Shallow reefs in the SeaFlower Biosphere Reserve, even at the remotest bank atolls, are showing a steady decline in coral cover overall health condition during the last 20 years (Sánchez et al., 2019b). Mesophotic Coral Ecosystems (MCEs), located between 30 and >150 m of water depth, may act as a refuge of coral populations due to favorable conditions in this less altered environment (Bongaerts et al., 2010). Particularly, in our study area, San Andres Island, populations of corals reaching the lower (>60 m) mesophotic zone, 40–90 m, such as Agaricia undata exhibit genetic connectivity throughout its depth range (Gonzalez-Zapata et al., 2018a), supporting this zone as a major reef-building coral refuge. It has been suggested that depending on the type of endosymbiont, corals can acclimatize to deeper depths (Ziegler et al., 2015), which was in fact observed in the bacterial population from A. undata in San Andrés Island (Gonzalez-Zapata et al., 2018a). In addition, many species of fish, corals, and other invertebrates from shallow reefs are also found in mesophotic reefs and it is proposed that these populations could contribute to the recovery of affected populations in shallower areas following a disturbance (Kramer et al., 2019).

There are potential new species of corals and associated species, including common shallow-water fauna, in mesophotic reefs (Luck et al., 2013; Petrescu et al., 2014), which urges studies surveying coral diversity at these depths. However, these reefs have been rarely explored below 60 m in water depth. The dataset presented in this study, provide the first exploration of the mesophotic zone (40–120 m deep) in an oceanic barrier reef complex (SeaFlower Biosphere Reserve), San Andrés Island, Southwestern Caribbean. The dataset presented here includes the collection information, and community composition of corals sensu lato (stony corals, hydrocorals, black corals, and octocorals). The ultimate goal was to contribute to the understanding of sensible and vulnerable environments, in the SeaFlower Biosphere reserve, in which San Andrés Island is immersed.

We concentrated the study in the fore-reef slope of San Andrés Island barrier reef complex near the location called “Trampa de Tortugas” or “Trampa Tortuga,” which offered a number of logistic advantages. The site bears a shelter to anchor the supporting boat despite its location on the fore-reef terrace. In addition, this site provides the only accessible glimpse of the oldest slope of the barrier-reef complex of San Andrés (Geister, 1975; Díaz et al., 1995; Diaz et al., 1996). We explored the reef using Close-Circuit Rebreather (CCR) (Megalodon, Inner Space Systems) and hypoxic trimix techniques (e.g., 11% Oxygen and 60% Helium) with complete bail-out support for each diver. At the site, we installed a mooring block at 24 m as a gas station that had high oxygen bail-outs (O₂ 96%) and from where we tied a 200 m long reel down to a depth of 114 m. The reel was used to safely explore down the site and to have an easy return to shallower waters. Seven dives were planned with a maximum of 20–30 min of bottom time and the longest dives spanned 133–328 min including decompression stops. The sampling included digital imagery (Nikon^TM D7000, Nikkor micro 60 mm lens, Sea & Sea^TM YS-D1 strobe, and Aquatica^TM AD7000 housing) and 113 voucher specimens (dry and ethanol 96%), which were deposited at Museo de Historia ANDES (Bogotá, Colombia) (Table 1).

The dataset included 113 specimens from 33 species collected below 40 m (8 black corals, 1 lace coral, 8 scleractinian corals, and 16 gorgonian corals: Table 1). Exploring “Trampa de Tortugas,” we noticed the disappearance of most reef-building corals and zooxanthellate octocorals at different depths. Some reef-building corals, notably Mycethophyllia reesi, Agaricia undata, A. fragilis, and Madracis sp., distributed well into the lower mesophotic zone (~90 m) and are characterized by an increase in the presence of the euendolithic algae Ostreobium, which is clearly observable at the colony surface (Gonzalez-Zapata et al., 2018b). These colonial corals are replaced below 80 m by azooxanthellate cup corals, including Javania cailleti, Phacelocyathus flos, Balanophyllia cyathoides, and Thalamophyllia riisei forma solida (Figure 1). A noteworthy observation was the presence of Ostreobium at the basal portion of T. riisei cup-corals.

The lower mesophotic reef is also the habitat of many unique black corals (e.g., Rhipidipathes colombiana and Tanacetipathes hirta) and hydrocorals (Stylaster duchassaingi) but the most abundant group are azoxanthellate octocorals mostly from the Plexauridae family (Sánchez, 2017; Sánchez et al., 2019a). The species replacement is enhanced by short terraces intertwined with abrupt steps at every 10 m starting at 60, 90, 100, and 115 m, at “Trampa Tortuga” reef in San Andrés Island. On the leeward side of the Island, there is parallel slope sand corridor (mostly from Halimeda copiosa flocks) following the reef slope, which end at about 60–80, where the reef growth continues. The deepest zooxanthellate gorgonian coral was Antillogorgia hystrix (65 m), followed by Eunicea pinta (55 m). Occasionally, Muricea laxa, A. bipinnata, and A. americana were seen at depths of about 40 m. These zooxanthellate octocorals share habitat with some azooxanthellate octocorals such as Iciligorgia schrammi and diverse ellisellids (Sánchez et al., 2019a).

Black corals (Antipatharia) are also common from 30 m down to the lower mesophotic area (Bo et al., 2019). The most abundant black corals are Antipathes furcata (Figure 1A), A. caribbeana, A. atlantica (Figure 1B), Plumapathes pennacea, Stichopathes lutkeni, and S. occidentalis. Wire corals, Stichopathes spp., with colonies reaching more than 2 m long were seen in high densities of to have more than 10 colonies per square meter (Figures 1C,D). Below 70 m, the aforementioned black corals are less seen and other species emerge such as Rhipidipathes colombiana, first seen off the Colombian coast (Opresko and Sánchez, 2005), and Tanacetipathes hirta. There is also a great amount of wire corals from species we could not identify and probably comprise new undescribed species. Despite the clear characters of S. lutkeni and S. occidentalis under the electron microscope, there were specimens, Stichopathes sp., with conspicuously smaller spines not found in any other species described for this region (Figure 1H).

The most unexpected finding comprised a number of new records for several deep-sea corals, which have been usually found on deeper waters. Stylaster duchassaingi Pourtalès, 1867, a hydrocoral (Stylasteridae) was observed from 80 to 115 m forming seafan colonies up to 40 cm in height (Figures 1I–K). This is the southernmost record of the species and one of the shallower observations in its range. Stylaster roseus is commonly observed in the same reef but above 40 m (JAS, personal observation). San Andrés Island is the only coral reef complex so far in the Caribbean with two documented species of Stylaster.

As expected, the exploration of the lower mesophotic zone uncovered a great amount of new species records and potential discoveries (e.g., Stichopathes sp.). In addition, this is the first time that many of the species have been ever seen and photographed in their natural environment (Sánchez et al., 2019a). Continuing research in this environment will enrich the ecology, systematics, and conservation of understudied corals such as cup corals. For instance, the Thallamophyllia riisei cup coral found in San Andrés is extremely different to the reported T. riisei from the Colombian coast (Flórez et al., 2010), which is colonial with great differences in morphological traits. It is worth mention that this is the product of only seven dives (and <140 min of total bottom time) for San Andres Island.

The specimens collected and properly curated (deposited and IN-ANDES in Bogota, Colombia) comprise a valuable resource for further systematic studies in several groups of corals, which could comprise new species. In addition, it is important to mention that the specimens in this data report have not been monitored in the past, giving the logistic constraints of deep-sea diving. As the interest in MCEs increases biodiversity data becomes crucial for comparisons.

The datasets for this study can be found at https://ipt.biodiversidad.co/cr-sib/resource.do?r=0359_mesofoticos_20190729, titled “Biodiversidad y Conectividad de los arrecifes mesofóticos (30–120 m) de la costa Caribe colombiana”. The data presented here corresponds to coral specimens (Cnidaria: Anthozoa andHydrozoa) collected between 40 and 115 m in the mesophotic corals ecosystems from San Andrés Island (SeaFlower Biosphere Reserve).

JS, LD, JA, and NB conceived the study. JS, JA, and NB collected the data. FG-Z, AS, DV, AP-V, and JS identified and processed the material. JS wrote the report with the help of LD, FG-Z, and NB. All authors read and accepted the manuscript.