Revealing the Mediterranean Monk Seal (Monachus monachus)’s Cave Preference in Gökova Bay on the Southwest Coast of Türkiye

Abstract

The first cave-monitoring studies to be carried out on the southwest (SW) coast of Türkiye on endangered Mediterranean monk seals using camera traps occurred between 2017 and 2021 in five marine caves within Gökova Bay. The visual data obtained from the monitoring studies were evaluated to reveal the Mediterranean monk seals’ seasonal and diel cave use and identify the individual seals who were using the caves. Moreover, the necessary features and measurements of the identified caves were recorded to determine whether there were any correlations between the determined variables and monthly cave use by the monk seals. The results showed that cave use occurred mainly nocturnally, with the seals showing a diurnal activity pattern in the area. We evaluated 108,280 images/videos in total and identified 18 individuals using five caves in Gökova Bay. Three of these caves provided suitable characteristics for pupping, and two of them were used for pupping. A beta regression model revealed that the monthly cave use ratios varied seasonally, with more use in the fall season. Furthermore, the trends in annual cave use ratios, seasonality, wind speed, size of the wet area, luminance, number of cave-entry paths, and human activity were the best variables with which we could forecast the cave preferences of the seals.

1. Introduction

With the Caribbean monk seal becoming extinct in the twentieth century, and with its closest relative, the Hawaiian monk seal, currently belonging to another genus (Neomonachus), the Mediterranean monk seal (Monachus monachus) has become the sole representative of the genus Monachus [1]. As a result, it is now even more critical that research and conservation studies on the Mediterranean monk seal be carried out. The species is categorized as endangered globally, with fewer than 800 individuals remaining in the eastern Mediterranean Sea [2] and the northeastern Atlantic Ocean [3]. These pinnipeds face significant anthropogenic threats, such as habitat loss and degradation, negative interactions with fishers, the bycatch of subadult individuals, declining prey due to overfishing, marine pollution, including macro- and microplastics, and the effects of climate change [2,4].

Previous studies conducted on the Mediterranean monk seal utilized direct observations, as well as interviews with fishers and other observers, to gather seal-sighting records and determine the species’ status, distribution, and habitats using field surveys along the Turkish coast [5,6,7,8,9,10,11]. The decrease in Mediterranean monk seal numbers has been attributed to intentional killing by fishers, the entanglement and drowning of juvenile seals in fishing gear, uncontrolled coastal development leading to habitat loss, human disturbance, and overfishing [5,6,7,8,9,10,11]. Mursaloğlu [12,13] examined the biology and ethology of the species and identified the factors contributing to the population decline indicated above.

Mediterranean monk seals have previously been studied in two specific regions of Gökova Bay: the Bodrum and Datça–Bozburun Peninsulas. In addition to the identification of the monk seal’s habitats, threats to the species have been identified, and the importance of conservation measures to protect the Mediterranean monk seal population in these regions has been highlighted [6,7,8,10].

It was reported using seal observation data that there were approximately one hundred individuals living in Turkish waters between 1994 and 1998, including twelve and seven sightings of a minimum of two Mediterranean monk seals in the Bodrum and Datça Peninsulas, respectively [14].

Between 2009 and 2010, cave research expeditions were carried out to identify monk-seal habitats, and seal observation records were obtained from local people and fishers in the Gökova Special Environmental Protection Area (SEPA). As a result of this study, it was proposed that the borders of the Gökova SEPA be expanded to include important Mediterranean monk seal habitats near the northern border [15].

As one of the most important Mediterranean monk seal habitats, Gökova hosted its first cave-monitoring study, which was conducted between 2017 and 2021. In this study, we aimed to determine the species’ seasonal and circadian cave use and to identify individuals using the determined marine caves [16,17]. Prior to this, between 2016 and 2017, coastline surveys were carried out by using a boat to identify potential monk seal caves in Gökova Bay and to set up camera traps to reveal monk seal cave use.

More information is needed to control and reduce the threats to the Mediterranean monk seal, understand its critical habitats, and assess the level of threat to which this species is exposed in these areas. Although the Turkish Mediterranean coast is one of the strongholds of the Mediterranean monk seal and hosts the essential research cited above, there are still data gaps on cave use by the species. The lack of Mediterranean monk seal monitoring studies using camera traps in the marine caves in this area is the reason we prioritized studies in this region.

Türkiye hosts many international conventions that include Mediterranean monk seal conservation, such as the Bonn (Appendix I and II), Barcelona (fourth protocol species), Bern (Appendix II), Biodiversity (Eligible species), and CITES (Appendix I) conventions [14]. Gökova Bay contains two different SEPAs, and the Gökova SEPA management plans include activities to protect and monitor monk seals [18]. Furthermore, the species is protected by national legislation, including the Act on Hunting No. 3167, related Decisions of the Central Hunting Commission since 1977, and the Act on Aqua-Product Fishing No. 1380 [19]. However, the conservation initiatives for the Mediterranean monk seal and its habitats exist mostly on paper. Therefore, this study aims to use monitoring data as proof of monk seal cave use in the area to respond to the critical need for the designation and enforcement of core protection zones around marine caves used by seals for pupping.

2. Materials and Methods

2.1. Study Area

On the southwest Mediterranean coast of Türkiye, Gökova Bay (Figure 1) refers to the area behind the virtual line connecting two points on the Bodrum Peninsula’s Hüseyin Cape in the north and the Datça Peninsula’s Knidos Cape in the south [20]. The Bay is located within the Mediterranean Basin global biodiversity hotspot and the World Wildlife Fund Global 200 Ecoregion [21,22]; it covers a marine area of 1851 km² with rich biodiversity. Gökova Bay is home to multiple endangered species, such as the Mediterranean monk seal (Monachus monachus), the Dusky Grouper (Epinephelus marginatus), and the Sandbar shark (Carcharhinus plumbeus) [22].

Gökova Bay embodies one of the largest marine and coastal protected areas (MCPAs), 1092.79 km² in Türkiye, including seven no-fishing zones (NFZs) where all types of commercial fishing activities are prohibited, according to Declaration no. 5/1 on the regulation of Commercial Fisheries (Declaration No. 2020/20) (Official Gazette no. 31221). In addition, there is a 270 km² area designated as a no-trawling and no-purse-seining zone in Gökova Bay, according to the coordinates given in Official Gazette no. 31221 [22]. From east to west, Gökova Bay is approximately 92 km long, with a width of approximately 20 km from north to south in the middle of the bay (Figure 1) [23].

2.2. Field Surveys and Data Collection

At the beginning of the study, face-to-face interviews were conducted with local fishers in Gökova Bay to extract local information on the locations of possible monk seal caves. In light of local knowledge, cave research expeditions were then completed along most of Gökova Bay’s 322 km coastline (Figure 1). The aim was to identify marine caves suitable for monk seal use as part of the Mediterranean Monk Seal Monitoring Program conducted by the Mediterranean Conservation Society (MCS) between 2016 and 2017 [24].

The team began surveying locations around the field station using the MCS’s 4.2 m long inflatable boat with a 50 HP outboard motor, and explorations were subsequently resumed using a 22 m long gullet as a base. An MCS inflatable boat was also used to explore the coastline of Gökova Bay in great detail.

The marine caves were examined via skin diving in order to determine whether they were suitable for monk seal use, as well as for the camera trap installations.

During the cave research expeditions, the characteristics of marine caves were examined to determine their suitability as potential pupping caves, as described in previous studies [25,26]. In addition to this, indicators such as odor, sleeping depressions, tracks, fur, and scats on cave beaches/rocky platforms were examined to understand monk seal usage. Since most of the evidence of seal presence was not in the splash zone, the remains and tracks in the dry zones were used only as an indicator of cave use [16].

The team identified five potential monk seal caves (C1–C5) with surface entrances, including the cave with an artificially implemented dry ledge [24], and installed camera traps with PIR-based motion detectors. One to three camera traps were set up inside depending on the cave’s size and the number of suitable dry areas for monk seal use. A total of eight camera traps were installed in the designated marine caves. Two of these caves were in the Gökova SEPA, and three of them were in the Datça–Bozburun SEPA (Figure 1). The caves were identified on different research expeditions, so the monitoring did not begin simultaneously for each cave (

Table 1. Monitoring periods of camera traps in each marine cave.

Cave Code—Camera Trap No	Monitoring Period
C1—1	February 2017–December 2021
C1—2	January 2018–December 2021
C2	February 2017–December 2021
C3—1	October 2017–December 2021
C3—2	October 2017–December 2021
C3—3	April 2019–December 2021
C4	January 2017–December 2021
C5	June 2019–December 2021

).

During the cave-monitoring study, camera traps were set to record three images and a 20 s video each time the camera was triggered. The interval between the two consecutive triggers was set at 5–10 min. The team conducted routine cave checks every two or three months to collect visual data and change the camera trap batteries. If necessary, the damaged cameras were replaced with new ones. Considering that not all caves could be visited for a routine cave check in each field study due to their locations, a total of seventy field trips were made during the monitoring study.

In October 2019, the team took the initial step towards using a better monitoring technique and built a solar-powered, real-time, no-glow infrared, totally waterproof video monitoring system consisting of two cameras, a recorder, hard drives, Wi-Fi, and a GSM transmitter installed in a cabinet on top of cave C1 for the first time in Türkiye. The footage was recorded by the system and relayed through GSM transmitters.

With this new system of recording 4K-quality video in the cave, it enabled the differentiation of individual seals, easily identifying their sex and age group without the team needing to enter the caves for routine cave checks.

Traditional camera traps are not well suited to marine cave environments, where high humidity, salinity, and intense wave activity prevent proper imagery, drain batteries, and eventually break the equipment. However, this new monitoring system provides decisive results in the surveillance of Mediterranean monk seals [27,28].

2.3. Data Evaluation

After the completion of each routine cave check, we extracted the date and time data of the footage and sorted them chronologically according to their trigger factors. The durations of cave use periods were then calculated as the time difference between the first and the last monk seal footage trigger. If the time of exit of the seals from the caves was at the turning of the season or month, the duration of cave use was split and evaluated within the following month or season for that cave entry or exit.

The cave use period was divided into dark and light hours, considering each day’s sunrise and sunset times (Boğaziçi University, Kandilli Observatory, and Earthquake Research Institute, Astronomy Laboratory website dataset was obtained on 15 January 2021) when the images were recorded to reveal diel cave use periods.

We evaluated 108,280 images/videos, including single events, from eight camera traps in five marine caves. In total, 22,543 of them showed seal presence, but single events were excluded from the evaluation because the length of stay in the cave could not be calculated. The footage recorded between October 2019 and December 2021 by the real-time video monitoring system was not included in these numbers since it was a 24/7 continuous recording system.

Moreover, the necessary features and measurements of the identified caves were recorded [25,26] by using a laser distance meter to reveal whether there was any correlation between the determined factors and the duration of seals’ cave use periods via statistical analyses.

For the identification of individual seals, seal recordings were grouped according to sex, body size, and maturity stage [29,30]. Furthermore, distinctive features of individual seals, including predominantly dorsal natural marks, scars, and fur colorations (mostly on the lateral part of the body), were used for differentiating individuals. After identifying the individuals, we confirmed whether they used any of the other marine caves under study. The distances between the caves (

Table 2. Distances between the identified monk seal caves.

Caves/Distance (km)	C1	C2	C3	C4	C5
C1	-	16.7	49.1	48.3	68.1
C2	-	-	64.3	63.5	83.7
C3	-	-	-	0.8	19.5
C4	-	-	-	-	20.3
C5	-	-	-	-	-

) were measured by the “add path” feature of Google Earth Pro [31] to understand the distance between the caves used by the identified individuals [32].

Statistical Analyses on Monk Seal Cave Preference

There are significant factors affecting Mediterranean monk seal cave preferences, especially for pupping [25,26]. The previously determined variables (

Table 3. Physical and environmental variables, which may have an impact on the cave use of the monk seals.

Cave-Morphology-Related Variables	Description	Expected Direction of Effect
Entrance width (m)	Distance at water level between two walls of the cave entrance	Negative
Entrance height (m)	Vertical distance between the center of the entrance and the ceiling of the cave	-
Entrance depth (m)	Vertical distance between the center of the entrance and the bottom of the sea	Positive
Number of entry paths	Total number of paths leading into the interior of the cave	Positive
Corridor length (m)	Direct distance (m) at sea level between the center of the entrance and the middle of the main beach	Positive
Total dry area (m²)	Size of all dry areas inside the cave	Positive
Main used dry area (m²)	Size of the dry area used by monk seals inside the cave	Positive
Wet area (m2)	Size of the water surface area in the interior of the cave	Positive: having a protected pool inside the cave; negative: decreasing dry area
Weather-related variables	Description	Expected direction of effect
Wind speed	Wind speed (kph and bft) and direction	Negative
Wind speed and direction and cave entrance direction	Wind speed (bft) and direction in relation to the direction of the cave entrance (low, medium, and high)	Negative (in cases where the cave entrance direction and the wind direction overlap)

and

Table 4. Species-based variables, which may have an impact on the cave use of the monk seals.

Variables Related to Biological and Behavioral Needs of the Species	Description	Expected Direction of Effect
Beach visibility	Ability to see the dry surface in the interior of the cave from 30 m outside the entrance of the cave (not visible—0; partially visible—50; visible—100)	Negative
Luminance	Amount of daylight inside the cave (low—0; medium—50; high—100)	Positive (monk seal preference for pupping is that caves should be neither completely dark nor completely luminous)
Main beach substrate	Type of substrate of the main dry area (sand/pebbles, stones, boulders, and rock platform)	For pupping caves, soft substrate is preferable, especially during parturition period
Substrate size	Sand/pebbles—12.5; stones—37.5; boulders—67.5; rock platform—100	Especially for pupping caves, as the size of the substrate particle decreases, the likelihood of preference increases
Human activity	Intensity of the human activity within a 2 km radius from the cave’s entrance (low—settlement further than a 2 km radius; medium—settlement within a 2 km radius; high—settlement closer than a 2 km radius)	Negative

) [25] were evaluated via beta regression using the software Stata 17 [33]. During the data evaluation process, we distributed these factors into the following groups: cave morphology, biological and behavioral needs of the species, and weather-related variables (Table 3 and Table 4).

We assessed the cave use periods of monk seals and considered monthly cave use proportions in the statistical analyses to avoid misjudgments that could be caused by the monitoring study not starting at the same time for each cave. In addition, considering that each camera trap had failed for certain periods, or the batteries had run out earlier than the field visit to replace them, we accepted that such problems were homogeneously scattered among the caves during the monitoring study and were an unavoidable feature of working in these conditions (Appendix A). The proportion of the duration of the monk seals’ stay in the caves during the entire period of cave use was taken as the dependent variable. Annual trends and caves were included in the models as independent variables.

The monitoring data collected from the caves were analyzed via beta regression, which is commonly used by practitioners to model variables that assume values in the standard unit interval (0,1). Beta regression is used to model the relationship between these values and one or more predictor variables. The predictor variables can be continuous or categorical, and the model can be extended to include interaction terms [34,35].

However, the coefficients of the variables cannot be directly interpreted using a beta regression. The signs of the coefficients indicate the effect of increasing or decreasing in the dependent variable. The magnitude of this effect should be calculated as the marginal effect [36].

The beta regression model is based on an alternative parameterization of the beta density in terms of the variate mean and precision parameters. The beta density is usually expressed as follows:

$$f(y;p,q)=\frac{\Gamma (p+q)}{\Gamma \left(p\right)\Gamma \left(q\right)}{y}^{p-1}{(1-y)}^{q-1},0<y<1,$$

where p and q > 0 and Γ(·) is the gamma function. The scientists [34] proposed a different parameterization by setting μ = p/(p + q) and φ = p + q:

$$f(y;\mu ,\phi )=\frac{\Gamma \left(\phi \right)}{\Gamma \left(\mu \phi \right)\Gamma \left(\right(1-\mu \left)\phi \right)}{y}^{\mu \phi -1}{(1-y)}^{(1-\mu )\phi -1},0<y<1,$$

with 0 < μ < 1 and φ > 0. We write y~B(μ,φ). Here, E(y) = μ and VAR(y) = μ(1 − μ)/(1 + φ). The parameter φ is known as the precision parameter since, for fixed μ, the larger the φ is, the smaller the variance of y is. φ⁻¹ is a dispersion parameter [35,36].

Stepwise beta regression was used in the statistical modeling to select a subset of predictor variables for inclusion in the model. The reasons that we selected stepwise beta regression were as follows: to identify a subset of predictors that are most strongly related to the outcome variable and reduce the number of variables in the model to improve the interpretability of the results; to consider the statistical significance of each predictor variable and only include those that are most strongly related to the outcome; because once a subset of predictors is selected, the resulting model may be more accurate and easier to interpret; and, finally because stepwise beta regression is easy to implement and is available in many statistical software packages.

The main used dry area by seals was specific parts of the cave, which we were certain about based on camera trap footage and indicators such as sleeping depressions, tracks, fur, and scats. During the monitoring study, the total dry area in the cave was considered as the land area that remained dry, except when there was strong wave activity, which also covered the main used dry area. Finally, the wet area was interpreted as the sea area, which was constantly covered with water (Table 3). Regarding exposure to wind, when a horizontal 180-degree arc is imagined at the entrance of the cave, it is predicted that wind can blow from all directions in this arc and create wave activity that can have an effect inside the cave (Table 3). Regarding human activities, these were defined as 1 if they were high and 0 if they were low and medium, based on a dummy variable (Table 4).

3. Results

3.1. Seasonal and Diel Cave Use by the Mediterranean Monk Seals

Figure 2 and Figure 3 show the current state of the raw data obtained from the monitoring study. The seasonal cave use of the Mediterranean monk seals across the years in relation to the number of haul-outs and seals using the caves is shown in Figure 2.

During the study, the number of seals generally increased from winter to the end of spring. During the winter season, the maximum number of seals using cave C4 was a group of seven seals—three pups and four adults. This was the highest recorded at any point in the duration of the study period (Figure 2). We obtained footage of newborns in the caves suitable for pupping (C3 and C4).

In Gökova Bay, we found that the monk seals used the marine caves more intensively in the hours between sunset and sunrise, mainly nocturnally, and the seals showed diurnal activity patterns like the seals in the Cilician colony [16]. Out of the 197 haul-outs, 89 of them started after sunset, 64 of them started before sunrise, and 44 of them started during the daytime (Figure 3).

Moreover, the visual data obtained from the camera traps in the designated marine caves were evaluated to reveal the monthly cave use ratios by the Mediterranean monk seals to eliminate the differences between the start dates of monitoring for each cave. These data were used for statistical analyses (

Table 5. Mean and standard error of the monthly cave use ratios for each monk seal cave (annual number of haul-outs for each cave is shown in Figure 2. Empty cells indicate that there is no record for that year in that cave).

Cave/Year	2017	2018	2019	2020	2021
Cave 1	0.3411 ± 0.1382	0.2306 ± 0.0955	0.3592 ± 0.0938	0.2790 ± 0.0506	0.1048 ± 0.0661
Cave 2	0.3333 ± 0.0576		0.3343 ± 0.3329	0.0683 ± 0.0641
Cave 3	0.2279 ± 0.1768	0.3336 ± 0.1660	0.1753 ± 0.0560	0.6545 ± 0.3255	0.1458 ± 0.0479
Cave 4	0.0208 ± 0.0167	1.0000 ± 0.0000	0.1372 ± 0.0850		0.0476 ± 0.0137
Cave 5	NA	NA		0.5217 ± 0.1871	0.2300 ± 0.0636

NA indicates that there was no monitoring study in Cave 5 between 2017 and 2018.

).

3.2. Identification of the Mediterranean Monk Seals: Photo—ID Study

Between January 2017 and December 2021, eighteen monk seal individuals were photographed in Gökova Bay’s endorsed marine caves. During the study period, we identified seven individuals in 2017, one individual in 2018, three individuals in 2019, one individual in 2020, and, six individuals in 2021. The evaluation of monk seal recordings revealed that five of the identified individuals were pups, and one of them was a juvenile on the first recording date [29,37]. The results show that six identified seals used more than one cave in the area, and two of the seals can be seen in Figure 4.

Four of the identified individuals were recorded in C1 and C2 from only one to three haul-outs in 2017 and 2018, after which we did not obtain any footage of them using the caves.

Only two of the caves under study, C3 and C4, which are separated by about a km, were where the scientists found pups, and the mothers and pups were seen using both caves. According to [25], within 10 days after parturition, mother and newborn pup pairs were observed travelling distances of up to 2 km to reach a more protected marine cave against intense wave activity [38]. The marine cave C4 was used by seven individuals, including mother–pup pairs at the same time (Figure 2 and Figure 5). Among the adult seals using the designated caves, only two of them were found to be male.

One of the seals that we identified as a molted unweaned pup in December 2017 was found dead on the C4 beach in November 2018, estimated to be 1 to 1.5 years old. The cause of death was unknown since a necropsy could not be performed due to advanced decay (Figure 6).

For the cave with a manually implemented artificial dry ledge (C5), a juvenile used it for resting initially in 2020 [24], after which we recorded two other seals using the cave in October 2021. Given the impact of sea level rise that could be caused by climate change in the future [2], the implementation of the artificial dry platform as an alternative habitat for the Mediterranean monk seal is thought to offer the species potentially important support in pupping and resting.

3.3. Evaluation of the Monk Seal Cave Use Preference

Stepwise beta regression analysis was performed in order to obtain the most significant variables affecting the monthly cave use proportions of seals (Table 3 and Table 4).

In the beta regression estimation, a positive coefficient of a variable indicates that an increase in the predictor variable is associated with an increase in the mean of the dependent variable. During the data evaluation process, all the variables in Table 3 and Table 4 were included in the stepwise and customized beta regression models. The final estimation results from the analyses are shown in

Table 6. Stepwise beta regression results and marginal effects.

	Estimation Results				Marginal Effects
Cave Use Rate	Coefficient	Std. err.	z	P > |z| **	dy/dx *	P > |z| **
Cave use rate Trend in annual cave use ratio	−0.1712323	0.0819284	−2.09	0.037	−0.0391161	0.037
Winter	−0.9463345	0.2548635	−3.71	0.000	−0.2114705	0.000
Spring	−0.2954965	0.3266665	−0.90	0.366	−0.065301	0.348
Summer	−0.1040194	0.3210316	−0.32	0.746	−0.0235163	0.743
Wind speed_bft ≤ 3	−0.6664194	0.2790405	−2.39	0.017	−0.1460101	0.012
Wind speed_3.01≤ bft ≤ 4	−0.6257136	0.2817903	−2.22	0.026	−0.1343311	0.017
Luminance_High	−0.4766574	0.2411808	−1.98	0.048	−0.1083527	0.046
_constant	1.151355	0.5260535	2.19	0.029
scale _constant	−0.1475534	0.0851323	−1.73	0.083

* dy/dx is for a discrete change in the dummy variable from 0 to 1; ** p < 0.05 indicates statistical significance.

and

Table 7. Customized beta regression results and marginal effects.

	Estimation Results				Marginal Effects
Cave Use Rate	Coefficient	Std. err.	z	P > |z| **	dy/dx *	P > |z| **
Cave use rate Number of entry path	14.11748	5.995442	2.35	0.019	3.235325	0.019
Human activity_High	−4.383445	2.072883	−2.11	0.034	−0.3892708	0.000
Luminance	−1.695895	0.8018865	−2.11	0.034	−0.388651	0.035
Wet area	−0.0795416	0.0340877	−2.33	0.020	−0.0182287	0.020
_constant	−5.824936	2.105663	−2.77	0.006
scale _constant	−0.2539095	0.0824504	−3.08	0.002

* dy/dx is for a discrete change in the dummy variable from 0 to 1; ** p < 0.05 implies statistical significance.

.

According to the results of the analyses, the trends regarding the annual cave use ratio, seasonality, wind speed on a bft scale, size of wet areas inside the caves, luminance (amount of daylight inside the cave), the intensity of human activity, and the number of paths entering the caves were the best variables that could predict the cave preference of monk seals.

In comparison with the fall season, the monthly cave use proportion in winter was 21.1% less, and fall did not show any significant difference compared to the spring and summer months (Table 6).

Due to the distribution of data, we categorized wind speed variables as bft ≤ 3, 3.01 ≤ bft ≤ 4, and 4 < bft, and these influenced the monthly cave use proportion of monk seals. According to the 4 < bft category, the wind speed in the category bft ≤ 3 reduced the monthly cave usage rate by 14.6%, whereas the wind speed of the category 3.01 ≤ bft ≤ 4 caused a decrease of 13.4% in the monthly cave use proportion. Moreover, high luminance caused a decrease of 10.8% in the proportion of monk seal cave use compared to low or medium luminosity. Finally, the proportion of monk seal cave use proportions decreased by 3.9% every year (Table 6).

According to the customized beta regression results, it was found that the monthly cave use rates increased by 3.23 times when the number of entry paths to the cave increased by one unit. The high intensity of human activity showed a 38.9% decrease in cave use proportion in comparison to medium- or low-intensity human activity. Luminance was found to cause a 38.8% decrease in cave utilization rates in every luminance level increase from low to high. Finally, when the size of the wet areas inside the caves increased by 1 m², the cave usage proportion decreased by 1.8% (Table 7).

4. Discussion

To evaluate and then eliminate the threats to Mediterranean monk seals, it is necessary to identify the location, dimensions, and features of the marine caves, which are the current habitat of the species due to human pressure, and to reveal the purpose and frequency of use by the species to inform potential conservation actions. With this study, we aimed to cover this gap in Gökova Bay where no previous monitoring study of Mediterranean monk seals in marine caves had been conducted.

According to the monitoring study results, based on the morphological characteristics of newborns recorded in the caves, we determined that births generally took place in October and November in Gökova Bay on the southwest coast of Türkiye, aligning with populations in other areas of the Eastern Mediterranean [3,39]. Cave C4 hosted the largest group, with three pups and four adult females observed during the winter months of 2021. The presence of a shallow and protected pool in cave C4, as well as both an open beach and a dim cavity with a ceiling height of about 70 cm within the cave, can be determined as the factors important for seals in selecting this cave for nursing due to its sheltered structure. In this case, it can be inferred that cave C3, about a km away and about five times larger than the neighboring cave, is not as preferred as cave C4 since it was dominated by adult females. As previous studies have described [40], mother–pup bonding and the social bonding of pups, vocalizations for recognition, and caring for pups by other females were also observed during the study period.

In cave C3, one of the identified female adults was recorded remaining in the cave continuously for 11 days between the 9th and 20th of January 2021 without interruption, and footage of this female with a newborn pup on the 12th of October 2021 led us to conclude that the adult female may have been in the early stage of the prenatal period at that time since the gestation period lasts approximately 9–11 months [3].

After the evaluation of the cave characteristics, weather-related factors, and the factors related to the biological and behavioral needs of the species, we conducted statistical analyses to find out which variables are more important in the cave preference of the monk seals. As a result of the stepwise and customized beta regression analyses, the following variables became prominent: trend in annual cave use ratio, seasonality, wind speed on a bft scale, size of wet areas inside the caves, luminosity, number of entry paths, and intensity of human activity. This aligns with the results of a study conducted in the National Marine Park of Alonnisos, Northern Sporades (NMPANS) [25], where the variables luminance, human activity, number of entry paths, and wind speed were presented as the main determinants of suitable breeding habitats for the species.

The statistical analysis results of five years of cave monitoring showed that the cave use proportions of the seals varied seasonally; it was determined that in comparison to fall, the cave use proportions of the spring and summer months stayed the same, but there was a decrease in the winter months. Considering that the pupping season of Mediterranean monk seals in the Eastern Mediterranean occurs in a period extending from fall to winter [2], with a longer whelping season from July to December in the Northern Sporades [37], the results we obtained were supported by others conducted in the Eastern Mediterranean. Moreover, the result of monthly cave use ratios decreasing by 3.9% each year could be because the caves were never used by seals in some months. This can be interpreted as the seal individuals going to other regions in that period of time, and this requires further investigation, such as the movements of seals to adjacent regions, to shed light on this decrease.

In Gökova Bay, the distance between the five designated caves varies from ~1 km to 84 km. A previous study [41] recorded an identified individual covering max. ~78 km of straight distance and travelling approx. 280 km distance in total over a three-month period in Greece. This justifies the recording of the same individuals using more than one cave during the study period.

The disturbance caused by wave activity related to wind speed and direction during the cave use of the seals affects mother–pup pairs especially and may force them to abandon the pupping cave and move to another more sheltered cave [16]. Unlike the study conducted in Greece [25], we considered not only the data for October but also the data for all months to assess the effects of weather-related factors on the cave use ratio of seals all year round by using the statistical analysis model. We also considered a wind speed of less than four on the bft scale due to the distribution of data, unlike this previous study [25].

The results of the beta regression model indicated that a decrease in wind speed on the bft scale caused a decrease in cave use proportions. This implies that the caves served as a shelter against adverse weather conditions during cave use periods. Since the wind mostly blew in directions that would not affect the study caves when the seals were using them, the interaction of the cave entrance direction and the monthly prevailing wind direction did not seem significant, according to the statistical analyses. Since more wet areas in the cave mean less space for dry areas, especially during bad weather conditions, as in the statistical analysis results, a larger wet area correlates with a decrease in the cave utilization rates of the species. We found that Mediterranean monk seals prefer dry areas for resting and nursing within the caves. The existence of the dry area closed off from intense wave activity is important in cave selection, especially due to the risk of pup washout. The size of dry areas inside the caves, therefore, influences cave preference, especially by lactating females, for the survival of newborns [25,26,42].

Another factor related to the cave preference of monk seals was luminance. Since the species previously used open beaches for resting and pupping, the requirement for daylight within the cave is not surprising [25]. The species, therefore, appears to need a certain level of daylight by its nature, but based on its behavior in this study, it does not choose areas of high luminosity. This may be because an increase in luminance increases the visibility of the cave interior and may increase the risk of attracting attention [25].

As the number of entry paths to the caves increases, it appears that Mediterranean monk seals may feel safer, possibly due to the presence of multiple escape routes [25]. Our results also supported this approach and pointed to the existence of a directly proportional increase in the number of entry paths and the cave use ratios.

The other distinctive factor that has a negative impact on the cave preference of seals is the intensity of human activity close to the caves. Considering the most critical period being the first six months for pups following parturition, human disturbance in the proximity of the pupping sites can cause the caves to be abandoned, can result in the separation of mother–pup pairs, or can even result in miscarriage [16,25,28,43].

In Gökova Bay, No Fishing Zones (NFZs) were designated in 2010, and these areas have been actively enforced against illegal fishing through the MCS’s voluntary marine ranger system to support the Turkish Coast Guard Command since 2012. This concerted effort has yielded positive results, as it has led to an increase in fish stocks due to the spillover effect [20,22]. The subsequent increase in fish biomass has the potential to reduce negative interactions between fishers and monk seals and increase the seal’s predation in the area. As a result, these NFZs could have become more favorable habitats for monk seals. Additionally, the increased number of NFZs could provide crucial protection within the seals’ range, further reducing the risks of by-catch and entanglement not only within Gökova Bay but also in other areas.

Considering the significant threats identified to Mediterranean monk seals, including habitat loss and human disturbance in marine caves, particularly the drowning of pups by accidental entanglement in fishing gear, there is strong evidence that human activities beyond essential scientific research should be prevented by the declaration of core protection zones containing the pupping caves and the surrounding areas in their vicinity, like the other stronghold areas of the species [44,45].

In addition to the legislative aspect, as in the Cabo Blanco Peninsula and the Desertas Islands Nature Reserve on the archipelago of Madeira [46,47], core protection zones to be established should be monitored continuously using solar-powered, real-time monitoring systems that will report when human activity is detected using artificial intelligence (AI), and these areas, enforced by regular and instant patrols, will thus become a disincentive, meaning they play a crucial role in the survival of this species.