Effect of water flow on lesion regeneration of Acropora
Palmata transplants used in coral reef restoration
around St. Eustatius.
A Biology BSC Thesis
NOVEMBER 2018
UNIVERSITY OF AMSTERDAMAbstract
Coral reef restoration projects are trying to save coral reefs in decline. For Restoration of Ecosystem Services and Coral Reef Quality, a coral restoration project in the Dutch Caribbean, a research was set out to see how the reef restoration method of direct transplantation could be improved. The
experiment of three months would focus on the effect of different water flow strengths and the orientation to the dominant water flow on the wound regeneration process of transplants used for reef restoration around St. Eustatius. Two bays differing in water flow strength were the location for placement of fragments of Acropora Palmata, the dominant stony reef building coral in the
Caribbean which has seen a decline of more than 95%. We observed that the regeneration rate differed between the two restoration sites, the higher water flow strength was beneficial for wound regeneration. We also found that fragments with wounds parallel to the dominant water flow would have a lower maximum amount of regenerated lesion after the experiment. At last we observed an effect of the rotation of fragments and the regeneration rate and maximum amount of regenerated wound. This can be due to the fact that the rotation makes the effective surface for photosynthesis smaller which would have as a consequence that less nutrients would be present for the coral provided by zooxanthellae. These findings can be implemented in future restoration projects to increase the success rates of such projects
Introduction
Coral reefs have experienced a decline worldwide. One thing that played a big role in this decline is white band disease (WBD). WBD, fuelled by global warming (Randall & Van Woesik, 2015), was responsible for the decline of almost 95% of elkhorn (Acropora palmata) , a stony reef building species, in the Caribbean. Elkhorn was one of the dominating corals in the Caribbean from the reef crest down to approximately 5 meters depth (Pandolfi, 2002). The decline of the dominating elkhorn resulted in the loss of the benefits it provides for nature and humans. Benefits such as coastal protection, providing a foundation for other corals, creating habitats for lots of coral dwelling species due to their uniquely branching framework, and their carbon dioxide uptake out of the ocean are disappearing together with the stony reef building corals.
Coral restoration projects are needed because the prospect for their natural recovery process seems unpromising due to corals being impeded in every life stage by human driven changes in their environment. A complex interplay of factors like for example rising water temperature, ocean acidification and increasing algal abundance, impedes corals larval availability, successful settlement, and post-settlement survival and growth, which are all necessary for the recovery of the coral reefs (Ritson-Williams et al., 2009).
One of these restoration projects for the Caribbean reefs is Restoration of Ecosystem Services and Coral Reef Quality (RESCQ)1 . RESCQ focuses on restoring the reefs with the aquaculture of fast growing, sturdy, stony reef building corals as elkhorn corals. The aquaculture of elkhorn relies on fragmentation; the asexual reproduction method of some corals on which elkhorn emphasize rather than sexual reproduction (Highsmith, 1982). Fragmentation consists of fragments breaking off, which create new colonies elsewhere when they successfully attach themselves to the substrate.
Aquaculture of elkhorn consists of actively cutting fragments and cultivating them in conditions favourable to the coral species. So-called nurseries are being set up, which are structures where fragments are being grown and can then be split up in more fragments. Fragments can then be outplanted back on the reef. Another method is direct transplantation. This consists of cutting fragments of healthy colonies and placing them immediately back on the reef. The difference
between these methods is that the nursery method needs to take a fragment from a colony once, this fragment can then be grown and provide a stock which can be outplanted to the reef without needing to harm the original colony again. RESCQ carries both methods out in coral restoration projects in corporation with St Eustatius National Parks Foundation (STENAPA)2 around St. Eustatius. This research will focus on direct transplantation due to the fact that hurricane Irma and a big swell last spring destroyed every nursery placed by RESCQ around Statia (Dutch Caribbean Nature Alliance, 2017). The tools and materials needed for creating new nurseries will take months to arrive on the island. Meanwhile, a first step of restoration will be done by direct transplantation.
Research has shown that direct transplantation is suitable for elkhorn and that elkhorn has shown to be more resilient against storms and swells, which were the reasons that previous restoration projects around St. Eustatius have been unsuccessful (Forrester, Ferguson, O’Connell-Rodwell, & Jarecki, 2014). Besides this is elkhorn specialized in fragmentation which is the reproductive method on which the concept of direct transplantation is modelled, research showed how adaptative strategies in their morphology, which is an uniquely branching framework, and the habitat selection, the high-energy shallows exposed to waves and strong water flow, are favourable for fragmentation (Highsmith, 1982).
Long term experiments have shown that the direct transplantation method resulted in high rates of mortality (low success rate) for the fragments, mostly due to dislodgement and dying off in place. The reason for the low success rate of direct transplantation is unclear and asks for more research (Forrester et al., 2014).
The aim for this research is to investigate how to reduce the high mortality of direct transplantation. Investigating differences between source sites and the restoration sites with respect to
environmental factors could provide information about preferred habitats for fragments. Research showed that matching the environment of the source sites of the fragments with the restoration site creates a lower mortality for the transplants, i.e. fragments of coral transported and placed/bound on restoration site. In this research the environments where described in factors of water flow exposure, depth, maximum tidal waves and water clarity (Forrester, Taylor, Schofield, & Maynard, 2013).
This research will investigate the effects of different water flow rates and the orientation of the water flow relative to the transplants; which we will call the azimuth hereinafter. That water flow has effects on elkhorn has been shown repeatedly, but there is a lack of knowledge how it effects coral restoration of elkhorn. Research has shown that an increased water flow comes with an increased exposure to harmful sedimentation (Comeau, Edmunds, Lantz, & Carpenter, 2014), and higher photosynthesis and calcification rates (Dennison & Barnes, 1988). The transport of corals to a decreased water flow reduces the calcification rate (Kuffner, 2002). Increased water flow increases the efflux of oxygen and so reduces the oxidative stress for corals during their maximum
photosynthesis rate (Finelli, Helmuth, Pentcheff, & Wethey, 2006).This is of important for elkhorn because elkhorn corals have to endure a high light intensity for a long duration every day around St. Eustatius (Kuffner, 2002).
Water flow is also an important factor for the success for fragmentation of elkhorn corals. Research has shown that elkhorn corals have adaptive strategies in their morphology to cope with the high wave energy (Highsmith, 1982). The branching framework comes with strategies in the angle of attack of the water on the branches (1), variation of branch size and shape (2), strategies in the
colony eccentricity and alignment (3) and azimuths (4) relative to the direction of the water flow (Richard R. Graus, John A. Chamberlain Jr., 1977). The specific morphology shows that although swell or water flow direction and wave energy varies through time, the elkhorn endures on average one direction most intensely to which it will align itself in a specific manner. We think for this reason that the water flow can be a factor to which colonies adapt themselves locally and can be of influence on the fragments their health when they are being moved to a location with a different water flow. Due to the importance of the water flow for the elkhorn colonies, we want to investigate differences in water flow rate (1) and orientation to the water flow (2) on elkhorn transplants. We want to see if these factors can efficiently be integrated into coral restoration projects which make use of
fragments, e.g. direct transplantation and nursery transplantation, to increase the success of such projects.
Due to the short time period of the experiment (3 months) we choose to focus on the effect on the regeneration of the wound of the transplants. The side of the fragment which is cut we call
henceforth a lesion. Research has shown that elkhorn focuses nutrients and energy on regenerating lesions (Bak, 1983; Meesters, Pacuhli, & Bak, 1997). Some of the reasons for this are that lesions are the locations on coral colony most susceptible to diseases and coral recruitment. Corals are in competition with algae for space and research has shown that algae can be reservoir for coral diseases (Gignoux-Wolfsohn, Marks, & Vollmer, 2012; Nugues, Smith, Van Hooidonk, Seabra, & Bak, 2004; Sweet, Bythell, & Nugues, 2013). Though it is known that the effects of algae on corals are species specific and often the competition is not so intense that healthy corals can be outcompeted by algae, it has been shown that when corals have undergone stress and/or damage algae can become harmful (Sweet et al., 2013). This is exactly the case for the transplants. The transplants have undergone stress from the transplantation method and have a big lesion to coral surface ratio. The lesions of a coral are the place susceptible to algal recruitment and where also possible diseases can be transmitted which are present in the water column, e.g. the white band disease
(Gignoux-Wolfsohn et al., 2012) . So it’s important for healthy corals to regenerate damaged tissue and for transplants even a bigger importance. Algae have a high recruitment rate and growth rate, which is a reason why coral reefs have turned into algal reefs when the healthy colonies went into decline and coral recruitment was impeded. So we think that it’s important for elkhorn transplants to regenerate their lesions fast and completely. Some evidence has been offered by former research that showed allocation of nutrients to lesions and that growth is impeded when coral damage is present
(Rinkevich, 2005). Research showed that regeneration of lesions is shape and size specific, i.e. the higher perimeter to surface ratio of the lesion and the smaller the wound the faster the
regeneration. Also research showed that bigger wounds don’t get fully recovered; after a while the corals stop regenerating their lesions (Lirman, 2000; Meesters, Pauchli, & Bak, 1997).
We will focus on the effect of the water flow rate and orientation on the regeneration rate and maximum amount of lesion regeneration of transplants. As noted before are elkhorn corals mostly found in the high-energy shallows. We suspect a high water flow rate will be more suitable for the elkhorn transplants and that the lesion regeneration process will be faster and that the amount of surface regenerated will be higher. Though this higher water flow can also come with more sedimentation which could impede the regeneration process. There is evidence found for this in earlier research (Sabine, Smith, Williams, & Brandt, 2015)
We suspect that when fragments their lesions are more oriented towards being parallel to the water flow shear stress and erosion will contribute to impeding the regeneration process. We expect to see
that lesions in a parallel orientation to the water flow will have a lower maximum regeneration than other fragments.
To test this hypothesis a three months long in-situ experiment will be set up in which two bays, differing in water flow rate, will be supplied with fragments of elkhorn. The fragments will be randomly placed in orientation to the water flow. The lesion regeneration process will be recorded and the orientation to the water flow will be measured. This will provide the information needed for the influence the water flow rate and azimuth on the regeneration process of elkhorn corals and at the same time will be a first step in restoring elkhorn corals around St. Eustatius.
Material & Methods
We carried out a comparative experiment wherein two shallow bays around St. Eustatius, in which Acropora Palmata have disappeared, were being repopulated with fragments taken off healthy Acropora Palmata found around the island. The two areas being repopulated with fragments are henceforth being called restoration sites. The colonies providing the fragments will be called donor colonies hereinafter.
We looked for donor colonies in the shallow depths between 3-10 meters . The fragments taken off a healthy donor colony did not make up more 10% of the whole colony. A total of 55 fragments were cut off from 11 different donor colonies and randomly divided over two restoration sites. The restoration sites for the transplants were two bays with difference in water flow rate while being easily reachable from the harbour but are not a commonly visited by fishermen or tourists. The selection came on Jenkin’s Bay and Crook’s Castle, for the low and high water flow rate respectively. From May to July in 2018 fragments, 3 – 16 cm long, were being collected and randomly distributed over the dead elkhorn and volcanic reef patches on a depth between 2-5 meters. The pieces were bonded to the substrate with either tie wraps or epoxy. The reef patches were cleared of algae before placing the fragments and the algae growing back were removed weekly. Pictures and notes regarding the status of the fragments were taken every week. The fragments were measured with a ruler in terms of their maximum length and their middle bisector length, the length perpendicular to the centre of the former maximum length. These two measurements together provided the surface area. Besides this, with the help of a diving a computer, also the depth at donor colony and
restoration site, transport time from donor colony to restoration site were recorded and examined in the statistical procedure. We assumed the shape of the lesions to be elliptic. A horizontal and vertical length of the lesion were measured to calculate the lesion surface.
Figure 1ab These are the dominant currents (left) and wind patterns (right) around St. Eustatius, the island is marked with the green circle in the left figure (the island itself is not designated). In the right figure the red star marks the restoration site at Jenkin's Bay and the black star marks the restoration site at Crook's Castle.
Difference in waterflow strength
The fragments were divided over two restoration sites. Jenkin’s Bay played the role of the restoration site with a lower water flow strength and Crook’s Castle was the restoration site with a higher water flow strength. At Crook’s Castle 28 fragments were set out and the 26 other fragments were
outplanted at Jenkin’s Bay, one was lost during transport. Figure 1 shows the ocean currents and wind patterns around St. Eustatius, also the location of the two restoration sites are shown, the wind patterns come from the almost always present trade winds. Because the experiment is set out in the shallower depth both swell and wind-generated surface waves are affecting the sites. Jenkin’s Bay on the Northern part of St. Eustatius is on the lee side of the mountain Boven as one can determine from Figure 1b. Besides this are the dominant currents coming from South-East towards St. Eustatius as is seen in Figure 1a. Therefore Jenkin’s Bay is a calmer bay than Crook’s Castle, which we also experienced during the dives.
Orientation to dominant water flow
A fragment at the donor colony undergoes the force of the dominant waterflow. We assumed for this experiment that this flow comes mostly from one direction.
The water flow azimuth was determined using a neutrally buoyant rope recorded from directly above, next to the donor colony. The azimuth is the angle between the growth direction of the fragment and the incoming waterflow in the plane parallel to the ocean surface. So it’s the angle around the axis going vertical from the fragment to the surface. Transplants also could be rotated over the other two axes. For clarification will the two other axes be called the “horizontal axis” and the fragment their “own axis”. The own axis is the axis going through the fragment, perpendicular to the lesion. A rotation over the horizontal axis of 90 degrees would be a fragment which was first oriented parallel to the ocean surface towards the water flow but at the restoration site would be pointed towards the ocean surface. So the horizontal axis could be imagined parallel to the lesion surface Pictures taken from the fragments on the donor colony provided information to determine the azimuths in ImageJ3. For the transplants at the restoration site the same procedure was carried out. When the water flow was too weak and the rope method was not suitable, the dominant orientation of the sea fans was used to determine the dominant direction of the water flow. Sea fans are mostly found perpendicular oriented towards the dominant water flow. Elkhorn corals undergo
the water flow from the incoming and outcoming swell so a piece turned around for 180 degrees at the restoration site would still undergo the same direction of water flow as on the donor colony, so there is no change in azimuth. So a shift of 90 degrees is the biggest change in orientation. That is why the azimuths were transformed into an angle from 0 - 90 degrees, i.e. from the same orientation - perpendicular to the old orientation.
Output values
Due to the fact that our stay was short and that it can take sometimes a month for the first signs of growth to show, we used the lesion half-regeneration time as a value to examine the influence of the water flow strength and orientation. Pictures taken from the lesion of the fragments, if reachable, will be examined in ImageJ to track the regeneration process. Every week a picture showed how much of the lesion was regenerated and using ImageJ the ratio of remaining visible lesion to the total initial lesion gave the percentage of the remaining lesion to close. In excel graphs were made of the regeneration process and the time value of 50% was calculated. The time it takes to regenerate its lesion for 50% was used as a value for the statistics. Also the initial horizontal and vertical length of the lesion were recorded from which the lesion surface area was deduced. The ratio of the lesion surface and fragment surface could then also be examined for the different treatments. For reasons as that some fragments died, were lost due to dislodgement, or that their lesion was not able to be measured, e.g. the lesion was in the epoxy or the lesion was towards the reef, the total number of fragments for the statistical analysis became 33 fragments. All data was recorded in a google datasheet and all pictures were stored in a google drive open to public4. All statistical analysis were carried out in Rstudio.5
Results
Of a total of 33 fragments we were able to monitor their regeneration process. The regeneration process of a fragment is shown in pictures in Figure 2. Figure 3 shows the regeneration over time of the fragments.
4 https://drive.google.com/drive/u/1/folders/1u5SVuBQ4KUcxXrfDSxEEs5igByseYp5k is open to public where all pictures are found and all datasheets are placed. The datasheet and script used for this paper will be placed in the Appendix
Figure 2a-c These three pictures show transplant F6 at Jenkin's Bay after 27, 49 and 58 days and captures the process of its lesion regeneration
Figure 3 This graph shows the amount of regenerated lesion surface as a function of time for 33 fragments
The angle measurement tool in ImageJ provided us with the azimuths at donor colony and restoration site. In the pictures the dominant water flow direction would be presented in the compass and the line perpendicular to the lesion would be the reference line to determine the azimuth. An example can be seen in Figure 4.
Figure 4 Here are fragment D2 and F6 with the dominant water flow direction presented on the compass. Note that the rope is not in the same direction as the direction of the line in the compass this is due to the fact that this is only a moment and the water flow direction was determined using videos.
Effect of restoration sites with difference in water flow strength on lesion
regeneration
The half-regeneration time of the transplants on restoration site Crook’s Castle were significantly lower (13.6, n = 12) then those on Jenkin’s (19.7, n = 21) (one-way ANOVA, p < 0.05). When simulations of the 95% confidence limits were carried out, the means only differed on a significance level of p < 0.1. As is shown in Figure 5 that the 95% confidence limits of the half-regeneration time overlap for the different restoration sites.
Shown also in figure 5 is that the lesion to fragment surface ratio was significantly lower for Jenkin’s bay than for Crook’s Castle. The data showed a significant difference between the two restoration sites with respect to lesion to fragment surface area ratio (lm-test, p= 0.001115). Crook’s Castle (mean= 0.252, n = 12) had a on average fragments with a higher ratio than Jenkin’s Bay (mean = 0.103 , n = 21).
Figure 5 Shown here are the means of half-regeneration time and lesion to fragment ratio for the two restoration sites. The 95% confidence limits are presented as the horizontal and vertical segment lines.
The data didn’t show a significant correlation between the half regeneration time and the surface ratio (lm-test, p = 0.374, n = 33). It is summarized in Figure 6.
Figure 6 This graph shows the half-regeneration time as function of the lesion to fragment surface area ratio. The fragments at Crook's Castle and Jenkin's Bay are shown in black and red, respectively.
Other environmental factors were being monitored as well. The data showed a negative correlation between the transport time, the time it took from cutting a piece until it was placed at a restoration site, and the half-regeneration time and is shown in Figure 7 (lm-test, p = 0.03866)
Figure 7 This figure shows the half-regeneration time as a function of the transport time with the 95% confidence limits presented as dashed lines.
Though the transport time didn’t significantly differ between the fragments going to Crook’s Castle (83 min) or going to Jenkin’s Bay (108 min) (lm-test, p = 0.4042). This is presented in Figure 8 wherein the 95% confidence limits of the two groups are shown and overlap.
Figure 8 This graph shows the transport time and half-regeneration time for the two restoration sites with their 95% confidence limits presented as the horizontal and vertical segment lines respectively.
The results of the statistical test of the rest of the factors which have been examined are summarized in Table 1. We monitored the factors:
Fragment length and surface Lesion surface and shape
Depth of fragments on the donor colony, at the restoration site and the change between the donor colony and the restoration site
Azimuth at donor colony, restoration site and the change between those azimuths. The rotations over the three spatial axes and those rotations combined.
Table 1 ; The distribution of different factors over the two restoration sites are summarized. The means, F-value, p-value and confidence limits are given. The last collum shows if the simulated confidence limits of the restoration site are seperated or overlap when there was a signififcant difference observed between the two restoration sites. * = (p < 0.1) ** = (p< 0.05), *** = (p<0.01) Factor Crook’s Castle Jenkin’s Bay F(df) p-value Crook’s 95% interval Jenkin’ s 95% interva l Interval separation in 100 simulation s Fragment Dimensions Fragment Initial Maximum Length (cm) 8.2 10.2 (1,32) = 5.425 0.02633** 6.991-9.564 9.088-11.174 8% separated Fragment Surface Area (cm2) 10.42 13.56 (1,32) = 1.295 0.2636 7.332-14.158 10.124-16.578 No Lesion Surface Area (cm2) 5.18 3.90 (1,31) = 3.997 0.05441* 4.240-6.267 3.086-4.747 0% separated Lesion Shape (-) 4.74 5.24 (1,31) = 0.1658 0.6866 2.36-6.77 4.02-6.66 no
Depth (m) Fragments depth at Donor Colony (m) 3.63 3.24 (1,28) = 2.341 0.1372 3.276-3.991 2.926-3.516 no Fragments depth at Restoration site (m) 3.09 3.57 (1,30) = 5.182 0.03013** 2.77-3.42 3.31-3.81 2% separated Change in Depth (m) -0.58 0.43 (1,26) = 11.01 0.00269** * -0.991 – -0.002 0.004-0.738 97% separated Orientation to Water Flow (degrees)
Water flow Azimuth at Donor Colony (°) 26.84 32.62 (1,32) = 0.4896 0.4891 13.51-40.27 22.00-41.70 No Water flow Azimuth at Restoration Site (°) 35.33 38.60 (1,30) = 0.1017 0.7521 21.38 – 50.83 24.69 -52.76 No Azimuth Change (°) 23.50 29.65 (1,30) = 0.5921 0.4476 13.152-34.444 18.776 – 39.724 No Horizontal Axis Rotation (°) 6.5 22.5 (1,30) = 5.575 0.02491** 4.29 -18.09 13.74 – 30.13 12% separated Own Axis Rotation (°) 32.08 7.5 (1,30) = 7.391 0.01079** 16.916 – 46.693 -3.166-18.118 53% separated Combined Rotation (-) 0.158 0.165 (1,30) = 0.0305 0.857 0.0859 – 0.2314 0.1211 – 0.2047 No
Table 1 shows the following observations: Fragment Dimensions
There a significant difference between the two restoration sites their fragments with respect to the initial maximum length and the lesion surface area; the fragments on Crook’s Castle were on average shorter fragments (8.2 cm) and had bigger lesions (5.18 cm2) than the fragments on Jenkin’s Bay (10.2 cm, 3.9 cm2).
Depth
As noted in table 1 was there a significant difference between the two restoration sites with respect to the depth the fragments were placed and to the change in depth of the fragments between donor colony and restoration site. The fragments on Crook’s Castle were placed at a shallower depth (3.09 m) than the fragments on Jenkin’s bay (3.57 m). Also were the fragments on Crook’s Castle placed on a shallower depth compared to the depth of their donor colony (-0.58 m) while the fragments on Jenkin’s bay were placed on a deeper depth compared to the depth of their donor colony (+0.43 m). Orientation
The rotation over the horizontal axis and the own axis differed between the two restoration sites their fragments. The fragments at Crook’s Castle were more rotated over the own axis (mean = 32.05 degrees) and less twisted over their horizontal axis (mean = 6.5 degrees) than the fragments at Jenkin’s Bay were twisted over their own axis (mean = 7.5 degrees, lm-test, p = 0.02491) ) and horizontal axis (mean = 22.5, lm-test, p = 0.01079)
The correlation of the factors that differed over the restoration sites (initial maximum length, lesion surface area, depth at restoration site and change in depth) with half-regeneration time are shown in Figure 9. The results of the lm-tests over the four factors and half-regeneration time are shown in Table 2. None of the factors showed a significant correlation with half-regeneration time.
Figure 9 The correlation of the half-regeneration time with the initial maximum length (top left), the lesion surface area (top right), the depth at restoration site (bottom left), the change in depth (bottom right). With the fragments for Crook’s Castle and Jenkin’s Bay marked with black and red points respectively.
Table 2 The results of LM-tests on half-regeneration time and the four factors which differed over the two restoration sites. Shown is the size sample, multiple r-squared, F-value, p-value, slope and intercept.
Factor df R2 F p-value Slope Intercept
Initial Maximum Length 31 0.03215 1.03 0.3181 0.4244 14.7397 Lesion Surface Area 30 0.02378 0.7309 0.3994 -0.6313 21.7250 Fragments Depth at 29 0.004158 0.1211 0.7304 0.3912 15.34595
Restoration Site Change in
Depth 24 0.0136 0.7769 0.3868
2.056 18.181
When testing the correlation of the horizontal axis rotation and the own axis rotation with the half-regeneration time, lm-test gave no significant effect as is shown in table 3.
Table 3 Results of lm-tests of different factors on the half-regeneration time. Sample size (N), Multiple R-squared, F and p value, slope and intercept are given.
Factor df R2 F p-value Slope Intercept
Water flow azimuth at donor colony 31 0.04692 1.526 0.226 -0.0786 21.153 Water flow azimuth at restoration site 29 0.002544 0.07395 0.7876 -0.0155 18.01162 Azimuth change 29 0.005483 0.1599 0.6922 0.0268453 16.6727 Horizontal axis rotation 29 0.085262 2.715 0.1102 0.1125 16.9325 Own axis rotation 29 0.05618 1.726 0.1992 -0.075217 18.6891 Combined rotation 29 0.004782 0.1393 0.7117 5.658 17.896
Results of orientation on lesion regeneration rate
The second part of this research focused on fragments their orientation to the water flow. In figure 10 is the distribution shown of azimuths found at the donor colony.
Figure 10 This histogram shows the frequency of found azimuth of fragments at the donor colony.
As noted before Table 3 shows the correlation between the different orientation on the water flow on three axes. Lm-tests didn’t show a significant correlation between the azimuth at donor colony and at restoration site, azimuth change and the rotation over each of the three spatial axes and the combination of those didn’t show a significant correlation. The lm-tests of these 6 factors with the half-regeneration time are presented visually in Figure 11-16.
Figure 11-12 Here is the half-regeneration time presented as a function of the azimuths at the donor colony (left, p-value = 0.226) and at the restoration site (right, p-value = 0.7876) . The 95% confidence limits are shown with the dashed lines with the fragments for Crook’s Castle and Jenkin’s Bay marked with black and red points respectively.
Figure 13-14 Here is the half-regeneration time presented as a function of the rotation over the horizontal axis (left, p-value = 0.1102) and the rotation over own axis (right, p-value = 0.1992). The 95% confidence limits are shown with the dashed lines with the fragments for Crook’s Castle and Jenkin’s Bay marked with black and red points respectively.
Figure 15-16 Here is the half-regeneration time presented as a function of the azimuths difference (left, p-value = 0.6922) and the three rotations combined (right, p-value = 0.7117). The 95% confidence limits are shown with the dashed lines with the fragments for Crook’s Castle and Jenkin’s Bay marked with black and red points respectively.
Amount of lesion regenerated
As noted before lesions tend not to be fully regenerated always. The regeneration process reaches an asymptote as can be seen in Figure 3. Our data didn’t show a correlation of the amount of
regenerated lesion surface with initial lesion size (lm-test, p = 0.8796) , initial fragment size (lm-test, p = 0.9538) or the lesion to fragment surface area ratio (lm-test, p = 0.582).
Between the two restoration sites was no difference found in the maximum regeneration of the transplants (one-way ANOVA, p =0.675). The average maximum regeneration of the fragments is 78% (n=32).
We did find a significant correlation with the maximum regeneration and transport time. The transport time showed a negative correlation with the maximum regeneration as is shown in figure (lm-test, p<0.01)
Figure 16 This figure shows the maximum regeneration as a function of the transport time with the 95% confidence limits presented as dashed lines.
The azimuth at the restoration site also showed a negative correlation with the maximum regeneration (lm=test, p<0.05). This is shown in figure 17.
Figure 17 This figure shows the maximum regeneration as a function of of the azimuth found at the restoration sites with the 95% confidence limits presented as dashed lines.
Also the rotation on the horizontal axis showed a negative correlation with the maximum regeneration as showed in figure 18 (lm-test, p<0.01)
Figure 18 This figure shows the maximum regeneration as a function of the rotation over the horizontal axis with the 95% confidence limits presented as dashed lines.
Table 4 shows the results of an analysis of variance of the different fitted models which can be made from the three factors found to have a correlation with the maximum regeneration. As is shown in the Table 4 is that there is no significant difference between the fitted models of:
1. the maximum regeneration described as a function of the transport time and rotation on the horizontal axis
2. the same model as above but then with the addition of the azimuth at the restoration site (p=0.1636)
Though the residual sums of squares is the lowest for the model describing the maximum regeneration as a function of all three the factors.
Table 4 This table shows the analysis of variance and if they differ significantly. A nested factor gets a second factor added which is than compared to just the fitted model with just the first factor. After this the third factor is added and this is compared to the model with the two factors. Before this an analysis of variance is done between for the interaction of the factors in the model. For example an analysis of variance was done also to check the interaction of the factors. The interaction was for every two compared fitted models never significant (p>0.1). ^ = for models to be compared they have to have the same amount of observations, so that’s why the F, df and p differ for the transport time with figure. * p < 0.1, ** = P < 0.05, *** p < 0.01.
Nested Factor RSS F(df) p-value
Transport Time
5845.3 (1,28)^ =
11.81 0.001858***
+2nd factor
Azimuth Restoration Site
5143.4 (2,27) =
3.6848 0.06553 * + 3
rd factor
Rotation Horizontal Axis
4039.0 (3,26) =
7.1091 0.01301 **
+2nd factor
Rotation Horizontal Axis
4358.4 (2,27) =
9.2118 0.005271*** + 3
rd factor
Azimuth Restoration Site
4039.0 (3,26) =
2.0557 0.1636 Azimuth Restoration Site
6963.8 (1,28) = 5.416 0.02741 ** + 2nd factor Transport Time 5143.4 (2,27) = 9.5561 0.004588*** + 3 rd Factor
Rotation Horizontal Axis
4039.0 (3,26) =
7.1091 0.01301**
+ 2nd factor
Rotation Horizontal Axis
5764.4 (2,27) = 5.618 0.02518** + 3 rd factor Transport Time 4039.0 (3,26) = 11.107 0.002588*** Rotation Horizontal Axis
6489.2 (1,28) = 7.86 0.009076 *** +2nd factor Transport Time 4358.4 (2,27) = 13.2 0.001158*** + 3 rd factor
Azimuth Restoration Site
4039.0 (3,26) =
2.0557 0.1636
+2nd factor
Azimuth Restoration Site
5764.4 (2,27) = 3.395 0.0764 * + 3 rd factor Transport Time 4039.0 (3,26) = 11.107 0.002588 ***
Discussion
We hypothesized that elkhorn corals would regenerate slower at locations with a low water flow strength. We found that at Jenkin’s bay, which is the restoration site with a low water flow strength, there was a higher half-regeneration time, i.e. the regeneration process took longer there. To see if the difference between the two restoration sites comes from the water flow strength we tried to examine as many factors which could influence the regeneration of fragments and see if there influence was present.
We found on average smaller lesion to fragment ratio’s (smaller lesions, bigger fragments) at Jenkin’s bay. This would suspect that that a faster regeneration would correlate with a smaller lesions or bigger fragments but we found the opposite. Fragments at Jenkin’s Bay were placed deeper compared to Crook’s Castle but the difference of only half a meter should not be much on an influence cause elkhorn corals used to dominate the complete shallows up until 10 meters. Our data also shows that fragments were placed between 3 and 5 meters depth at the restoration sites and there was no effect to be found on the regeneration process as we would expect on this small range of depth differences.
Research from the Global Coral Reef Monitoring Network-Caribbean, a coral reef monitoring project of the Caribbean , showed that there is no difference in turbidity between Jenkin’s Bay and Crook’s Castle which is used as a measurement for the intensity of sedimentation (Kitson-Walters, 2017). So this could also not be creating the difference between the two sites.
The second part of this research focused on effect of the orientation to the water flow on the regeneration rate. Our experiment showed that there was no showed no significant effect of the orientation on the regeneration rate.
Though we did find that more fragments we cut off of the donor colony tended to have a smaller azimuth as was shown in Figure 9. This could be due to the fact that the fragments cut off were mostly chosen on the efficiency of cutting; branched parts of the colony could probably be cut off easier. Branches are long in one direction but short in the perpendicular direction, so a big piece could easy be collected by only having to cut a small distance, that’s why they are preferable to cut. But this then implies that the branching morphology of elkhorn indeed elongates their branches in directions against or with the dominant water flow. So maybe the orientation of the fragments could influence the growth of the fragments, but for the regeneration rate of lesions we did not observe a direct effect.
The maximum amount of lesion regenerated was the second aspect of lesion regeneration this research focused on. For the maximum regeneration was no difference observed between the two restoration sites though we hypothesized that at the site with the lower water flow rate the
maximum regeneration would be lower. This observation is maybe due to the fact that whatever the conditions influencing the wellbeing of the fragments, the wound has the highest priority to be closed as much as possible to overcome the immediate threat of algae and diseases.
We did find a negative correlation of (1) a rotation over the horizontal axis and of (2) the azimuth at the restoration site. The bigger the azimuth (2), the lower the maximum regeneration, this is also what we hypothesized. We think this comes from the fact that the passing water flow creates stress and erosion on the lesion when the lesion is parallel to the water flow, i.e. of the incoming and outgoing water flow. Also when the lesion is perpendicular to the water flow it only undergoes the force of the water flow of the incoming or of the outgoing flow, not both.
The influence on the horizontal axis (1) can come from the fact that when fragments are rotated over this axis the consequence is that the amount of sunlight for photosynthesis becomes less. Elkhorn corals show in their morphology that it grows preferably in branching sheets which are parallel to the ocean surface as can be seen in Figure 19. A rotation over the horizontal axis does not fit this preference. It is comparable to when one uses his hand to block the incoming sunlight from the face. This is most effective with the palm of the hand being in juxtaposition to the incoming sunrays. When one rotates his hand 90 degrees and his fingertips are now pointing towards the sun, there is less sunlight being blocked. Photosynthesis carried out by the zooxanthellae provides the coral with
nutrients, which can be a reason why fragments rotated over the horizontal axis are doing worse than non-rotated fragments.
Figure 19 This figure shows the morphology of elkhorn from directly above. It is clear that the growth orientation of the branching sheets is parallel to the ocean surface and creates a big surface for catching light.
Though there was no significant statistical correlation found between the horizontal rotation and the half-regeneration time, we think there would still be an effect present of the horizontal rotation on the half-regeneration time. The sample size is rather small which could be the reason for not finding a significant effect. At the same time it is a factor in which the transplants differed significantly between Jenkin’s bay and Crook’s Castle. Crook’s Castle had a bigger rotation over the horizontal axis which could correspond to the slower regeneration that was observed there. Besides this was the p-value quite small (0.1105), the smallest p-value of the factors corresponding to orientation to water flow. We suspect a bigger sample size would have shown this influence of rotation of the horizontal axis on the half-regeneration time as well as it was observed on the maximum regeneration.
The small sample size could be influencing more of our findings. As former research showed that there is an effect of the lesion size on the lesion regeneration but we didn’t find this in our data. We often found significant differences between the two sites shown by statistical tests, but simulation tests didn’t show the difference between the two restoration sites. More often is this the case for factors influencing the half-regeneration time, e.g. fragments initial length and lesion surface area as shown in table 2. This could also be the reason why we see in our data sometimes strong tendencies of influence on the half-regeneration time but cannot make the claim on statistical tests. We
expected to find, as earlier research has observed, that the bigger the lesion size compared to the fragment the longer it will take to heal those lesions, but we didn’t observe this correlation (lm-test, p=0,375, figure 5). We think this is due partly to the small sample size (around 33 fragments). Unfortunately a lot of fragments their lesion regeneration process couldn’t be recorded, for reasons as that the lesions were not in sight. For the maximum regeneration the sample size was even smaller due to the fact that some fragments were not at their maximum of regenerated lesion surface, so we decided to exclude those for that part of the research.
We observed that a longer transport time had a negative correlation with the half-regeneration time, i.e. a longer transport time with a faster regeneration. At the same time we observed a negative correlation with the maximum regeneration, i.e. a longer transport time correlates with a smaller lesion surface being regenerated. One would suspect that the transport of the fragments is not
happening in optimal conditions for fragments and a long transport time would be expected to be harmful to the fragments. We also showed with fitted models that the observations for maximum regeneration are best explained by the rotation over horizontal axis and transport time and that the addition of the azimuth at restoration site is not necessary (p=0,1636). So here the findings
correspond to our expectations. The reason why the half-regeneration time is lower, i.e. faster regeneration, is unknown to us. It could be that the correlation observed it not a real effect, the transport time is a relative short period of time compared to the whole experiment. But then again the transport time correlation with the maximum regeneration seems very strong.
We conclude that the orientation of fragments to the water flow and the water flow strength can influence the lesion regeneration process of elkhorn fragments being used for coral reef
rehabilitation. These findings can be implemented in the continuing projects of coral reef restoration in the Caribbean carried out by RESCQ. We already started using a method of placing the fragments with their wound in the epoxy during the experiment, so the fragment doesn’t have to regenerate the wound at all and can start growing earlier after placement. But this method often went hand in hand with a rotation over the horizontal axis. And this research has shown that this rotation will most probably be of harm to the fragments. Though this research focused on the wound regeneration, one would suspect that less effective photosynthesis by the rotation will influence the fragment on every aspect of its wellbeing. A strong water flow seemed more beneficent for the fragments but as the elkhorn coral used to be the dominant coral in most of the shallows around the Caribbean it will probably be quite resilient to different water flows.
For further research we advise to look for influence of more environmental factors which can impede the coral restoration projects. And while the reef restoration projects are necessary, the coral reefs are still in decline and without the coral reefs, many other organisms will disappear, the dominant factors being the reason for the loss of corals are still out there. Human actions are the reason behind climate change and this comes with rising ocean temperatures which has shown to be devastating to coral reefs. Coral reefs worldwide undergo massive bleaching events which comes with death as the final result for these organisms. To create a future that is safe for the coral reefs and their ecosystem a change in the ways of living of humans is necessary.
Acknowledgements
I am grateful for Erik Meesters and for his guidance through my thesis process and for the
opportunity to have been part of the RESCQ project. Also I’d like to thank STENAPA for having me as an intern and making me feel home on St. Eustatius. The enthusiasm and willingness to help of Francois Miller made the days fly by. At last I want to express my gratitude for Haye Geukes for being my always present dive-buddy, research partner and friend.
Literature
Bak, R. P. M. (1983). Neoplasia, regeneration and growth in the reef-building coral Acropora palmata.
Marine Biology, 77(3), 221–227. https://doi.org/10.1007/BF00395810
Comeau, S., Edmunds, P. J., Lantz, C. A., & Carpenter, R. C. (2014). Water flow modulates the response of coral reef communities to ocean acidification. Scientific Reports, 4, 1–6. https://doi.org/10.1038/srep06681
Dennison, W. C., & Barnes, D. J. (1988). Effect of water motion on coral photosynthesis and calcification. Journal of Experimental Marine Biology and Ecology, 115(1), 67–77. https://doi.org/10.1016/0022-0981(88)90190-6
St Eustatius. BioNews 3. Retrieved from
http://www.dcnanature.org/wp-content/uploads/2017/10/BioNews_2017_3_Nature-funding-projects-coral-restoration.pdf Finelli, C. M., Helmuth, B. S. T., Pentcheff, N. D., & Wethey, D. S. (2006). Water flow influences oxygen
transport and photosynthetic efficiency in corals. In Coral Reefs (Vol. 25, pp. 47–57). Springer-Verlag. https://doi.org/10.1007/s00338-005-0055-8
Forrester, G. E., Ferguson, M. A., O’Connell-Rodwell, C. E., & Jarecki, L. L. (2014). Long-term survival and colony growth of Acropora palmata fragments transplanted by volunteers for restoration.
Aquatic Conservation: Marine and Freshwater Ecosystems, 24(1), 81–91.
https://doi.org/10.1002/aqc.2374
Forrester, G. E., Taylor, K., Schofield, S., & Maynard, A. (2013). Colony growth of corals transplanted for restoration depends on their site of origin and environmental factors. Marine Ecology, 34(2), 186–192. https://doi.org/10.1111/maec.12000
Gignoux-Wolfsohn, S. A., Marks, C. J., & Vollmer, S. V. (2012). White Band Disease transmission in the threatened coral, Acropora cervicornis. Scientific Reports, 2(1), 804.
https://doi.org/10.1038/srep00804
Highsmith, R. (1982). Reproduction by Fragmentation in Corals. Marine Ecology Progress Series, 7, 207–226. https://doi.org/10.3354/meps007207
Kitson-Walters, K. (2017). ST. EUSTATIUS GCRMN Caribbean Final Report 2017. Oranjestad. Retrieved from https://www.icriforum.org/sites/default/files/St. Eustatius GCRMN 2017 Report LNV CNSI.pdf
Kuffner, I. B. (2002). Effects of ultraviolet radiation and water motion on the reef coral, Porites compressa Dana: A transplantation experiment. Journal of Experimental Marine Biology and
Ecology, 270(2), 147–169. https://doi.org/10.1016/S0022-0981(02)00031-X
Lirman, D. (2000). Lesion regeneration in the branching coral Acropora palmata: effects of colonization, colony size, lesion size, and lesion shape. Marine Ecology Progress Series, 197, 209–215. Retrieved from https://www.int-res.com/articles/meps/197/m197p209.pdf
Meesters, E. H., Pacuhli, W., & Bak, R. P. M. (1997). Predicting regeneration of physical damage on a reef-building coral by regeneration capacity and lesion shape. Marine Ecology Progress Series,
146, 91–99. Retrieved from https://www.int-res.com/articles/meps/146/m146p091.pdf
Meesters, E. H., Pauchli, W., & Bak, R. P. M. (1997). Predicting regeneration of physical damage on a reef-building coral by regeneration capacity and lesion shape. Marine Ecology Progress Series,
146(1–3), 91–99. https://doi.org/10.3354/meps146091
Nugues, M. M., Smith, G. W., Van Hooidonk, R. J., Seabra, M. I., & Bak, R. P. M. (2004). Algal contact as a trigger for coral disease. Ecology Letters, 7(10), 919–923. https://doi.org/10.1111/j.1461-0248.2004.00651.x
Pandolfi, J. M. (2002). Coral community dynamics at multiple scales. Coral Reefs, 21(1), 13–23. https://doi.org/10.1007/s00338-001-0204-7
Randall, C. J., & Van Woesik, R. (2015). Contemporary white-band disease in Caribbean corals driven by climate change. Nature Climate Change, 5(4), 375–379.
https://doi.org/10.1038/nclimate2530
Richard R. Graus, John A. Chamberlain Jr., A. M. B. (1977). Structural Modification of Corals in
Relation to Waves and Currents: Reef Biota. In SG 4: Reefs and Related Carbonates--Ecology and
http://archives.datapages.com/data/specpubs/carbona1/data/a045/a045/0001/0100/0135.ht m#purchaseoptions
Rinkevich, B. (2005). Conservation of coral reefs through active restoration measures: Recent approaches and last decade progress. Environmental Science and Technology.
https://doi.org/10.1021/es0482583
Ritson-Williams, R., Arnold, S., Fogarty, N., Steneck, R. S., Vermeij, M., & Paul, V. J. (2009). New perspectives on ecological mechanisms affecting coral recruitment on reefs. Smithsonian
Contributions to the Marine Sciences, (38), 437–457.
https://doi.org/10.5479/si.01960768.38.437
Sabine, A. M., Smith, T. B., Williams, D. E., & Brandt, M. E. (2015). Environmental conditions influence tissue regeneration rates in scleractinian corals. Marine Pollution Bulletin, 95(1), 253–264. https://doi.org/10.1016/j.marpolbul.2015.04.006
Sweet, M. J., Bythell, J. C., & Nugues, M. M. (2013). Algae as Reservoirs for Coral Pathogens. PLoS
ONE, 8(7), e69717. https://doi.org/10.1371/journal.pone.0069717
Appendix
The following link will be the directory with the Rstudio script, the pictures for the monitoring of the fragments, the datasheets and the ImageJ files all used for this thesis.
