Effectiveness of coral relocation as a mitigation strategy in Kāne‘ohe Bay, Hawai‘i

Donor site

Ten permanent survey sites were strategically placed in the channel to allow full documentation of changes following transplantation. These sites were marked with stainless steel pins and coded cable ties to allow for future resurveys. Initial sites established in 2005 were resurveyed in 2008, 2012, and 2016. An Olympus 5050 camera with underwater housing attached to a monopod for stability, consistent distance from the bottom, and constant image size that covered a 50 cm × 63 cm area. Images were analyzed using PhotoGrid (Bird, 2001). Proportion of coral cover was transformed using a square-root transformation to meet the assumption of normality and equal variance for statistical tests. A General Linear Model (GLM) (MINITAB 17) was used to examine the differences in mean total coral cover between years followed by Tukey Pairwise comparisons with simultaneous 95% Confidence Intervals (CI). Percent changes (Excel 2010) were calculated to compare temporal changes in coral cover and composition between years. The differences in median cover of two coral species, Montipora capitata and Porites compressa were examined separately using a Kruskal–Wallis test.

Relocation site

Fish populations were recorded using standard visual belt transects (Brock, 1954). Divers using SCUBA swam along one 50 m × 5 m transect (125 m²) recording species, abundance, and total fish length. All fishes were identified to the lowest taxon possible and estimated to the nearest centimeter. Length-weight fitting parameters obtained from the Hawai‘i Cooperative Fishery Research Unit and Fishbase (http://www.fishbase.org) were used to convert total length to biomass (Kg m⁻²). Trophic levels for fish species were determined using published data. This research was conducted under the Hawai‘i Department of Land and Natural Resources-Division of Aquatic Resources Special Activity Permit No. 2005-25.

Donor site

There was a significant effect of year on total coral cover in the cleared channel at the donor site (F_3,34 = 9.53, p = 0.000, $R_{\mathrm{adj.}}^2=40.9\text{\%}$) (Fig. 2). The average percent cover of all corals was the highest in 2012 (Table 1). Most of the coral increase in the channel at the donor site occurred in the first three years between 2005 and 2008 (527.5%, Table 2). Subsequently, between 2008 and 2012 there was a further increase of 57.2% in total coral cover. The differences of means between 2005 and 2008 (95% CI [0.042–0.543]) as well as 2005 and 2012 (95% CI [0.218–0.707]) were statistically significant. Between 2012 and 2016 however, total coral cover decreased by 62.7% (Table 2) and there was a significant difference of mean between these years (95% CI [−0.545–−0.070]). When differences in percent cover were analyzed by species, M. capitata and P. compressa, were not statistically significant between survey years due to high variability indicated by the coefficient of variation (Table 1).

Relocation site

The invasive orange key-hole sponge, Mycale grandis doubled in percent cover from initial placement (1.0%) to the 2012 resurvey (2.3%).

Fish abundance increased 4,700%, mean fish length increased by over 400%, and fish biomass increased over 436,000% since the original survey in 2005. Average fish length more than tripled from 2005 to 2008 and steadily increased until 2016. Number of species has also increased with the largest increase (sevenfold) observed between 2005 and 2008. The number of species has nearly doubled since 2008. While the total number of species has increased from 2008 to 2016, the percentage of endemics in 2016 (19%) decreased since 2008 (29%). No introduced species were recorded in 2016 although in 2012 they made up 9% of the total. Herbivores absent in 2005 prior to the coral relocation currently (2016) make up over 80% of total fish at the relocation site (Table 3).

Donor site

Erftemeijer et al. (2012) reviewed 35 documented case studies of dredged and areas near dredged coral reefs from around the world. While they acknowledged that these cases represent only a fraction of coastal development projects that affect coral reefs, the evidence shows that dredging causes turbidity and sedimentation that is very likely to result in coral cover decline (Erftemeijer et al., 2012). Large areas of dredged patch reefs may not recover because sand and fine sediment accumulation blocks coral settlement (Uchino, 2004). However, these sandy reef flats are capable of supporting a functional coral reef community if seeded with large relocated corals.

Coral transplantation has limitations and can have negative effects on donor and transplanted colonies, i.e., reduced fecundity (Szmant-Froelich, 1985; Edwards & Clark, 1999; Okubo, Motokawa & Omori, 2007) and low diversity at the transplant site (Horoszowski-Fridman & Rinkevich, 2015; Okubo & Onuma, 2015). Corals transplanted into new areas can carry flora and fauna that may disrupt the functioning ecosystem and may also render corals more susceptible to disease at the new location (Casey, Connolly & Ainsworth, 2015). These downfalls can be avoided by using larger, reproductively competent fragments or colonies, a large number of species with different life history strategies (Kotb, 2016; Darling et al., 2012), by removing any attached species that are not present at the relocation site (as in the present study), and by choosing transplantation sites that have similar environmental conditions as the donor sites (Muñoz Chagín, 1997). Lowered fecundity at restoration sites is less of a concern when relocating entire colonies; however, in these cases, donor sites are not expected to recover rapidly.

The rapid increase in coral cover at the donor site in the HIMB channel, in fact, was not anticipated. Unlike the findings by Uchino (2004), removal of the large carbonate structures and subsequent conditions in the dredged channel did not prevent recovery. Instead, rapid growth of the remaining fragments occurred (Fig. 2). Richmond & Hunter (1990) determined that asexual reproduction in habitats with unconsolidated substrata is highly important and Kolinski (2004) showed that asexual reproduction through fragmentation is the primary method of propagation by the reef coral Montipora capitata. The concept of rapidly increasing coral cover through spreading of fragments in Kāne‘ohe Bay (Maragos, 1974) also provides explanation for the increase observed during this study. Under the conditions of high water flushing and high irradiance found in the channel the corals would be expected to increase in radius by 1 cm to 3 cm per yr. (Jokiel & Tyler, 1993). A fragment with a radius of 2 cm in 2005 (area = 7 cm²) growing at a rate of 1.5 cm yr⁻¹ would have increased to a radius of 6.5 cm (133 cm²) in 2008 for a 1900% increase in area. Our results show an overall mean increase in total coral cover of >500% between 2005 and 2008.

Since 2005, this dredged channel bottom at the donor site has served as a primary collection site for corals to be used in experiments at the marine laboratory. This was considered to be an excellent way of conserving the coral resources while keeping the channel clear. However, the amount of material removed for research purposes (1–2 m²) has turned out to be trivial in comparison to the documented population increase and therefore did not significantly reduce further accretion of corals in the channel.

The overall decrease (63%) in coral cover between 2012 and 2016 can be attributed to widespread bleaching events that occurred in 2014 and 2015. No bleaching was noted between 2005 and 2012, however, in 2014 an average of 45% of corals bay wide were bleached but mortality was very low (<1%). A different pattern emerged in 2015; bleached corals in the south bay, where the donor and relocation sites are located, reached 61% with significantly higher mortality (33%; Bahr, Rodgers & Jokiel, 2015). The northern and central portions of Kāne‘ohe Bay experienced high recovery after bleaching in 2014 and 2015, potentially due to their stronger connections with the open ocean but the south bay experiences much higher water residence times (1–2 months) than the northern (∼1 day) and central zones (∼6 days; Lowe et al., 2009). Inadequate nutrient exchange and transport of waste during these events likely exacerbated the bleaching situation for corals in the south bay where the donor site is located.

Relocation site

Ecological restoration aims to return damaged ecosystems to a healthy status. A healthy ecosystem is generally defined as one that is able to maintain structure and function while remaining resilient to stressful conditions that may occur over time (Costanza & Mageau, 1999). Healthy, resilient coral reefs are defined by a number of important ecological factors that vary greatly by location but are generally defined as having high coral diversity, habitat complexity, herbivore biomass, recruitment, tidal mixing and low nutrients and sedimentation (McClanahan et al., 2012). After the dredging of the relocation site in the late 1930s, habitat complexity was greatly diminished and the reef remained depauperate in benthic and fish species for 60 years. Following the relocation of the large coral heads, spatial complexity was greatly enhanced (Fig. 3) and the increased rugosity immediately led to an increase in numbers of fish and fish species. Fish number, biomass, and diversity are highly associated with spatial relief (Friedlander & Parrish, 1998). Large areas of barren carbonate rock on the sides of the transplanted coral colonies have been overgrown by coral tissue, further increasing bottom complexity. Increased substrate provides habitat for benthic invertebrates, which serve as the main diet of many species of fishes, which in turn are utilized at other trophic levels to further increase diversity. Habitat heterogeneity provides refuge for fishes from predation and competition and expands the availability of resources and their rate of production. Coral cover is also associated with obligate corallivores, so higher coral coverage increases corallivore populations.

The effects of successful coral transplantation have also been shown to have a positive impact on fish populations in Sulawesi, Indonesia (Newman & Chuan, 1994) and the Philippines (Shaish et al., 2010; Gomez et al., 2014) and in Tanzania (Mbije, Spanier & Rinkevich, 2013) where coral-associated invertebrates have also increased. However, fish populations are highly variable due to mobility, cryptic ability of some fishes, and diurnal shifts. A large number of transects would have to be conducted to quantify actual populations. Conversely, changes in relative populations between years can be highly evident with few transects when there are massive differences in number and biomass as in this study.

Observations of relocated corals show fragments and branches that fell on hard substratum attached to the bottom and larger colonies placed on sand were able to avoid smothering by sediment or abrasion by sand scour. Rapid coral growth is evident by overgrowth of wires used to secure tags to the corals (Fig. 4A) and large branched corals placed on their sides that changed direction of growth and formed new vertical branches (Fig. 4B). These observations are consistent with the results from other transplantation projects in Kāne‘ohe Bay (Maragos, 1974; Kolinski & Jokiel, 1996; Jokiel & Brown, 1998; Jokiel et al., 1999; Jokiel & Naughton, 2001), Mexico (Tortolero-Langarica, Cupul-Magaña & Rodríguez-Troncoso, 2014), and the Red Sea (Kotb, 2016) demonstrating transplanted corals can grow and spread across the substratum into large thickets over a period of only a few years.

Rapid reef recovery has been well documented throughout the Hawaiian Islands: in Kāne‘ohe Bay in response to sewage outfall removal (Banner, 1974; Evans, Maragos & Holthus, 1986), at Kaho‘olawe as a result of ungulate removal and soil stabilization (Jokiel, Cox & Crosby, 1993), on the island of Kaua‘i in response to high wave energy from Hurricane Iwa and Iniki (Jokiel, 2008), and on the Hamakua coast of the island of Hawai‘i that received insult from the sugar industry’s discharge of bagasse and sediment (Grigg, 1985). However, this successful relocation project is an exception when compared to most historical accounts of coral relocation in the main Hawaiian Islands and the larger Pacific region (Jokiel et al., 2006). Corals relocated nearby into favorable wave-protected areas with similar depth and environmental factors experience low mortality. Marginal habitats, however, that cannot support high coral cover will eventually suffer high initial mortality or slow decline due to wave damage, eutrophication, sedimentation or predation (Naughton & Jokiel, 2001; Jokiel, 2008). Crown of Thorns sea star predation severely reduced relocated coral colonies in Piti Bay, Guam (Kenchington & Kelleher, 1992). Stochastic wave events can abrade, displace, bury, or break corals as demonstrated at Kawaihae Harbor, Hawai‘i (Cox & Jokiel, 1995) and in Mexico (Tortolero-Langarica, Cupul-Magaña & Rodríguez-Troncoso, 2014). In other areas eutrophication can increase algal competition (Sparks, Stone & White, 2010; Banner, 1974) and sedimentation can prevent settlement, smother colonies, and block light (Jokiel et al., 2008; Erftemeijer et al., 2012). In these cases, the emphasis should be on reducing or eliminating the impact to allow natural recovery, but the major threat to relocated corals as well as to established coral reefs is from global impacts, mainly ocean temperature increase (Bahr, Rodgers & Jokiel, 2015; Ainsworth et al., 2016).

While scientists worldwide are attempting to restore currently degraded reefs a new global climate change reality is starting to become the norm. Seawater temperatures are increasing, ENSO events are having devastating effects, and degraded coral reefs are increasing in spatial extent. Coral restoration projects are better designed and implemented today than decades ago, but they may take the focus off the underlying problems. We need to reduce pollution, prevent erosion, and reduce carbon emissions. Restoration efforts on reefs vulnerable to poor land management, pollution, and/or continued severe bleaching may render restoration efforts futile. Effective translocation and management plans should include reduction or elimination of watershed stressors, establishment of marine reserves, development of integrated coastal management systems, and establishment and enforcement of regulations that protect coral reefs.