Discovering the mesophotic fauna of Malpelo Ridge: biodiversity insights in a remote marine sanctuary

Malpelo Island is the emergent summit of the Malpelo Ridge, a NE-SW-trending oceanic volcanic ridge approximately 330 km long and 100 km wide, associated with the Cocos-Nazca volcanic province (Marcaillou et al., 2006). Numerous submerged rocky pinnacles, ridges and cones characterize the ridge. These geological features, together with the island's geographic isolation and local oceanographic conditions, provide habitat for diverse benthic communities and contribute to the development of distinct biological assemblages within the Eastern Tropical Pacific Ocean (ETPO), a region recognized for its high biodiversity, including endemic, resident, and migratory species (Enright et al., 2021). As a result, several world-renowned marine protected areas (MPAs) have been established in the region, including Malpelo Flora and Fauna Sanctuary (FFS), which has been designated as a World Heritage Site (UNESCO, 2026), and together with the Galapagos, as Particularly Sensitive Sea Areas (PSSAs) (IMO, 2026). However, most available information on marine biodiversity in the region is limited to shallow ecosystems within the photic zone, with comparatively little data on mesophotic and deeper habitats.Mesophotic Coral Ecosystems (MCEs) are typically defined as occurring at depths between 30 and 150 m depth and host diverse biological communities, often with high levels of endemism (Sinniger et al., 2016). These environments are characterized by low light availability, lower temperatures, reduced dissolved oxygen concentrations, and increased hydrostatic pressure, favoring specialized biological communities (Slattery et al., 2011;Baker et al., 2016). Although advances in underwater technologies, including remotely operated vehicles (ROVs), autonomous platforms, and mixed-gas diving, have improved access to mesophotic and deep-sea environments (Clark et al., 2006;Roberts et al., 2009), these habitats in the ETPO remain poorly documented due to logistical and technical constrains (Baker et al., 2016). In the Colombian Pacific, where deep-sea environments dominate much of the seafloor, biological and habitat mapping efforts remain limited, particularly around remote oceanic ridges and submarine elevations. Although exploratory initiatives have increased during the last decade (Cedeño-Posso et al. 2022b;Bessudo et al., 2022;Dueñas et al., 2024;Rodriguez-Toscano et al., 2025;Rodriguez-Toscano et al., 2026), many submarine features remain insufficiently characterized.Although Malpelo is widely recognized as an important aggregation site for pelagic species, including fishes, sharks, marine mammals, sea birds and sea turtles (Palacios et al., 2012), information on its mesophotic ecosystems (30-150 m depth) remain scarce. A review of datasets available through the Marine Biodiversity Information System of Colombia (SIBM), OBIS and GBIF revealed only 75 historical records for these depth ranges, primarily concentrated between 40-50 m and dating back to the early 2000s. These records include four phyla: Cnidaria (80%), Mollusca (9.3%), Chordata (8%), and Echinodermata (2.7%), indicating limited taxonomic representation and spatial coverage (NMNH, 2001;Baranova et al., 2009;NIWA, 2016;Montoya-Cadavid and Bohórquez, 2021;Bermúdez and Chaverra, 2024;Catania and Fong, 2024; and 59 GBIF datasets accessed via https://doi.org/10.15468/dl.df3kq5).In this context, the dataset presented here provides a standardized compilation of occurrence records and associated metadata describing the diversity and distribution of mesophotic species inhabiting submarine elevations and deep terraces within the Malpelo FFS. The dataset integrates imaged-based observations obtained through remotely operated vehicle (ROV) surveys, scientific diving, and biological samples collected during the 2025 Mesophotic Scientific Expedition to the Malpelo FSS, and follows international biodiversity data standards to ensure interoperability and reuse (Wieczorek et al., 2012;Appeltans et al., 2026). It provides baseline information on mesophotic biodiversity in the Colombian Pacific and supports future research, monitoring and ecosystem-based management, in MPAs.High-resolution multibeam bathymetric data were acquired from 0-200 m depth during the first phase of the expedition (2-6 March 2025), prior to the ROV operations. Preliminary onboard processing and geomorphological interpretation of the bathymetric data enabled the identification submarine elevations, rocky terraces, and other seabed anomalies potentially associated with mesophotic habitats. In total, 34 bathymetric anomalies were identified, of which 20 were selected and priorized as target stations for ROV surveys and biological characterization.ROV surveys. From 8 to 12 March 2025, video transects were conducted using the remotely operated vehicle (R.O.V) FOIII Eloy V, owned and operated by INVEMAR. The ROV is equipped with three thrusters, a forward-facing high-definition digital video camera (Full HD/4K), high-power LED lights mounted on a tilting bar, and paired parallel lasers spaced 10 cm apart for scale reference. Video recordings were obtained under two configurations: (1) Full HD for high-resolution analysis and (2) low-resolution video with on-screen display (OSD), including compass heading, vehicle azimuth, and depth. Linear video transects were conducted as the research vessel drifted over the target sites, covering a depth range of 30 to 161 m. A total of 18 ROV video transects were completed, yielding approximately 6.3 h of video footage (Fig. 1). ROV positions were estimated indirectly from geographic position of the research vessel, combined with ROV depth and azimuth data recorded in the OSD system. Consequently, spatial coordinates associated with image-based records represent approximate positions along the transect and may include positional uncertainty related to vessel drift, tether angle and current conditions. Scientific diving surveys and specimen collection. Two submarine hills (sites B2 and B3) were accessible by autonomous diving (Fig. 1). Scientific diving surveys were conducted on 8 March 2025 at site B3 and on 13 March 2025 at site B2. At each site, two scientific divers performed a single scuba dive, covering two distinct depth strata 37.5 and 40 m depth. During these dives, photographs and video recordings were obtained, and physical samples were collected, including rocky substrate fragments and cnidarian colonies harboring associated invertebrates. A total of 125 invertebrate specimens were separated, including echinoderms (sea urchins and ophiuroids), mollusks, crustaceans annelids, bryozoans, and other macrofaunal groups. Collected specimens were fixed in 96% ethanol. All specimens were identified to the lowest possible taxonomic level using standard taxonomic approaches and deposited in the reference collections of the Museum of Marine Natural History of Colombia (MHNMC-Makuriwa, INVEMAR), where they were catalogued and assigned unique collection numbers linked to the dataset through Darwin Core fields.Image extraction, annotation and database construction. A total of 628 Still images (JPEG format) were extracted from 6.3 hours of high definition ROV video transects at 30 s intervals using GOM Player software to reduce spatial autocorrelation within transects. Each extracted image was automatically assigned a unique filename by the software, including the transect number, date, time, observation minute. Images that were excessively dark, out of focus, or recorded at excessive camera altitude were classified as "Not Determined" (ND) and exclude from biological annotation. All extracted images were incorporated manually into a relational database, containing associated metadata, including transect ID, date, time, depth, azimuth, geographic position and macro-habitat classification. Macro-habitats were classified following the deep-sea macrohabitat scheme proposed by Cedeño-Posso et al. (2022a). Frames containing visible biota were subsequently annotated and screened for biological records. Mega and macrobiota were initially classified into broad taxonomic and morphological categories following the CATAMI classification scheme (CATAMI, 2014). Subsequently, taxonomic identifications were refined to the lowest possible level by specialized taxonomists from the Museum of Marine Natural History of Colombia (MHNMC-Makuriwa, INVEMAR), when diagnostic features were visible in the imagery. When species-level identification was not possible, records were assigned to higher taxonomic categories (family, order, or class), using standardized identification qualifiers following Sigovini et al. (2016). Taxonomic nomenclature was validated against the World Register of Marine Species (WORMS Editorial Board, 2026). Data processing, standardization, and quality control. Data documentation and preparation followed internationally recognized biodiversity data standards (Wieczorek et al., 2012;Appeltans et al., 2026). All data tables were structured according to Darwin Core (DwC) standard for species occurrence data, ensuring compatibility with the Ocean Biodiversity Information System (OBIS), the Colombian Biodiversity Information System (SiB Colombia), and the Global Biodiversity Information Facility (GBIF). The dataset was published and validated using the Integrated Publishing Toolkit (IPT) of the OBIS Colombia node. During this process, a multi-step quality control workflow was implemented, including: (i) curatorial validation, based on verification of physical specimens and associated records; (ii) taxonomic validation, ensuring consistency with current nomenclature standards; (iii) spatial validation, including verification of geographic coordinates and sampling locations; and (iv) documentation validation, ensuring completeness and consistency of associated metadata. Standardization procedures were further assessed using IPT validation tools and institutional quality indicators to ensure compliance with OBIS, SiB Colombia, and GBIF requirements. Metadata completeness and structure were validated according to the Ecological Metadata Language (EML) standard. A final validation step was conducted prior to publication to ensure dataset integrity and interoperability. Image-derived data underwent a structured processing workflow, including basic image editing (e.g., cleaning and labeling), extraction of occurrence information, and cross-validation between image records and preliminary taxonomic identifications. Image metadata was generated and documented in the MHNMC-Makuriwa digital repository (Flickr platform) and linked to the dataset through Darwin Core extensions for multimedia resources.The dataset complies standardized occurrence records derives from ROV image-based observations and specimens-based sampling conducted during the 2025 Mesophotic Scientific Expedition to Malpelo FFS. The resource contains presence-only data associated with 20 sampling stations linked through an event core structure. A total of 588 occurrence records are included, and additionally, 821 associated records describing complementary variables (e.g., organism counts, organism type, life stage, vitality, sampling equipment, and research vessel) are provided through the Extended Measurement or Facts (eMoF) extension. Supporting multimedia documentation includes 70 annotated images available through Simple Multimedia extension.Organisms recorded in the dataset belong to thirteen phyla: Cyanobacteria, Rhodophyta, Porifera, Cnidaria, Bryozoa, Annelida, Platyhelminthes, Nematoda, Nemertea, Mollusca, Arthropoda, Echinodermata, and Chordata (Fig. 2). A total of 64 benthic megafaunal taxa and 33 fish taxa and were identified from image-based analyses, encompassing multiple benthic groups with varying taxonomic resolution depending on image quality and the visibility of diagnostic features. Among primary producers, representatives of Rhodophyta (including crustose coralline algae, Corallinales), and Cyanobacteria were recorded across different depth ranges (33 and 166 m). Porifera represented one of the most conspicuous sessile groups, with 15 sponge morphotypes recognized between 37 and 161 m depth, including a morphotype assigned as aff. Neophrissospongia galapagoensis. Cnidarians were represented by 19 morphotypes belonging to Hydrozoa, Scyphozoa, and Anthozoa, including octocorals (Muricea, Leptogorgia, Pacifigorgia), antipatharians (Antipathes cf. galapagoensis), and scleractinians (Tubastraea spp.). A coronate scyphozoan was also documented. Echinoderms were represented by 31 morphotypes, including first records of Linckia columbiae and Eucidaris galapagensis. Fish assemblages were frequently associated with mesophotic rocky slopes and substrates, including large groups of olive grouper (Hyporthodus cifuentesi), misty grouper (Hyporthodus mystacinus), and yellowtail amberjack (Seriola lalandi), suggest Malpelo FFS may sustain higher fish biomass than previously estimated (Quimbayo et al., 2017), and extends well into deeper reef-associated habitats. In addition, six fish species are new records for the sanctuary, as they have not been reported in peer-reviewed publications on Malpelo ichthyofauna: Prognathodes carlhubbsi, Hyporthodus cifuentesi, Hyporthodus mystacinus, Liopropoma fasciatum, Pronotogrammus multifasciatus, and Enchelycore lichenosa, the latter also representing a new record for Colombian Pacific.Among the 125 benthic invertebrates collected through diving, multiple macrofaunal groups were represented, including Annelida (Polychaeta, Oligochaeta, and Sipuncula), Nematoda (Enoplea and Chromadorea), Nemertea (Monostilifera), Platyhelminthes (Polycladida and Tricladida), as well as Arthropoda (Decapoda, Isopoda, Amphipoda, Scalpellomorpha, Cyclopoida, Podocopida, Tanaidacea and Pantopoda), Mollusca (Bivalvia, Polyplacophora and Gastropoda) and Bryozoans. Within these groups, 54 families, 53 genera, and 56 species and morphospecies were identified. Most macrobenthic records associated with mesophotic environments in Malpelo FFS represent new records (>90%), both at family level and at species and morphospecies level.All biological specimens collected during this study were deposited in the reference collections of the MHNMC-Makuriwa, ensuring long-term preservation and accessibility. Those specimens are linked to the dataset through catalog numbers, facilitating taxonomic verification and reproducibility. The dataset constitutes a valuable resource for future research on taxonomy, systematics, biogeography, and biodiversity patterns in mesophotic ecosystems of the Colombian Pacific, a region where the topic remains poorly studied. The inclusion of multimedia records linked through Darwin Core extensions enhances its applicability for image-based ecological analyses and machine learning approaches. Additionally, the dataset contributes to improving the representation of mesophotic biodiversity in global repositories such as OBIS and GBIF.