Transplanting Coral Fragments in Close Contact Enhances Their Survival and Growth on Seawalls

: Accelerated urbanisation has replaced many natural shorelines with coastal defences, resulting in the loss of natural habitats. However, structures such as seawalls can support some biotic assemblages, albeit of lower species richness. Ecological engineering techniques such as coral transplantation can enhance biodiversity on these artiﬁcial structures, but its success is circumscribed by high costs. Little is known about the fusion of discrete coral colonies that could potentially improve coral transplantation success on seawalls, particularly for the slow-growing massive species that are generally well-adapted to living on seawalls. Here, we investigated the feasibility and cost-effectiveness of transplanting Platygyra sinensis on seawalls by comparing the survivability and growth of fragments transplanted adjoining with those transplanted further apart. Fragments (approximately 3 cm diameter; n = 24) derived from three individuals were randomly grouped into two treatments, transplanted at 0.5 cm and 5 cm apart. Fragments in the former treatment came into contact with each other after three months. We observed that in all cases, the contact zones were characterised by a border of raised skeletal ridges without tissue necrosis, often termed nonfusion (=histoincompatible fusion). The adjoining transplants showed better survival (75 vs. 43%) and grew at a rate that was signiﬁcantly higher than fragments transplanted 5 cm apart (3.7 ± 1.6 vs. 0.6 ± 1.1 cm 2 month − 1 ). Our projections demonstrated the possibility of reducing transplantation cost (USD cm − 2 ) by 48.3% through nonfusion. These ﬁndings present nonfusion as a possible strategy to increase the overall cost-effectiveness of transplanting slow-growing massive species on seawalls.


Introduction
The world's natural coastlines are increasingly replaced with artificial structures such as groynes, breakwaters and more commonly, seawalls that primarily serve to defend the coast [1]. These structures mitigate the risks of flooding [2] while protecting infrastructures from coastal erosion and wave impact [3,4]. However, this has also resulted in the extensive loss of coastal habitats such as seagrass meadows, mangroves and coral reefs, along with the associated biodiversity and ecosystem functions. In spite of these impacts, studies have recently documented how artificial structures can provide habitats to molluscs [5], corals [6][7][8][9] and fish [10][11][12], although some biotic assemblages can be less diverse compared to their natural counterparts [9,12,13]. As it is crucial to preserve the myriad ecosystem services that human communities are reliant on, strategies such as ecological engineering, which aim to enhance biodiversity on artificial structures such as seawalls, are increasingly being studied see [14][15][16][17]. One of the developing approaches in ecological engineering is coral transplantation see [16][17][18], a technique that is commonly employed in reef restoration [19,20].
A primary goal of coral transplantation is to quickly increase the amount of live coral cover at degraded reefs [21]. Thus, although massive corals are more resilient to physical impacts [22][23][24] and thermal stress [25][26][27], most transplantation efforts disproportionately focus on branching coral species [22] such as acroporids and pocilloporids, which are fastgrowing and can be easily propagated to generate large areas of coral cover [28,29]. Among coral species that were transplanted on seawalls, branching and plating corals did not seem to survive well [17,19], while transplants of massive species on both subtidal [17,30] and intertidal seawalls [19] were observed to survive better than their branching counterparts. Greater fluctuations in environmental parameters such as temperature [31,32] and wave motion [33,34] in the vicinity of seawalls appear to be more stressful to branching corals. Recent studies have shown that more massive species were found naturally on seawalls than branching species [8,35,36]. Nevertheless, knowledge gaps exist in the enhancement of coral cover on seawalls. It is therefore important to develop and refine techniques that improve the growth and survivorship of massive coral transplants, so as to maximise the success of ecological engineering efforts [18].
Fusion, defined as segments of coral colonies in contact [37], could represent a scalable method of enhancing coral cover on seawalls since studies have shown that it can accelerate coral growth [22,37,38]. Fusion can be divided into two groups, histocompatible and histoincompatible reactions. The former refers to complete fusion: characterised as continuous skeleton and tissue across the contact area, while the latter is subdivided into nonfusion and rejection [39]. Nonfusion occurs where tissues of corals in contact are demarcated by a border of sutures (raised skeletal ridges) without tissue necrosis [40]; rejection occurs where tissue necrosis is present at the zone of contact, which is subsequently colonised by algae [40,41]. Fusion between isogeneic (i.e., same parent) juveniles resulted in increased colony size and a greater number of polyps than allogeneic (i.e., different parent) pairs, thus bolstering both the survivorship and growth of small coral colonies [38]. Additionally, isogeneically fused fragments of Acropora millepora (Ehrenberg, 1834) [42], Pocillopora damicornis (Linnaeus, 1758) [38] and Stylophora pistillata (Esper, 1797) [39] exhibited faster growth and lower mortality than individual fragments. Additionally, juveniles of Montipora capitata (Dana, 1846) that aggregated were more resilient against bleaching than those that were solitary [39], suggesting that physical contact between conspecific coral fragments could confer benefits beyond increased growth, survivorship and bleaching resilience. Research on fusion has largely focused on coral spats or juveniles of branching coral genera such as Pocillopora and Seriatopora see [38,40,43]. In comparison, less is known on the effects of fusion on the fragments of massive coral species, as well as whether such an approach has the potential to augment coral transplant growth, and consequently, ecological engineering outcomes.
Because transplantation can be costly [17,[44][45][46], it is also important to assess the variability in biological responses that arise from employing a new ecological engineering technique such as coral fusion, so that the cost-effectiveness of such an approach can be optimised [46]. Toh et al. showed that providing Artemia to juvenile corals resulted in improved post-transplantation growth and survivorship, as well as greater cost-effectiveness [44]. There are, however, few studies projecting or comparing cost breakdowns between alternative strategies in ecological engineering [17,[44][45][46]. For example, it was demonstrated that labour cost was substantially reduced through volunteer-driven coral nursery maintenance [17] and that transplanting only small coral fragments (2-4 cm) could result in greater return-on-effort given limited coral source material [30]. These thought experiments can aid in identifying strategies that represent the best use of limited resources [45] and are increasingly crucial given the growing acceptance and implementation of ecological engineering techniques in coastal development see [15][16][17][18]45,[47][48][49].
This study seeks to evaluate the feasibility of histoincompatible coral fusion as a technique to improve the success of transplanting fragments of Platygyra sinensis (Milne Edwards and Haime, 1849), a massive species frequently found on seawalls [8] and commonly used in reef restoration and ecological engineering initiatives [17,30]. Here, we compared the survivorship and growth of fragments of P. sinensis transplanted in close contact ("adjoining") with those transplanted apart ("separate"). We hypothesised that fusion between adjoining transplants would lead to significantly greater growth and survivorship than transplants that were placed separately. What-if scenarios were also conceived in this study to assess how transplantation protocols and decision-making could be augmented through cost estimates in ecological engineering [50]. Subsequently, we estimated the costeffectiveness of coral fusion to determine its economic viability for coral transplantation efforts. The findings of this study will augment ecological engineering initiatives, especially for slow-growing massive coral species.

Materials and Methods
Twenty four fragments of P. sinensis that were reared for at least six months in flow-through aquarium tanks at the St. John's Island National Marine Laboratory were transplanted subtidally on a sloping seawall at Lazarus Island, Singapore (1 • 13.37 N, 103 • 51.08 E; Figure 1). The fragments originated from three parent colonies that were at least 15 m apart to ensure genetic diversity. All fragments were approximately 3 cm in maximum diameter. In order to investigate the effect of fusion on growth and survivorship of P. sinensis transplants, 12 fragments were each assigned to two treatments: adjoining (transplanted approximately 0.5 cm apart from two other fragments in a cluster) and separate (See Figure 2a,c) (transplanted 5 cm apart from two other fragments in a cluster). Each treatment comprised four clusters of three fragments. Fragments in each cluster were all from different parent colonies. The initial area of fragments did not differ significantly between treatments (F = 0.48, p > 0.05). Transplants were secured onto granite boulders on the seawall at approximately −2 m (Chart Datum), using marine epoxy.

Statistical Analysis
After fragments in the adjoining treatment came into contact over the course of the study, the live tissue area of fragments from both treatments was subsequently measured as a single mass for each cluster (n = 4). Survivorship of the transplants in both treatments was examined with Kaplan-Meier survival analysis and compared with log-rank (Mantel-Cox) tests. To compare differences in growth between adjoining and separate treatments, one-way ANOVA and post-hoc tests using the Tukey's test and Bonferroni correction were performed, after fulfilling the assumptions of normality and homogeneity of variance. All analyses were conducted using SPSS Statistics (Version 21, IBM). Survivorship and live tissue area of the transplants were monitored monthly for 16 months by photogrammetry, a simple and accurate method for tracking two-dimensional changes in coral growth, as previously described by Kikuzawa et al. [51]. Each transplant was individually photographed from a top-down angle with a scale bar, using an Olympus TG-5 camera, and subsequently processed with the ImageJ software (NIH). The increase in live tissue area was monitored as a proxy of coral growth. Transplants that were detached or had >95% dead tissue were considered "dead".

Statistical Analysis
After fragments in the adjoining treatment came into contact over the course of the study, the live tissue area of fragments from both treatments was subsequently measured as a single mass for each cluster (n = 4). Survivorship of the transplants in both treatments was examined with Kaplan-Meier survival analysis and compared with log-rank (Mantel-Cox) tests. To compare differences in growth between adjoining and separate treatments, one-way ANOVA and post-hoc tests using the Tukey's test and Bonferroni correction were performed, after fulfilling the assumptions of normality and homogeneity of variance. All analyses were conducted using SPSS Statistics (Version 21, IBM).

"What-If?" Scenarios for Improved Cost Effectiveness
"What-if?" scenarios were formulated to predict the estimated cost of live coral area generated when fragments were transplanted based on two hypothetical scenarios: adjoining and separate. Each scenario entailed the collection of 1000 cm 2 of coral material, fragmentation into 100 fragments of 9 cm 2 each with the assumption of up to 10% of coral material lost as wastage see [30], transplantation of fragments in adjoining and separate designs respectively, as well as monitoring and maintenance of the transplants for a year. The projected coral tissue area after one year was computed: Projected live tissue area = (G + I) × N × S where G = tissue growth in a year (cm 2 ), I = initial tissue area (cm 2 ), N = number of fragments or clusters, and S = mean survivorship. The calculation for the cost per unit area of coral was adapted from Toh et al. [16], as the transplantation technique used was similar. The projected total live tissue area after one year was based on the results from the current study (see Supplementary Table S1). Costs were estimated in Singapore dollars (SGD) prior to conversion to USD at the rate of SGD 1.33 = USD 1.

Results
In the adjoining treatment, fragments all came into contact during the third month and continued to increase in size until the end of the study (Figure 2). No signs of rejection (i.e., tissue mortality) were observed after contact ( Figure 3). Upon close examination, fragments in the adjoining treatment were regarded as exhibiting a nonfusion contact response since this response was previously reported [40,41]. On the eighth month, sutures had begun to form at the contact areas. In the 15th month, some of these sutures became raised and progressed to grow over another fragment ( Figure 3).

"What-If?" Scenarios for Improved Cost Effectiveness
"What-if?" scenarios were formulated to predict the estimated cost of live coral area generated when fragments were transplanted based on two hypothetical scenarios: adjoining and separate. Each scenario entailed the collection of 1000 cm 2 of coral material, fragmentation into 100 fragments of 9 cm 2 each with the assumption of up to 10% of coral material lost as wastage see [30], transplantation of fragments in adjoining and separate designs respectively, as well as monitoring and maintenance of the transplants for a year. The projected coral tissue area after one year was computed: Projected live tissue area = (G + I) × N × S where G = tissue growth in a year (cm 2 ), I = initial tissue area (cm 2 ), N = number of fragments or clusters, and S = mean survivorship. The calculation for the cost per unit area of coral was adapted from Toh et al. [16], as the transplantation technique used was similar. The projected total live tissue area after one year was based on the results from the current study (see Supplementary Table S1). Costs were estimated in Singapore dollars (SGD) prior to conversion to USD at the rate of SGD 1.33 = USD 1.

Results
In the adjoining treatment, fragments all came into contact during the third month and continued to increase in size until the end of the study (Figure 2). No signs of rejection (i.e., tissue mortality) were observed after contact (Figure 3). Upon close examination, fragments in the adjoining treatment were regarded as exhibiting a nonfusion contact response since this response was previously reported [40,41]. On the eighth month, sutures had begun to form at the contact areas. In the 15th month, some of these sutures became raised and progressed to grow over another fragment (Figure 3).
After 16 months, fragments in the adjoining and separate treatments registered 75.0% (n = 4) and 42.7% survivorship respectively (n = 12; Figure 4). While the survivorship of both treatments decreased sharply during the eighth month, survivorship in the separate treatment subsequently decreased again during the 11th month. After a year, the survivorship of both adjoining and separate fragments stabilized until the end of the study period ( Figure 4).  After 16 months, fragments in the adjoining and separate treatments registered 75.0% (n = 4) and 42.7% survivorship respectively (n = 12; Figure 4). While the survivorship of both treatments decreased sharply during the eighth month, survivorship in the separate treatment subsequently decreased again during the 11th month. After a year, the survivorship of both adjoining and separate fragments stabilized until the end of the study period ( Figure 4). Although the mean live tissue area of clusters in both treatments increased, the average monthly increase in the live tissue area of adjoining transplants was more than five times that of the separated transplants ( Table 1). The average monthly increase in the live tissue area of adjoining transplants (3.74 ± 1.58 cm 2 mth −1 ) was significantly greater (F = 7.66, p = 0.05) than that of separated transplants (0.63 ± 1.13 cm 2 mth −1 ). The total live tissue area of adjoining and separated transplants after 16 months was 291.4 cm 2 and 160.1 cm 2 respectively (Table 1). Between Month 5 to Month 8, monitoring was suspended due to unexpectedly high levels of turbidity and sedimentation at the study site, resulting in poor visibility and large deposits of silt on the transplants. Interestingly, we observed that less sediment had accumulated on the adjoining transplants than on the separated transplants ( Figure 5). The survivorship of separated transplants decreased sharply after Month 8 and continued to decline, while that of adjoining transplants remained constant (Figure 4). Tissue damage from Drupella spp. (Gastropoda) or Scaridae was not observed in this study. Although the mean live tissue area of clusters in both treatments increased, the average monthly increase in the live tissue area of adjoining transplants was more than five times that of the separated transplants ( Table 1). The average monthly increase in the live tissue area of adjoining transplants (3.74 ± 1.58 cm 2 mth −1 ) was significantly greater (F = 7.66, p = 0.05) than that of separated transplants (0.63 ± 1.13 cm 2 mth −1 ). The total live tissue area of adjoining and separated transplants after 16 months was 291.4 cm 2 and 160.1 cm 2 respectively (Table 1). Between Month 5 to Month 8, monitoring was suspended due to unexpectedly high levels of turbidity and sedimentation at the study site, resulting in poor visibility and large deposits of silt on the transplants. Interestingly, we observed that less sediment had accumulated on the adjoining transplants than on the separated transplants ( Figure 5). The survivorship of separated transplants decreased sharply after Month 8 and continued to decline, while that of adjoining transplants remained constant (Figure 4). Tissue damage from Drupella spp. (Gastropoda) or Scaridae was not observed in this study.  The "what-if" scenarios showed that the cost per cm 2 of coral transplanted could be minimised to US$ 6.49 if all transplants were adjoining to promote nonfusion among the P. sinensis fragments ( Table 2). In scenario 1 (adjoining), the cost per cm 2 of coral projected was reduced by 48.3% as compared to scenario 2 (separate). The "what-if" scenarios showed that the cost per cm 2 of coral transplanted could be minimised to US$ 6.49 if all transplants were adjoining to promote nonfusion among the P. sinensis fragments ( Table 2). In scenario 1 (adjoining), the cost per cm 2 of coral projected was reduced by 48.3% as compared to scenario 2 (separate). Table 2. Cost estimates (in USD) of transplanting coral fragments adjoining to promote nonfusion (Scenario 1, "adjoining") versus transplanting fragments apart (Scenario 2, "separate"). Cost per cm 2 of coral transplanted (US$ cm −2 ) 6.49

12.55
Cost estimates for collecting 1000 cm 2 of Platygyra sinensis followed by transplanting 100 fragments (9 cm 2 ) each, with subsequent monthly monitoring and maintenance for a year. The cost per unit area was calculated using total cost of production following Toh et al. (2017) and the projected total coral tissue area (from this study) after one year. Costs were estimated in Singapore dollars (SGD) prior to conversion to USD at the rate of SGD1.33 = USD1.00. Detailed breakdown of cost estimates is provided in Supplementary Table S1.

Discussion
Ecological engineering is necessary and effective in augmenting biodiversity on coastal artificial structures [16,18,52,53] but is relatively novel and costly [17,[47][48][49]. It includes techniques such as retrofitting [54] and coral transplantation [55]. A growing body of work shows that coral transplantation is a viable ecological engineering approach [17,47,49], but there are several unknowns to consider when increasing coral cover on artificial structures, such as the appropriate size of transplants and suitability of species. Therefore, it is important to adopt best practices from other fields including reef restoration. Here we showed that transplanting allogeneic fragments helps to hasten live coral growth and that the technique is cost-effective. We also determined that direct contact between allogeneic P. sinensis transplants resulted in nonfusion (i.e., formation of raised skeletal sutures without tissue necrosis). We did not observe complete fusion, which is characterised by continuous tissue and skeleton across the zone of contact [40,41].
The improved growth and survivorship of the adjoining transplants may be explained by the rapid increase in coral tissue area within the cluster, conferring benefits such as increased resources to deal with stress. Our findings corroborated with studies that showed that larger fragments grew faster than smaller fragments [56][57][58]. As smaller fragments possess fewer resources, they are more likely to be overwhelmed by physical and environmental stresses [59] such as fragmentation and sedimentation, leading to reduced growth and survivorship. Conversely, in larger colonies, the proportion of healthy coral tissue bordering the injury generated from disturbance is greater than in smaller colonies, thus contributing to greater regeneration and consequently higher survivorship [60]. Furthermore, studies have suggested that resources could be shared between adjoining transplants upon contact, where additional energy reserves from neighbouring healthy coral polyps can help improve both tissue regeneration [60,61]. In addition, the high sediment load of Singapore's marine environment ranging between 5 to 20 mg cm −2 day −1 [62], could have led to considerable stress on both adjoining and separated transplants. However, massive corals such as P. sinensis are known to be efficient in removing sedimentation through mucus [63,64], ensuring access to light for optimal photosynthetic performance [63,65,66]. This effect is likely more pronounced amongst adjoining transplants than separated transplants due to an increase in colony size as the tissue comes into contact. Larger corals with greater energy reserves can produce more mucus and expand their tissues to remove sediment and reduce smothering [66,67], hence adjoining transplants were able to both survive and grow better than the separated transplants. Our observations suggest that adjoining transplants seemed to deal better with sedimentation stress than the separated transplants. While there was no evidence of coenosarc tissue fusing between adjoining transplants, detailed studies investigating the physiological reactions between adjoining colonies could provide insights into how the growth and survivorship of these colonies are affected.
At 0.63 ± 1.13 cm 2 mth −1 , the mean monthly areal growth of P. sinensis transplanted separately was lower than that of another ecological engineering project in Singapore, where P. sinensis fragments were also transplanted apart on seawalls (3.92 ± 1.15 cm 2 mth −1 ) [17]. The latter growth rate was more comparable to that of the adjoining transplants in our study (3.74 ± 1.58 cm 2 mth −1 ). Although the mean initial sizes of fragments used in Toh et al. [17] were greater by approximately five times, recent work by Sam et al. [30] showed that there was no significant influence of initial fragment size on the post-transplantation growth of P. sinensis. The differences in growth rates could be influenced by varying local environmental conditions such as water motion and sedimentation as previous studies have reported the negative influence of high water motion [68,69] and high sedimentation [70,71] suppressing coral growth. Nevertheless, we demonstrated that adjoining transplants grew significantly faster than separated transplants.
High mean survivorship of adjoining P. sinensis transplants (75%) demonstrated the suitability of this species for transplantation on seawalls. This is similar to that observed in other studies also involving P. sinensis being transplanted onto seawalls and monitored for between 6 and 18 months (97% and 65% survivorship respectively) see [17,30]. Taken together with our results, they suggest that the survivorship of transplants generally stabilises after one year, underscoring the importance of long-term monitoring (>1 year) for coral transplantation. Although environmental conditions on the seawall in this study were not measured, they were likely suboptimal due to acute incidences of high sedimentation (pers. obs.). However, the higher mean survivorship of adjoining transplants compared to those transplanted separately suggests that the potential benefits exist when transplanting corals in close contact, especially in a sedimented environment.
The coral transplantation methods applied in this study are adopted from reef restoration. As reefs remain amongst one of the costliest ecosystems to restore [45], current ecological engineering techniques are likely expensive as well, however, our proposed method can reduce overall project cost. Our "what-if" scenarios showed that the overall cost per cm 2 of coral transplanted could be reduced by 48.3% if transplants were adjoining to promote nonfusion, instead of being placed further apart. Rather than lowering costs by reducing manpower or materials used during the transplantation process, our approach reduces ongoing maintenance costs by ensuring higher transplant survivorship. Thus, we recommend that fragments of P. sinensis should be transplanted in close contact to promote nonfusion, so as to augment growth and survivorship, and consequently improve cost-effectiveness. To enhance ecological engineering outcomes, this strategy should also be tested on other coral species so that its overall effectiveness may be assessed. Findings from this study can potentially help advance coastal management, as transplanting corals on coastal artificial structures has been limited, and poses different challenges than transplantation on reefs. By transplanting coral fragments in close contact to promote nonfusion, colonies could collectively overcome the stressful conditions on seawalls, resulting in improved survivorship and growth, thus augmenting biodiversity while curtailing the project costs.
In conclusion, this study has shown that transplanting P. sinensis fragments adjoining on seawalls is potentially a practical and cost-effective ecological engineering approach.
Additional studies are needed to verify if this technique is also applicable for other coral species, especially those with massive and foliose growth forms. Future studies should also be carried out to discern the physiological factors that have led to the increased performance of adjoining transplants.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/jmse9121377/s1, Table S1: Detailed cost estimates of transplanting corals adjoining to promote nonfusion ("what-if" scenario 1) or separate ("what-if" scenario 2) [72].  Institutional Review Board Statement: All applicable national and institutional guidelines for animal testing, animal care and use of animals were followed by the authors. All necessary permits for sampling and observational field studies were obtained by the authors from the National Parks Board (NParks; NP/RP17-037-2).

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets generated during the current study are available from the corresponding author on reasonable request.