic method, initially proposed by Cornell (1968) and later commercialized by McGuire (1976), that introduces the usage of a seismotectonic model.\u003c/p\u003e\n\u003cp\u003eGiven the high seismicity of Greece, many seismic hazard assessment studies have been conducted, dating as far back as 1985 (Makropoulos and Burton, 1985) and continuing to the present (Bonatis, 2020; Bonatis et al., 2021; Pavlou et al., 2021; Kaviris et al., 2022a, 2022b, 2023). The Ionian Islands are widely recognized as an area of high seismic activity and are among the most seismically active regions in Europe. Therefore, a constantly updated with recent data seismic hazard assessment is crucial for urban planning to mitigate seismic risk. \u0026nbsp;Earthquakes in the region have produced high PGA values, for example, the two 2014 Cephalonia events produced ground acceleration of about 560 cm/s\u003csup\u003e2\u003c/sup\u003e and 735 cm/s\u003csup\u003e2\u003c/sup\u003e, respectively (Kassaras et al., 2017). These values well exceed the regulations proposed by the 2003 National Building Code (EAK, 2003), highlighting the need for an updated seismic hazard model. It is worth noting that previous studies have been conducted for Cephalonia and Ithaca, as evidenced by the work of cited researchers (Bonatis, 2020; Bonatis et al., 2021; Sakkas et al., 2022). Nevertheless, in line with Cornell\u0026apos;s insights, we investigate the possibility of reducing the considerable uncertainties regarding seismogenic sources and seismic wave attenuation.\u003c/p\u003e\n\u003cp\u003eIn this work, we assess seismic hazard by estimating PGA and PGV for Cephalonia and Ithaca, and Sa exclusively for the most populated localities of these islands, i.e. Lixouri, Argostoli, and Vathi. To mitigate uncertainties, we have developed an extensive and \u0026ldquo;non\u0026ndash;trivial\u0026rdquo; logic tree decision diagram for estimating IMT. In addition, we adopt a stochastic statistical sampling method to capture a substantial portion of the Probability Density Function (PDF) for each IMT. By executing these techniques, the goal of this study is to effectively reduce the prevailing uncertainties and establish an accurate seismic hazard model that will have the potential to be applied in more precise seismic risk assessments for this highly seismically active study area.\u003c/p\u003e"},{"header":"2. Methodology","content":"\u003cp\u003eThe PSHA results of the approach introduced by Cornell (1968) and McGuire (1976) are in the form of annual probabilities of surpassing specific random values of acceleration, velocity, or spectral acceleration. Subsequently, the final output consists of PGA, PGV, or Sa levels corresponding to the selected return period. A pioneering aspect of this methodology, at the time of its inception, was the incorporation of a source model to depict the seismotectonic attributes of each study region. These models are classified according to the knowledge of the existing active faults within a region. For instance, a model may include the actual fault sources of the study area and report the annual exceedance rates for each magnitude bin per fault. This requires a good awareness of the dynamic characteristics of each fault. On the other hand, a source area model takes the form of polygons where seismicity is distributed, following a normal distribution within each polygon, and all attributes are consistent. This model type is typically employed in regions where the occurrence rates of all potential active faults are not well documented or are unmapped. This is partly the situation in our study area, Cephalonia and Ithaca, where the majority of active faults are offshore, and their dynamic characteristics are not known in detail. Consequently, the source area type of model was adopted for this study.\u003c/p\u003e\n\u003cp\u003eThe use of a single model may introduce high epistemic uncertainties regarding the seismological attributes of the broader study area. A conventional strategy for addressing epistemic uncertainties in PSHA involves introducing variability in the input data (Bommer and Scherbaum, 2008; Atkinson et al., 2014; Marzocchi et al., 2015; Kaviris et al., 2023). To address this concern, we integrated three models into our computational framework, i.e., the European Seismic Hazard Model 2013 (ESHM13) by Woessner et al. (2015), its subsequent update, ESHM20, developed by Danciu et al. (2021), and the local Ionian\u0026ndash;Island source area model formulated by Bonatis (2020), herein referred to as BON20. Visual representations of these models are available in Figures 2a\u0026ndash;c. ESHM13, ESHM20, and BON20 were treated as equivalent in our analysis, as each one exhibits distinct advantages and disadvantages in comparison to the others, thus preventing in designating one as superior. This decision was substantiated by examining the surface area of each source zone within these models. Small polygons, like the ones of ESHM13 and BON20 in the Central Ionian Islands (Figures 2a and 2c), provide a more precise description of the seismotectonic features within the area, whereas larger polygons, as those of ESHM20 (Figure 2b), are less accurate in describing these features, but with a higher number of earthquakes for statistical analysis. The latter provides insight into comparing the seismicity parameters obtained for each source model. BON20 (Figure 2c) has smaller zones providing a more accurate description of the seismotectonics of the Ionian Islands, when compared to the other two models. However, the required seismicity data of the small source areas around Cephalonia and Ithaca will result in a complex and non-smooth spatial distribution of the computed PGA and PGV values, which may hinder drawing conclusions about the seismic hazard of the study area. Consequently, implementing both local and non-local models is the best option to address this problem.\u003c/p\u003e\n\u003cp\u003eThe selection of an earthquake catalogue is a crucial step, as it is the basis for extracting events in each source area and conducting subsequent statistical analyses to derive essential seismicity parameters. Therefore, it is imperative to rely on a trustworthy and validated data source. In Greece, the instrumental earthquake catalogue of Makropoulos et al. (2012) is specifically tailored for seismic hazard studies, as it has a threshold magnitude of M\u003csub\u003es\u003c/sub\u003e = 4.0 and M\u003csub\u003ew\u003c/sub\u003e = 4.1. It covers the period from 1900 to 2009, with the significant limitation of not including earthquakes since 2010. This lack of data is noteworthy because, as mentioned in the introduction, events such as those that occurred in 2014 and 2015 would be omitted from the computational framework. To address this issue, we extended the earthquake catalogue to 2019 in a consistent manner, following the same methodology as Makropoulos et al. (2012), i.e. incorporating reviewed events from the Bulletin of the International Seismological Centre (ISC). A common practice in PSHA is to decluster earthquake catalogues in order to retain only the parent earthquakes, removing foreshocks and aftershocks. However, this practice is debatable, as there have been reports suggesting that aftershocks can generate significant ground motions and are often responsible for additional damage (Marzocchi and Taroni, 2014; Taroni and Akinci, 2021). In this study, we opted not to decluster the earthquake catalogue, as this would result in loss of valuable data, especially considering that Cephalonia and Ithaca are regions characterized by frequent earthquake sequences.\u003c/p\u003e\n\u003cp\u003eThe seismicity parameters of each source area required for PSHA encompass the magnitude of completeness (M\u003csub\u003ec\u003c/sub\u003e), the annual rate of Mc exceedance (\u0026lambda;M\u003csub\u003ec\u003c/sub\u003e), the maximum expected magnitude (M\u003csub\u003eu\u003c/sub\u003e), and the b\u0026ndash;value of the Gutenberg\u0026ndash;Richter Frequency\u0026ndash;Magnitude Distribution (FMD) (Gutenberg and Richter, 1944). We herein employ two methods for calculating M\u003csub\u003ec\u003c/sub\u003e and the b\u0026ndash;value. The first is the classical maximum curvature method (MAXC), which was originally introduced by Wiemer and Wyss (2000) and identifies M\u003csub\u003ec\u003c/sub\u003e by pinpointing the maximum curvature of the FMD curve through the calculation of its first derivative\u0026apos;s highest value. Subsequently, the b\u0026ndash;value is determined using the maximum likelihood method introduced by Aki (1965). Overall, MAXC is highly reliable and robust and has undergone validation in several PSHA studies (Zhou et al., 2018; Pavlou et al., 2021; Kaviris et al., 2023). The second method, proposed by Godano and Petrillo (2023), offers a rapid and straightforward estimation of M\u003csub\u003ec\u003c/sub\u003e. It relies on the observation that the Gutenberg\u0026ndash;Richter distribution exhibits an exponential behavior for magnitudes exceeding M\u003csub\u003ec\u003c/sub\u003e and a linear behavior for the smaller ones. Consequently, the average magnitude value (M\u003csub\u003ea\u003c/sub\u003e) exhibits linear increase as the threshold magnitude (M\u003csub\u003eth\u003c/sub\u003e) increases and the deviations from this behavior allow for an accurate M\u003csub\u003ec\u003c/sub\u003e computation, while the linearity of M\u003csub\u003ea\u003c/sub\u003e versus M\u003csub\u003eth\u003c/sub\u003e facilitates the b\u0026ndash;value estimation. In this study, we seize the opportunity to also explore this new method for M\u003csub\u003ec\u003c/sub\u003e and b\u0026ndash;value estimation, aiming to comprehend its advantages and limitations for future assessments. Variability was also considered for the M\u003csub\u003eu\u003c/sub\u003e parameter, because of the absence of a definitive estimation technique. Specifically, three techniques are utilized to estimate M\u003csub\u003eu\u003c/sub\u003e: the first, yielding the lowest possible M\u003csub\u003eu\u003c/sub\u003e, is based on the maximum observed earthquake magnitude (M\u003csub\u003emaxobs\u003c/sub\u003e) within each source area. An intermediate M\u003csub\u003eu\u003c/sub\u003e value is estimated through the Robson\u0026ndash;Whitlock\u0026ndash;Cooke (RWC) technique, as described by Robson and Whitlock (1964) and Cooke (1979). RWC introduces a small positive factor based on the second maximum earthquake magnitude (M\u003csub\u003emax\u003c/sub\u003e\u003csub\u003en\u0026ndash;1\u003c/sub\u003e). The highest potential M\u003csub\u003eu\u003c/sub\u003e is determined by adding the positive factor 0.5 to M\u003csub\u003emaxobs\u003c/sub\u003e. Finally, \u0026lambda;M\u003csub\u003ec\u003c/sub\u003e is computed using the maximum likelihood estimator technique outlined in the works of Kijko and Sellevoll (1989) and Kijko and Smit (2012). Summarizing, a strong variability has been included regarding the adoption of source models and the calculation of seismicity parameters.\u003c/p\u003e\n\u003cp\u003eThe Ground Motion Prediction Equation (GMPE) plays a pivotal role in PSHA by providing estimations of peak ground motions that take into account the earthquake magnitude, distance from the site, focal mechanism type, and soil conditions. It is important to emphasize that the prediction of ground motions resulting from an earthquake may introduce errors, especially in the near\u0026ndash;field, hence GMPEs are empirical relationships that possess notable epistemic uncertainties. In this context, variability is a critical aspect, allowing us to incorporate a range of PGA, PGV, and Sa values for the same independent variables, such as earthquake magnitude and distance. In our study, we employed GMPEs that have undergone rigorous testing, validation, and ranking in recent PSHA studies (Pavlou et al., 2021; Kaviris et al., 2022a, 2022b; Sakkas et al., 2022; Kaviris et al., 2023). For PGA, we utilized the GMPEs proposed by Skarlatoudis et al. (2003) [SKA03], Danciu and Tselentis (2007) [DAT07], Sakkas (2016) [SAK16] and Chousianitis et al. (2018) [CHO18]. For PGV, the GMPEs of Skarlatoudis et al. (2007) [SKA07], Danciu and Tselentis (2007) [DAT07] and Chousianitis et al. (2018) [CHO18] were selected and for Sa the GMPE of Danciu and Tselentis (2007) [DAT07] was chosen. For PSHA purposes, it is essential not to assign to the mentioned GMPEs only one type of focal mechanism (normal or non\u0026ndash;normal) to each source area, which can lead to overestimation or underestimation of peak ground motions. To address this challenge, we utilized the focal mechanism catalogue proposed by Kapetanidis and Kassaras (2019) to determine the accurate percentages of normal and non\u0026ndash;normal focal mechanisms for each source area. We then applied those percents to the selected GMPEs to ensure accurate proportions between the normal and non\u0026ndash;normal versions of each GMPE. This approach allows to avoid the need for interpolating the type of focal mechanism for each source area. The percentages can be found in Table S1.\u003c/p\u003e\n\u003cp\u003eThe concept of logic tree diagrams, originally introduced by Kulkarni et al. (1984), has consistently proven to be a reliable method for mitigating uncertainties (Bommer and Scherbaum, 2008; Atkinson et al., 2014; Marzocchi et al., 2015). In PSHA, a logic tree comprises multiple branches, each representing a potential seismic hazard outcome, reflecting the associated uncertainties. These branches are created to account for different choices that the analyst deems feasible and, to express the level of confidence in each one, every branch is assigned a normalized weight. Within the context of this paper, the logic tree visually represents the steps discussed (Figure 3). The process commences with the source models, where three branches are established, one for each model, with equal weighting. Subsequently, for every source area of each source model, additional branches are introduced to accommodate variables, such as b\u0026ndash;values, M\u003csub\u003ec\u003c/sub\u003e, and \u0026lambda;(M\u003csub\u003ec\u003c/sub\u003e), determined using the MAXC technique and the newly proposed method by Godano and Petrillo (2023). Moreover, variations in M\u003csub\u003eu\u003c/sub\u003e are considered by incorporating three branches, representing low, intermediate, and high M\u003csub\u003eu\u003c/sub\u003e levels. Furthermore, the selected GMPEs are integrated into the analysis, each one associated with a minor logic tree that accounts for the percentages of normal and non-normal focal mechanism types within each source area. It is important to note that this complexity refers to a single source area only. Consequently, the total number of logic tree samples becomes extraordinarily large, akin to the situation in PSHA for Canada (Kolaj et al., 2020). In such cases, it is essential to implement statistical sampling methods, as emphasized by Pagani et al. (2014). For our specific needs, the Latin Hypercube Sampling (LHS) method, initially proposed by McKay et al. (1979), is deemed the most suitable. Unlike the random sampling of the Monte Carlo technique, LHS divides the input ranges into equal intervals, ensuring that only one value is selected from each interval. This approach offers a more systematic and representative way to sample the various branches of the logic tree, thereby capturing the full range of scenarios. The number of samples used for the sampling process was determined through a trial-and-error approach, yielding stable results for PGA, PGV and Sa. It was decided that 10,000 samples would be employed, consistent with the procedure followed in ESHM13 and ESHM20.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003eThe results are presented in the form of PGA and PGV spatial distribution maps for Cephalonia and Ithaca (Figures 4 and 5, respectively), considering return periods of 475 (Figures 4a and 5a) and 950 years (Figures 4b and 5b). Additionally, PGA\u0026ndash;hazard curves were generated for the three most densely populated towns in our study area: Lixouri, Argostoli, and Vathi (Figure 6a). Those were produced to demonstrate PGA values for a wide range of return periods. Moreover, for the same sites, Uniform Hazard Spectra (UHS) were constructed by utilizing Sa levels, which correspond to their natural periods (Figure 6b).\u003c/p\u003e\n\u003cp\u003eBased on the spatial distribution of both PGA and PGV for return periods of 475 and 950 years, the lowest values are observed in the southeastern portion of the study area. As we move towards the north and west, the values increase, reaching their peak mainly in Paliki and a small area east of the Myrtos Gulf. The pattern observed may be attributed mainly to the small distance of Paliki peninsula from the Cephalonia segment of the CLTFZ, characterized by high seismicity, as well as to the onshore faulting system, where the 2014 Cephalonia earthquakes occurred. The peak near Myrtos Gulf, where intermediate magnitude events have occurred due to the activation of secondary smaller structures transverse to the axis of the CLTFZ (Sakkas et al., 2022), may be aleatoric due to the small contour area.\u003c/p\u003e\n\u003cp\u003eThe lowest PGA value for return period of 475 years is approximately 460 cm/s\u0026sup2;, and the highest around 580 cm/s\u0026sup2; (Figure 4a). The PGA range is about 120 cm/s\u0026sup2;, with the most intense variation occurring along the southeastern edge of Cephalonia. This indicates that intermediate to high PGA values are prevalent throughout most of the study area. Similar observations can be made for the return period of 950 years (Figure 4b), where the difference between the highest (around 700 cm/s\u0026sup2;) and the lowest (approximately 560 cm/s\u0026sup2;) PGA is about 140 cm/s\u0026sup2;. Again, the high PGA variation mainly occurs along the south easternmost edge of the study area.\u003c/p\u003e\n\u003cp\u003eRegarding PGV, for a return period of 475 years (Figure 5a), some differences can be noticed compared to PGA. In particular, the spatial distribution is smoother, with the lowest PGV value approximately 33 cm/s, while the highest is around 36 cm/s. This relative stability in PGV values can be attributed to the smaller number of GMPEs used for PGV computation, as SAK16 does not propose a PGV model. This implies that the logic tree has significantly fewer total branches, resulting in a lower complexity of the spatial distribution. Continuing to the results of PGV for the return period of 950 years (Figure 5b), we find a similar situation. The highest value is ~49 cm/s, and the lowest is about 45 cm/s, with a small deviation of approximately 4 cm/s.\u003c/p\u003e\n\u003cp\u003eContinuing to the site\u0026ndash;specific analysis and the PGA\u0026ndash;hazard curves (Figure 6a), we can observe the PGA levels across a range of probabilities of exceedance in 50 years. Lixouri and Argostoli exhibit similar values, with Lixouri\u0026apos;s curve being slightly higher than that of Argostoli by approximately 10 cm/s\u0026sup2; for all probabilities of exceedance. This small difference can possibly be attributed to their distance from the CLTFZ, as Lixouri is closer to the fault than Argostoli. Vathi displays the lowest hazard curve, which aligns with the spatial distribution of PGA values for both return periods, as Vathi is situated in a region characterized by intermediate PGA values. It is worth noting that even for the highest presented return periods, ground motions do not exceed 1 g.\u003c/p\u003e\n\u003cp\u003eRegarding the UHS for the same towns (Figure 6b), Lixouri exhibits the highest Sa levels compared to Argostoli and Vathi across the entire range of natural periods. This observation is in agreement with the previously mentioned results. Furthermore, the UHS provides information about the natural period of the single\u0026ndash;degree freedom system that experiences the highest Sa value, which in our case is 0.25 s. However, it is essential to acknowledge a slightly lower peak at 0.45 s, which should be considered, especially for Lixouri which experiences Sa values that are nearly identical for these two periods.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe primary objective of this study is to conduct a reassessment of seismic hazard in Cephalonia and Ithaca by involving the integration of new statistical techniques aimed at reducing epistemic uncertainties related to source models, seismicity data and GMPEs. The first set of results pertains to the spatial distribution of PGA and PGV for return periods 475 and 950 years across Cephalonia and Ithaca and the second one focuses exclusively on the three most densely populated localities, namely Lixouri, Argostoli, and Vathi. For these towns, PGA hazard curves and UHS were developed to illustrate the variation in PGA values over a wide range of probabilities of exceedance and the Sa distribution across various natural periods, respectively.\u003c/p\u003e\n\u003cp\u003eThe study suggests that the area of higher seismic hazard is the Paliki peninsula in western Cephalonia. This is mainly due to the vicinity of Paliki to the Cephalonia segment of the CLTFZ, maybe the most seismically active structure in Greece. In addition, high PGA values in Paliki are also influenced by the onshore local faults related with the 2014 Cephalonia earthquakes (M\u003csub\u003ew\u003c/sub\u003e=6.1 and M\u003csub\u003ew\u003c/sub\u003e=5.8), which also caused local ground deformation (Sakkas et al., 2022). Therefore, future infrastructure or seismic retrofitting in Paliki require special attention. Moreover, the area east of Myrtos Gulf has a high level of seismic hazard, although this finding may contain a considerable level of uncertainty. In the work of Sakkas et al. (2022), it was demonstrated that the post-seismic activity of the 2014 earthquake sequence primarily migrated northward, with clusters also located within the Myrtos Gulf. The site-specific results show that Lixouri has the highest PGA hazard curve and UHS, while Argostoli has intermediate curves, similar with those of Lixouri. In contrast, Vathi has the lowest maximum expected ground motions among the towns. Lixouri\u0026apos;s proximity to the CLTFZ, in comparison to Argostoli and Vathi, may be the reason for this distribution.\u003c/p\u003e\n\u003cp\u003eThe findings for both return periods offer valuable insights for structural design and engineering purposes. Specifically, the information obtained by the spatial distribution of PGA and PGV serves as a crucial reference for engineers, enabling them to design infrastructure capable of withstanding the maximum anticipated ground motions. Furthermore, for existing buildings, the results can guide seismic retrofitting efforts by developing a more accurate seismic risk assessment of the area, ensuring that they meet safety standards. Such practices are essential in regions characterized by high seismic activity, as they aid to ensure the resilience of structures against infrequent but potentially destructive earthquakes. The outcomes of the site\u0026ndash;specific analysis for the three towns offer valuable insights into the anticipated maximum ground motions over a 50\u0026ndash;year timeframe, taking into account varying probabilities of occurrence. Additionally, these findings help in identifying measures to prevent resonance phenomena linked to the prevailing soil period (which was at 0.25 s).\u003c/p\u003e\n\u003cp\u003eAs previously mentioned, there have been several studies conducted to assess seismic hazard in Cephalonia and Ithaca. The one of Bonatis (2020) employed the same PSHA methodology as the one outlined here. However, differences exist between the source models and GMPEs used in the herein proposed PSHA. Specifically, Bonatis (2020) utilized a single seismotectonic model, BON20, and calculated PGA for a 475\u0026ndash;year return period using various GMPEs, which, however, were not combined through a logic tree technique. The PGA results ranged from 200 to 900 cm/s\u003csup\u003e2\u003c/sup\u003e, while the herein obtained values are in a much narrower range. Nonetheless, the spatial distribution of PGA remained consistent, particularly regarding the high variability observed at the southern edge of Cephalonia. In this case, the PGA values ranged from 50 to 450 cm/s\u003csup\u003e2\u003c/sup\u003e, and PGV from 0 to 25 cm/s. However, it is important to note that these results cannot be directly compared to those of the current study, given the fundamental differences in the methodologies employed. In their recent work, Sakkas et al. (2022) used the ESHM13 source model and the GMPE developed by Danciu and Tselentis (2007), only for non-normal focal mechanisms. The MAXC method was used to obtain their seismicity parameters. Therefore, due to differences in the preprocessing part, variations in the results are expected between their work and this study. The logic tree used by Sakkas et al. (2022) had significantly fewer branches compared to the one used in this PSHA, resulting in a more smoothed spatial distribution of PGA. In their computational grid, Sakkas et al. (2022) included all the Ionian Islands. Therefore, for visualization clarity, the values in Cephalonia and Ithaca are depicted as a single value of approximately 500 cm/s\u003csup\u003e2\u003c/sup\u003e for a return period of 475 years.\u003c/p\u003e\n\u003cp\u003e\u0026Tau;he present PSHA was performed aiming to reduce the epistemic uncertainties, lowering them when compared to the aforementioned studies. This was achieved via a sophisticated logic tree approach and the application of a reliable sampling technique. The logic tree contains a very large number of different seismic hazard outcomes, considering the sublogic trees employed for each source area. In the process of estimating the b-value and Mc, two methods were considered, one of which is the new technique proposed by Godano and Petrillo (2023). There is a limitation in this, as it tends to estimate higher Mc values compared to other techniques (for instance, the MAXC that was also utilized in this study). The generation of high Mc values could potentially result in a limited number of data points for the estimation of the b-value, thereby leading to higher uncertainty in the regression model. The lack of a reliable b-value is a considerable drawback, given its crucial role in characterizing the seismicity for each source area. Nevertheless, the technique of Godano and Petrillo (2023) can be chosen in regions of high seismicity, where an adequate number of earthquakes certainly exists. Consequently, this method can be selected for the region of Cephalonia and Ithaca. Given the high degree of uncertainty regarding the Mu, we chose to designate three levels for each source area (low, intermediate, and high) in order to capture a broad range of seismic hazard outcomes related to this aspect. Furthermore, by incorporating precise weights into both the normal and non-normal versions of each GMPE at each source zone, we were able to manage the uncertainty related to the extrapolation of the focal mechanism. This is a critical step in preventing the overestimation and underestimation of the maximum expected ground motions. It is worth noting that a reliable sampling technique was required for this extensive logic tree. The LHS was chosen due to its non-memoryless nature, which is particularly significant in this context, as it effectively samples a considerable portion of the entire distribution of logic tree branches.\u003c/p\u003e\n\u003cp\u003eThis PSHA has certain limitations, for instance, it does not account for soil conditions, which can influence the results through amplification or attenuation phenomena. In the absence of information on true soil conditions, the study of Allen and Wald (2009), which relies only on topographic data can be used as a proxy for seismic site conditions, but it was not chosen for this PSHA in order to avoid additional potential uncertainties. Another limitation is the absence of a GMPE ranking system in the analysis, as it could provide valuable insights into which one of the selected empirical models best matches the recorded strong motion values and their relative weights for inclusion in the logic tree approach. However, it is worth noting that a recent study by Kaviris et al. (2023) conducted a GMPE ranking system for the same GMPEs as those selected in the herein computational framework, and the results indicated that the relative weights were very close to each other. This suggests that the PGA, PGV and Sa obtained here may not significantly differ from those generated using a non-equal-weighted logic tree approach.\u003c/p\u003e\n\u003cp\u003eFuture research endeavors could involve conducting fieldwork to capture ambient noise in three spatial dimensions at various sites. This data collection would facilitate the determination of the fundamental resonance frequency of the ground across Cephalonia and Ithaca using a spatial grid. Understanding this parameter is of great significance in earthquake engineering, as it provides valuable insights into site\u0026ndash;specific effects. Another avenue for exploration is the development of a new GMPE that specifically predicts PGA and PGV for the vertical component of ground motion. Such information would be particularly beneficial for the construction of bridges, especially in regions with high seismic activity. Furthermore, a seismic risk assessment could be undertaken, exploiting the herein proposed PSHA as an input to determine the maximum expected ground motions.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThe aim of this study is to re\u0026ndash;evaluate the seismic hazard for Cephalonia and Ithaca, taking into account new data and statistical methodologies. It is recognized that seismic hazard assessment is a field characterized by significant epistemic uncertainty, especially in terms of source models, seismicity parameters, and GMPEs (Cornell, 1968). Therefore, in regions like the one of the present work, which are subject to high seismic hazard, as highlighted by the current National Building Code (EAK, 2003), it is essential to update the maximum expected ground motions in order to have a reliable input for future seismic risk studies (Cornell, 1968).\u003c/p\u003e\n\u003cp\u003eA well stablished and dependable approach to addressing epistemic uncertainty involves the use of a comprehensive logic tree decision graph. This method generates various seismic hazard outcomes, all taken into account in the final results. The most prevalent sources of bias and uncertainty are the seismotectonic models, seismicity parameters and GMPEs. In the herein proposed PSHA, we incorporate three groups of source areas, i.e., two European and one local for Cephalonia and Ithaca. These models are considered with equal weights in the computational schemes due to the lack of specific criteria indicating one model\u0026apos;s superiority over another. Significant attention is given to the computation of seismicity parameters, where two methods were implemented for calculating the pair of Mc and b\u0026ndash;value, and three for estimating Mu. Due to the absence of adequate criteria, we do not assign different weights among these techniques; they are treated as equal in the logic tree. The most critical aspect is addressing the uncertainty regarding GMPEs, as these empirical relations predict the maximum expected ground motions (PGA, PGV, and Sa). We select several GMPEs that consider the epicentral type of distance, since we do not have faults, but source areas, and the type of focal mechanism. In PSHA, it is common to extrapolate the focal mechanism type for each source area, leading to the selection of GMPEs for one specific faulting type only. In this updated PSHA, we create a sub\u0026ndash;logic tree for each GMPE that considers the relative percentages of normal and non\u0026ndash;normal types of focal mechanisms. When considering all possible seismic hazard outcomes, it becomes clear that the final number of branches is large, making a complete enumeration almost impossible. Therefore, we opt for a sampling of the logic tree using reliable techniques to capture a significant portion of it. Consequently, we believe that this study\u0026apos;s results have thus far achieved the lowest degree of uncertainty.\u003c/p\u003e\n\u003cp\u003eThe findings of this study are presented as spatial distributions of PGA and PGV for return periods of 475 and 950 years, along with site\u0026ndash;specific results for Lixouri, Argostoli, and Vathi (PGA\u0026ndash;hazard curves and UHS). The highest ground motions are observed in the western portions of Cephalonia and Ithaca, with significant variability in the southeastern edge of Cephalonia. Among the towns studied, Lixouri exhibits the highest level of seismic hazard, while Vathi the lowest. All three towns have a dominant frequency of 0.25 s. The PGA results for the first return period can be compared to the reference value proposed by EAK (2003), which divides Greece into three seismic hazard zones, each being attributed a specific PGA for bedrock conditions. For Cephalonia and Ithaca, this reference value is approximately 360 cm/s\u0026sup2;, the highest in the country. However, the maximum PGA from our proposed PSHA model is 580 cm/s\u0026sup2;, underscoring the need to update the reference value. The results for the second return period can aid the construction of critical structures, such as medical centers. The site\u0026ndash;specific analysis results provide insights on the estimated fundamental resonance frequency and the maximum expected spectral acceleration.\u003c/p\u003e"},{"header":"6. References","content":"\u003col\u003e\n\u003cli\u003eAki, K., 1965. Maximum likelihood estimate of b in the formula log N = a-bM and its confidence limits. Bull. Earthq. Res. Inst. 43, 237\u0026ndash;239.\u003c/li\u003e\n\u003cli\u003eAlbini, P., Locati, M., Rovida, A., Stucchi, M., 2013. European Archive of Historical EArthquake Data (AHEAD). Istituto Nazionale di Geofisica e Vulcanologia (INGV). https://doi.org/10.6092/ingv.it-ahead\u003c/li\u003e\n\u003cli\u003eAllen, T.I., Wald, D.J., 2009. On the Use of High-Resolution Topographic Data as a Proxy for Seismic Site Conditions (VS30). Bulletin of the Seismological Society of America 99, 935\u0026ndash;943. https://doi.org/10.1785/0120080255\u003c/li\u003e\n\u003cli\u003eAnderson, H., Jackson, J., 1987. Active tectonics of the Adriatic Region. Geophys. J. Int. 91, 937\u0026ndash;983. https://doi.org/10.1111/j.1365-246X.1987.tb01675.x\u003c/li\u003e\n\u003cli\u003eAtkinson, G.M., Bommer, J.J., Abrahamson, N.A., 2014. Alternative Approaches to Modeling Epistemic Uncertainty in Ground Motions in Probabilistic Seismic‐Hazard Analysis. Seismological Research Letters 85, 1141\u0026ndash;1144. https://doi.org/10.1785/0220140120\u003c/li\u003e\n\u003cli\u003eBommer, J.J., Scherbaum, F., 2008. The Use and Misuse of Logic Trees in Probabilistic Seismic Hazard Analysis. Earthquake Spectra 24, 997\u0026ndash;1009. https://doi.org/10.1193/1.2977755\u003c/li\u003e\n\u003cli\u003eBonatis, P., 2020. Strong ground motion simulation in the Central Ionian Islands using a hybrid (deterministic and stochastic) approach. Master Thesis, School of Geology, Aristotle University of Thessaloniki. 129.\u003c/li\u003e\n\u003cli\u003eBonatis, P., Akinci, A., Karakostas, V., Papadimitriou, E., Kaviris, G., 2021. Near-Fault Broadband Ground Motion Simulation Applications at the Central Ionian Islands, Greece. Pure Appl. Geophys. 178, 3505\u0026ndash;3527. https://doi.org/10.1007/s00024-021-02825-9\u003c/li\u003e\n\u003cli\u003eChousianitis, K., Del Gaudio, V., Pierri, P., Tselentis, G.-A., 2018. Regional ground-motion prediction equations for amplitude-, frequency response-, and duration-based parameters for Greece. Earthquake Engng Struct Dyn 47, 2252\u0026ndash;2274. https://doi.org/10.1002/eqe.3067\u003c/li\u003e\n\u003cli\u003eCooke, P., 1979. Statistical inference for bounds of random variables. Biometrika 66, 367\u0026ndash;374. https://doi.org/10.1093/biomet/66.2.367\u003c/li\u003e\n\u003cli\u003eCornell, C., 1968. Engineering Seismic Risk Analysis. Bulletin of the Seismological Society of America. 58, 1583\u0026ndash;1606.\u003c/li\u003e\n\u003cli\u003eDanciu, L., Nandan, S., Reyes, C., Basili, R., Weatherill, G., Beauval, C., Rovida, A., Vilanova, S., Sesetyan, K., Bard, P.-Y., Cotton, F., Wiemer, S., Giardini, D., 2021. ESHM20 - EFEHR Technical ReportThe 2020 update of the European Seismic Hazard Model - ESHM20: Model Overview. EFEHR European Facilities of Earthquake Hazard and Risk. https://doi.org/10.12686/A15\u003c/li\u003e\n\u003cli\u003eDanciu, L., Tselentis, G.-A., 2007. Engineering Ground-Motion Parameters Attenuation Relationships for Greece. Bulletin of the Seismological Society of America 97, 162\u0026ndash;183. https://doi.org/10.1785/0120050087\u003c/li\u003e\n\u003cli\u003eEAK, 2003. Greek seismic code edited by: Earthquake planning and protection organization. Athens, Greece.\u003c/li\u003e\n\u003cli\u003eGanas, A., Oikonomou, I.A., Tsimi, C., 2013a. NOAfaults: a digital database for active faults in Greece. geosociety 47, 518\u0026ndash;530. https://doi.org/10.12681/bgsg.11079\u003c/li\u003e\n\u003cli\u003eGanas, A., Marinou, A., Anastasiou, D., Paradissis, D., Papazissi, K., Tzavaras, P., Drakatos, G., 2013b. GPS-derived estimates of crustal deformation in the central and north Ionian Sea, Greece: 3-yr results from NOANET continuous network data. J. Geodyn. 67, 62\u0026ndash;71. https://doi.org/10.1016/j.jog.2012.05.010\u003c/li\u003e\n\u003cli\u003eGanas, A., 2023. NOAFAULTS KMZ layer Version 5.0 (V5.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.8075517\u003c/li\u003e\n\u003cli\u003eGodano, C., Petrillo, G., 2023. Estimating the Completeness Magnitude mc and the b-Values in a Snap. Earth and Space Science 10, e2022EA002540. https://doi.org/10.1029/2022EA002540\u003c/li\u003e\n\u003cli\u003eGumbel, E., Lieblein, J., 1954. Statistical Theory of Extreme Values and Some Practical Applications: A Series of Lectures. National Bureau of Standards, US Government Printing Office, Washington. 33.\u003c/li\u003e\n\u003cli\u003eGutenberg, B., Richter, C.F., 1944. Frequency of earthquakes in California. Bulletin of the Seismological society of America 34, 185\u0026ndash;188.\u003c/li\u003e\n\u003cli\u003eKapetanidis, V., Kassaras, I., 2019. Contemporary crustal stress of the Greek region deduced from earthquake focal mechanisms. Journal of Geodynamics 123, 55\u0026ndash;82. https://doi.org/10.1016/j.jog.2018.11.004\u003c/li\u003e\n\u003cli\u003eKarakostas, V., Papadimitriou, E., Mesimeri, M., Gkarlaouni, C., Paradisopoulou, P., 2015. The 2014 Kefalonia Doublet (MW6.1 and MW6.0), Central Ionian Islands, Greece: Seismotectonic Implications along the Kefalonia Transform Fault Zone. Acta Geophys. 63, 1\u0026ndash;16. https://doi.org/10.2478/s11600-014-0227-4\u003c/li\u003e\n\u003cli\u003eKarastathis, V.K., Mouzakiotis, E., Ganas, A., Papadopoulos, G.A., 2015. High-precision relocation of seismic sequences above a dipping Moho: the case of the January\u0026ndash;February 2014 seismic sequence on Cephalonia island (Greece). Solid Earth 6, 173\u0026ndash;184. https://doi.org/10.5194/se-6-173-2015\u003c/li\u003e\n\u003cli\u003eKassaras, I., Papadimitriou, P., Kapetanidis, V., Voulgaris, N., 2017. Seismic site characterization at the western Cephalonia Island in the aftermath of the 2014 earthquake series. Geo-Engineering 8, 7. https://doi.org/10.1186/s40703-017-0045-z\u003c/li\u003e\n\u003cli\u003eKaviris, G., Zymvragakis, A., Bonatis, P., Kapetanidis, V., Spingos, I., Mavroulis, S., Kotsi, E., Lekkas, E., Voulgaris, N., 2023. A Logic-Tree Approach for Probabilistic Seismic Hazard Assessment in the Administrative Region of Attica (Greece). Applied Sciences 13. https://doi.org/10.3390/app13137553\u003c/li\u003e\n\u003cli\u003eKaviris, G., Zymvragakis, A., Bonatis, P., Kapetanidis, V., Voulgaris, N., 2022a. Probabilistic and Scenario-Based Seismic Hazard Assessment on the Western Gulf of Corinth (Central Greece). Applied Sciences 12. https://doi.org/10.3390/app122111152\u003c/li\u003e\n\u003cli\u003eKaviris, G., Zymvragakis, A., Bonatis, P., Sakkas, G., Kouskouna, V., Voulgaris, N., 2022b. Probabilistic Seismic Hazard Assessment for the Broader Messinia (SW Greece) Region. Pure and Applied Geophysics 179, 551\u0026ndash;567. https://doi.org/10.1007/s00024-022-02950-z\u003c/li\u003e\n\u003cli\u003eKijko, A., Sellevoll, M.A., 1989. Estimation of earthquake hazard parameters from incomplete data files. Part I. Utilization of extreme and complete catalogs with different threshold magnitudes. Bulletin of the Seismological Society of America 79, 645\u0026ndash;654. https://doi.org/10.1785/BSSA0790030645\u003c/li\u003e\n\u003cli\u003eKijko, A., Smit, A., 2012. Extension of the Aki‐Utsu b‐Value Estimator for Incomplete Catalogs. Bulletin of the Seismological Society of America 102, 1283\u0026ndash;1287. https://doi.org/10.1785/0120110226\u003c/li\u003e\n\u003cli\u003eKolaj, M., Adams, J., Halchuk, S., 2020. The 6th generation seismic hazard model of Canada, in: 17th World Conference on Earthquake Engineering. pp. 1\u0026ndash;12.\u003c/li\u003e\n\u003cli\u003eKulkarni, R., Youngs, R., Coppersmith, K., 1984. Assessment of confidence intervals for results of seismic hazard analysis, in: Proceedings of the Eighth World Conference on Earthquake Engineering. San Francisco, pp. 263\u0026ndash;270.\u003c/li\u003e\n\u003cli\u003eLagios, E., Sakkas, V., Papadimitriou, P., Parcharidis, I., Damiata, B.N., Chousianitis, K., Vassilopoulou, S., 2007. Crustal deformation in the Central Ionian Islands (Greece): Results from DGPS and DInSAR analyses (1995\u0026ndash;2006). Tectonophysics 444, 119\u0026ndash;145. https://doi.org/10.1016/j.tecto.2007.08.018\u003c/li\u003e\n\u003cli\u003eLekkas, E.L., Mavroulis, S.D., 2016. Fault zones ruptured during the early 2014 Cephalonia Island (Ionian Sea, Western Greece) earthquakes (January 26 and February 3, Mw 6.0) based on the associated co-seismic surface ruptures. J. Seismol. 20, 63\u0026ndash;78. https://doi.org/10.1007/s10950-015-9510-3\u003c/li\u003e\n\u003cli\u003eMakropoulos, K., Kaviris, G., Kouskouna, V., 2012. An updated and extended earthquake catalogue for Greece and adjacent areas since 1900. Natural Hazards and Earth System Sciences 12, 1425\u0026ndash;1430. https://doi.org/10.5194/nhess-12-1425-2012\u003c/li\u003e\n\u003cli\u003eMakropoulos, K.C., Burton, P.W., 1985. Seismic hazard in Greece. II. Ground acceleration. Tectonophysics 117, 259\u0026ndash;294. https://doi.org/10.1016/0040-1951(85)90274-4\u003c/li\u003e\n\u003cli\u003eMarzocchi, W., Taroni, M., 2014. Some Thoughts on Declustering in Probabilistic Seismic‐Hazard Analysis. Bulletin of the Seismological Society of America 104, 1838\u0026ndash;1845. https://doi.org/10.1785/0120130300\u003c/li\u003e\n\u003cli\u003eMarzocchi, W., Taroni, M., Selva, J., 2015. Accounting for Epistemic Uncertainty in PSHA: Logic Tree and Ensemble Modeling. Bulletin of the Seismological Society of America 105, 2151\u0026ndash;2159. https://doi.org/10.1785/0120140131\u003c/li\u003e\n\u003cli\u003eMavroulis, S., Lekkas, E., 2021. Revisiting the Most Destructive Earthquake Sequence in the Recent History of Greece: Environmental Effects Induced by the 9, 11 and 12 August 1953 Ionian Sea Earthquakes. Appl. Sci. 11, 8429. https://doi.org/10.3390/app11188429\u003c/li\u003e\n\u003cli\u003eMcGuire, R.K., 1976. FORTRAN computer program for seismic risk analysis (Report No. 76\u0026ndash;67), Open-File Report. https://doi.org/10.3133/ofr7667\u003c/li\u003e\n\u003cli\u003eMcKay, D., Beckman, R., Conover, W., 1979. A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code. Technometrics 21, 239\u0026ndash;245. https://doi.org/10.2307/1268522\u003c/li\u003e\n\u003cli\u003eMcKenzie, D., 1972. Active Tectonics of the Mediterranean Region. Geophys. J. R. Astron. Soc. 30, 109\u0026ndash;185. https://doi.org/10.1111/j.1365-246X.1972.tb02351.x \u003c/li\u003e\n\u003cli\u003ePagani, M., Monelli, D., Weatherill, G., Danciu, L., Crowley, H., Silva, V., Henshaw, P., Butler, L., Nastasi, M., Panzeri, L., Simionato, M., Vigano, D., 2014. OpenQuake Engine: An Open Hazard (and Risk) Software for the Global Earthquake Model. Seismological Research Letters 85, 692\u0026ndash;702. https://doi.org/10.1785/0220130087\u003c/li\u003e\n\u003cli\u003ePapadimitriou, E., Karakostas, V., Mesimeri, M., Chouliaras, G., Kourouklas, C., 2017. The Mw6.5 17 November 2015 Lefkada (Greece) Earthquake: Structural Interpretation by Means of the Aftershock Analysis. Pure Appl. Geophys. 174, 3869\u0026ndash;3888. https://doi.org/10.1007/s00024-017-1601-3\u003c/li\u003e\n\u003cli\u003ePapadimitriou, P., 1988, Etude de la structure du manteau superieur de l\u0026rsquo;Europe et modelisation des ondes de volume engendrees par des seismes Egeens, PhD thesis. University of Paris, Paris.\u003c/li\u003e\n\u003cli\u003ePapadimitriou, P., Kaviris, G., Makropoulos, K., 2006. The Mw = 6.3 2003 Lefkada Earthquake (Greece) and induced stress transfer changes. Tectonophysics 423, 73\u0026ndash;82. https://doi.org/10.1016/j.tecto.2006.03.003\u003c/li\u003e\n\u003cli\u003ePapadimitriou, P., Voulgaris, N., Kouskouna, V., Kassaras, I., Kaviris, G., Pavlou, K., Karakonstantis, A., Bozionelos, G., Kapetanidis, V., 2014. The Kefallinia Island earthquake sequence January\u0026ndash;February 2014. In Proceedings of the Second European Conference on Earthquake Engineering and Seismology (2ECEES), Istanbul, Turkey, 24\u0026ndash;29 August 2014.\u003c/li\u003e\n\u003cli\u003ePapadimitriou, P., Kapetanidis, V., Karakonstantis, A., Spingos, I., Pavlou, K., Kaviris, G., Kassaras, I., Sakkas, V., Voulgaris, N., 2021. The 25 October, 2018 Zakynthos (Greece) earthquake: seismic activity at the transition between a transform fault and a subduction zone. Geophys. J. Int. 225, 15\u0026ndash;36. https://doi.org/10.1093/gji/ggaa575\u003c/li\u003e\n\u003cli\u003ePapazachos, B.C., Papazachou, C.B., 2003. The Earthquakes of Greece. Ziti Publications, Thessaloniki, 273 p. (In Greek).\u003c/li\u003e\n\u003cli\u003ePavlou, K., \u0026Kappa;aviris, G., Kouskouna, V., Sakkas, G., Zymvragakis, A., Sakkas, V., Drakatos, G., 2021. Minor seismic hazard changes in the broader area of Pournari artificial lake after the first filling (W. Greece). Results in Geophysical Sciences 100025. https://doi.org/10.1016/j.ringps.2021.100025\u003c/li\u003e\n\u003cli\u003eRobson, D.S., Whitlock, J.H., 1964. Estimation of a truncation point. Biometrika 51, 33\u0026ndash;39. https://doi.org/10.1093/biomet/51.1-2.33\u003c/li\u003e\n\u003cli\u003eSakkas, G., 2016. Calculation and analysis of the seismic motion rotational components in Greece, Ph.D. Thesis. Geophysics-Geothermics Department, Faculty of Geology, University of Athens, Greece (in Greek). 278. http://dx.doi.org/10.12681/eadd/39773\u003c/li\u003e\n\u003cli\u003eSakkas, G., Kouskouna V., Makropoulos, K., 2010. Seismic hazard analysis in the Ionian Islands using macroseismic intensities. Hell. J. Geosci. 45, 239\u0026ndash;248\u003c/li\u003e\n\u003cli\u003eSakkas, V., Lagios, E., 2015. Fault modelling of the early-2014 ~M6 Earthquakes in Cephalonia Island (W. Greece) based on GPS measurements. Tectonophysics 644\u0026ndash;645, 184\u0026ndash;196. https://doi.org/10.1016/j.tecto.2015.01.010\u003c/li\u003e\n\u003cli\u003eSakkas, V., Kapetanidis, V., Kaviris, G., Spingos, I., Mavroulis, S., Diakakis, M., Alexopoulos, J.D., Kazantzidou-Firtinidou, D., Kassaras, I., Dilalos, S., Vassilakis, E., Kotsi, E., Tselentis, G., Lekkas, E., Voulgaris, N., 2022. Seismological and Ground Deformation Study of the Ionian Islands (W. Greece) during 2014\u0026ndash;2018, a Period of Intense Seismic Activity. Applied Sciences 12. https://doi.org/10.3390/app12052331\u003c/li\u003e\n\u003cli\u003eScordilis, E.M., Karakaisis, G.F., Karacostas, B.G., Panagiotopoulos, D.G., Comninakis, P.E., Papazachos, B.C., 1985. Evidence for transform faulting in the Ionian sea: The Cephalonia island earthquake sequence of 1983. Pure Appl. Geophys. PAGEOPH 123, 388\u0026ndash;397. https://doi.org/10.1007/BF00880738\u003c/li\u003e\n\u003cli\u003eSkarlatoudis, A.A., Papazachos, C.B., Margaris, B.N., Theodulidis, N., Papaioannou, Ch., Kalogeras, I., Scordilis, E.M., Karakostas, V., 2007. Erratum to Empirical Peak Ground-Motion Predictive Relations for Shallow Earthquakes in Greece. Bulletin of the Seismological Society of America 97, 2219\u0026ndash;2221. https://doi.org/10.1785/0120070176\u003c/li\u003e\n\u003cli\u003eSkarlatoudis, A.A., Papazachos, C.B., Margaris, B.N., Theodulidis, N., Papaioannou, Ch., Kalogeras, I., Scordilis, E.M., Karakostas, V., 2003. Empirical Peak Ground-Motion Predictive Relations for Shallow Earthquakes in Greece. Bulletin of the Seismological Society of America 93, 2591\u0026ndash;2603. https://doi.org/10.1785/0120030016Stiros, S.C., Pirazzoli, P.A., Laborel, J., Laborel-Deguen, F., 1994. The 1953 earthquake in Cephalonia (Western Hellenic Arc): coastal uplift and halotectonic faulting. Geophys. J. Int. 117, 834\u0026ndash;849. https://doi.org/10.1111/j.1365-246X.1994.tb02474.x\u003c/li\u003e\n\u003cli\u003eSokos, E., Kiratzi, A., Gallovič, F., Zahradn\u0026iacute;k, J., Serpetsidaki, A., Plicka, V., Jansk\u0026yacute;, J., Kosteleck\u0026yacute;, J., Tselentis, G.A., 2015. Rupture process of the 2014 Cephalonia, Greece, earthquake doublet (Mw6) as inferred from regional and local seismic data. Tectonophysics 656, 131\u0026ndash;141. https://doi.org/10.1016/j.tecto.2015.06.013\u003c/li\u003e\n\u003cli\u003eStucchi, M., Rovida, A., Gomez Capera, A.A., Alexandre, P., Camelbeeck, T., Demircioglu, M.B., Gasperini, P., Kouskouna, V., Musson, R.M.W., Radulian, M., Sesetyan, K., Vilanova, S., Baumont, D., Bungum, H., F\u0026auml;h, D., Lenhardt, W., Makropoulos, K., Martinez Solares, J.M., Scotti, O., Živčić, M., Albini, P., Batllo, J., Papaioannou, C., Tatevossian, R., Locati, M., Meletti, C., Vigan\u0026ograve;, D., Giardini, D., 2013. The SHARE European Earthquake Catalogue (SHEEC) 1000-1899. J. Seismol. 17, 523\u0026ndash;544. https://doi.org/10.1007/s10950-012-9335-2\u003c/li\u003e\n\u003cli\u003eTaroni, M., Akinci, A., 2021. Good practices in PSHA: declustering, b-value estimation, foreshocks and aftershocks inclusion; a case study in Italy. Geophysical Journal International 224, 1174\u0026ndash;1187. https://doi.org/10.1093/gji/ggaa462\u003c/li\u003e\n\u003cli\u003eUnderhill, J.R., 1989. Late Cenozoic deformation of the Hellenide foreland, western Greece. Geol. Soc. Am. Bull. 101, 613\u0026ndash;634. https://doi.org/10.1130/0016-7606(1989)101\u0026lt;0613:LCDOTH\u0026gt;2.3.CO;2\u003c/li\u003e\n\u003cli\u003eWang, Z., 2005. A Clear Definition of Seismic Hazard and Risk: A Basis for Hazard and Risk Assessment, Communication, and Management. Presented at the AGU Fall Meeting Abstracts, pp. S53B-1110.\u003c/li\u003e\n\u003cli\u003eWiemer, S., Wyss, M., 2000. Minimum Magnitude of Completeness in Earthquake Catalogs: Examples from Alaska, the Western United States, and Japan. Bulletin of the Seismological Society of America 90, 859\u0026ndash;869. https://doi.org/10.1785/0119990114\u003c/li\u003e\n\u003cli\u003eWoessner, J., Laurentiu, D., Giardini, D., Crowley, H., Cotton, F., Gr\u0026uuml;nthal, G., Valensise, G., Arvidsson, R., Basili, R., Demircioglu, M.B., Hiemer, S., Meletti, C., Musson, R.W., Rovida, A.N., Sesetyan, K., Stucchi, M., The SHARE Consortium, 2015. The 2013 European Seismic Hazard Model: key components and results. Bulletin of Earthquake Engineering 13, 3553\u0026ndash;3596. https://doi.org/10.1007/s10518-015-9795-1\u003c/li\u003e\n\u003cli\u003eZhou, Y., Zhou, S., Zhuang, J., 2018. A test on methods for MC estimation based on earthquake catalog. Earth and Planetary Physics 2, 150\u0026ndash;162. https://doi.org/10.26464/epp2018015\u003c/li\u003e\n\u003c/ol\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":"journal-of-seismology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jose","sideBox":"Learn more about [Journal of Seismology](http://link.springer.com/journal/10950)","snPcode":"10950","submissionUrl":"https://submission.nature.com/new-submission/10950/3","title":"Journal of Seismology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"PSHA, PGA, PGV, UHS, Sa, Logic tree, Epistemic uncertainty","lastPublishedDoi":"10.21203/rs.3.rs-3991269/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3991269/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe Central Ionian Islands of Cephalonia and Ithaca belong to the most seismically active Greek region, mainly due to the presence of the dextral Cephalonia-Lefkada Transform Fault Zone. The study area has experienced strong earthquakes in the 20\u003csup\u003eth\u003c/sup\u003e century, including the destructive 1953 sequence with maximum intensity 9.0. The Paliki peninsula, western Cephalonia, hosted two strong earthquakes (M\u003csub\u003ew\u003c/sub\u003e= 6.1 and 5.8) in 2014, with ground acceleration reaching ~560 cm/s\u003csup\u003e2\u003c/sup\u003e and 735 cm/s\u003csup\u003e2\u003c/sup\u003e, respectively. This study updates the seismic hazard evaluation in Cephalonia and Ithaca using new data and computational techniques to reduce epistemic uncertainties. The probabilistic approach of Cornell and McGuire was used, and the uncertainties are reduced through data variability of the source models, seismicity data, and Ground Motion Prediction Equations using a logic tree approach, sampled by implementing the Latin Hypercube Sampling method. The spatial distribution of Peak Ground Acceleration and Peak Ground Velocity for return periods of 475 and 950 years indicates low variation in the entire study area and that the Paliki peninsula possesses the highest level of seismic hazard. Additionally, site-specific analysis across the three main towns, Lixouri and Argostoli in Cephalonia and Vathi in Ithaca, reveals that Lixouri has the greatest level of seismic hazard, while Vathi the lowest.\u003c/p\u003e","manuscriptTitle":"A Logic-Tree Based Probabilistic Seismic Hazard Assessment for the Central Ionian Islands of Cephalonia and Ithaca (Western Greece)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-29 17:03:12","doi":"10.21203/rs.3.rs-3991269/v1","editorialEvents":[{"type":"communityComments","content":1},{"type":"decision","content":"Revision requested","date":"2024-08-05T08:54:21+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-03T20:22:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"49478129118309440655573614323993892969","date":"2024-05-03T08:13:28+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-03-04T13:03:33+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-02-27T03:54:01+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-02-27T03:54:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Seismology","date":"2024-02-26T14:39:27+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-seismology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jose","sideBox":"Learn more about [Journal of Seismology](http://link.springer.com/journal/10950)","snPcode":"10950","submissionUrl":"https://submission.nature.com/new-submission/10950/3","title":"Journal of Seismology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"4f2d67ac-cc03-4bf1-a475-a10594144925","owner":[],"postedDate":"February 29th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-09-09T16:12:41+00:00","versionOfRecord":{"articleIdentity":"rs-3991269","link":"https://doi.org/10.1007/s10950-024-10242-3","journal":{"identity":"journal-of-seismology","isVorOnly":false,"title":"Journal of Seismology"},"publishedOn":"2024-09-06 16:05:28","publishedOnDateReadable":"September 6th, 2024"},"versionCreatedAt":"2024-02-29 17:03:12","video":"","vorDoi":"10.1007/s10950-024-10242-3","vorDoiUrl":"https://doi.org/10.1007/s10950-024-10242-3","workflowStages":[]},"version":"v1","identity":"rs-3991269","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3991269","identity":"rs-3991269","version":["v1"]},"buildId":"-HB7Z8yhvgn0wM9Nzuekk","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}