1. Introduction
Adequate quantification of tsunami risk on populated seashores is a theoretical and practical problem in the Earth sciences. The consequences of the tsunamis in the Indian Ocean in 2004, Tohoku, Japan, in 2011, Sulawesi, Indonesia, in 2018, and the Aegean Sea in 2020 showed that the underestimating of this hazard has led to catastrophic long-term effects for coastal populations, the economy, and the environment [
1,
2,
3,
4]. The lack of reliable historical data on tsunami occurrences, characteristic of many coasts, stimulated the development of methods for assessing the tsunami hazard and the risk based on information about the seismic activity described in the scientific literature [
5,
6,
7], as well as magnitude scales for tsunamis [
8]. Modern methods for calculating the maximum heights of tsunami waves with a particular frequency are based on the statistics of known tsunami events, instrumental measurements, historical data, and oral accounts of eyewitnesses.
The analysis of tsunamis in the Black Sea region has been the subject of scientific interest in the past several decades due to the increasing concentration of population, the construction of infrastructure, and tourist sites in low-lying coastal areas [
9,
10,
11,
12]. Although there have been no observations of strong or catastrophic effects of tsunami waves in the Black Sea compared to other parts of the world, such as the Pacific and Indian Oceans, in the last century, there are historical tide gauge records of tsunami events at specific coastal locations in the study area. Such events are, for example, due to offshore earthquakes in 1927 and 1966 close to the northern and northeastern coast of the Black Sea [
12]. Another tsunami phenomenon occurred after the very strong earthquake on land known as the Erzincan earthquake in 1939, when the sea receded in the coastal cities of Ordu, Fatsa, and Uniye (Northern Turkey) by about 15 m, 50 m, and 100 m, respectively, and then it advanced to the land [
13].
Experimental research through numerical modeling and simulation under different scenarios is a necessary step to clarify the tsunamigenic potential in the Black Sea region due to the small number of tsunamis documented in the chronicles and the limited number of recorded post-earthquake events through continuous tide-gauge observations along the coast [
14,
15,
16,
17]. Many studies have been published in the scientific literature using different software programs for single or multiple-event numerical model simulations, different input parameters for earthquakes or other sources generating tsunamis (subaerial or submarine landslides, meteorological phenomena, etc.), various scenarios, and different spatial and temporal resolutions of the data used [
18,
19,
20,
21,
22,
23,
24,
25]. However, targeted studies of seismogenic faults as a major source that can trigger tsunamis are very rare. For example, such analyses have been conducted for specific coastal locations, such as those in [
26,
27,
28,
29,
30]. In the case of commercial oil and natural gas exploration, e.g., [
31], the data are difficult to access for scientific research.
In scientific studies, modeling and simulations of tsunamigenic phenomena in the Black Sea region are necessary to consider different tsunamigenic geometric dimensions, correctly interpret different responses at distant and local sources, and reliably determine the input parameters of the source zones. The lack of digital and analog pen-and-paper records from tide gauge stations along the coast further limits achieving a certain accuracy of the estimated wave parameters. For this reason, recent tsunami hazard and risk assessment studies of the entire Black Sea region are usually performed with a dataset of varying quality and completeness regardless of what source they are generated from and are based on previously published data or research [
9,
12,
16,
22,
24,
32]. More than 20 tsunami events, established from the recorded and described historical data in the scientific literature, were mainly caused by earthquakes in the coastal zone or on land near the coastline of the Black Sea. Therefore, the availability of a complete and comprehensive earthquake catalog for a specific study area is a prerequisite for performing a qualitative and thorough probabilistic seismic and tsunami hazard analysis. Furthermore, each seismic catalog is an essential source of information about the background earthquakes and earthquake sequences in the studied region and its local sub-areas, with specific seismotectonic settings [
6].
Several publications on historical Black Sea seismicity and tsunami catalogs report earthquakes of a magnitude greater than 7 (M ≥ 7) [
8,
22,
32], most of which have generated tsunamis in the past. From 1650 to the present, evidence of five moderate intensity 4 ≤ K < 7 tsunamis on the 12-point tsunami intensity scale has been identified [
8]. According to the study of Triantafyllou et al. [
32] for tsunami intensities 5, 6, and 7 in the Black Sea, the return periods in this basin are 46, 128, and 1660 years, respectively.
In order to determine the seismogenic zones in the Black Sea region that can cause tsunamis, it is necessary to analyze a lot of geological, tectonic, and seismological information, which is assumed to be complete, accurate, and reliable for scientific conclusions. Therefore, a homogeneous data set of seismic events in the study area related to activated tectonic faults or their segments and properly defined earthquake parameters is needed to determine the seismogenic sources. These requirements are often difficult to meet from the available observational data. Although tectonic fault databases have been developed, such as those presented in [
33,
34], a comprehensive study of individual seismogenic zones in the Black Sea Basin with a detailed exploration of active faults related to their tsunamigenic potential is still limited or at an initial stage, as described above.
Following this trend, this study aims to analyze the spatial and temporal seismic activity from 1905 to 2022 in the Black Sea region and adjacent coastal territories as potential tsunami sources supplementing the research presented by Dimitrova et al. [
35]. As a first step for a detailed assessment of the tsunamigenic potential of fault zones in the Black Sea region,
Section 2 presents the study region, data, and methods used to compile a regional seismic catalog called the Black Sea Earthquake (BSE) catalog, which is described in
Section 3. The catalog has been compiled using earthquake bulletins published by the international data centers ISC, GCMT, USGS, EMSC, Romanian and Turkish catalogs, and earthquake bulletins extracted from scientific publications. In addition, the newly compiled Bulgarian national catalog by Solakov et al. [
36] has also been added to the produced earthquake catalog. The unification of the seismic events from the used catalogs was performed after converting the different magnitude scales into moment magnitude for five distinct geographical polygons. The obtained results polygon by polygon are given in
Section 4. Next, the completeness and quality of a newly compiled catalog have been assessed and discussed in
Section 5. In addition, we investigated the distribution of the strongest earthquakes (Mw ≥ 5) that occurred offshore and inland in the Black Sea region. With focal mechanisms and spatial distribution of the
b-values assessed from a newly compiled regional seismic catalog of completeness, an interpretation of future seismic activity is suggested.
Section 6 reviews the newly compiled catalog and its applications for other seismicity and hazard assessment investigations in the Black Sea region. Finally, some inferences related to the tsunamigenic potential of earthquake sources are summarized in the conclusion section.
5. Homogenization and Evaluation of the Completeness of the Compiled BSE Catalog
We compiled a new regional declustered catalog, including the earthquakes in the Black Sea region since the beginning of the past century. The complete catalog comprises the earthquakes with moment magnitudes Mw, if available in the downloaded catalogs used, and Mw-proxy magnitudes produced with the obtained relations, which express the best data fit. According to the maximum earth crust in the region [
68], the depth of the listed earthquakes is limited to 40 km. As a result, the catalog of 54,647 events was obtained for the region 27° E–42° E; 40° N–47° N covering the period 1905 to 2022. The data of these earthquakes are saved in the CSV format and proved via Zenodo on the following link:
https://zenodo.org/badge/DOI/10.5281/zenodo.8070409.svg (accessed on 22 June 2023). The catalog has been released under an open license and integrates data from global and local earthquake catalogs based on each of the five polygons.
Figure 6 shows the cumulative curve of the catalog events from 1905 to 2022. It is seen that the number of registered earthquakes exponentially increased after 1970.
The BSE catalog contains the following information:
Date: year, month, day;
Origin time in standardized GMT: hour, minute, and second;
Epicenter coordinates: Latitude φ° N and Longitude λ° E;
Focal depth in km;
Moment magnitude, Mw.
To assess the catalog data quality, we estimate the magnitude of completeness Mc, which, theoretically, is the lowest magnitude above which 100% of the earthquakes in the given region and time range are detected [
69]. The minimum magnitude of completeness Mc for a given catalog is an essential parameter for many studies of the seismic regime and the earthquake hazard assessment in a particular region (see, e.g., [
70,
71]). Before calculating the completeness magnitude [
66,
72], the catalog must be declustered. Applying Reasenberg’s declustering algorithm [
62], the earthquakes are grouped into clusters according to spatial and temporal interaction zones in the Black Sea region, and the aftershocks and foreshocks are identified and removed. Thus, the catalog includes 18,528 unique events with Mw ≥ 1.0. Using a non-instrumental catalog or a catalog created by merging non-instrumental and instrumental often results in a higher value of Mc due to undersampling [
65,
72]. The estimated magnitude of completeness Mc is equal to 5 (Mc = 5). Therefore, the estimated value of Mc shows that there are earthquakes below 5, omitted in the catalog 1905–2022.
The Mc value obtained can be explained by incompleteness in the used catalogs in this study or incorrect magnitude estimations, especially around the 70s of the last century. This assumption is supported by the distribution of magnitude estimates over time for the declustered 1905–2022 catalog depicted in
Figure 6b. This Figure shows that till the middle of the last century, in the studied region, stronger and moderate earthquakes (Mw ≥ 4) have been predominantly registered. Over time, the recorded magnitude gradually and steadily decreases to 2.5. The number of events in the magnitude intervals is as follows: 49,040 are in 3 > Mw; 3876 are in 3.0 ≤ Mw < 3.5; 949 are in 3.5 ≤ Mw < 4, 468 are in 4.0 ≤ Mw < 4.5; 190 are in 4.5 ≤ Mw < 5.0, 72 are in 5.0 ≤ Mw < 5.5, 24 are in 5.5 ≤ Mw < 6.0, and 28 earthquakes with magnitude greater than 6 (Mw ≥ 6).
After the first half of the past century, significant progress has been made in increasing the number of seismic stations and earthquake detection capabilities in the Black Sea region, as seen in
Figure 6a. The sensitivity of national networks for recording weak earthquakes through installing modern broadband seismic stations has increased significantly, as well as the refinement of procedures and software for locating and estimating the magnitude of small to moderate earthquakes. Thus, the earthquake detection capabilities of the networks have been significantly improved. As a result, the number of recorded events has considerably increased, especially after 1980 (
Figure 6a,b), and the minimum magnitudes in the catalogs have been significantly reduced. The years around 1980 can be considered as a dividing line for non-instrumental and instrumental seismic observation. As a rule, the Mc is calculated for the instrumental earthquake catalog, which is more comprehensive than the non-instrumental catalog, predominantly based on the macroseismic observations.
Based on these circumstances and after several approaches made to calculate the magnitude of completeness for different starting times, we considered that 1977 could be accepted as the starting year from which the seismic data catalogs could be used for in-depth analysis, comparison and use for seismic hazard and tsunami hazard assessment. The declustered catalog 1977–2022 comprises 18,295 earthquakes within the magnitude range from 2.5 to 6.1 (2.5 ≤ M ≤ 6.1) and a maximum depth of 40 km (h ≤ 40 km).
Figure 7a shows the distribution of the number of earthquakes over the years in 1/12-year-long bins. The maximum number of earthquakes was in 1999, when two earthquakes of magnitude greater than 7 (Mw > 7) and many events of smaller magnitude occurred in SP.
Figure 7b shows the magnitude distribution over time for the declustered catalog 1977–2022. It is shown that there are no missing estimates over the entire magnitude range from 2.5 to 6.1.
The number of earthquakes distributed in bins with size 0.1 is shown in
Figure 7b. Most of the events in the studied region are in a magnitude range of 2.5–4.5, accounting for about 95% of the total number of events in the catalog. The largest part comprises the earthquakes with a magnitude of around 2.5 (2.6 ≥ M ≥ 2.5), and the number of the stronger earthquakes exponentially reduces.
This trend is evidenced by the distribution of earthquakes’ depths (
Figure 7d), especially for depths over 5 km. The hypocenters of events in the compiled catalog are distributed, respectively: 4133 events (7.6%) within the range 1–5 km, 30,041 (55%) are from 5 km to 10 km, 17,446 events (32%) are from 10 km to 20 km, and 3027 (5.5%) are below 20 km, as shown in
Figure 7d. Moreover, 85% (105) of earthquakes with Mw ≥ 5 were identified in the 10–40 km depth range.
The magnitude of completeness Mc is estimated using the maximum curvature method with bootstrapping [
65,
66]. The frequency–magnitude distribution (FMD) of the earthquakes listed in the declustered catalog from 1977 to 2022 is modeled using the Gutenberg–Richter power law shown in
Figure 8a. The FMD of the data fits the synthetic distribution well, plotted as a red line. The more considerable differences are in magnitudes above 5.1 (Mw > 5.1), where a simple power law cannot adequately explain the observed FMD. The calculated parameters of the values of
b and
a are 1.17 ± 0.01 and 7.192, respectively. Therefore, considering the completeness of Mc,
b-value, and
a-value, we can conclude that all earthquakes with a magnitude above 2.5 (Mw ≧ 2.5) are reliably recorded and presented in the newly compiled catalog.
The right panel of
Figure 8b presents the variations of the magnitude completeness Mc over the years. The temporal variation of Mc is estimated using a window with 500 events and a step of 25 events. The standard deviation of the Mc is depicted with a black dashed line, and it is not more than 0.2 (Mc ± σ: Mc ± 0.1). Considering
Figure 8b, the completeness magnitude is very stable over the years.
Figure 9a shows the spatial distribution of the completeness magnitude within the studied region. The spatial distribution of Mc is performed on a 20 km grid and at least 100 events greater than the magnitude (Mw > 2.5) within a 100 km radius for each grid node. To avoid the statistical errors arising from under-sampling and inhomogeneous data, we calculated the spatial–temporal distribution of the Mc and
b-value according to Amorese et al. [
73]. The completeness magnitude varies from 2.5 to 4.2.
The compiled catalog is complete for the NW, W, S, and E polygons, and the magnitude of completeness is within the interval 2.5–3.0. The higher value of Mc is likely due to the poor coverage of the local seismic networks and the not registered earthquakes with smaller magnitudes.
The value of Mc in the NP increases above 3 (Mc > 3) as in a few parts reaches 4.0 where there is not enough data in the catalog to fulfill the criterion for calculating the Mc. The white zones in the polygon are due to a lower level of seismicity.
Figure 9b shows the spatial distribution of the
b-value, which variations are calculated using the maximum likelihood method of Aki [
74]. The map is drawn with the same parameters as the map of the magnitude of completeness, i.e., executed on a 20 km grid size and at least 100 events within a 100 km radius for each grid node. Such parameters are sufficient to avoid erroneous
b-value statistical results due to undersampling and inhomogeneous subsets [
75,
76,
77,
78,
79,
80,
81,
82,
83,
84,
85]. As a result, relatively low
b-values are obtained for the WP and the EP (see
Figure 1 for the outlined polygons), while relatively high
b-values characterize the SP and NWP.
The temporal variation of the
b-value is shown in
Figure 9c. The standard deviation is drawn with black dashed curves. and it is within ± 0.1 (σb= ± 0.1). The shape of this curve largely repeats the curve of the completeness magnitude distribution over the years (
Figure 8b). In both curves, the variations of the calculated
b-value and Mc magnitude are strongly dependent on the maximum and minimum magnitude of realized earthquakes as well as the number of recorded earthquakes of the corresponding magnitude (
Figure 7a,b).
The depth distribution of the
b-value is given in
Figure 9b. Variations in depth are calculated over 5 km depth layers that contain at least 150 events. The maximum
b-value is at a depth of 7 km, where most of the events in the catalog are realized. There are four other peaks in the curve at depths of 5 km, 8 km, 10 km, and 12 km, and two distinct minima at 11 km and below 30 km depth.
Figure 9b and
Figure 7d prove that the prevailing part of the earthquakes occurs at the depth range from 5 km to 15 km, and there are no seismically inactive layers of the Earth’s crust in the study region.
6. Discussion
The Black Sea is located within the Anatolian sector of the Alpine Himalayan orogenic system. The geodynamic processes in this region are related to the northward movement of the African and Arabian plates, which collide with the Eurasian tectonic plate. The Anatolian block is moving westward with a rotational motion focused northward from the Sinai Peninsula [
86]. These processes delineate specific zones of concentration of seismic epicenters along tectonic units [
35,
87,
88]. As previously described by Oynakov et al. [
89], the seismic events in the Black Sea region could be clustered into the strongest earthquakes and outlined the main seismic zones: Shabla-Kaliakra, the Crimean Peninsula, three clusters located on the eastern shores of the Black Sea (west coasts of Georgia and Russia), and one on the southern coast near Bartin, Northern Turkey. A large number of earthquakes are related to the Vrancea seismic zone. The present study demonstrates the same seismic pattern in the Black Sea region.
Figure 1 shows the distribution of the denoted by grey dot epicenters. According to the compiled catalog, the hypocentral depth of the earthquakes in the study region is not more than 40 km. The strongest earthquakes above 5 (Mw ≥ 5) are mainly realized along the Northern Black Sea coast of Turkey (offshore and on land), the Western Black Sea coast of Georgia (offshore and on land), the offshore of the Northern Bulgaria Black Sea coast and the Vrancea seismic zone.
In order to give a quality estimation of the compiled in the study catalog, we compare its parameters Mc,
b-value, and
a-value to such ones calculated by Dimitrova et al. [
35]. The produced magnitude of completeness Mc is equal to 2.5. The value is smaller than those in [
35] (hereafter, the parameters related to the catalog by Dimitrova et al. [
35] will be marked with a suffix 21), namely Mc21= 3.0. The relatively lower Mc value is the result of the addition of the national catalogs of the Black Sea countries in the current catalog. As a rule, the local networks can detect and process the smaller magnitude (M ≤ 3) events included in the local catalogs. Thus, the Mc of the newly compiled catalog was reduced to 2.5. Nevertheless, the FMD of both compiled catalogs is linear with parameters as follows:
b21-value = 1.09 and
b-value = 1.17 (
Figure 8a);
a21-value = 7.279 and
a = 7.192 (
Figure 8a). The spatial distribution of the completeness magnitude for both catalogs follows the model of higher values in the Northeastern Black Sea region and smaller values in the west and south parts. In addition, the present study gives the spatial distribution of the Mc in more detail (
Figure 9a) due to the above-explained reasons.
The spatial distribution of the
b-values, calculated for both catalogs, has a similar behavior; the highest values are mapped in the eastern part of the Southern Black Sea coast. For the current catalog, there are higher
b-values in the southwestern and northwestern parts of the region and distinctly low values in the eastern and western parts (
Figure 9b). This behavior could be explained with more comprehensive local catalogs (Turkish, Romanian, Bulgarian).
There are several studies on the temporal and spatial variations of
b-values; they reflect stress conditions [
81,
82,
83,
84] or in strong connections with the different types of faulting [
84] and are considered as a precursor of upcoming small or strong events [
84,
85]. It was found that a high
b-value indicates a higher proportion of events with smaller magnitudes and vice versa—a small
b-value indicates earthquakes with larger magnitudes. In general, our results conform with those obtained by other authors cited above.
The variations of the temporal and depth distribution curves of the
b-value (see
Figure 9a,b) are the same as those of the corresponding curves for the
b21-value. The curve in
Figure 9a gives a more detailed temporal distribution of the
b-value, and the curve in
Figure 9b is smoother than that of the
b21-value. The appearance of the two curves is due to the more comprehensive national catalogs included.
The most significant earthquakes in the Black Sea region, which could affect the population, infrastructure, and tourist sites in low-lying coastal areas. have a magnitude above 5 (Mw≥ 5). According to the compiled catalog from 1977 to 2022, these earthquakes and their parameters extracted from relevant sources are listed in
Table 7. The focal mechanisms of the listed earthquakes are given in column 11, and the source is given in column 12 of the table. Unfortunately, the information on the focal moment tensor mechanisms of several earthquakes before 2000 is missing or not freely available.
Most of the earthquakes listed in
Table 7 have occurred in Northern Turkey and are associated with the branches of the North Anatolian Fault Zone [
86,
87,
88]. Four earthquakes (with numbers 1, 5, 8, and 25) are realized offshore in the aquatorium of the Black Sea, respectively: 1 is close to the NW coast of Turkey; 5 and 25 are close to the west coast of Georgia; 8 are close to the NE coast of Bulgaria (near Balchik). Two earthquakes (3 and 30) are related to the Vrancea seismic zone. The strongest event has a magnitude of Mw = 6.1 and is realized in the aquatorium of the Marmara Sea [
85].
The fault plane solutions of earthquakes in the Black Sea (near the coasts of NE Bulgaria and West Georgia) have strike-slip faulting. The reverse faulting component dominates for the Southwestern Black Sea (event 1) and Western Georgia (event 4). The focal mechanism of the events realized on the mainland of Northern Turkey (eastern, central, and western parts) has a predominantly strike-slip component. Therefore, we can conclude that the focal mechanisms of the earthquakes with Mw ≥ 5.0 in the Black Sea region agree with the primary fault neotectonics.
Several studies of the
b-value show that its variations are in strong connections with the different types of faulting [
81,
82,
83,
84,
85]. It is concluded that the areas where faults are located and subjected to greater stress (thrust faults) demonstrate a lower
b-value; those with seismic events associated with normal faults and strike-slip faults have the highest and intermediate
b-values, respectively. We do not notice such behavior of the
b-value in the presented study. To accurately determine spatial and temporal variations in the
b-value, high-precision seismic catalogs of all determined polygons are required, which is the goal of the next study.
The disruptive seismic events triggered tsunamis in the past fall within the main seismic zones outlined in the Black Sea region. According to the BSE catalog, from 1977 to 2022, a strong earthquake occurred in the offshore region of NW Turkey in the Marmara Sea in 1983 and had a magnitude Mw = 6.1 (see
Table 7, first column, event 4).
In 1999, the strongest earthquakes with magnitudes Mw = 7.5 and Mw = 7.1 were realized on the land of NW Turkey (events 19 and 38 in
Table 7). In 2000, an earthquake with magnitude Mw = 6 (event 47 in
Table 7) was realized in N Turkey. However, there was no evidence for the realization of the tsunami waves.
Based on the spatial distribution of the
b-value in the Black Sea region, which is mapped in
Figure 8b, the earthquake catalog, and the information on the geodynamic potential of the faults in the Black Sea region, we could conclude that stronger earthquakes can occur in the region in the near future. Nevertheless, the tsunami frequency in the Black Sea is rated as low [
9,
32], and the estimated tsunami hazard in the Black Sea is low to moderate; it cannot be neglected.
The recent Triantafyllou et al. [
32] study, regardless of the tsunami generation mechanisms, shows that more realistic results for a probabilistic tsunami risk for the Mediterranean and connected seas can be obtained using complete data sets than incomplete and uncertain historical impact records. Specifically for the Black Sea, the authors found that the tsunami risk is lowest compared to the other parts of the studied area, examining 22 tsunami events—10 historical (fifth century BC to 1899 AD) and 12 for the instrumental period (1900–2021). It should be noted that for the instrumental period of registered seismic events from the beginning of the last century, there are a small number of registered tsunami events by tide gauge observations along the Black Sea coast, especially for the western part. Also, a comprehensive tsunami hazard and risk analysis are difficult due to the incompleteness and uncertainties of all data files. Therefore, as an initial step to clarify the tsunamigenic potential of the seismic zones in the Black Sea region, the parameters of the active or potentially active faults that can generate earthquakes with a magnitude greater than M > 5.0 should be clarified. So far, geophysical and seismic studies, apart from some exceptions, are rare in this region. The exchange of seismic data between coastal countries is one of the obstacles to a comprehensive and in-depth analysis of the seismic activity in the region as a whole. Furthermore, the ambiguous attachment of specific focal mechanisms of earthquakes to active faults or their segments leads to contradictory results, complicating the subsequent analysis and interpretation of seismic and tsunami hazards and risks. Therefore, conducting detailed interdisciplinary research for local or site-specific assessments is necessary, including collecting data and evidence from past tsunami events.
We compiled the BSE catalog based on statistical data processing and calculation procedures with the specialized software ZMAP [
65], analyzing many seismic events realized at different seismic sources and registered by different sensors. Therefore, the results of our research have certain limitations of a different nature. For example, some of them are due to the completeness, accuracy, and reliability of the data sets contained in the analyzed catalogs. Regarding data completeness, it should be noted that the analyzed data may not include all seismic events, especially those that were of low magnitude or unreported, due to different procedures used by seismic data centers. Also, data on seismic events can be subject to various sources of error, such as sensor issues or data transmission disruptions. Due to the absence or poorly presented seismic data in all analyzed datasets, there are certain incompletenesses in the compiled BSE catalog, which is evident for the Northern and Eastern Polygons in the studied region. In complex cases or ambiguous seismic events, subjective decisions may affect the inaccurate representation of seismic events in the studied region. Also, we derived linear dependences for transformation from different magnitude scales into moment magnitude Mw for the corresponding geographic polygon using different numbers of events. The number of seismic event pairs available to derive these linear relationships is critical in estimating the parameters of the corresponding linear regression. Other factors that should be considered are the calculation procedures and software algorithms, where expert knowledge is used in analyzing and interpreting the results, for example, in the procedure for declustering, applying the Reasenberg algorithm [
62].
Further work is needed (1) to include seismic events, which are not so far included in the BSE catalog, to complement and improve it; (2) for detailed local studies to identify potentially active faults triggering earthquakes with tsunamigenic potential; (3) to derive unambiguous empirical dependences when transforming between different magnitude scales into moment magnitude for the identified seismically active seismic zones, and (4) to derive maximum-likelihood parameters used as inputs in modeling and simulations of earthquake-triggered tsunamis in the study area.
7. Conclusions
We used freely available data catalogs of four international data centers, namely USGS, ISC, EMSC, and five national catalogs, respectively, Romanian, Turkish, Bulgarian, and published catalog information on events on the territory of Russia to compile a new catalog of the Black Sea earthquakes. Initially, the summary catalog of earthquakes from 1905 to 2022 in the specified region 27° E–42° E and 40° N–47° N included 54,918 earthquakes. First, we calculated numerically empirical model dependencies for five geographically defined polygons to convert magnitudes from heterogeneous locally reported magnitude scales into standard proxy magnitude scales equivalent to the moment magnitude Mw. Second, a declustering procedure was applied to remove duplicate events. As a result, a catalog of precisely defined earthquakes in the Black Sea region was compiled, comprising 54,647 events, and published in CSV format under an open license. The created catalog can be helpful to seismologists, geologists, and other specialists studying seismic phenomena in the Black Sea region. The BSE catalog will be freely available and can be regularly updated, including recent earthquakes registered by the available and updated national seismic networks in the studied region. In addition, the part of the homogenized catalog from 1977 to 2022 related to the instrumental monitoring of the seismicity of the Black Sea region is studied. A declustering procedure was applied to remove duplicate events in the catalog. The catalog comprises as many as 18.528 unique earthquakes within the magnitude range from 2.5 to 6.1 (2.5 ≤ Mw ≤ 6.1) and realized at a depth not more than 40 km. Then, the quality of the newly compiled catalog was assessed by evaluating the magnitude of completeness Mc and its uncertainty. According to the Gutenberg–Richter law, the frequency–magnitude distribution is calculated. The magnitude of completeness is equal to 2.5 (Mc = 2.5), and the b-value and the a-value are 1.17 (±0.01) and 7.192, respectively. Based on the three parameters’ estimations, we can conclude that the new BSE catalog is representative and can be used in the presented study and for further seismic and tsunami hazard analyses of the Black Sea region.
The spatial distribution of the magnitude of completeness shows as follows: the Northwestern, Western, and Southern Polygons demonstrate the lowest values of Mc-below 3; the regions of Southern Crimea and Western Russia (parts of the NP and the EP) have completeness magnitudes around 3.5, for which, the most likely reason is a lack of free available catalogs. The spatial distribution of the b-value is obtained and plotted on the map, delineating areas of low values in the WP and the EP, and areas of relatively higher values in the SP and the NWP. Making a separate spatio-temporal analysis for specific low-lying locations along the Black Sea coast is necessary to obtain more meaningful results in studying the tsunamigenic potential of seismically active fault zones. Finally, the strongest earthquakes with a magnitude above 5 (Mw ≥ 5), which are realized in the studied region, are estimated based on their focal mechanisms and spatial–temporal distribution.
In this study, newly collected and analyzed information about the seismic regime during the last century in the Black Sea region shows that the probability of a strong earthquake in the future should not be underestimated. Tsunamis in the Black Sea are not as large, destructive, and frequent as in other tsunami-prone areas. However, the tsunami risk for some local Black Sea coastal areas such as the Balchik–Kaliakra region (Northeastern Bulgaria), Northwestern Turkey, the southern coast of the Crimean Peninsula, and the western coast of Georgia is relatively high compared to others and should be studied in detail.