Mass Mortality of Shallow-Water Temperate Corals in Marine Protected Areas of the North Aegean Sea (Eastern Mediterranean)

: Coral mortality is a global phenomenon of increasing magnitude, correlated with climate change. Prolonged marine heatwaves have particularly affected the north Aegean Sea in summer 2021, threatening shallow-water stony corals, such as Balanophyllia europaea and Cladocora caespitosa . To assess their population status, ten coastal, rocky-bottom stations dispersed in Natura 2000 sites of Chalkidiki (north Aegean) were surveyed using non-destructive techniques in autumn 2021. At each station, corals’ abundance was estimated in situ, by counting the number of B. europaea polyps within randomly placed 50 × 50 cm quadrats, and the number of C. caespitosa colonies along three replicate belt transects 1 × 10 m. The status of corals was qualitatively assigned as healthy, bleached (partially or complete), or in necrosis (partial or complete). B. europaea was found in 80% of stations; in total, 58.17% of the coral specimens were affected by necrosis. C. caespitosa was found in 30% of stations; in total, 27.49% of the coral colonies were partially bleached and 11.32% in necrosis. Another nine sessile invertebrates (sponges, bivalves, and ascidians) were observed in necrosis. These results highlight the need to establish monitoring programs on vulnerable sessile invertebrate populations along the Aegean Sea to assess climate change impacts.


Introduction
Climate change constitute a major threat to marine biodiversity [1,2]. Increasing seawater temperature, acidification, and ice melting degrades the marine environment and cause devastating events, often lethal to a variety of life forms. Among marine biogeographic realms [3], the Mediterranean Sea seems to be especially affected, as is becoming warmer during the past decades and, within the entire basin, temperature is increasing much more rapidly (20%) than global mean [4]. This warming phenomenon is especially relevant for the eastern Mediterranean, which is suffering from the so-called "tropicalization" effect [5][6][7][8]. As a result, the frequency of appearance, the magnitude, and the duration of marine heatwaves, i.e., extreme rises in seawater temperature over an extended period, have significantly increased [9]. The Aegean Sea, and especially its northern part, is among the most impacted Mediterranean sub-regions [9]. Mean surface seawater temperature increases by a rate of over 0.60 • C per decade, which is much higher than the that relevant (0.40 • C/decade) for the entire eastern Mediterranean basin [10]. As a result, over 35 marine heatwave events (>300 days) occurred during the recent decade, and accordingly, the area seems to currently constitute a "hot spot" of marine heatwaves [10].
Increased seawater temperature and marine heatwaves severely affect marine biota, long-lived, sessile, epibenthic invertebrates, such as sponges, corals, bivalves, and ascidians, in particular [11][12][13][14]. Gorgonians and stony corals are among the most impacted species [15][16][17][18]. Focusing on stony corals, bleaching, i.e., the loss of endosymbiotic zooxanthellae from the host coral, is the first signal of thermal distress, which, if prolonged, leads to the necrosis of polyps, and subsequent erosion of corallites. Mass mortality events temperature associated with a number of intense and prolonged (over 20 days) marine heatwaves [10]. This prolonged warming may have detrimentally affected sessile marine biota, stony corals in particular. Shallow-water stony corals, due to their sensitivity to marine warming [15,16,21,23,24,31,33], may represent appropriate climate change descriptors to assess the resilience of temperate MPAs under ongoing marine warming.
Considering all the above, the present work aims to assess the population status of the most prominent stony-coral species, namely B. europaea and C. caespitosa in the shallow rocky shores of the five SACs included in Chalkidiki peninsula (north Aegean Sea) [47], right after the particularly hot summer of 2021, by applying non-destructive assessment techniques.
One up to four sampling stations were randomly established within each SAC; the number of stations was defined according to the cover area of SACs ( Figure 1). Overall, two stations (S1, S2) were set in Thermaikos Gulf, four (S3-S6) in Toroneos Gulf, and another four (S7-S10) in Siggitikos Gulf. All the sampling stations were set on the rocky coastline, with the exception of S7, which was set on the small islet of Diaporos. The sea bottom consisted of moderately to strongly inclined rocky cliffs (S3, S5, S7, S8, and S9) or of large boulder reefs (S1, S2, S4, S6, and S10) surrounded by dense meadows of the seagrass species Posidonia oceanica (class Magnoliopsida). At all stations, shallow water (>20 m) benthic communities were dominated by algal species with sponges, hydrozoans, corals, mollusks, and ascidians scattered in-between [36,37], as described in the Introduction.
Samplings were carried out in September (Kassandra, [10][11][12][13] and October (Sithonia, 20-22) 2021 with SCUBA diving in the shallow sublittoral zone, i.e., up to 15 m depth. Visual census included a combination of belt transects and haphazardly placed quadrats [37] to assess the health status of the populations of the prominent stony coral species in the area, namely B. europaea and C. caespitosa. Both surveyed species were found between two and 12 m depth (depth standardized according to mean sea level). B. europaea, was particularly abundant around 3-6 m, whereas C. caespitosa around 7-10 m; therefore, populations were surveyed at these depth levels, respectively.
In the case of B. europaea, population density was estimated in situ by counting all individuals present in the randomly placed ten replicate quadrate frames (50 × 50 cm) at each station. The condition of individuals, i.e., alive (normally pigmented polyps), bleached (white or transparent polyps), or dead (white skeleton denude of coral tissue) was also recorded. In the case of C. caespitosa, the number of colonies were counted along three replicate transects 1 × 10 m each, at each sampling station. The status of coral colonies was qualitatively assigned in five categories [48], as: (0) healthy, when all the polyps of the colony, or almost all, were normally colored, (1) partially bleached, when several of the polyps of the colony were transparent, (2) completely bleached, when all the polyps of the colony, or almost all were transparent, (3) in partial bleaching and necrosis, when a part of the colony had transparent polyps and the rest was dead, and (4) in recent necrosis, when none, or almost none of the colony's polyps were alive. In situ recorded data were used to assess the mortality or bleaching rate per sampling station, per SAC, and over the surveyed area. part of the colony had transparent polyps and the rest was dead, and (4) in recent necrosis, when none, or almost none of the colony's polyps were alive. In situ recorded data were used to assess the mortality or bleaching rate per sampling station, per SAC, and over the surveyed area. Analysis of variance (ANOVA) was used to test for spatial differences in the health status of stony coral species in Chalkidiki peninsula. More specifically, the effect of SAC (fixed factor) and sampling station (fixed factor nested to SAC) on the abundance of healthy and affected coral polyps or colonies (data were pooled over partially bleached to necrosis categories of corals' health status) was assessed through a general linear model [49]. The Fisher LSD test was used for post-hoc comparisons. ANOVAs were performed using the SPSS software package (IBM SPSS statistics version 19). Analysis of variance (ANOVA) was used to test for spatial differences in the health status of stony coral species in Chalkidiki peninsula. More specifically, the effect of SAC (fixed factor) and sampling station (fixed factor nested to SAC) on the abundance of healthy and affected coral polyps or colonies (data were pooled over partially bleached to necrosis categories of corals' health status) was assessed through a general linear model [49]. The Fisher LSD test was used for post-hoc comparisons. ANOVAs were performed using the SPSS software package (IBM SPSS statistics version 19).
At each sampling site, the presence of different invertebrate species affected by necrosis was also recorded during visual census. Seawater temperature, pH, and dissolved oxygen were also measured in the water column and up to 15 m depth with an autographic recorder (Hach Be Right TM) at the same time as visual census.
At each sampling site, the presence of different invertebrate species affected by necrosis was also recorded during visual census. Seawater temperature, pH, and dissolved oxygen were also measured in the water column and up to 15 m depth with an autographic recorder (Hach Be Right TM) at the same time as visual census.
Nested ANOVAs revealed significant spatial differences in the abundance of both healthy and affected B. europaea polyps (Table 2), with significantly increased abundance of impacted corals in GR1270008 and of healthy corals in GR1270002 ( Figure 4). This pattern was mainly due to the effect of S4 in GR1270008 and of S7 in GR127002. Nested ANOVAs revealed significant spatial differences in the abundance of both healthy and affected B. europaea polyps (Table 2), with significantly increased abundance of impacted corals in GR1270008 and of healthy corals in GR1270002 ( Figure 4). This pattern was mainly due to the effect of S4 in GR1270008 and of S7 in GR127002.  Table 2. ANOVA results on the effect of SAC (5-level fixed factor) and sampling station (8-level fixed factor nested in SAC) on the abundance of Balanophyllia europaea healthy and affected (in necrosis) polyps in Chalkidiki peninsula in autumn 2021.

Cladocora caespitosa
Cladocora caespitosa ( Figure 5) was only found at three of the ten surveyed stations (30%, Table 3). The population density of C. caespitosa ranged from 1.3 colonies/m 2 (S3) to 2.9 colonies/m 2 (S7) with a mean at 2.2 ± 0.82 colonies/m 2 (considering only the sites where the species were recorded) and an overall mean at 0.66 ± 1.13 (considering zero-coral sites, as well). The density of healthy corals (category 0) ranged from 0.8 colonies/m 2 (S3) to 1.3 colonies/m 2 (S7 and S8) with an overall mean 1.2 ± 0.3 colonies/m 2 . The density of partially bleached corals (category 1) ranged from 0.2 colonies/m 2 (S3) to 1.1 colonies/m 2 (S7) with an overall mean 0.7 ± 0.4 colonies/m 2 . No completely bleached colonies (category 2) were observed, and the density of colonies in partial bleaching and necrosis (category 3) was particularly low (0.06 ± 0.07 colonies/m 2 detected in S7 and S8). Finally, the corals in recent necrosis (category 4) ranged from 0.1 colonies/m 2 (S6) to 0.3 polyps/m 2 (S7 and S8) with an overall mean 0.2 ± 0.1 colonies/m 2 . Accordingly, about one-third (27.49%) of the observed colonies were partially bleached and 11.32% were in partial or in complete necrosis. A very dense population of C. caespitosa exist in the marine protected area of Vourvourou (GR1270002), which seems to have suffered from increased levels of bleaching and successive mortality (Table 3, Figure 3). The presence of colonies in old necrosis with strongly eroded corallites was also noticed during samplings (see Figure 5). Table 3. Total number of colonies, healthy colonies (0), affected colonies (1 = in partial bleaching, 2 = in complete bleaching, 3 = in partial bleaching and partial necrosis, 4 = in recent necrosis), and percent of partial bleaching and of mortality (colonies in partial and complete necrosis) of Cladocora caespitosa per sampling station at the surveyed SACs in Chalkidiki peninsula in autumn 2021.

Chalkidiki
Natura Kassandra  Kassandra The population density of C. caespitosa ranged from 1.3 colonies/m 2 (S3) to 2.9 colonies/m 2 (S7) with a mean at 2.2 ± 0.82 colonies/m 2 (considering only the sites where the species were recorded) and an overall mean at 0.66 ± 1.13 (considering zero-coral sites, as well). The density of healthy corals (category 0) ranged from 0.8 colonies/m 2 (S3) to 1.3 colonies/m 2 (S7 and S8) with an overall mean 1.2 ± 0.3 colonies/m 2 . The density of partially bleached corals (category 1) ranged from 0.2 colonies/m 2 (S3) to 1.1 colonies/m 2 (S7) with an overall mean 0.7 ± 0.4 colonies/m 2 . No completely bleached colonies (category 2) were observed, and the density of colonies in partial bleaching and necrosis (category 3) was particularly low (0.06 ± 0.07 colonies/m 2 detected in S7 and S8). Finally, the corals in recent necrosis (category 4) ranged from 0.1 colonies/m 2 (S6) to 0.3 polyps/m 2 (S7 and S8) with an overall mean 0.2 ± 0.1 colonies/m 2 . Accordingly, about one-third (27.49%) of the observed colonies were partially bleached and 11.32% were in partial or in complete necrosis. A very dense population of C. caespitosa exist in the marine protected area of Vourvourou (GR1270002), which seems to have suffered from increased levels of bleaching and successive mortality (Table 3, Figure 3). The presence of colonies in old necrosis with strongly eroded corallites was also noticed during samplings (see Figure 5).
Nested ANOVAs revealed significant spatial differences in the abundance of partially bleached and of affected by necrosis C. caespitosa colonies, in contrast with healthy ones (Table 4). Significantly increased abundance of impacted corals were observed in both stations of GR1270002 ( Figure 6).
Nested ANOVAs revealed significant spatial differences in the abundance of partially bleached and of affected by necrosis C. caespitosa colonies, in contrast with healthy ones (Table 4). Significantly increased abundance of impacted corals were observed in both stations of GR1270002 ( Figure 6). Table 4. ANOVA results on the effect of SAC (2-level fixed factor) and sampling station (3-level fixed factor nested in SAC) on the abundance of Cladocora caespitosa healthy, partially bleached, and affected (in partial or complete necrosis) colonies in Chalkidiki peninsula in autumn 2021.

Various Other Affected Sessile Invertebrates
Apart of the two stony coral species described above, another nine (9) sessile invertebrate species exhibited signs of partial or complete necrosis within the study area. These were the sponges Aplysina aerophoba (Nardo, 1833) Figure 7). Although no quantitative data were collected for the above species, S. foetidus, I. variabilis, and S. gaederopus were by far the most severely affected over the study area.

Seawater Temperature, pH, and Dissolved Oxygen
The seawater temperature was homogenous at vertical (0-15 m depth) scale. In horizontal scale (i.e., across the spatially dispersed sampling stations), water masses were also homogenous at all the sites sampled in the same time, i.e., Kassandra stations, sampled in 10-13 September, and Sithonia stations, sampled about a month later, in 20-22 October. Mean temperature was 24.75 °C in September and 20.06 °C in October samplings. The pH was extremely stable at both studied spatial scales, varying around 8.22. Dissolved oxygen ranged from 8.14 mg/L to 9.14 mg/L; accordingly, seawater masses were highly saturated.

Discussion
The rocky shores of the north Aegean Sea constitute a highly diverse habitat that host a large number and variety of benthic invertebrate species [36,37], justifying their inclusion in Natura 2000 network as SAC. These SACs have been implemented in the study area (Chalkidiki peninsula) since 2011, and yet, no active management plans have been established. Marine biodiversity remains largely understudied, as only very recently (2021-2023) the relevant Management Unit completed the first basal surveys on these

Seawater Temperature, pH, and Dissolved Oxygen
The seawater temperature was homogenous at vertical (0-15 m depth) scale. In horizontal scale (i.e., across the spatially dispersed sampling stations), water masses were also homogenous at all the sites sampled in the same time, i.e., Kassandra stations, sampled in 10-13 September, and Sithonia stations, sampled about a month later, in 20-22 October. Mean temperature was 24.75 • C in September and 20.06 • C in October samplings. The pH was extremely stable at both studied spatial scales, varying around 8.22. Dissolved oxygen ranged from 8.14 mg/L to 9.14 mg/L; accordingly, seawater masses were highly saturated.

Discussion
The rocky shores of the north Aegean Sea constitute a highly diverse habitat that host a large number and variety of benthic invertebrate species [36,37], justifying their inclusion in Natura 2000 network as SAC. These SACs have been implemented in the study area (Chalkidiki peninsula) since 2011, and yet, no active management plans have been established. Marine biodiversity remains largely understudied, as only very recently (2021-2023) the relevant Management Unit completed the first basal surveys on these SACs.
In the shallow sublittoral zone of Chalkidiki's SACs, the most prominent macroscopic invertebrates are sponges and stony corals, among which are several of the currently reported species in partial or complete mortality. By focusing on stony corals, B. europaea and C. caespitosa are by far the most common species [37,47], as also observed during the performed underwater surveys.
Successive mortality events of growing magnitude have affected B. europaea populations in several Mediterranean locations [17,20,21,26,27] causing localized declines and reducing fitness [28]. However, no relevant events have been reported in the past from the Aegean Sea. The population of B. europaea has been thoroughly studied in the Mljet MPA of the Adriatic Sea for over a decade with annual surveys and seawater temperature monthly records [17]. In Mljet, the reported mortality of the species ranged from 80% to 100% and it was highly correlated to thermal anomalies and marine heatwaves during the summertime. The southern sector of the MPA was particularly affected; the coral has been assigned as locally extinct from several sites, and its recovery is uncertain [17].
According to the presented results, the 2021 mass mortality event had severely affected the population of B. europaea in Chalkidiki peninsula, as, overall, 58% of the observed coral specimens suffered from necrosis. The species mortality was especially high in Toroneos Gulf (GR1270008, Kassandra), where necrosis affected locally over 80% of the surveyed population. Within the same SAC, the condition of B. europaea was much better in S3, which is the most exposed site over the surveyed stations. S3 is located at the edge of Kassandra peninsula, at cape Kalogria ("meeting point" of Thermaikos and Toroneos gulfs), an area of very steep bathymetry, influenced by strong currents; this marine area has been less affected by the 2021 marine heatwaves due to its geomorphology and prevailing hydrological conditions [10]. On the contrary, S4, which is located rather close to S3, but in a shelter and rather shallow bay of Kassandra peninsula having front to Toroneos Gulf, was extremely affected. The SACs of Sithonia were impacted in a similar degree, as mortality affected around 57% of the coral's population in GR1270002 and GR1270007; in GR120009, less than half of B. europaea population were assessed in necrosis. The latter SAC, which is located close to the edge of Sithonia peninsula and in the front of Siggitikos Gulf, represents an exposed marine area, less impacted by the 2021 marine heatwave [10].
Severe bleaching and subsequent mass mortality events of C. caespitosa have been widely reported from the Mediterranean Sea [13,15,16,19,20,28,[50][51][52][53] since 1997 [14]. Focusing into the Aegean Sea, only one mass mortality event has been previously reported from the North Turkish coasts, where about 15% of the coral colonies were in necrosis [54]. Due to the long lifespan and slow dynamics, the recovery abilities of affected C. caespitosa populations were low [28]. Furthermore, the species exhibit different physiological responses, according to the maximum values of the seawater temperature and the duration of the thermal anomaly, as well as to habitat features and other abiotic factors [16,31]. For example, in the Ligurian Sea [20] and the Spanish Columbretes Islands [31], C. caespitosa colonies were not bleached but died due to progressive tissue necrosis, and this response has been attributed to the thermal tolerance of its symbiotic zooxanthellae of the genus Symbiodinium [31,55]. On the contrary, bleaching of colonies has been observed in the Adriatic Sea [16]. Totally bleached colonies failed to recover in Mljet MPA, in contrast with the Piran MPA, where the majority of bleached colonies did not suffer from tissue necrosis and recovered, a response attributed to the higher food availability in the latter area [16].
The 2021 mass mortality event has considerably affected C. caespitosa as, overall, 27.49% of the observed colonies were partially bleached and another 11.32% showed partial or complete necrosis. Mortality rate was higher in GR1270002 SAC (Sithonia peninsula). This marine area forms a kind of "interrupted atoll" due to the presence of several islets in irregular crescent shape to the shoreline. Due to the peculiar geomorphology, currents are rather strong ensuring food availability. Dense C. caespitosa population has been reported within this SAC, both as scattered colonies and as a large bank of over 2 m high, which is unique in the Aegean Sea [41,56]. Therefore, this marine area represents a biotope of local importance for the survival of the species. Moreover, considering that C. caespitosa colonies create a complex biogenic habitat that hosts a highly diverse assemblage of macrobenthic invertebrates, over 286 invertebrate species have been reported to live in association with the coral colonies [41]. Since algae rapidly cover the dead coral colonies [16], mass mortality events may have detrimental cascade effects on the biodiversity of associated invertebrate communities, as well.
Balanophyllia europaea was the most widespread stony coral species over the studied area, as it has been found in eight out of the ten surveyed stations; on the contrary, C. caespitosa was recorded at only three sites. The tooth coral (B. europaea) was not found at S5-not even as eroded remnants of polyps in necrosis-a station where it has been previously recorded as abundant (authors' personal data concerning the period 2000-2010). The Mediterranean pillow coral (C. caespitosa) was absent from the same station (S5), as well as from another two stations (S4 and S9) where it has been previously reported as abundant [36,37], and authors' personal data; there are no data for its presence on the other four surveyed stations. This result implies the local extinction of B. europaea and C. caespitosa from one and three stations, respectively, during about the last decade raising concerns on their future viability. These stations are located on rocky areas of increasing touristic development and are highly frequented by SCUBA divers, and recreational and professional fishers. Although it is difficult to assess possible anthropogenic impacts, mainly due to the absence of systematic surveys, the collection of C. caespitosa colonies for decorative reasons has been witnessed, whereas the detachment and mechanical damage (i.e., fragmentation) of colonies from small-scale fisheries has been documented in the North Aegean Sea [57,58].
Both coral species were also absent from S2 (located in Loutra Agia Paraskevi Kassandra). S2 constitutes a particular case due to the presence of gas emission vents in shallow depth (0.5-1 m and 5 m), with diffuse bubbling and aligned bubble trains of low or medium flux [59]. White deposits are visible around the orifices and the substrata contains consolidated sand and stones, which are indicative of hot-water emission. The vents contain H 2 S, which is, however, below the analytical detection limit (<10 ppm) [59]. Although gas exchange between bubbles and seawater seems to be limited [59], the peculiar facies of hydrothermal oozes with impoverished fauna and dense populations of nematodes dominate the benthic community [60]. Therefore, environmental conditions may be unfavorable for the development of stony corals in S2. This is the first documented mass mortality event of stony corals in the Hellenic Seas, as only anecdotal reports on affected colonies exist. According to the presented results, the mass mortality event has largely affected the population of B. europaea-58% of coral specimens suffered from necrosis-and to a lesser degree of C. caespitosa-27.49% and 11.32% of coral colonies suffered from partial bleaching and partial or complete necrosis, respectively. Persisting marine heatwaves that have affected the north Aegean Sea during summer 2021 [10] have probably caused the reported event [15], which seems to have distress the entire Chalkidiki peninsula. It is also worth noting that the species were absent from sites where they were previously reported as abundant [36,37], indicating that further threats, such as the collection for decorative purposes or from gillnet fisheries, apart from climate change affect their survival. Coastal topography (shelter bays and depth) combined with seawater circulation patterns and mixing may explain the observed differences in mortality rates between the sampling Natura 2000 sites, as healthier populations were recorded at the most exposed sites. Differences in mortality rates of temperate corals in small spatial scales have been attributed to habitat variability with temperature being the triggering factor [16].
Overall, 11 long-live sessile invertebrates were affected over the study area, with stony corals and sponges (S. muscarum, in particular) being among the most impacted species, as previously reported from other Mediterranean regions [14]. The 2021 event was rather unique for the study area, which was severely affected by persistent marine heatwaves [10]. These findings highlight the need for implementing a strategic monitoring plan over SACs to safeguard their management utility in conserving fragile marine habitats and biodiversity. Obviously, the temperate stony corals B. europaea and C. caespitosa represent ideal descriptors for monitoring the impacts of marine warming and the resilience of MPAs under predicted climate crisis.

Conclusions
This is the first report of B. europaea mass mortality event and the second of C. caespitosa from the Aegean Sea. Our results demonstrated that this intense and prolonged marine heatwaves that affected the north Aegean Sea, had detrimental effect on shallow water stony corals by devastating 58% and 39% of B. europaea and C. caespitosa populations, respectively. Therefore, both the endemic Mediterranean solitary tooth-coral B. europaea and the colonial pillow coral C. caespitosa seem to be under serious threat in the study area, especially considering the predicted warming trend. Due to their sensitivity to warming, the species may serve as appropriate climate-change descriptors over the Mediterranean Sea to assess the resilience of MPAs. Moreover, seeing that other benthic species, such as sponges (S. muscarum, I. variabilis), bivalves (S. gaederopus) and ascidians (M. sabatieri) suffered as well, recursive monitoring using standard protocols are needed to detect future changes and impacts in larger spatial scales under climate crisis to safeguard the marine biodiversity.  Institutional Review Board Statement: Ethical review and approval were waived for this study due to the absence of a relevant framework for the usage of stony corals in marine research; moreover, the study was solely based on non-destructive sampling techniques (in-situ visual census and photography of coral specimens) without collecting or performing any experimental treatment that could torture the animals.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy reasons.