Gardening Pocillopora spp. fragments and their potential for rebuilding reef systems in the southern Mexican Pacific

The degradation of coral reefs during the last decades has turned attention toward management and restoration interventions. This study seeks to operationalize coral gardening of Pocillopora spp. fragments in low profile bottom‐anchored nurseries and to compare survival and growth patterns between sites, time, and fragment size in the southern Mexican Pacific. After 357 days, fragments showed high survival (96.4%), growth rates (4.35 cm/year), and an increase in coral coverage from 3.62 ± 1.3% (mean ± SD) to 17.42 ± 4.8% (approximately 480%). Coral growth rate did not show differences between sites (pseudo‐F[1,635] = 0.21, p = 0.63), but corals grew more significantly during the upwelling season in the Gulf of Tehuantepec. According to linear and local regression analysis, the extension rate in nurseries was significantly higher when corals were smaller (<7 cm in diameter) being of relevance for operationalizing coral gardening from donor colonies; nevertheless, coral shrinkage (13.1%), when corals were smaller (i.e. during the early phases of the coral gardening protocol), calls for precaution and close monitoring. Operationalizing coral gardening of Pocillopora spp. fragments, including successful metrics in upwelling areas, are relevant for reef restoration in the eastern tropical Pacific; nevertheless, lessons regarding shrinkage and differential growth rates related to coral size should also be considered.


Introduction
Anthropogenic activities (i.e.fishing, pollution, mechanical habitat destruction, species introductions, and climate change) in synergy with natural hydrometeorological events (i.e.storms and hurricanes) have caused an alarming degradation of coral reefs at an annual rate of 0.5-2% globally (Bruno & Selig 2007;De'ath et al. 2012).Reef deterioration registered during the last decades has turned attention toward management and coral reef restoration interventions (Epstein et al. 2003;Bongiorni et al. 2011;Taira et al. 2016).In general, assistance for the natural recovery of coral reefs can include passive restoration measures (i.e.establishment of marine protected areas or controlling coral predators), active restoration interventions such as in situ coral cultivation and transplantation, or both (Ammar 2009).
In the past 20 years, a great variety of active restoration techniques and approaches have emerged, from direct coral transplantation onto the natural substrate to coral aquaculture, being those based on asexual fragmentation the most widely employed (Bayraktarov et al. 2020;Boström-Einarsson et al. 2020).However, according to recent reviews, coral restoration projects have been mainly conducted in the Indo-Pacific and the Caribbean, particularly in the United States (Florida, Hawaii), Philippines, Indonesia, and Thailand (Boström-Einarsson et al. 2020), yet scarcely in the eastern tropical Pacific (ETP) where direct transplantation, coral gardening, and micro-fragmentation represents the handful of approaches conducted in the Pocilloporadominated ETP reefs (Bayraktarov et al. 2020;Tortolero-Langarica et al. 2020;Combillet et al. 2022).
The ETP represents a suboptimal zone for reef development due to its narrow continental shelf, high productivity, relatively cold waters, and seasonal upwelling (Glynn & Ault 2000).Within the ETP, the biological corridor Puerto Ángel-Bahías de Huatulco, located in the southern Mexican Pacific, is not only considered a vital coral reef area in terms of percent coral cover and framework development (Glynn & Leyte-Morales 1997) but also of pivotal relevance to maintain coral connectivity across the ETP (Lequeux et al. 2018).Nevertheless, human development in coastal areas of Huatulco resulted in severe coral mortality mainly due to sedimentation (Glynn & Leyte-Morales 1997;L opez-Pérez et al. 2002).This condition is aggravated by coral bleaching caused by warm (El Niño) and cold (La Niña) phases of El Niño-Southern Oscillation (ENSO) (Glynn & Leyte-Morales 1997;Reyes-Bonilla et al. 2002;L opez-Pérez et al. 2016), as well as meteorological events such as hurricanes and storms (Lirman et al. 2001) that caused a significant reduction in coral cover, resulting in the partial or total disappearance of reef patches in the whole area without signs of natural recovery.
The relevance of the coral reef area of Huatulco at a regional scale and its current conservation status motivate the need for exploring coral restoration as an active management program to assist in their recovery.Because of that, the aim of the current work is threefold (1) operationalizing the first step of the "gardening of bare reef area" protocol based on the direct transplantation of fragments of opportunity of Pocillopora spp., (2) not only monitoring survivorship and growth as a measure of gardening success, but also explore coral shrinkage of Pocillopora spp. as a proxy of stress that lead to improving measures of coral gardening, and (3) evaluate variations in coral growth across space, time, and individuals.Because the extent of the Huatulco reef tract is under the influence of the seasonal upwelling that experiences the Gulf of Tehuantepec (Trasviña et al. 1995), we expect that survivorship and coral growth are similar between locations but experiences decrements across time-related thermal stress during upwelling as observed in Central America (Combillet et al. 2022).

Study Area
The experiments were carried out in two sites in the Huatulco area, south of the Mexican Pacific.La Entrega (15 44 0 42.79 00 N, 96 7 0 41.78 00 W) and Punta Chahue (15 45 0 8.48 00 N, 96 7 0 23.44 00 W) are located on the western margin of the Gulf of Tehuantepec, southeast of the Huatulco National Park (Fig. 1).The area experiences intense periods of upwelling (October-April) as a result of seasonal winds called Tehuanos (Chapa-Balcorta et al. 2015).Under the influence of Tehuanos, sea surface temperature (SST) can drop 10 C from the annual average (28 C) at the center of the gulf (Trasviña et al. 1995), while productivity increases considerably (>0.3 mg/m 3 ) as a product of upwelling mainly from February to April (Lluch-Cota et al. 1997).
Natural fragmentation of Pocillopora colonies is commonly related to bioerosion and predation in Huatulco reefs, with mean densities of loose fragments ranging from 1.3 to 6.7, in La Entrega, where mechanical damage by tourism takes place up to 13.1/m2 can be reached (L opez-Pérez et al. 2007), but can exceed 26.2/m 2 following swells and intense hydrometeorological events (Lirman et al. 2001).Consequently, numerous Pocillopora fragments of opportunity are potentially available for nursery without affecting wild donor colonies.

Nursery Design and Installation
Because reefs located in Huatulco are seasonally exposed to high wave impact (Lirman et al. 2001), we avoid operating mid-water nurseries in favor of a bottom-anchored nursery system.Twelve low-profile nurseries were constructed of 1 Â 1 m polyvinyl chloride (PVC) frames with high-density polyethylene plastic mesh (18 Â 18 mm mesh size).These were fixed to the substrate using rebar (Fig. S1).In order to avoid the burial or the mechanical abrasion caused by other fragments, the nurseries were elevated 25 cm above the substrate.The nurseries (six in each locality) were haphazardly distributed in the reef area at depths between 5 and 8 m to avoid wave action at shallower depths and maximize coral growth (Medellín-Maldonado et al. 2016).
Fragments of the genus Pocillopora, ranging in size from 5 to 60 cm 2 in planar area and in diameter from 2.5 to 11.27 cm, were selected without evidence of bleaching, dead tissue, or sponge coverage.Currently, the Pocillopora genus not only displays discrepancies between the macro-morphology of the colonies and molecular identification (Johnston et al. 2018) but exhibits a high level of phenotypic plasticity promoted by different environmental conditions (Paz-García et al. 2015).In addition, the growth rate among common-most morpho-species is quite similar in the Huatulco area, ranging from 2.94 cm/year (AE 0.32) for P. damicornis to 3.42 cm/year (AE 0.32) for P. verrucosa (3.42 AE 0.32 cm/year) (Medellín-Maldonado et al. 2016).Interpreting morphological variation and the identifying species is challenging; consequently, fragments of opportunity were merged under the genus Pocillopora.

Data Collection and Analysis
Each nursery had a variable number (17-27) of Pocillopora spp.fragments haphazardly attached using nylon strings.A total of 255 fragments were measured five times between January 2016 and January 2017 (January [T0], March, April, September, and January 2017).Digital images of the fragments were used to evaluate survivorship (%) and coral growth.Aerial photographs from the same angle and the same calibrated scale were taken from a distance of 1.5 m from the nursery using GoPro sports cameras (Hero 4 and SJCAM5000).Images were analyzed using the software ImageJ (v.1.8.0).
Based on the nursery photographs, planar fragment area (cm 2 ) and "size" were assessed.The size was considered as the average perpendicular diameter of the fragment to compensate for any deviation of shape (elliptical rather than circular).Based on the size, fragment "growth" (cm/year) was calculated for T1 through T4, as the difference in size divided by the elapsed time (lapse 1, January-March 2016 = 56 days; lapse 2, March-April 2016 = 37 days; lapse 3, April-September 2016 = 149 days; lapse 4, September 2016-January 2017 = 115 days).In order to remove size-related bias, the growth rate was divided by fragment diameter.
Only coral fragments that survived the entire year (until January 2017) and did not suffer breakage were included for estimating fragment growth rates.Differences in coral coverage and growth rate between sites and times were evaluated by a two-way crosspermutational-based variance analysis (PERMANOVA) using Euclidean distance (Anderson et al. 2008).
We evaluate the relationship between coral growth rate and fragment size through linear regression.Because we detected a significant inverse relationship between growth and fragment size, we employed a nonparametric local regression analysis (loess regression) to fit a smooth curve to a scatterplot addressing the relationship between growth rate and coral size in order to identify approximate sizes at which, meaningful changes in growth rate occurred.Fitting curves were explored interactively (smoothing factor: 0.2-0.7),searching for a compromise between noise (over-fitting) and lack of fit.At the same time, a confidence interval (95%) was generated by bootstrapping the observed values (Cleveland & Devlin 1988).
PERMANOVA was carried out using PRIMER 6 & PER-MANOVA+, regression with Statistica (V.10), and local regression analysis was performed in PAST 4.08.

Results
By the end of the experiment, 94.7-97.1% of the corals survived in the nurseries at La Entrega and Punta Chahue, respectively.The average planar area of the initial Pocillopora spp.fragments were 23.53 AE 12.65 cm 2 (mean AE standard deviation [SD]) for both sites, which increased to 113.01 AE 40.80 cm 2 after 1 year, corresponding to a 480% increase in area (Fig. 2).PERMANOVA analysis showed no significant differences in coral coverage between La Entrega (7.8 AE 5.4%) and Punta Chahue (8.6 AE 6.1%; pseudo-F [1,50] = 1.11, p = 0.29) but significant differences occurred across times (pseudo-F [4,50] = 50.8,p = 0.0001).Temporal differences in coral coverage occurred between September 2016 and 4 January 2017 against January, March, and April 2016.No interaction between space and time was detected (F [3,50] = 0.58, p = 0.6).
Coral fragments grew at 3.91 AE 1.69 cm/year across both sites.According to the PERMANOVA analysis, no significant difference exists between La Entrega (3.81 AE 1.72 cm/year) and Punta Chahue (4.02 AE 1.67 cm/year; pseudo-F [1,635] = 0.21, p = 0.63), but significant differences occurred across times (pseudo-F [3,635] = 4.73, p = 0.002).Higher growth rates occurred during the first two time-lapse periods (January-March 2016, 3.87 AE 0.77 cm/year; March-April 2016, 6.48 AE 1.34 cm/year), after which the growth rates decreased, reaching lowest values during April-September 2016 (2.19 AE 0.43 cm/year).No interaction between space and time was detected (F [3,635] = 1.64, p = 0.18).The reported extension values represent horizontal growth rates as growth was based on the nursery photographs.Notwithstanding, unpublished data (Francisco Medellin, Universidad Nacional Aut onoma de Mexico) indicate that in the study area Pocillopora spp.colonies grew 11.48% larger on the vertical than on the horizontal axis.Therefore, Pocillopora fragments may have grown at an average rate of 4.35 cm/year in the nurseries.
Regression analysis indicates an inverse but significant change in growth rate with fragment size (r 2 = 0.24, p < 0.001, n = 643; y = 1.1314-0.0751x).Nevertheless, nonlinear local regression analysis indicates that the coral growth rate is relatively high when fragments are small (approximately 2.5 cm) but rapidly declines for coral fragments between 2.5 and 7 cm in diameter.When coral colonies reach 7 cm in diameter, their growth rates stabilize and remain around 3.99 AE 1.6 cm/ year (Fig. 3).

Discussion
Our cultured Pocillopora spp.fragments in low-profile nurseries resulted in high survivorship (>96%), growth (4.35 cm/year), and an increase in coral coverage (480%).These metrics are currently employed to evaluate restoration success It has been the nursery phase which has received the most attention from reef restoration experts, businesses, and civil organizations as numerous materials, shapes, sizes, and locations in the water column have been explored (Shaish et al. 2008;Bayraktarov et al. 2020;Boström-Einarsson et al. 2020).Clear survival patterns exist across nursery designs and locations (deeper > shallower) in the water column.However, all survival patterns are related to those designs that isolate or move away from stressful conditions or areas with a more favorable environment.For instance, Ruiz-Diaz et al. ( 2022) observed higher survival in deeper farms compared to mid and shallower farms arguing that corals in deeper zones are better protected from damaging water temperature and light-intensity levels, while Goergen et al. (2018) demonstrate that raising corals on floating structures is most successful when colonies are suspended away from the direct contact with fouling organisms.Similarly, as well as other bottom-anchored coral nursery designs (i.e.cinder or cement blocks, steel, PVC tables, etc.), ours was designed not only to avoid the mid-water stress caused by the wave impact and periodic strong swells that experience the Huatulco area (Lirman et al. 2001) but also to partially isolate coral fragments from marine substrate reducing burial and abrasion signaled as the primary source of mortality in restoration experiences (Luthfi et al. 2015;Ng et al. 2017;Ruiz-Diaz et al. 2022), and from biotic interactions (i.e.competition and predation) that potentially stresses coral growth on the restoration sites.
While the issue regarding the performance of a particular nursery design over others remains unsettled and future works via meta-analysis may disentangle the question at a large scale, the quest for searching designs that outperform others is only of relevance when results from comparison against onsite wild metrics.In this regard, our current gardening design may be considered adequate for the Huatulco area as it resulted in 53.9% larger growth rate values when compared with the onsite Pocillopora extension (2.02-3.42cm/year) (Medellín-Maldonado et al. 2016).Similarly, average nursery survival was 278% larger when compared with the survival of wild fragments (34.5%) in the area (L opez-Pérez et al. 2007).
According to the PERMANOVA analysis, the highest coral extension rates occurred from January to April 2016, but lower rates from May 2016 to January 2017.From January to April, strong upwelling occurs in the Gulf of Tehuantepec, during which SST is comparatively lower (approximately 1.8 C), and waters are more productive (approximately 4.9 g/m 3 ) when compared to the non-upwelling season (Trasviña et al. 1995;Lluch-Cota et al. 1997;Chapa-Balcorta et al. 2015).Our results opposed those observed by Combillet et al. (2022) while gardening Pocillopora spp. in an upwelling area of northern Costa Rica; according to their results, coral fragments grew faster during the non-upwelling season.However, during January 2016, SST was 1.4 C above historic temperature in the study area and remained hotter thorough the year associated with El Niño 2016-2017 (Khan et al. 2021).It may be argued that an anomalously hot period could reduce the upwelling cold water stress for corals and allow them to grow faster; in contrast, the hottest SST records could hamper coral growth in summer.Such STT pattern may produce artificially opposite results to that observed in Central America.While the whole argument is appealing to explain the observed growth pattern in Huatulco, tracking growth patterns for more extended periods is recommended to parse out confounding factors.The relationship between upwelling and coral growth can be complicated, particularly for field studies in which environmental effects that covary with upwelling (temperature, productivity, light, carbon intake, etc.) cannot be isolated.Because upwelling is a strong seasonal environmental driver of coral development in the ETP (Glynn et al. 2017), laboratory and field studies should carefully incorporate this to disentangle the observed patterns.
According to linear and local regression analysis, the extension rate was significantly higher when corals were smaller (i.e.following deployment to nurseries).However, the extension rate rapidly declined and remained steady once coral colonies reached ≥7 cm in diameter.Experiments searching for growing patterns among coral fragments of different sizes in Pocillopora had mixed results.Nonsignificant differences in the growth rate of Pocillopora spp.were detected between class sizes (2, 5, and 8 cm) for nursery-reared fragments on steel structures by Combillet et al. (2022) in northern Costa Rica.Lizcano-Sandoval et al. (2018), on their part, found that medium (4 cm) and large (7 cm) fragments of Pocillopora damicornis grew faster than small ones (1-2 cm) when transplanted to back-reef, reef-flat, and reef-crest of La Azufrada reef (Gorgona Island, Colombian Pacific).Curiously, Ishida-Castañeda et al. ( 2020) observed that among the nursery-reared fragments, the largest (4 to <8 cm) had the highest growth rates, while the tiniest fragments (<2 cm) had the highest among the directly transplanted fragments.Our results and those recorded by Ishida-Castañeda et al. (2020) for directly transplanted fragments agree with those obtained via the outplanting of coral micro-fragments (Forsman et al. 2006;Page et al. 2018;Tortolero-Langarica et al. 2020).As data suggest, the fastest growth rates in nurseries are attained by smaller fragments in the same way that micro-fragments grow faster than large fragments or whole colonies (Forsman et al. 2006;Page et al. 2018).While our results support that the deployment of fragments when colonies less than 7 cm in diameter promotes faster growth rates, paradoxically, mortality is commonly higher at small coral sizes, potentially compromising restoration efforts based on small fragments of Pocillopora (Lizcano-Sandoval et al. 2018;Ishida-Castañeda et al. 2020).
Although the extension rate was significantly higher when corals were smaller and hence supporting the deployment of fragments when colonies less than 7 cm (see previous paragraph), we detected that 13.10% of the colonies experienced shrinkage at the beginning of the experiment when coral fragments were small (January-March 2016 and March-April 2016 in both experimental sites) yet diminished and became rare when corals grew larger.Coral shrinkage is rarely reported in coral growth experiments, but current results suggest it may represent an important factor in restoration activities.Coral shrinkage relates to partial mortality and biotic interactions irrespective of coral size as observed during hydrometeorological events (Lirman et al. 2001;Mercado-Molina et al. 2018) or predation (Colgan 1987), but also resulting from coral stress and damage during manipulation (Edwards & Gomez 2007;Tortolero-Langarica et al. 2014;Ng et al. 2017) or related to vulnerability at small coral sizes (Penin et al. 2010;Ishida-Castañeda et al. 2020).As coral shrinkage is not currently reported and small-fragment size co-occurred at the beginning of the experiment, it is complicated to resolve whether shrinkage is related to coral manipulation or size.Therefore, we strongly suggest that future restoration studies report coral shrinkage concurrently with fragment size as this can be a better metric to address adequate fragment size for coral restoration.
Finally, our cultured Pocillopora spp.fragments in lowprofile nurseries resulted in high survivorship, growth, and an increase in coral coverage, suggesting a relatively adequate operationalizing of the first step of the "gardening of bare reef area" protocol in the Huatulco area.Remarkably, a higher growth rate during the upwelling season calls for a more thoughtful search to disentangle proper drivers of coral growth in upwelling areas of the ETP.In the same way, higher coral growth when fragments are smaller (<7 cm) might be of relevance for operationalizing coral gardening, especially in those cases when donor organisms represent the first step of gardening (see Combillet et al. 2022); nevertheless, coral shrinkage must be considered as a relevant variable as it may jeopardize the first steps of gardening or when nursery-reared fragments are small.

Figure 3 .
Figure 3. Loess regression.Standardized growth rate versus fragment size.Red = expected growth based on model, blue = 95% confidence interval.Blue band = higher growth rate, and green band = steady growth rate.