Effect of Substratum Structural Complexity of Coral Seedlings on the Settlement and Post-Settlement Survivorship of Coral Settlers

: The substratum structure is critical for facilitating settlement and increasing the survivorship of coral settlers. However, knowledge about its structural complexity is largely lacking. In this study, we examined the effect of complexity on the settlement and post-settlement survivorship of coral settlers using four types of structures: groove, using a CSD (Coral Settlement Device, 4.5 cm ϕ × 2.5 cm H, top-shaped ceramic); flat, using a CP (Ceramic Plate, 29.5 cm L × 3.1 cm W × 0.9 cm H, unglazed ceramic plate); linear, using a CN (Coral Net, mesh size 19 mm, biodegradable plastic net); and wrinkle, using a SS (Scallop Shell, 11.0 cm in shell length). The complexity was obtained from the ratio of the surface area to the vertically projected area of the substratum. The substratum sets were installed in the coral reef around the Ryukyu Islands every May from 2012 to 2014. After about 2 or 6 months of spawning, a certain number of substratum types were sampled, and the number of coral spats that settled on them was counted by taxa classified into Acropora, Pocilloporidae, Millepora, and Others. The larval settlement rate in the first set of samples and the survivorship of coral spats in the second set of samples were estimated. The mean settlement rate was, in order, the CSD; SS; CN; and CP, and the mean survivorship was, in order, the CSD; CP; SS; and CN, over three years. A positive correlation was found between the structural complexity, mean settlement rate, and mean survivorship. Our results show that the structural complexity of coral seedlings affects the settlement of coral larvae and the survivorship of coral spats.


Introduction
Coral reefs have been declining globally in recent years. In total, 19% of the world's coral reefs have already deteriorated, and more than 60% are under serious threat [1]. The causes include increased environmental loads due to human activities, such as sediments and nutrients in zooxanthellate corals (hereinafter referred to as corals); predation by crownof-thorns starfish; and coral diseases. Coral bleaching due to rising water temperatures caused by global warming also poses a long-term threat to coral reefs [2]. These impacts have put coral reefs in a potential extinction crisis in one-third of the world [3].
It is important to remove anthropogenic impacts to allow for the recovery of degraded coral reefs. However, as noted by Rinkevich (2005) [4], if the reefs have declined beyond their innate ability for recovery, then coral outplanting is a potentially significant method of reef restoration. Coral outplanting contributes to the recovery of ecosystems, the expansion of larval sources, and the rehabilitation of marine habitats by artificially restoring coral communities in areas where coral reefs cannot recover under their own power [5]. Due to rising water temperatures and ocean acidification caused by global warming, there is also concern about the survival of outplanted coral [6]. However, coral reefs that have  1 Height. 2 26.0 in 2013 and 2014 experiments. 3 Thick. 4 9.5 in 2013 experiment. 5 Average of 10 sets.  A top-shaped ceramic (brown color) with a diameter of 4.5 cm and a height of 2.5 cm containing about 50% steel slag. Radial grooves on the underside facilitate the settlement of coral larvae. The surface has rich, minute irregularities. A bundle of 11 CSDs connecting a lower leg with an upper hole was loaded into a plastic case [12]. Eight bundles with 88 CSDs were loaded into one case, which was fixed horizontally on a stainless steel frame.
• CP (Ceramic Plate): An unglazed ceramic (brown color) with a length of 29.5 cm, a width of 3.1 cm, and a thickness of 0.9 cm. Its surface has rich, minute irregularities. Note that the length of the plate installed in 2013 and 2014 was 26.0 cm. There is a groove for fragmentation with a depth of about 2 mm at intervals of about 3 cm on one side (about 2 cm for installations in 2013 and 2014), which can be divided in the field at the time of outplanting for seedlings. In 2012, 2 plates were made into one set, and 9 sets for 18 plates were fixed vertically on the frame. In 2013 and 2014, 2 sets of 20 plates were loaded in the same plastic case as the CSD and fixed vertically on the frame.

• CN (Coral Net):
A hard plastic mesh base with a mesh size of 19 mm, composed of biodegradable material. The surface shape is glossy and smooth. In 2012, two nets measuring 30 cm × 60 cm were fixed horizontally on the frame. In 2013, 15 nets measuring 10 cm × 10 cm were stacked to form a set, and 2 sets of 30 nets were horizontally mounted on the same case and fixed to the frame.
The central parts of several pairs of the bivalve Mizuhopecten yessoensis with lengths of 10-12 cm (11.0 cm on average), were skewered by an iron rod. Two sets of 28 shells in 14 pairs, for a total of 56 shells, were fixed vertically on the frame in 2012. In 2013, two sets of 12 shells in 6 pairs, for a total of 24 shells, were installed in the same case and fixed vertically on the frame.

Calculating Structural Complexity of the Substratum
Complexity was calculated with the following equation using the vertically projected area and surface area of each substratum shown in Table 1. Substrate surface roughness is known to physically affect planktonic larval settlement [13]. The ratio of the roughness spacing to the roughness height of the substrate has been suggested as a theoretical value for the roughness index [14]. In this study, we applied this theory to the substratum used and showed the complexity as the ratio of the surface area in a certain area of the substratum. Complexity = surface area/vertical projected area × 100

Study Sites
Based on the results of previous surveys for coral distribution, we selected three highcover areas covered with Acropora, which predominantly occurs in the Ryukyu Islands. Furthermore, we surveyed the distribution of juvenile coral at each site to confirm the recruitment of Acropora. The number of deployment sites was 4 in the Kunigami area, 4 in the Motobu area, and 6 in the Zamami area. Those in fringe and patch reefs were located at a depth of about 1 to 6 m ( Figure 2, Table 2). Based on the results of previous surveys for coral distribution, we selected three highcover areas covered with Acropora, which predominantly occurs in the Ryukyu Islands. Furthermore, we surveyed the distribution of juvenile coral at each site to confirm the recruitment of Acropora. The number of deployment sites was 4 in the Kunigami area, 4 in the Motobu area, and 6 in the Zamami area. Those in fringe and patch reefs were located at a depth of about 1 to 6 m ( Figure 2, Table 2).

Deployment and Sampling of the Substratum
Acropora is dominant in the Kerama Archipelago of Okinawa, and it spawns annually, starting around the full moon in May up to June [15], and once a year, it does so simultaneously [16]. The settlement rate reaches a maximum of 80% or more on the 8th day after spawning begins [7]. On Ishigaki Island in southern Ryukyu, more than 80% of Acropora settle 5 to 8 days after spawning [17]. Based on these data, we considered the fastest settlement period of Acropora to be 5 days after spawning. Coral settlement is inhibited by microalgae and favored by crustose coralline algae [18,19]. Since the adhesion of crustose coralline algae to the window of the underwater observatory in central Japan is noticeable around the 14th day after cleaning the glass [20], the substratum was installed at least 9 days before the full moon. There is a view that the earlier the installation time, the better. However, if the installation period is too long, the sessile of the microalgae will increase, meaning that frequent cleaning must occur. The substrata were installed around the full moon on 4 June 2012, 24 June 2013, and 13 June 2014, as shown in Table 2.

Deployment and Sampling of the Substratum
Acropora is dominant in the Kerama Archipelago of Okinawa, and it spawns annually, starting around the full moon in May up to June [15], and once a year, it does so simultaneously [16]. The settlement rate reaches a maximum of 80% or more on the 8th day after spawning begins [7]. On Ishigaki Island in southern Ryukyu, more than 80% of Acropora settle 5 to 8 days after spawning [17]. Based on these data, we considered the fastest settlement period of Acropora to be 5 days after spawning. Coral settlement is inhibited by microalgae and favored by crustose coralline algae [18,19]. Since the adhesion of crustose coralline algae to the window of the underwater observatory in central Japan is noticeable around the 14th day after cleaning the glass [20], the substratum was installed at least 9 days before the full moon. There is a view that the earlier the installation time, the better. However, if the installation period is too long, the sessile of the microalgae will increase, meaning that frequent cleaning must occur. The substrata were installed around the full moon on 4 June 2012, 24 June 2013, and 13 June 2014, as shown in Table 2.
In 2012, a stainless steel frame (60 cm × 60 cm) equipped with the substratum cases was installed by sticking iron stakes through its four corners so that the frame bottom was fixed 30-50 cm above the seabed. In 2013 and 2014, the frames were fixed to a steel pipe frame assembled on the seabed. Regarding the set angle of the substrata, it is known that vertical installation can obtain more settlement in shallow places with abundant light because this avoids the influence of sedimentation [21]. Therefore, the CPs and SSs with flat shapes were installed to make a wide, vertical surface. After installation, in every period between July and August (about 2 months after spawning) and between December and January (about 6 months after spawning), a certain number of substrata were sampled to measure the coral settlers (Table 2).

Measurement of Coral Settlers
After sampling, the substrata were divided into units to compare them as each outplanting unit: 1 piece of CSD; 10 pieces of CP in 2012 and 13 pieces in 2013 and 2014; a piece 8.7 cm × 9.5 cm of CN; and 4 pieces of SS; see Figure 1. Coral settlers on the units were classified into four types (Acropora, Pocilloporidae, Millepora, and Others), and we recorded the numbers with a loupe or stereomicroscope. In the first sampling, the number of settlers was measured; then, the settlement rate, defined as seedling collection rate in this study, was calculated (=number of units settled/number of units installed × 100). In seedling production, the number of substrata containing at least one coral is important [22]. In the second sampling, the same measurement as in the first sampling was performed, and the survivorship was calculated (=number of settlers in the second sampling/number of settlers in the first sampling × 100). Since Pocilloporidae coral spawns every month except in winter [23], the number of settlers in the second sampling likely included colonies spawned after the first sampling, having grown for about 4 months. Pocillopora damicornis was reared downward with a net to prevent grazing on the seabed after settlement in a tank, and it had a mean maximum diameter of 5 mm from 13 to 21 weeks after settlement [24]. Therefore, among the number of Pocilloporidae settlers in the second sampling, those with a diameter of 5 mm or less were excluded.

Structural Complexity of the Substratum
The complexity of each substratum was measured. The CSD had the highest complexity at 3.77, followed by the CP at 2.65, and there was no difference between the CN (2.00) and the SS (2.00) ( Table 3). Because the CSD had a three-dimensional structure and radial grooves formed on the underside, its complexity was high; on the other hand, the CP, CN, and SS had almost flat shapes. Table 3. Structural complexity of the substrata.  Table 4 shows the mean settlement rate for each study area in 2012. In the Kunigami area, the mean settlement rate order was the SS (28.3%) and CP (26.3%), followed by the CSD (13.6%) and CN (8.5%). The mean rate in the Motobu area was also the SS (17.3%), CP (13.8%), CSD (4.6%), and CN (4.3%), in that order. In the Zamami area, although the mean rate of the SS was as high as 38.5%, the order of the others was different compared with other areas: CSD (19.7%), CP (9.2%), and CN (8.3%). Even in an average of three areas, the SS was higher at 28.0%, followed by the CP (16.4%) and CSD (12.6%), and the CN was the lowest at 7.0%. In the settlement ratio of coral taxa settled on the substrata, we found that Acropora prevailed by over 50% in all substrata except in the CN in the Zamami area ( Figure 3). In 2013, the settlement rates recorded at site K5 in the Kunigami area showed that the CSD (78.4%) and CN (73.3%) were higher than the SS (47.9%) and CP (45.0%). At site Z6 in the Zamami area, the CN was also higher (33.3%), followed by the SS (28.1%), CP (25.4%), and CSD (19.3%). The highest average for the two areas was the CN (53.3%), followed by the CSD (48.9%), SS (38.0%), and CP (35.2%) ( Table 4). In the proportion of coral settlers by taxa, the ratio of Acropora was small in both areas. Although the ratio at K5 in the Kunigami area was higher than at site Z6 in the Zamami area, the ratio of Acropora was less than 30% in all substrata. Pocilloporidae dominated all substrata by 57% to 78% at site K5 in the Kunigami area. However, it only appeared at a rate of 71.0% in the CSD at site Z6 in the Zamami area; there were no dominant coral taxa in the other substrata ( Figure 3). The 2014 settlement rate was 64.8% in the CSD and 48.1% in the CP at site K7 in the Kunigami area, and it was 36.0% in the CSD and 24.6% in the CP at site M4 in the Motobu area. The CSD was clearly higher than the CP in both areas, and the average in both areas was 50.4% in the CSD and 36.4% in the CP (Table 4). In the proportion of coral settlers by taxa, Pocilloporidae appeared at a rate of 60%, greater than Acropora in both the CSD and CP in the Kunigami area. However, at site M4 in the Motobu area, Acropora dominated in both the CSD (58.1%) and CP (70.3%) (Figure 3).   The mean settlement rate after 3 years was estimated to be the highest on the CSD (37.3%), followed by the SS (33.0%), CN (30.2%), and CP (29.3%) by considering the data for all sampling years together ( Table 4). The coral taxa settlement ratio showed significant spatiotemporal variation, and the coral taxa scarcely revealed any tendency in settling on the substrata. A correlation between the complexity and settlement rate was so low that it showed no 5% significance level (p = 0.27) (Figure 4).  The mean settlement rate after 3 years was estimated to be the highest on the CSD (37.3%), followed by the SS (33.0%), CN (30.2%), and CP (29.3%) by considering the data for all sampling years together ( Table 4). The coral taxa settlement ratio showed significant spatiotemporal variation, and the coral taxa scarcely revealed any tendency in settling on the substrata. A correlation between the complexity and settlement rate was so low that it showed no 5% significance level (p = 0.27) (Figure 4).

Settlement Rate
The mean settlement rate after 3 years was estimated to be the highest on the CSD (37.3%), followed by the SS (33.0%), CN (30.2%), and CP (29.3%) by considering the data for all sampling years together ( Table 4). The coral taxa settlement ratio showed significant spatiotemporal variation, and the coral taxa scarcely revealed any tendency in settling on the substrata. A correlation between the complexity and settlement rate was so low that it showed no 5% significance level (p = 0.27) (Figure 4).

Survivorship
The number of coral settlers in the CSD in the first and second samplings in 2012 is shown in Figure 5. The survivorship was 50.0% at site K5 in the Kunigami area and 100% at site M4 in the Motobu area, although there was an increase in the number of settlers. No settlers were observed at site M6 in the Motobu area in either the first or second samplings. The mean survivorship for both areas is estimated at 75.0% (Table 5). For survivorship measured by taxa, the number of Acropora settlers did not change in either site. Pocilloporidae showed a decrease at site K5 in the Kunigami area, and it was replaced by other species at site M4 in the Motobu area.

Survivorship
The number of coral settlers in the CSD in the first and second samplings in 201 shown in Figure 5. The survivorship was 50.0% at site K5 in the Kunigami area and 10 at site M4 in the Motobu area, although there was an increase in the number of settl No settlers were observed at site M6 in the Motobu area in either the first or second s plings. The mean survivorship for both areas is estimated at 75.0% (Table 5). For surviv ship measured by taxa, the number of Acropora settlers did not change in either site. Po loporidae showed a decrease at site K5 in the Kunigami area, and it was replaced by ot species at site M4 in the Motobu area.  The number of coral settlers on each substratum in the first and second sampling 2013 is shown in Figure 5.  The number of coral settlers on each substratum in the first and second samplings in 2013 is shown in Figure 5. The survivorship was 32.7% in the CP, 28.3% in the CSD, 13.9% in the CN, and 10.9% in the SS at site K5 in the Kunigami area; in the Zamami area, the order was SS (39.8%), CSD (18.9%), CP (9.0%), and CN (0.0%). The mean survivorship in both areas was 25.4% in the SS, 23.6% in the CSD, 20.9% in the CP, and 7.0% in the CN (Table 5). Based on coral taxa, the survivorship of Pocilloporidae tended to be higher than that of Acropora at site K5 in the Kunigami area, and this taxon showed higher survivorship in the SS at site Z6 in the Zamami area. The number of coral settlers on each substratum in the first and second samplings in 2014 is shown in Figure 5. The survivorship was 56.6% in the CSD and 45.1% in the CP at site M4 in the Motobu area. No data were acquired at site K7 in the Kunigami area due to the loss of the substrata because of typhoons (Table 5). Based on coral taxa, Acropora did not change, as its highest value was 57.1%, and Pocilloporidae increased from 28.6% to 35.7% in the CSD. In the CP, Acropora also showed the highest rate, increasing from 71.4% to 75.0%, but Pocilloporidae decreased from 28.6% to 18.8%.
The mean survivorship of settlers 6 months after spawning in all areas from 2012 to 2014 was 51.7% in the CSD, the highest, followed by 33.0% in the CP, 25.4% in the SS, and 7.0% in the CN (Table 5). A correlation between the complexity and the survivorship was found (Figure 6), but there was no 5% significance level (p = 0.09). The coral taxa survivorship on the substrata showed few tendencies.
Oceans 2022, 3, FOR PEER REVIEW 9 ( Table 5). Based on coral taxa, Acropora did not change, as its highest value was 57.1%, and Pocilloporidae increased from 28.6% to 35.7% in the CSD. In the CP, Acropora also showed the highest rate, increasing from 71.4% to 75.0%, but Pocilloporidae decreased from 28.6% to 18.8%. The mean survivorship of settlers 6 months after spawning in all areas from 2012 to 2014 was 51.7% in the CSD, the highest, followed by 33.0% in the CP, 25.4% in the SS, and 7.0% in the CN (Table 5). A correlation between the complexity and the survivorship was found ( Figure 6), but there was no 5% significance level (p = 0.09). The coral taxa survivorship on the substrata showed few tendencies.

Effect of the Structural Complexity of the Substratum on the Settlement Rate
It is well known that coral larvae selectively settle on calcareous substrates in the field [25]. Based on this mechanism, it is believed that substances and biofilms derived from crustose coralline algae play a role in signaling larval settlement [26]. Since we installed substrata in the same areas at the same times, it was suggested that smothering the substratum material by crustose coralline algae in each area proceed in the same way. Therefore, we believe the structure of each substratum was the main factor in the differences between the coral larval settlements. Coral larvae prefer cryptic microhabitats and the undersides of complicated substrata [27]. Although biofilms are important as attractive factors for settlement, it has been pointed out that the structure of the substratum is more effective [28]. We found a trend that the settlement rate increased as complexity increased ( Figure 4); therefore, a complicated structure is significant for larvae settlement. The CSD did not show a high settlement rate despite the high complexity. This may be because the grooves on the underside (4 mm wide) were not adapted to the size of the larvae. They can detect micro-seafloor features smaller than their length (approximately 200 µm) and settle on them [29]. Despite the low complexity, the SS showed a higher settlement rate than those of the CN and CP, possibly due to the micro-wrinkle structure, which is not reflected in the complexity of the shell surface.
There have been several studies on the relationship between coral settlement and the structure of the substrate. One study found that two coral species settled selectively in the grooves of tiles with a depth and width of 2 mm [19]. Another study compared the number of coral settlers found between grooved tiles and flat tiles, showing that the number

Effect of the Structural Complexity of the Substratum on the Settlement Rate
It is well known that coral larvae selectively settle on calcareous substrates in the field [25]. Based on this mechanism, it is believed that substances and biofilms derived from crustose coralline algae play a role in signaling larval settlement [26]. Since we installed substrata in the same areas at the same times, it was suggested that smothering the substratum material by crustose coralline algae in each area proceed in the same way. Therefore, we believe the structure of each substratum was the main factor in the differences between the coral larval settlements. Coral larvae prefer cryptic microhabitats and the undersides of complicated substrata [27]. Although biofilms are important as attractive factors for settlement, it has been pointed out that the structure of the substratum is more effective [28]. We found a trend that the settlement rate increased as complexity increased ( Figure 4); therefore, a complicated structure is significant for larvae settlement. The CSD did not show a high settlement rate despite the high complexity. This may be because the grooves on the underside (4 mm wide) were not adapted to the size of the larvae. They can detect micro-seafloor features smaller than their length (approximately 200 µm) and settle on them [29]. Despite the low complexity, the SS showed a higher settlement rate than those of the CN and CP, possibly due to the micro-wrinkle structure, which is not reflected in the complexity of the shell surface.
There have been several studies on the relationship between coral settlement and the structure of the substrate. One study found that two coral species settled selectively in the grooves of tiles with a depth and width of 2 mm [19]. Another study compared the number of coral settlers found between grooved tiles and flat tiles, showing that the number of settlers in the grooved tiles was one digit higher than in the flat tiles [30]. In addition, as a result of placing a coral skeleton on the seabed and counting the number of settlers 4 months after spawning, another study found a correlation between the number of settlers and the irregularity of the skeleton surface [21]. These studies suggest that the complexity of the substrate plays a significant role in the settlement of coral larvae.

Effect of the Structural Complexity of the Substratum on the Survivorship
It is well known that the mortality of juvenile corals settled on the sea bottom is extremely high [7]. After 7-9 months, the survivorship of broadcast-spawning coral species settled on horizontally installed tiles was less than 2.8% on the east coast of Australia [31]. In Amakusa, western Japan, after 3-10 months, the survivorship of coral spats settled on tiles set horizontally was less than 12% [32]. The coral spats on these tiles are not only affected by sediments and algae but are also grazed by benthic invertebrates such as sea urchins [25,33]. On the reef bottom of the Red Sea, juvenile coral mortality after one year was found to be 27 to 33% [34]. The post-settlement mortality of juvenile coral in the first year is 30-99%, which is the most critical period in the coral life cycle [10]. Since predation by fish and invertebrates is a major factor in the post-settlement survivorship of Acropora, the complexity of the settlement base structure is important for reducing predation pressure. It has been reported that the complexity of the settlement base structure, which provides cryptic microhabitats and refuge, reduces the grazing pressure of predators and greatly contributes to the improvement of juvenile coral survivorship [30,33]. Therefore, at all sites for 3 years, we examined the relationship between the complexity of substrata and mean survivorship after 6 months of settling ( Table 5). As a result, we found that complicated seedling substratum structures contribute to increased post-settlement survivorship ( Figure 6). Although there was no difference in complexity between the SS and CN, a significant difference in survivorship was found. This is likely due to the micro-wrinkle structure of the shell surface, which was not reflected in the complexity.
Concerning the effect of complex settlement base structures on decreases in survivorship due to predation, the survivorship of coral settlers was up to 12% on tiles with micro-crevices on the surface, installed on the seabed for one year, compared with 0% for tiles without micro-crevices [35]. The survivorship of coral settlers after 2 years on tiles with projections on the surface was higher in the most closely spaced projection bases compared with other bases. This showed the effect of preventing benthic invertebrates such as sea urchins and snails from grazing [36]. After 29 days, the survivorship rate of the two species of coral juveniles attached to tiles with large and small grooves was higher than that of juveniles attached to exposed substrate, regardless of the size of the groove. It has been reported that this is likely due to differences in predation among coral-eating fish [37].
Sedimentation and algal thriving also affect juvenile coral survivorship, in addition to predation pressure. It is known that the survivorship of early coral settlers on artificial substrate is lowest on the upper surface and highest on the undersurface. This is because the undersurface is less affected by sedimentation and algal overgrowth [24,38]. Since the CSD was installed horizontally, many larvae settled in the groove on the undersurface so that the effect of sedimentation and algal growth can be avoided [39], which is also a factor in high survivorship.

Conclusions
We showed that the complexity of the seedling substratum is a critical factor in creating adequate coral seedlings with high larval settlement and post-settlement survivorship, although the small number of cases did not show any statistical significance. More studies can resolve this point. This study also showed that the CSD not only has a high settlement function but also a high survival function because of the structure's greater complexity. We continue further study on the settlement rate using substratum with different micro-features to clarify a more adequate material composition for larval settlement. The shape of the sampling units, the number of sampling units per frame, and how the number and distribution of units within each frame might alter the hydrodynamic conditions around the frame. Because we have qualitatively obtained knowledge that increasing the space between the CSD bundles, cases, and frames improves hydrodynamic conditions and the settlement rate, we continue to study to obtain quantitative data. Those results will help develop coral restoration techniques using sexual reproduction. Studies on the production of thermal-tolerant coral seedlings are also desirable to prevent degradation due to coral bleaching.