Taxonomic Diversity of Decapod and Stomatopod Crustaceans Associated with Pocilloporid Corals in the Central Mexican Paciﬁc

: Many crustacean species are obligate associates of pocilloporid corals, where they feed, reproduce, and ﬁnd shelter. However, these coral-associated crustaceans have been poorly studied in the eastern tropical Paciﬁc. Determining the crustacean richness and taxonomic distinctness could help in comparing different coral reefs and the potential effects of degradation. This study evaluated the spatio–temporal variation of the taxonomic diversity and distinctness of coral-associated crustaceans in four ecosystems of the Central Mexican Paciﬁc (CMP) with different conditions and coral cover. In all ecosystems, 48 quadrants were sampled during the summer and winter for two years. A total of 12,647 individuals belonging to 88 species, 43 genera, and 21 families were recorded. The sampling effort yielded 79.6% of the expected species richness in the study area. Species rarity had 19% singletons, 4% doubletons, 22% unique, and 9% duplicate species; two species represented new records for the Mexican Paciﬁc, and six were new to the CMP. This study recorded most of the symbiotic crustacean species in pocilloporid corals previously reported in the CMP. The taxonomic diversity and distinctness differed signiﬁcantly between coral ecosystems and seasons, which was also visualized by nMDS ordination, showing an evident spatio–temporal variation in the taxonomic beta diversity. F.A.R.-Z. M.A.-P.; formal analysis A.A.-D. and F.A.R.-Z.; resources, F.A.R.-Z.; data curation, A.A.-D. M.A.-P.; writing–original draft preparation, A.A.-D. and F.A.R.-Z.; writing—review and editing, F.A.R.-Z., M.A.-P., E.R.-J., M.d.C.E.-G., M.E.H. and O.V.-P.; project administration and funding acquisition, F.A.R.-Z.


Introduction
Invertebrates are frequently associated with scleractinian corals of the genus Pocillopora [1]. Macrocrustaceans are the most representative coral-associated fauna. Among these, diverse assemblages find shelter among the Pocillopora branches [2,3]. Different taxa, including shrimps, crabs, isopods, and copepods, have been described as coral symbionts [4], presenting different degrees of specialization in form and function [5]. Several of these species are obligate symbionts, always and permanently associated with specific hosts, while other species are facultative symbionts that can also survive outside their host, usually on non-living substrates [1,5]. As a general trend, the coral-associated fauna especially symbiotic species. Evidence has shown that the Pocillopora-associated fauna has a spatio-temporal variation due to environmental drivers [3,6,21,37]. We hypothesized that the sites with the most discontinuous coral cover and highest human intervention (local tourism, fishing, etc.) would have the lowest richness and taxonomic distinctness, along with a high abundance of coral-associated fauna due to the low coverage and greater isolation of the host coral colonies.

Study Area
The study area included four coral ecosystems in the Central Mexican Pacific (CMP): (i) Chamela and (ii) Cuastecomate-Punta Melaque in southern Jalisco, and (iii) Carrizales and (iv) Punto B in Colima ( Figure 1). The CMP is part of the eastern tropical Pacific ecoregion spanning from Baja California to northern Peru and the Galapagos Islands, Ecuador [38]. In the summer, the CMP is influenced by the California Current, the Cabo Corrientes Upwelling, the Mexican Warm Pool, and the Costa Rica Coastal Current. However, these currents have a weaker effect during the winter and spring due to the cold water from the California Current and the warm water from the Cortés Current [39]. Furthermore, the Mexican Warm Pool is part of the Western Hemisphere Warm Pool, which induces an important annual climatic variation in the water temperature to develop the El Niño-Southern Oscillation (ENSO) [40,41]. These currents provide the CMP with species from different biogeographic provinces [38]. Hurricanes, tropical storms, and upwellings also significantly impact the coral colony structure and associated fauna [18].
Diversity 2022, 14, x FOR PEER REVIEW 3 of 13 protected, most of the coral ecosystems are not [18]. Corals are very susceptible to environmental changes and natural and anthropogenic impacts. These changes affect associated fauna, especially symbiotic species. Evidence has shown that the Pocillopora-associated fauna has a spatio-temporal variation due to environmental drivers [3,6,21,37]. We hypothesized that the sites with the most discontinuous coral cover and highest human intervention (local tourism, fishing, etc.) would have the lowest richness and taxonomic distinctness, along with a high abundance of coral-associated fauna due to the low coverage and greater isolation of the host coral colonies.

Study Area
The study area included four coral ecosystems in the Central Mexican Pacific (CMP): (i) Chamela and (ii) Cuastecomate-Punta Melaque in southern Jalisco, and (iii) Carrizales and (iv) Punto B in Colima ( Figure 1). The CMP is part of the eastern tropical Pacific ecoregion spanning from Baja California to northern Peru and the Galapagos Islands, Ecuador [38]. In the summer, the CMP is influenced by the California Current, the Cabo Corrientes Upwelling, the Mexican Warm Pool, and the Costa Rica Coastal Current. However, these currents have a weaker effect during the winter and spring due to the cold water from the California Current and the warm water from the Cortés Current [39]. Furthermore, the Mexican Warm Pool is part of the Western Hemisphere Warm Pool, which induces an important annual climatic variation in the water temperature to develop the El Niño-Southern Oscillation (ENSO) [40,41]. These currents provide the CMP with species from different biogeographic provinces [38]. Hurricanes, tropical storms, and upwellings also significantly impact the coral colony structure and associated fauna [18].  Some of the general characteristics of the sampled sites are as follows: (1) Chamela (CH) is formed by different small islands and islets; its coral ecosystems are patchy and isolated, and the benthos has a high coverage of rubble, sand, and dead coral. This site is important for fishing and local tourism. (2) Cuastecomate-Punta Melaque (CT) has a discontinuously distributed high coral cover, characterized by small reef patches with fleshy macroalgae stands, sand, and rocks. (3) Carrizales (CA) is located in Ceniceros Bay; it is a short beach defined by two small and fringing coral reefs on each side of the shore with~100% live coral cover. (4) Punto B (PB) is located in Santiago Bay near the Julualpan Lagoon's mouth and is considered a highly touristic area. Its coral community has scarce, isolated coral colonies but great coverage of sponges, calcareous algae, and sandy and rocky substrates.
Three samples of live Pocillopora corals were collected in each coral ecosystem using randomly placed 0.25 m 2 quadrants. Each sample position was marked using a global positioning system (GPS). A total of 48 samples were collected during September 2017, January and September 2018, and January 2019. All samplings were obtained by scuba diving at a 10 m depth. Each coral sample was covered with a plastic bag to avoid losing organisms and detached using a hammer and chisel. Subsequently, the coral was carefully fragmented to collect all the organisms between and within the Pocillopora branches. All live decapods and stomatopods were fixed with 70% ethylic alcohol. Samples were identified to the most precise taxonomic level possible in the Molecular Ecology, Microbiology, and Taxonomy Laboratory (LEMITAX), Universidad de Guadalajara. The specialized literature for identification included Rathbun [42], Haig [43], Abele and Kim [44], Castro [45], Anker et al. [46], Hendrickx et al. [47], Ayón-Parente [48], García-Madrigal and Andréu-Sánchez [49], Hermoso-Salazar [50], Salgado-Barragán and Hendrickx [51], and Hiller and Lessios [52]. A presence/absence matrix was constructed to perform the ecological analysis.

Data Analysis
The spatial and temporal variation of the taxonomic diversity was evaluated with a three-way experimental design with crossed factors expressed as: where Y is the variable under analysis (taxonomic diversity), and µ is the mean of the analyzed variable. The year factor (Ye i ) had two levels (years), and each year was composed of two seasons (dry and wet seasons), so the first year included September 2017 and January 2018, and the second included September 2018 and January 2019. The season factor (Se j ) had two levels: wet (September 2017 and 2018) and dry (January 2018 and 2019). The site factor (Si k ) had four levels corresponding to the studied coral ecosystems. Finally, ε ijk represented the accumulated error. All factors were considered as fixed effects (model type I).
The sampling effort was evaluated using sample-based rarefactions at three levels (i.e., site, season, and year) with the observed species richness and the expected richness estimated using non-parametric estimators Chao 1, Chao 2, Jackknife 1, Jackknife 2, ICE, and ACE. These estimators were based on rare species; they estimated the number of potential species considering the incidence and abundance data recorded in the samplings [53]. The total observed richness (S Obs ) was calculated for each ecosystem with the Mao Tao function. Then, the coral ecosystems were compared in pairs with individual-based rarefactions and 95% confidence intervals. All rarefaction curves were built with 10,000 randomizations without replacement. Species rarity was also calculated (singletons, doubletons, unique, and duplicate species), and the species were identified. These analyses were performed in the software EstimateS 9.1 [54]. The absolute density of each species (represented as the number of individuals per m 2 ) and their absolute frequency were also estimated.
The taxonomic diversity analysis considered each site's taxonomic differences and singularities regarding the seasonal variation and the years analyzed. Thus, the average taxonomic distinctness (∆ + ) analysis was performed to evaluate the species' distribution and in- cidence as well as their taxonomic relations [28]. This analysis also measured the taxonomic distance between two species and its variation (Λ + ), according to the following equations: where S represents the number of species, and ω ij denotes the assigned weight of each supraspecific taxonomic level. An eight-level taxonomic aggregation matrix was built, including species, genus, family, subfamily, suborder, order, subclass, and class. According to Warwick and Clarke [55], the taxa were weighted as follows: ω = 1, species within the same genus; ω = 2, species within the same family but different genus; ω = 3, species within the same subfamily but in a different family; and so on. The ∆ + and Λ + were estimated for each site, season, and year. The models were created with a 95% confidence interval, and the statistical significance was tested with 10,000 permutations. The ∆ + analysis was followed by a taxonomic dissimilarity analysis (Γ + ), which is described as: where Γ + denotes the gamma + taxonomic dissimilarity, S 1 represents the number of species in the first sample, S 2 is the number of species in the second sample, and ω ij denotes the path length between species i and j.
A non-parametric multidimensional scaling (nMDS) and a cluster analysis were performed using the taxonomic dissimilarity (Γ + ) matrix to explore the crustacean taxonomic differentiation patterns across the spatio-temporal experimental design (site, season, and year). The cluster analysis was built with the average group linking method and similarity profile analysis (SIMPROF) to assess group formation using 10,000 permutations. Therefore, nMDS ordination was coupled with the cluster analysis outputs. All analyses (i.e., ∆ + , Λ + , Γ + , nMDS, and cluster analysis) were performed in PRIMER 7.0.21 and PERMANOVA +1 [56].

Results
A total of 12,647 specimens were collected, representing 21 families, 43 genera, and 88 species (Supplementary Material, Table S1). For each quadrant, the number of species collected ranged from 13 to 38, and the number of individuals ranged from 36 to 705. The most diverse families were Alpheidae (21 species) and Porcellanidae (20 species). Ten families (47%) were represented by a single species (Supplementary Material, Table S1). The sample-based rarefaction showed that the sampling effort had an adequate representativity (79.6%) of the species richness expected by chance (Supplementary Material, Figure S1). The sampling effort ranged between 77.9% and 91.6% of representativity for all sites. The seasons showed 84.6% of representativity during the dry season and 85.7% in the wet season (Supplementary Material, Table S2). The representativity for year one was 79.5%, and for year two it was 88% (Supplementary Material, Table S2).
Individual-based rarefactions in pairwise comparisons showed that the species richness between sites was similar because their confidence intervals (95%) overlapped (Supplementary Material, Figure S2). An exception was Chamela and Carrizales, which had the highest and lowest number of species, respectively (Supplementary Material, Figure  S2). The highest total species richness and abundance recorded over the sampling period were as follows: for Chamela, 69 species and 2371 individuals; for Cuastecomate-Punta Melaque, 64 species and 3266 individuals; for Carrizales, 58 species and 2752 individuals; and for Punto B, 68 species and 4258 individuals (Supplementary Material, Table S1). The total species richness was similar between years and between seasons (Supplementary Material, Figure S3 The average taxonomic distinctness (∆ + ) analysis at the site level showed that the ∆ + values for all the sites fell inside the probability funnel or within the 95% confidence intervals (p > 0.05). Chamela had the lowest ∆ + values despite having the greatest number of species (Figure 2). Punto B had the highest ∆ + values above the global ∆ + of the model. However, the Λ + values for all sites fell within the probability funnel, indicating that the sampled sites were representative of the taxonomic diversity of the area. The seasons had different ∆ + values because the wet season fell within the probability funnel, but the dry season was outside the funnel (p < 0.05). The Λ + values for the dry season were outside the funnel, so the taxonomic representativity during the dry season was lower than expected by chance (Figure 2). The ∆ + and Λ + values between years were similar and fell inside the probability funnel ( Figure 2).
The nMDS ordination showed that the taxonomic dissimilarity (Γ + ) differed among the sites (Figure 3). The cluster analysis based on the SIMPROF procedure confirmed a group constituted by the southern sites (i.e., Carrizales and Punto B) and two separate entities (i.e., Chamela and Cuastecomate-Punta Melaque). This was also observed in the nMDS ordination. Carrizales and Punto B shared several species (e.g., Pseudosquillisma adiastalta, Pomagnathus corallinus, and Synalpheus arostris) and genera (e.g., Trapezia, Liomera, and Pomaghnathus), and they had almost the same families, except for the Pinnotheridae, which was only present in Punto B (and Cuastecomate-Punta Melaque). Conversely, Cuastecomate-Punta Melaque had a different crustacean fauna compared to the other sites and showed a mixture of taxa shared with Chamela and Punto B. Cuastecomate-Punta Melaque presented 34 genera and 17 families; these families were the same as Punto B, except for Panopeidae, which was found exclusively in this site, and Pseudosquillidae, which was absent. Chamela is the northernmost site and the most distant from the others. It was different because it had one superfamily (Parthenopoidea) not found elsewhere and two absent superfamilies (Eriphioidea and Pinnotheroidea). In Chamela, two families that were not found in other sites were collected (Parthenopidae and Lysmatidae), and four families (Panopeidae, Oziidae, Pseudosquillidae, and Pinnotheridae) were absent.

Discussion
This study recorded most of the Pocillopora obligate symbiotic crustacean species reported by previous studies, including Trapezia bidentata, T. corallina, T. digitalis, T. formosa,

Discussion
This study recorded most of the Pocillopora obligate symbiotic crustacean species reported by previous studies, including Trapezia bidentata, T. corallina, T. digitalis, T. formosa,

Discussion
This study recorded most of the Pocillopora obligate symbiotic crustacean species reported by previous studies, including Trapezia bidentata, T. corallina, T. digitalis, T. formosa, Alpheus lottini, Hapalocarcinus marsupialis, and some species of Synalpheus. However, we did not find some species known to be associated with Pocillopora, such as Fennera chacei, Alpheus sulcatus, Palaemonella holmesi, Stenorhynchus debilis, Thor algicola, and Petrolisthes galathinus, which had been previously reported in the study area [24,25]. Nonetheless, we obtained two new records for the Mexican Pacific and six new records for the CMP (Supplementary Material, Table S1), increasing the known information regarding regional crustaceans.
Several species collected during this study, i.e., Tumidotheres margarita, Typton sp., and Pontonia sp., have been reported as endosymbionts of sponges, ascidians, or bivalves. These hosts are frequently associated with pocilloporid corals, and these decapods might be recognized as having a secondary association with pocilloporid corals. Tumidotheres margarita is an endosymbiont of the bivalves Barbatia reevaena, Limaria pacifica, and Pinctada mazatlanica [57], which are known as Pocillopora-associated mollusks in the Mexican Pacific [58]. Typton tortugae and T. serratus have been recorded as being associated with sponges living on corals [59]. In this study, some sponges were found to be associated with corals, and a similar association could exist in the cases of T. hephaestus and T. granulosus. Shrimps of the genus Pontonia are reported as obligate symbionts of the bivalves Pinna spp. and P. mazatlanica [60]. We assumed that the Pontonia specimens collected during this study were dislodged from their host during the collecting process or after the samples were preserved.
Our study increased the inventory of crustaceans associated with Pocillopora coral in the Mexican Pacific from 59 [20][21][22][23][24] to 88 species. Comparatively, in Huatulco, Oaxaca, a method similar to the one used here (0.25 m 2 quadrants) recorded 47 species of brachyuran crabs in pocilloporid corals [21]. In La Paz and Loreto Bay, Baja California Sur, 44 species of decapods were recorded [22]. Furthermore, a study covering almost the entire Mexican Pacific, from the Gulf of California to Oaxaca, recorded 36 crustacean species associated with pocilloporids [24]. The difference between the number of species reported herein and by Hernández et al. [24] may be a consequence of the visual census they performed. With this method, some close species are easily confused (e.g., Synalpheus spp., Trapezia spp., and Alpheus spp.) or overlooked (e.g., Hapalocarcinus marsupialis). The expected species richness estimated by the sample-based rarefaction was 20% higher than the observed richness due to the large number of rare species collected. Expected species richness is a good indicator of the potential species expected in the area. The sample-based rarefaction confirmed that the sampling effort was sufficient to elucidate the actual number of crustacean species associated with the Pocillopora coral in the CMP.
Decapod crustacean fauna associated with Pocillopora coral has been studied in many tropical and subtropical regions of the world's oceans. The species diversity recorded in this study is superior to the 36 species associated with Pocillopora off the Arabian coast in the Red Sea [61]. However, it is lower than the diversity reported from Oahu (Hawaii), where 127 species were found associated with Pocillopora damicornis [62], and then the 91 species reported more recently in 751 colonies of P. meandrina, also in Oahu [3]. For the northern Great Barrier Reef, Australia, 102 species were found in 50 colonies of P. damicornis [63]. It is important to mention that the obligate symbiotic composition observed in our study is similar to what has been reported for the Red Sea and the Great Barrier Reef, i.e., all three studies share the same brachyuran crabs (Trapezia bidentata, T. digitalis, and Domecia hispida) and caridean shrimps (Alpheus lottini, Synalpheus charon, and Harpiliopsis depressa).
A previous study indicated that the number of species present in coral ecosystems depends on the size of the coral colony [64]. The authors reported species richness ranging from 3 to 22 per colony (1500 cm 3 size) in the Gulf of Panama; in Costa Rica, 20 cm diameter colonies had 18 species [65]. Despite using quadrants of the same size, this study collected 13-18 species and 36-711 individuals per 0.25 m 2 of coral sample. These differences in abundance and richness are substantial and cannot only be attributed to colony size.
To predict the species richness or abundance in colonies with stable conditions, some authors considered coral complexity (e.g., inter-branch space, penetration depth, and size of living space) [9], but in some cases, this factor was unable to explain the changes between different colonies [63]. For example, species such as the symbiotic Trapezia are not limited by coral complexity; they only need a healthy coral fragment for their survival [66]. Other characteristics, such as the percentage of live tissue and habitat degradation, could also influence the richness and abundance shifting. The species richness and abundance increase when the proportion of live coral tissue cover decreases [7,63]; this might happen because coral loss allows other species to move to new colonies. Moreover, coral mortality increases the abundance in single colonies [15], which may occur for two reasons: (1) the death of symbionts allows for other opportunistic species to move to more stable colonies, or (2) coral loss induces migrations of individuals looking for new space to live [7,9]. This situation could be happening in Punto B, where the coral colonies are isolated, fewer colonies are available, and the ecosystem is subject to anthropogenic pressure [18]. Symbiont loss does not seem to be a problem in Punto B because of the abundant obligate symbionts found in all samples.
The average taxonomic distinctness (∆ + ) varied between sites and seasons. The ∆ + values fell inside the 95% probability funnel, meaning they were a good representation of the taxonomic diversity of decapods and stomatopods associated with pocilloporid corals. However, Chamela had a lower ∆ + value despite having the greatest species richness among the four sites. This contrast occurred because Chamela featured the fewest supraspecific taxonomic hierarchies since many species belonged to the same families, i.e., Alpheidae (17 species) and Porcellanidae (18 species). In contrast, Punto B had the highest ∆ + value above the global ∆ + of the model and sustained almost the same species richness as Chamela. Punto B shared the taxonomic hierarchies with other sites and did not present any exclusive hierarchy.
Regarding the temporal variation of the taxonomic diversity, the ∆ + values fell outside the probability funnel in the dry season, meaning a relatively low taxonomic diversity change during this season; six genera (Areopaguristes, Aniculus, Daldorfia, Bottoxanthodes, Pontonia, and Pseudosquillisma), two families (Parthenopidae and Pseudosquillidae), and one superfamily (Parthenopoidea) were not recorded in this season. In contrast, the taxonomic diversity was better represented during the wet season, when the Parthenopoidea superfamily was present, portrayed by Daldorfia trigona, a species not collected in the dry season. Moreover, 15 species and 6 genera were exclusively collected during the wet season (Supplementary Material, Table S1). Years one and two had similar species richness and taxonomic structure. Likewise, both ∆ + values fell into the probability funnel close to the global ∆ + level, demonstrating that the studied years adequately represented the taxonomic diversity estimated by the global ∆ + model.
The nMDS ordination coupled with cluster analysis showed that Chamela had the highest taxonomic dissimilarities (Γ + ) among the studied sites. Chamela-the northernmost site-was the most different with the highest taxonomic dissimilarity, the lowest ∆ + , and the highest species richness. The Chamela samples contained one superfamily, two families, and four genera exclusive to this site, but several superfamilies, families, and genera present in the other sites were absent. It has been suggested that a low taxonomic distinctness can indicate a loss in the taxonomic diversity due to anthropogenic stress [30]. However, in this study, Punto B was the most anthropogenically affected site and displayed the highest ∆ + values. Despite moderate disturbances, symbiotic species tend to stay in their host for a long time [9,63]. Nevertheless, some symbionts (e.g., Trapezia) can migrate to other coral colonies in search of more suitable habitat [66,67]. Limited habitat availability makes them pile up in the colony, increasing species richness and abundance. This phenomenon could affect the ∆ + values in Punto B, increasing the values higher than the global ∆ + of the model. The low levels of taxonomic diversity in Chamela might be attributed to other variables, including the spatial process [68], benthic heterogeneity, habitat availability [69], or habitat type [70]. In addition, it is important to remember that the variety of microhabitats is one of the main factors driving the diversity and abundance of coral-associated crustaceans [7].
In conclusion, the sampling effort in this study allowed for obtaining more than 70% of the expected species, indicating a good taxonomic representativity. The species richness and the taxonomic distinctness were within the expected values, despite being lower during the dry season. Most of the expected coral-obligated symbionts were collected, except for Fennera chacei, a small species frequently living in the coral base, which probably escaped during the collecting process. In contrast with the initial hypothesis, the sites with the most discontinuous coral cover and the largest human intervention did not have the lowest taxonomic distinctness (Punto B). However, as expected, the greatest abundance was observed in Punto B; this can be explained by the low coral availability, environmental variables, or anthropogenic stress. The present study should be complemented with α, γ, and β diversity analysis to assess the spatio-temporal differences in this particular species assemblage. It is also important to consider the influence of environmental variables, reef structural complexity, and human impact on the richness and abundance of these crustacean species, particularly in the obligate coral-symbiotic species. This study helped us to understand the crustacean assemblage associated with corals in the CMP and the spatiotemporal variations in their taxonomic diversity. Furthermore, it increased the taxonomic inventory of the coral-associated species in the studied region and the Mexican Pacific.

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary Material. Likewise, the data are available upon request from the corresponding author.