Invasion and current distribution of the octocoral Carijoa riisei (Duchassaing & Michelotti, 1860) in the Ecuadorian coast (Eastern Tropical Pacific)

Carijoa riisei is a snowflake coral that has aggressively spread across many coastal habitats in the Atlantic and Pacific oceans, threatening a number of tropical ecosystems worldwide, including coral reefs. The aim of this work was to evaluate the distribution and provide an estimation of abundance of the invasive octocoral C. riisei along the Ecuadorian coast, as well as the relationship between its abundance and different environmental variables. In a field survey, high abundances of C. riisei colonies were reported growing over corals and sessile communities at 5 of 43 sampling sites. The areas with highest relative abundance were found in the Manabí province, at two sites in Jama: Bajo Londres (44.57% coverage) and Vaca Brava 1 (20.25%). Results of ordination and grouping statistical analyses showed no significant differences between invaded and not invaded sites as regards community composition or environmental characteristics, suggesting neither biotic nor abiotic factors could be limiting C. riisei dispersal along the Ecuadorian coast. Results from a bibliographic survey covering occurrence data up to 2020 were in accordance, showing that in that period C. riisei became an established species to the Ecuadorian coast, being present in at least 22 of the 43 sites, including various sites in Marine Protected Areas. Based on these findings, recommendations are made to promote urgent monitoring programs to detect C. riisei in new areas along the coast of Ecuador and in the Galapagos Marine Reserve, in order to develop a mitigation program and to take actions to conserve the ecosystems affected by this invasion.


Introduction
Carijoa riisei (Duchassaing & Michelotti, 1860), commonly known as snowflake coral, is an octocoral (Octocorallia: Cnidaria) native to the Indo-Pacific, currently present in the Atlantic, Pacific and Caribbean Seas (Sánchez and Ballesteros 2014). This species was reported for the first time, as an invasive species in 1972, in Hawaii . It has been reported in areas with rich organic matter, where it inhabits rocky and artificial substrates (Sánchez 1994;Venkataraman et al. 2016). It forms dense, monospecific aggregations, carpeting hard substrata. This octocoral is an aggressive competitor that monopolizes food and space, feeding mainly on phytoplankton, diatoms and zooplankton larvae (Lira et al. 2009). It has few natural predators, therefore no species has been identified as potential for biocontrol (Wagner et al. 2009). This species is usually found to compete successfully with black corals and invertebrates, impacting coral reefs (Kahng and Grigg 2005;Barbosa et al. 2014;Sánchez and Ballesteros 2014). This process can lead to major shifts in species interactions and changes in both nutrient cycles and energy flow, which can result in unpredictable cascading effects (Carlton 2001).
Carijoa riisei was reported for the first time in Ecuador in 2010, in the Machalilla National Park. One year later it was observed in the northern province of Esmeraldas, in the Galera San Francisco Marine Reserve (Instituto Nazca 2013; Keith and Martínez 2017). In 2012, it was detected on the central coast in El Islote Los Ahorcados, Manabí (Instituto Nazca 2013) and south central coast in El Pelado Marine Reserve (Cárdenas-Calle et al. 2018a;Instituto Nazca 2013). In 2013, it was observed at a shallow water site known as El Cuarenta, known for its high diversity of benthic and fish communities (Cárdenas-Calle et al. 2018a). In a short time, this species was reported at several coastal sandy sites as well as in an open water area that is characterized by rocky reefs (Cárdenas-Calle et al. 2018b).
Ecuador is the smallest among the so-called "megadiverse" countries, harboring a wide range of ecosystems and many endemic species distributed in small geographic ranges (Cuesta et al. 2017;Mestanza-Ramón et al. 2019). The Ecuadorian coast is approximately 2,900 km in length, and has been considered a hotspot of species richness due to its location in the great Panamic-Pacific zoogeographic province (Olsson 1961). For example, in a recent survey, a total of 479 macroinvertebrate species and 133 seaweeds were recorded, only in the subtidal (Cárdenas-Calle et al. 2020). Therefore, it is imperative not only to document the extent of the distribution of C. riisei along the Ecuadorian coast, but also to analyze the environmental factors that could drive its distribution, in order to have a better understanding of the implications of this invasion and its possible limits. In previous works, C. riisei was morphologically identified in the field and in the laboratory, however, molecular identification of the species through the analysis of mitochondrial markers has been suggested as a way to strengthen its identification as well as to have some insight into its possible geographical origin (Concepción et al. 2010).
Therefore, the aims of this study were: 1) to confirm the identity of the species through molecular analysis; 2) to evaluate the present distribution and provide an estimation of abundance of C. riisei along the Ecuadorian coast; 3) to generate an updated list of sites colonized by C. riisei 4) to evaluate the relationship between different environmental variables and the presence of C. riisei.

Study area
The Ecuadorian coast is influenced by the Ecuadorian Coastal Current (ECC) (Allauca 1990). The coastal continental zone, excluding the Galapagos archipelago, is home to two sub-regions. The northern coast corresponds to the Panama Bight marine ecoregion which extends from Azuero Peninsula in Panama to Bahía de Caráquez in Ecuador, while the southern coast lies within the Guayaquil marine ecoregion that starts in Bahía de Caráquez and ends in the Peninsula de Illescas in northern Peru (Miloslavich et al. 2011;Sullivan-Sealey and Bustamante 1999).
A total of 43 sites were sampled from April 2015 to October 2015, along the four coastal provinces of Ecuador (Esmeraldas, Manabí, Santa Elena and El Oro). The study area covered 1860 km 2 and the selected sites were chosen considering the representativeness of ecosystems and degree of human intervention. The Marine Protected Areas (MPAs) were: Galeras San Francisco Marine Reserve (RMGSF), Pacoche Wildlife Refuge (RVSMCP), Machalilla National Park (PNM), Bajo Cope Marine Reserve (RMBC), El Pelado Marine Reserve (REMAPE), Puntilla de Santa Elena Fauna Production Reserve (REMACOPSE) and the Santa Clara Island Wildlife Refuge (RVSISC). In addition, three non-protected areas (Jama, Canoa and Ayampe) were sampled for this study. Further details are provided in Supplementary material Table S1.

Taxonomic identification by molecular analysis
Three samples of the octocoral morphologically identified as C. riisei were collected by divers in "La Pared" at El Pelado Marine Reserve, the colonies were collected and identified morphologically (Collin et al. 2005). Samples from C. riisei were then immersed in liquid nitrogen for rapid freezing, then crushed with the mill "Retsch MM400" to obtain a fine powder and stored in a freezer at −80 °C prior to molecular analysis (Weber et al. 2017).
DNA isolation was performed in the Molecular Biology Laboratory at CIBE-ESPOL with the CTAB-based protocol (Doyle and Doyle 1990). Two pairs of primers were used to amplify fragments of the cytochrome oxidase I (COI) gene, and one pair for NADPH dehydrogenase subunit 6 (ND6) gene, according to Concepción et al. (2008). The primers used are shown in Table S2. Endpoint PCR reactions were carried out in a Mastercycler Gradient S No. 5345 (Eppendorf), in 50 μl reactions containing 25 μl of 2x-GoTaq® Green Master Mix (Promega), 0.5 μM of each primer, 60 ng (3 μl of a 20 ng/μl solution) DNA template, and the corresponding volume of sterile water for reaching 50 μl. For PCR runs, the following conditions were followed: 3 min at 94 °C (initial denaturation); 36 cycles of [30 s at 94 °C, 30 s at 58 °C (COIF/COIR; COI2F/COI2R) or 30 s at 50 °C (ND6/ND3), 1 min at 72 °C]; and 5 min at 72 °C (final elongation). Five microliters of PCR products were loaded onto agarose gels and electrophoresed to confirm band presence and expected molecular weight, and the remaining 45 μl were purified with the Wizard SV Gel and PCR Clean-Up System kit (Promega). PCR products were stored at −20 °C until sequencing. Sequencing was performed at Macrogen Inc. (Korea) with forward and reverse primers. COI phylogenetic trees were constructed with MEGAX (Kumar et al. 2018), using Maximum Likelihood for tree construction and GTR+I+G as evolutionary model (McFadden and van Ofwegen 2012). For ND6 fragments, a haplotype analysis was carried out based on the work of Concepción et al. (2010). A total of 525 nucleotide positions were analyzed, corresponding to the amplified fragment from this work. A pairwise distance matrix was constructed using MEGAX, which was used to distinguish 100% identical variants or haplotypes and match them to the previous work. As the sequence fragments from this work were shorter, some of the haplotypes from Concepción et al. (2010) were not able to be distinguished, and they were grouped for this analysis (24 haplotypes out of 30 identified in Concepción et al. (2010), with an average of 1.6 nucleotide identity among them).

Present distribution
Data were collected in the subtidal zone using SCUBA at a depth of 5-15 m. Each sample was taken by two divers moving along a 50 m transect parallel to the coast conducting visual censuses, one on each side of the transect (Edgar et al. 2011). An area of 2.5 m 2 was monitored using a PVC quadrat of 0.5 × 0.5 m (0.25 m 2 ). Each quadrat had a grid of 5 × 5 cm, constructed with polypropylene twine with 81 intersection points. Quadrats were placed systematically every 5 m along the same transect. The abundance of mobile species and coverage of sessile species were estimated per quadrat, averaged per site, and transformed to presence/absence data at each site for further analyses (see below). The list of species used is published elsewhere (Cárdenas-Calle et al. 2020). To explain the patterns of occurrence of C. riisei over sites in relation to community composition, a Non-Metric Multidimensional Scaling (n-MDS) of all sampled sites was carried out based on presence/ absence data of all other species (excluding C. riisei). Maps were made to show the sites sampled during 2015 and the types of sediments in accordance with the subtidal system proposed by Instituto Nazca (2005). Coverage of C. riisei was superimposed as bubbles with diameters proportional to percent coverage, using the statistical package vegan and other tools in R environment (R Core Team 2020). Additionally, the databases Scopus, Web of Sciences, ScienceDirect and Google Scholar as well as technical reports were reviewed to generate an updated list of sites colonized by C. riisei along the coast of Ecuador from 2010 until 2020. The search was done in google scholar, web of science and scopus. The terms of search were "Carijoa riisei" "Ecuador" "continental". The search was done as for March 2020. In this case, presence or absence was directly used as an indicator variable of C. riisei distribution at each site.

Relationship between environmental variables and the presence of C. riisei and estimation of colonizable sites
Seventeen environmental variables were examined to determine the characteristics of the areas inhabited by C. riisei using remote sensing from MODIS-AQUA (available in oceancolor.gsfc.nasa.gov and neo.sci.gsfc. nasa.gov) (Table S3). Geomorphic depth, slope and rugosity were derived from in situ measurement and downloaded from GEBCO (https://www. gebco.net/) and analyzed using QGIS version 3.8. As variables differ temporarily, monthly values from 2003 to 2016 were subtracted. Principal Component Analysis (PCA) was used in order to visualize the relation of the presence/absence of C. riisei and the environmental variables at each site. In the analysis, both the data reported during field work and the data extracted from other studies since 2010 to 2020 were included. PCA routines were assessed using R and the basic package stats (R Core Team 2020). For visualization of the analysis the package factoextra was used (Kassambara and Mundt 2020). All variables were normalized as all of them obtained a p < 0.05 according to the Shapiro-Wilk test.
Finally, for the estimation of colonizable sites, a grouping analysis contrasting the environmental conditions resulting from the PCA was carried out, using as grouping factor Presence or Absence of C. riisei. The rationale of this analysis is that if the circles do not have intersected areas, the groups differ from each other, and there are significant differences between conditions where C. riisei is present or absent. This can lead to the conclusion that the octocoral is not likely to expand further from its present distribution. To confirm the visual results, an analysis of similarities (ANOSIM) was run to test if differences were significant.

Taxonomic identification by molecular analysis
The molecular analysis of three samples of the octocoral collected in "La Pared" (Site 30) confirmed the morphological identification. A clear grouping The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.5553)). The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 51.06% sites). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 23 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. All positions with less than 95% site coverage were eliminated. That is, fewer than 5% alignment gaps, missing data, and ambiguous bases were allowed at any position. There were a total of 450 positions in the final dataset. The outgroup used was the cytochrome oxidase I sequence from Cornularia cornucopiae. Evolutionary analyses were conducted in MEGA X.
was observed between the sequences from this work and other Carijoa sequences (Figure 1). No resolution was observed inside the cluster, in accordance with results reported for longer COI fragments as well as for other genes (mtMutS and 28S rDNA) (McFadden and Ofwegen 2012).
As regards ND6 gene, haplotype analysis showed that the sequence from this work was identical in the analyzed region to haplotypes 17, 18, 19 and 26 from those of Concepción et al. (2010) (Table S3). Of these, haplotype 18 is the most commonly found, being dominant in the Atlantic zone (Concepción et al. 2010).

Present distribution
Colonies of C. riisei were determined morphologically and registered along the Ecuadorian coast during the 2015 surveys. Dense colonies with orange  Table S1). Bubbles correspond to С. riisei coverage values, diameters are proportional to percent coverage. coloration and white polyps were found on rocky outcrops, cave exteriors, under ledges, on flat rocky areas and on walls. The colonies were observed between 5 and 15 m (average depth 9.32 ± 3.09 m, mean ± SD) whilst the bottom temperature ranged from 22 °C to 27 °C (26.23 ± 1.27 °C) (Table S1). Dense aggregations were observed on the rocky reefboth in exposed and sheltered areas, where they began to grow as mats on the rocky edges and progressively extended on to rocks and sessile organisms. Overgrowth of Carijoa riisei was observed on corals and sessile macroinvertebrates communities such as: Pocillopora spp, Pinctada mazatlanica, Hyotissa solida, Muricea plantaginea, Megabalanus (Balanus) peninsularis and Aplysina sp. However, Tubastraea coccinea did not seem to be affected by C. riisei ( Figure S1).
The highest coverage of C. riisei was in in the Province of Manabi, in the area of Jama. The site Bajo Londres had the highest, with 44.6% coverage, followed by Vaca Brava 1, with 20.2%. In contrast, low coverage of C. riisei was found at sites in two MPAs: La Pared, in El Pelado Marine Reserve (REMAPE) and Punta Alta, in Galeras San Francisco Marine Reserve (RMGSF) (Figure 2, Table S1).
Accordingly, the literature review conducted for C. riisei in Ecuador from 2010 until 2020 showed the successful establishment and expansion of this species along the coast of Ecuador. The results showed that this invasive  Table S1.
species is now present at 22 sites covering four of the five coastal provinces (Figures 3 and 4).  Table S1.

Relationship between different environmental variables and the presence of C. riisei, and estimation of colonizable sites
PCA ordination allowed the visualization of sites according to their similarity in environmental data ( Figure 5). PC1 and PC2 explained ~ 64% of the variation. Sites mostly influenced by the PC1 are positively correlated with standard error of the sea surface temperature (SE_SST), solar radiation and means photosynthetically available radiation (PAR). Contrastingly, the PC1 was negatively correlated with means of sea surface temperature (SST), precipitation (Rain) and standard error of euphotic zone (SE_EUZ). On the other hand, PC2 was mostly explained by positive correlation with the increase of backscattering (Back), organic and inorganic particulate carbon (POC and PIC). A complete list of the environmental variables is shown in Table S4.
A grouping analysis allowed to unravel areas where C. riisei had not been reported and compare to those areas where C. riisei is currently behaving invasively. The areas where C. riisei is present or absent were mixed and superimposed ( Figure 5). This implies that the environmental conditions between sites invaded and not invaded do not differ from each other. The ANOSIM revealed that there were no significant differences between groups (R = 0.031, p = 0.279).

Discussion
The results from this study showed the presence of Carijoa riisei and its rapid expansion in four coastal provinces along the Ecuadorian coast. High densities were found in turbid waters with high levels of sedimentation and organic matter similar to that recorded in the Mexican Pacific (Galván-Villa and Ríos-Jara 2018). The results of the molecular analysis confirmed the identity of the species and showed genetic similarity with haplotypes found in the Atlantic Ocean (Concepción et al. 2010). However, further studies should be carried out to confirm these results, including more individuals, longer sequences and include other more variable genes such as the nuclear SRP54 (Concepción et al. 2010).
Our study showed that, by 2020, a total of 22 sites have been invaded by this species. Therefore, C. riisei represents a great threat for the coral reefs communities mainly in the El Pelado Marine Reserve and Machalilla National Park, where it was seen to rapidly spread, suggesting its high colonization capacity. This aggressive invasive behavior was also found in other places in the Eastern Tropical Pacific (Gutiérrez 2012;Sánchez and Ballesteros 2014;Galván-Villa and Ríos-Jara 2018). Carijoa riisei has been recorded from the north coast to the south of Ecuador, suggesting that there are more colonized sites but that they have not yet been reported. For this reason, it is necessary to implement a monitoring program at the harbors and MPAs to understand the growth of colonies, estimate the growth rate, its  Table S4. B) Grouping analysis, discriminating sites where C. riisei is present (red triangles) from sites where it is absent (blue circles). Bibliographic survey data from 2010 to 2020 were used for the construction of the matrix. ecological effect on sessile organisms, to know the type of secondary metabolites that it produces to compete successfully for space and to determine strategies of mitigation that avoid the loss of coral areas and diversity of the microbenthic communities, the decrease of tourist attractions in the dive sites and of the economic sustenance that they provide to the coastal populations that live from dive tourism.
In continental Ecuador, the colonies of C. riisei tolerate a wide range of habitats and conditions: shaded ledges with direct light and high current flow, rocky walls, as well as flat surfaces. Carijoa riisei was observed growing on rocks, on substrates covered with sediment and even on other organisms such as: sponges, scleractinian corals, octocorals, bivalves and balanus. These have shown to be important substrates for colonization and could drive in a short term the native isolated coral patches on small rocky bottoms actually present in Mainland Ecuador (Cabanillas-Terán et al. 2016). Contrary to what was observed in Hawaii specimens are restricted to cryptic habitats and turbid waters, avoiding direct sunlight (Concepción et al. 2008). The authors, based on the results, suggest that C. riisei is mostly opportunistic and the environmental conditions of the Ecuadorian coast do not represent an important driver. Notably, these sites are characterized by high density of native corals (Glynn 2003;Rivera and Martínez 2011) which are highly vulnerable to environmental variability, especially increasing temperature (Glynn et al. 2001). The weakening of native coral reefs in areas with rising temperatures-for example due to climate change-could accelerate the C. riisei invasion scenario.
The globalization of maritime traffic plays a key role in the introduction and spread of species since many of these organisms are moved between regions in ballast water, on ships hulls, in sea chests, on chains and even anchors (Kolar and Lodge 2002;Hulme 2009;Seebens et al. 2013). Eighty percent of Ecuador's foreign trade is carried out by sea, there are ports for oil and commercial cargo in Esmeraldas, Manta, La Libertad, Puerto Bolívar (Torres 2017) and Guayaquil; with several fishing and tourist piers along the coast. These larger ports could receive species from around the world and serve as hubs for the transport of non-native species to nearby smaller ports or sites that would not normally receive such organisms (Lockwood et al. 2009). Carijoa riisei has been reported as a fouling organism (Padmakumar et al. 2010), making it possible for this species to have been introduced by marine traffic in Ecuadorian coasts.
Due to how this species has behaved and expanded along the coast of Ecuador in a short period of time since its first record, there is a high risk of this species overtaking native species and altering the marine ecosystem as well as the ecosystem services that it provides. However, it is not only the coast of mainland Ecuador that is at risk, the unique marine biodiversity of the Galapagos Marine Reserve is also in danger due to the likelihood of this species being introduced to the Galapagos Islands because of the connectivity that exists between mainland Ecuador and the archipelago through marine traffic routes and current regimes (Carlton et al. 2019). Currently all vessels that enter the Galapagos Marine Reserve, be it from national or international destinations, must have their hull cleaned prior to entry and inspected on arrival by the Galapagos Biosecurity Agency, marine unit (Agencia de Regulación y Control de la Bioseguridad y Cuarentena para Galápagos 2018). Extreme care should be taken during inspections in order to minimize the risk of possible introduction of this species.
The data presented in this work could be used as a reference to continuing monitoring the spreading of C. riisei along the Eastern Pacific in order to measure the expanding distribution and abundance of this species. Biosecurity protocols should be socialized with dive centers, fishers and local boat owners as well as with cargo companies and the Ecuadorian Navy. A risk assessment of the impact that this species is causing should be presented to the authorities in order to consider management strategies such as real-time continuous monitoring.
Currently, the Ecuadorian Navy, through the National Directorate of Acuatic Spaces (Dirección Nacional de los Espacios Acuáticos, DIRNEA), is conducting a baseline survey of invasive Spaces in continental and insular Ecuador under funding from the International Maritime Organization (IMO) and the participation of the Permanent Commission of the South Pacific (CPPS).

Ethics and Permits
This paper contains partial information derived from an environmental consultancy carried out by the company BIOELITE for the Subsecretaría de Gestión Marina Costera of the Ministry of the Environment (MAE) during the period 2015-2016 and BIOELITE had the permits to carry out the investigation in the protected areas and coastal areas of Ecuador to carry out the research called "Quantitative subtidal and intertidal marine biodiversity inventories in six marine protected coastal areas and four areas of possible expansion" (CFC-001-2015) and Maritza Cárdenas obtained in 2016 the MAE's permission to use the data to scientific publications.
The following supplementary material is available for this article: Table S1. Sites included in the analysis. Table S2. Primers used to amplify the COI and NADPH genes. Table S3. Distance matrix of ND6 sequence fragments, showing haplotypes from this work and from the work of Concepción et al. 2010 .  Table S4. List of environmental variables used for Principal Component Analysis. Figure S1. Overgrowth of Carijoa riisei on invertebrates at Guayaquil Bioregion ("La Pared" on El Pelado Marine Reserve) during April 2015.