Coral Reefs

, Volume 22, Issue 3, pp 217–223

Coral reef rehabilitation through transplantation of staghorn corals: effects of artificial stabilization and mechanical damages

Report

DOI: 10.1007/s00338-003-0305-6

Cite this article as:
Lindahl, U. Coral Reefs (2003) 22: 217. doi:10.1007/s00338-003-0305-6

Abstract

In order to develop and test a low-cost method of coral reef rehabilitation, the staghorn corals Acropora muricata and A. vaughani were transplanted to a shallow site with unstable substrate. To avoid abrasion, dislodgement and transport due to water movement, the transplanted corals were tied to string sections, which were connected at the seabed to form a grid. This created stability and improved the survival of the corals. The average increase in weight of live coral over 1 year was 56%, eight times more than the control treatment with unattached coral branches. This difference was mainly due to a reduced partial mortality among smaller coral fragments in the stabilized treatment. Survival was positively related to initial size among the loosely placed coral branches, whereas the attached treatment showed a negative relation between size and relative increase in weight of the surviving parts of the coral branches. Coral fragments were not significantly affected by severe physical damage simulating the effects of handling.

Keywords

Coral reef rehabilitation Coral transplantation Coral growth Acropora muricata Acropora vaughani 

Introduction

Numerous experiments on coral transplantation have been carried out, aimed at assessing the feasibility of various methods of reef rehabilitation. Provided that the attachment is sufficient and that environmental factors, such as substrate and water quality, are favorable, a wide variety of coral species has been shown to survive transplantation well (Maragos 1974; Birkeland et al. 1979; Harriott and Fisk 1988a; Hudson and Diaz 1988; Guzman 1991; Kaly 1995; Berker and Mueller 1999; Tomlinson and Pratt 1999; Hudson 2000). Poor survival or loss of transplanted corals can often be explained by factors such as wave-induced dislodgement (Birkeland et al. 1979; Auberson 1982; Plucer-Rosario and Randall 1987; Harriot and Fisk 1988a; Newman and Chuan 1994; Clark and Edwards 1995; Bowden-Kerby 1997), human disturbances (Auberson 1982), burial and smothering by loose substrate (Auberson 1982; Harriott and Fisk 1988a; Bowden-Kerby 1997; Nagelkerken et al. 2000), or elevated temperatures (Yap and Gomez 1984; Yap et al. 1992; Lindahl et al. 2001). Coral reef rehabilitation is controversial, mainly because of the high costs involved, the damages to source populations, and the questionable long-term survival of the transplanted corals (Harriott and Fisk 1988b; Edwards and Clark 1999; Spurgeon and Lindahl 2000). One of the most difficult problems is to prevent dislodgement of the transplanted corals due to water movement. Different attachment methods can be used, depending on the substrate, wave exposure, and growth form of the coral.

Corals killed by natural or anthropogenic disturbances are often degraded to rubble (Carpenter and Alcala 1977; Alcala and Gomez 1987; Sano et al. 1987; Blanchon et al. 1997). This substrate is often inhospitable for natural re-colonization by corals (Alcala and Gomez 1979; Brown and Dunne 1988; Riegl and Luke 1998; Fox et al. 2000) and is therefore of primary interest for studies of reef rehabilitation. Corals transplanted to areas with unconsolidated sediment and exposure to water movements can be secured to artificial substrates (Maragos 1974; Schumacher and Schillak 1994; Clark and Edwards 1995). However, these structures are probably too expensive for a wider application in developing countries (Edwards and Clark 1999; Spurgeon and Lindahl 2000).

Methods for transplanting unattached fragments of staghorn corals (Acropora spp) to shallow backreef areas have been developed in order to keep the costs low (Bowden-Kerby 1997; Lindahl 1998). These methods are quick and do not require SCUBA divers or expensive materials and equipment. In sufficiently protected habitats, survival of unattached fragments can be nearly 100% (Bowden-Kerby 1997). However, water movements can cause severe mortality. When growing naturally, staghorn corals are mainly restricted to protected or moderately exposed habitats. Dense monoclonal thickets of staghorn corals, formed through vegetative reproduction and fragmentation, can often dominate the coral community in shallow back-reef areas. Individual colonies often gain stability and support by direct contact with surrounding colonies rather than by attachment to the substrate, and fusions between adjacent branches are common (Gilmore and Hall 1976; Highsmith 1982). This way of growth enables a thicket of staghorn corals to encroach over loose substrate, where the recruitment and growth of most other corals is prevented (Tunnicliffe 1981).

The aim of this study was to test a new low-cost method of artificial stabilization of staghorn coral fragments transplanted to loose substrate. In addition, the effects of mechanical damages on transplanted coral fragments were assessed.

Materials and methods

The study site

The study was carried out at Tutia Reef near Mafia Island, Tanzania (8°08′S, 39°40′E). The experimental site has a depth of 3 m and is situated in a sound between two reefs, approximately 100 m north of the crest of Tutia Reef. Oceanic waves entering the sound from the east are somewhat reduced before reaching the study site. Therefore, the site is never exposed to surf or breaking waves. The seabed is covered with a mix of coral rubble, sand, and rhodoliths. Tutia Reef is relatively unaffected by human disturbance, although some blast fishing and seine netting have occurred. The rubble fields are probably of natural origin and in the process of being colonized by adjacent thickets of branching corals (Acropora spp). Staghorn corals and other branching corals in the genus Acropora, of which approximately 80% were killed during the 1998 ENSO event, dominate the surrounding coral community.

Artificial stabilization of transplanted corals

Branches of the staghorn coral Acropora muricata were collected at 24 m depth near the study site at Tutia Reef, whereas A. vaughani were collected at about 5-6 m depth approximately 2 km away on the same reef. The coral branches were taken from several distinct thickets assumed to be clones. The average length of the collected branches was 34±9 cm (SD, n=364). The branches were transplanted in November 1998 using two different methods. They were either placed loosely on the seabed or attached using the "string-grid" method (Fig. 1). This means that 10 coral branches, all belonging to the same clone, were attached to a 1-m section of polythene string. The branches were attached at 10-cm intervals by two half-hitches on the polythene line. Several string sections were placed on the seabed and joined to form one large grid. The strings were not attached to the seabed. Seven clones of A. muricata and six clones of A. vaughani were used. From each clone, two groups of 10 coral branches were randomly assigned to the tied treatment and eight randomly selected branches were transplanted without attachment. Thus, 10+10+8 branches were transplanted from each clone, giving a total of 260 attached branches and 104 unattached branches. The branches were tagged with numbered cable ties and their weight and length were recorded before the transplantation and at the termination of the experiment, 1 year later. The branches were weighed with a spring scale out of the water, protected from direct sunlight. At the end of the experiment, all branches were inspected for partial mortality. Parts of the branches not covered by live tissue were regarded as dead, and the proportion of the dead parts was estimated visually to the nearest 10%. Branches that were not recovered were presumed dead, because live branches were relatively easy to spot, whereas dead branches were difficult to discern in the rubble substrate. Thus, the last careful search of the substrate that was carried out in order to recover the dead branches would certainly have revealed any remaining living branches. A large area surrounding the transplantation site was searched for tagged branches, but none was found more than 3 m away from the original site. Unrecovered branches were assumed to have kept their initial weight, since this was generally the case for branches that were recovered dead. Five live and untagged fragments were found, which could not be connected to any specific tagged branch. These fragments represented less than 0.2% of the total weight of live coral, and were not included in the analysis. For each branch three different variables were recorded or computed for statistical analysis:
Fig. 1.

Schematic illustration of the "string-grid" method of stabilization of transplanted corals. String sections with attached branches of staghorn corals have been placed on the seabed and tied together to form a square grid. The sides in each square in the grid are 1 m and branches are attached every 10 cm

  1. 1.

    "Partial mortality": the estimated proportion (by weight) of the parts of each branch that were regarded as dead, as defined above. Totally dead branches were regarded as having 100% partial mortality.

     
  2. 2.

    "Relative weight of whole coral": final total weight of each branch divided by its initial weight. Total weight includes all parts (dead and live) of the branch.

     
  3. 3.

    "Relative weight of surviving parts": the weight of the proportion that was regarded as alive on each branch divided by the initial weight of the branch.

     
The word "relative" indicates the relation between final weight and initial weight. The following example illustrates how these variables were computed.
  • Initial weight =100 g, final total weight = 150 g

  • "Partial mortality" = 20% (visually estimated)

  • "Relative weight of whole coral" = 150/100=1.5

  • "Relative weight of surviving parts" = 0.8×150/100=1.2

Mechanical damage

Fragments of A. muricata with an average weight of 83±27 g (SD, n=216) were collected from 12 different thickets, each assumed to be a distinct clone. The 18 fragments collected from each clone were randomly distributed to two treatments: "damage" and control (each treatment with 108 branches). Thus, for each of the 12 clones, nine branches were allocated to each of the two treatments. The damage treatment simulated the kind of injuries that occur during handling and transport of collected fragments, and involved forceful scraping with a knife twice along all main branches and trimming of all branch tips by 2 cm. The scraping removed the protruding corallites and soft tissue, creating a 3- to 5-mm-wide and 1-mm-deep scar. All branches were attached in vertical positions to a PVC-rack at 3 m depth, using cable ties. The weight of each branch was recorded immediately after the treatment and again after 8 months. Also partial mortality was estimated as in the experiment on artificial stabilization.

Statistical analyses

The results of the two experiments were analyzed with ANOVA, using the program Statistica (StatSoft Inc.). A separate ANOVA was done on each of the variables "partial mortality," "relative weight of whole coral," and "relative weight of surviving parts." The data were subjected to Cochran's test of homogeneity of variances prior to analysis. In order to get a balanced number of replicates, one randomly selected clone of A. muricata was excluded in the experiment on artificial stabilization. The average from all 10 branches from each string section was considered as one replicate, giving n=2. The unattached branches were randomly pooled into two groups per clone, and the average for each group was regarded as one replicate. Species was regarded as a fixed factor and the random factor clone was nested under species. The factors species and treatment were orthogonal. The partial mortality and relative change in weight of surviving parts on each of the transplanted branches were regressed on the initial weights of the branches. Regressions were carried out on each combination of species and treatment (attached or unattached) as well as on the two species pooled in each treatment. In order to facilitate comparison between the branch weights used in this paper and the more commonly presented type of results based on branch length, weight was regressed on length, yielding a polynomial curve for the relation between the two measurements.

Results

Artificial stabilization

The attached corals suffered less mortality and increased their relative weight of surviving parts at a significantly higher rate than those that were unattached (Fig. 2 and Table 1). There were no significant differences among species or clones for any of the three variables. The total weight change (i.e. including dead parts of the corals) did not differ significantly between species, clones, or treatments. No interactions between factors were found for any of the variables. Most corals showed signs of partial mortality 1 year after transplantation. The proportion of branches without dead tissue was 13% in the attached treatments and 7% among the unattached branches, whereas 3% of the attached branches and 13% of the loose branches were dead or missing. Among the attached corals, most partial mortality was found on parts of the branches that were embedded in, or in contact with, the sediment, whereas the loose branches had suffered much mortality also on parts that were above the sediment at the time of sampling. Adjacent branches attached to the same string were often fused to each other.
Fig. 2.

Effects of the two transplantation methods "attachment" and "loose" on partial mortality, relative weight of whole corals (= total weight), and relative weight of live corals (not including weight of dead parts) for A. muricata and A. vaughani pooled. Weight relations are calculated as final weight divided by initial weight. Error bars indicate SE, based on n=104 (loose branches) and 260 (attached branches)

Table 1.

ANOVA table of the effects of treatment, species, and clones on partial mortality and relative weight of whole corals and surviving parts in the experiment on artificial stabilization. Asterisks indicate p levels below 0.05

Effect

df effect

MS effect

df error

MS error

F

p level

Partial mortality

Treatment

1

5,007.4

10

184

27.179

0.0004*

Species

1

19.1

10

381

0.050

0.828

Clone (in species)

10

381.9

24

476

0.802

0.628

Treatment—species

1

4.2

10

184

0.023

0.883

Treatment—clone

10

184.2

24

476

0.387

0.940

Relative weight of whole coral

Treatment

1

0.389

10

0.180

2.168

0.172

Species

1

0.042

10

0.424

0.098

0.761

Clone (in species)

10

0.424

24

0.210

2.022

0.077

Treatment—species

1

0.100

10

0.180

0.561

0.471

Treatment—clone

10

0.180

24

0.210

0.856

0.583

Relative weight of surviving parts

Treatment

1

1.798

10

0.155

11.629

0.007*

Species

1

0.006

10

0.552

0.010

0.922

Clone (in species)

10

0.553

24

0.336

1.646

0.153

Treatment—species

1

0.019

10

0.155

0.121

0.735

Treatment—clone

10

0.155

24

0.336

0.460

0.899

Effects of initial weight

All regressions showed very low r2, indicating the influence of factors other than initial weight. The initial weight of the transplanted corals affected their development differently depending on treatment (Fig. 3). Small branches increased their relative weight of surviving parts more and had less partial mortality than larger corals provided that they were attached. Among the unattached corals, the larger branches had a reduced partial mortality and a weak trend towards greater increase in relative weight of surviving parts. The regressions carried out on the two species separately showed the same trends as the combined regressions. To examine the effect of the extremely large branches on the regressions, the two largest branches in the tied treatment and the three largest branches in the unattached treatment were removed from the regressions. This gave an insignificant p value (p=0.27) for partial mortality versus initial weight in the attached treatment, whereas the other significances remained. The weight of the transplanted branches (at the time of transplantation) was almost linearly related to the first order of their length (Fig. 4).
Fig. 3.

Linear regressions of partial mortality of individual branches on initial weight (a, b) and relative weight of surviving parts on initial weight (c, d). "Relative weight" indicates the relation between final weight and initial weight. Branches of A. muricata and A. vaughani are pooled in each regression. a n=104, r2=0.008, Y=−0.0335X+55; b n=260, r2=0.02, Y=0.0134X+15; c n=104, r2=0.01, Y=0.0282X+95; d n=260, r2= 0.1, Y=−0.0933X+190

Fig. 4.

Initial branch weight regressed on branch length on all transplanted branches (n=364). The equation for the best fit is Y=0.0006X3+0.035X2+8.1998X. r2=0.25

Mechanical damage

Due to dislodgement and/or breakage 11 of the 216 branches were lost, resulting in sample sizes between 7 and 9 for each combination of clone and treatment. Among the remaining branches the inflicted damages did not significantly affect the relative weight of surviving parts of the branches after 8 months of growth (Fig. 5). However, the different clones showed a very large variation in the rate of weight increase (p=0.001), with clonal averages of relative weight ranging between 1.52 and 3.71 after 8 months.
Fig. 5.

Final weight of live coral branches relative to initial weight in the experiment on mechanical damage. Each pair of bars represents the average for the damaged and control treatments for one clone of A. muricata. Error bars indicate SE. Due to losses of branches, sample sizes vary between 7 and 9

Discussion

Artificial stabilization

The tested method of artificial stabilization is intended to protect the transplanted fragments from two types of mortality: breakage and loss of live tissue. In a long-term field study such as this, growth and different kinds of mortality are difficult to measure separately, because these factors have a combined effect on the relative weight of surviving parts at the end of the experiment. However, this measure gives a good indication of the effectiveness of the stabilization method. The partial mortality found on the upper surfaces of the unattached coral branches was probably caused by contact with the sediment, indicating that they had been overturned by water movements during the course of the experiment. This significant mortality of coral branches close to an area where staghorn corals were proliferating naturally in dense thickets demonstrates the vulnerability of small, unattached corals. The study showed that staghorn coral branches could be successfully transplanted to a moderately wave-exposed habitat by using the "string-grid" method. The attachment to strings did not fix the corals entirely, but helped to maintain their orientation. Contacts between adjacent branches also prevented movement and frequently resulted in fusion, which added extra stability. The attachment resulted in a net increase in weight of surviving parts eight times greater than for the unattached corals. This is probably an effect of reduced partial mortality among the smallest, fast-growing fragments in the attached treatment.

Apart from the effect of burial and abrasion, the corals attached to strings were probably severely affected by competition between adjacent branches. Lindahl (1998) found that sparsely transplanted A. muricata had a greater growth rate than those that were placed in close proximity, indicating an adverse effect of intra-specific competition for light and space. Hence, there is a trade-off between the stability provided in dense populations and the increased availability of light and water circulation among more sparsely growing corals. The average increase in weight of surviving parts among the branches placed in racks (damage experiment) was 157% over 8 months. This would correspond to an annual increase considerably greater than the 82% that was found for branches of similar size among the corals tied to strings. Hence, it is clear that the placement on the seabed disturbed the corals even when they were attached to strings. A defensive response can be triggered through the contact of two colonies of different genetic origin (Hildemann et al. 1975). In order to avoid aggressive reactions, which may result in reduced rates of growth and reproduction (Rinkevich and Loya 1985), coral branches attached to the same string section should optimally originate from the same clone.

The combined weight of the corals attached to the string-grid was sufficient to anchor the structure and prevent deformation of the grid, without attachment to the seabed. Further experiments are necessary to assess the usefulness of this method on sites with stronger water movements, different substrate types or more steeply sloping seabed. It is important to note that this kind of attachment can only provide a temporary stabilization, since the strings will degrade with time. Even intact strings would be of less use after some time, as the transplanted corals grow and start to spread over the seabed. Since the natural proliferation of artificially created thickets is essential for the usefulness of the method, it should only be used in areas where thickets of staghorn corals are known to be able to grow naturally.

Size-effects

From the present results, it appears that transplanted corals in a wide range of fragment sizes will survive and grow well. Two effects of the initial size of transplanted coral branches remained when outliers had been removed from the regressions. First, larger initial size reduced the mortality among branches placed loosely on the seabed. Second, branches secured by strings or placed in racks showed a negative relation between relative weight of surviving parts and initial size. The low r2 values in the regressions indicate that several factors other than initial size influenced the survival and weight increase of the corals. These factors may include differences in growth form and genetic composition as well as external factors such as predation, overgrowth, diseases, and sediment interaction. Several studies have reported a positive correlation between fragment size and survivorship in corals (Highsmith et al. 1980; Hughes and Jackson 1985; Liddle and Kay 1987; Harriott and Fisk 1988a; Knowlton et al. 1988; Bowden-Kerby 1997; Smith and Hughes 1999), whereas others have not found such a correlation (Rogers et al. 1982; Lewis 1991; Bruno 1998). Most studies have not taken into account partial mortality. Had this not been done in the present study, no size-dependence would have been found, since very few corals were entirely dead. The positive relation between size and survival in the present study may be related to larger reserves and an increased capacity to repair damages (Connell 1973) or to increased resistance to water movements and ability to overgrow competitors. The greater relative growth rate for smaller fragments was probably related to a greater surface to volume ratio in smaller corals. This effect may have been present also among the unattached branches, but was probably overshadowed by the greater partial mortality among smaller branches. The relation between initial branch length and weight (Fig. 4) shows that the proportions of the transplanted branches were different depending on their length, since proportionality would have resulted in a linear relationship between the weight and the cube of the linear dimension (Maragos 1978). It is important to note that the branches used were fragments of colonies, and thus not representative of the natural growth form of the species.

Mechanical damage

The experiment on mechanical damage showed a more than five-fold difference in growth rate between some of the clones of A. muricata. Similarly, Rinkevich (2000) found a significant difference in growth rate between clones of Stylophora pistillata in the Red Sea. The magnitude of the variation in the present study was larger than expected and has important implications for coral reef rehabilitation. Clark and Edwards (1995) suggested that some corals may suffer from mortality and reduced growth as a consequence of the transplantation procedure. However, the present study did not show a significant effect of mechanical damage, even though the inflicted injuries were more severe than can be expected to occur during collection and transport. The capacity to repair damages is highly variable among coral species (Hall 1997), and the ability of A. muricata and other staghorn corals to withstand injuries is probably related to their strategy for reproduction through fragmentation (Tunnicliffe 1981; Kobayashi 1984).

Final notes

The string-grid method is extremely simple, and could easily be carried out by snorkelers down to a depth of 5-10 m after some basic training. The method is therefore well suited for community participation projects in developing countries, involving the local artisanal fishermen and other reef users. However, damages to source populations, recruitment limitation, and other ecological circumstances must be carefully considered by experts. Using artificial attachment allows transplantation of smaller fragments and results in greater rates of live-weight increase than would be possible with loosely placed branches in exposed conditions. There is no reason to assume that the method used in this study was optimal under the present circumstances. Therefore, future experimentation and new inventions are likely to produce better results. The string-grid method can be modified by altering the size and spacing of the corals and of the string sections, as well as by using other species or fast-growing clones. More research is also needed to assess the general applicability of the method in a wide range of habitats and to study effects of coral collection on source populations. Although transplanted staghorn corals have been shown to enhance fish abundance and diversity (Lindahl et al. 2001), it is important to bear in mind that transplantation of a few species of corals can never entirely replace a degraded coral reef ecosystem, and reef rehabilitation should never be viewed as an alternative to preserving existing coral reefs.

Acknowledgements

This study was carried out in co-operation with the Institute of Marine Sciences, Tanzania. I would also like to thank D. DeVilliers, K. Gallop, J. Greupner, H. Trattner, and M. Willson for assistance in the field. Dr. K. Johannesson, O. Lindén, J.-O. Strömberg, and anonymous referees gave valuable comments on the manuscript. Funding was provided by the Swedish International Development Co-operation Agency (Sida) through the Sarec Marine Science Program and by Göteborg University.

Copyright information

© Springer-Verlag 2003

Authors and Affiliations

  1. 1.Kristineberg Marine Research StationFiskebackskilSweden

Personalised recommendations