Species composition of Discomedusae jellyfish (Scyphozoa) in the coastal waters of Eastern Surabaya, East Java

The Indo-Pacific is recognized as a hotspot for marine diversity. The taxonomy of certain taxa, such as Discomedusae jellyfish, has been neglected, despite its importance in the fishery industry. This study documents the first records of Discomedusae for the Java Sea using an integrative approach and provide notes about its distribution in the area. We used up to 53 morphological and meristic characters and amplified one mitochondrial marker (COI). The comparison and assessment of these data resulted in the recognition of seven species of Discomedusae, from which five has been recorded for the Indo-Pacific area. Two other species require a taxonomic revision to confirm the species assignation. The distribution of jellyfish in the coast of Java Sea might be correlated with the jellyfish life history and species-specific ranges of tolerance, and not solely determined by the environmental parameters. These findings provide the foundations for extending the taxonomic research in the area; the description of the biodiversity will increase the understanding of the population dynamics and its implications in the fisheries.


Introduction
Scyphozoan jellyfish are one of the most common pelagic invertebrates which inhabit all the oceanic basins (Kramp 1961;Arai 1997). The class Scyphozoa includes the subclass Discomedusae (Order Rhizostomeae, shallow water, and Semaeostomeae, shallow and deep-water medusae) and Order Coronatae (shallow and deep-water medusae, Kramp 1961; Marques and Collins 2004;Bayha et al. 2010). Today, Discomedusae are increasing in frequency and abundance (Attrill et al. 2007;Condon et al. 2012;Brotz et al. 2012), rising the public and scientific interest due its detrimental ecological and economic consequences (Purcell et al. 2007;Richardson et al. 2009;Dong et al. 2010;Roux et al. 2013) and its growing demand and profits for the jelly-fishery industry (Kitamura and Omori 2010;Brotz 2016;Brotz et al. 2017). However, most of the studies do not address the genetic and taxonomic diversity (e.g., Condon et al. 2012), even for hotspot areas, which have been proven to account for a high biodiversity (e.g., Tropical Eastern Pacific, Gómez Daglio and Dawson (2017), in part Caribbean and Indo-Pacific, Abboud et al. 2018).
The Indo-Pacific Ocean has a well-documented marine diversity for invertebrates such as corals, mollusks, and crustaceans and fishes (Moretzsohn and McShane 2004;Barber and Bellwood 2005;Barber and Boyce 2006;Allen 2007;Hubert et al. 2012;Mihaljević et al. 2017). However, documentation of scyphomedusae diversity is limited to a few expeditions and taxonomic revisions made during nineteenth and early twentieth century (Haeckel 1880;Vanhöffen 1888;Maas 1907;Mayer 1910;Rao 1931;Stiasny 1940). The

Communicated by S. Piraino
Indo-Pacific is a heavily fishing area, and the jelly-fishery is a common activity in the area (Brotz 2016). Currently, only two studies describe some fished Discomedusae using an integrative approach (morphology and molecular tools, Strait of Malacca (Rizman-Idid et al. 2016), and Indian Ocean, Central Java (Nishikawa et al. 2014)). Hence, the jelly-fishery is managed without a reliable taxonomic identification of the targeted species, and estimates about its landing, abundance, and distribution are questioned (Kitamura and Omori 2010;Brotz 2016;Rizman-Idid et al. 2016).
Indonesia shows in the top three countries with the highest landing (~ 29,469 tons) and exports of jellyfish (Omori and Nakano 2001;Brotz 2016). Yet, there are 35 species of Discomedusae reported for Indonesia (Table S1; Kramp 1961;De Souza and Dawson 2018;Jarms and Morandini, 2019). However, the species identification and distribution range are uncertain for 13 of the 35 recorded species (Table S1). Only five confirmed records exist for the area: Crambione mastigophora (Asrial et al. 2015), Crambionella helmbiru (Nishikawa et al. 2014), Lobonemoides robustus (Kitamura and Omori 2010), Rhopilema esculentum (Kitamura and Omori 2010), and Versuriga anadyomene (Ohtsuka et al. 2010). Except for C. helmbiru, the identification and records rely on morphological characterization of the species, that can be problematic, due to the presence of cryptic species complexes within Discomedusae (Dawson and Jacobs 2001;Holland et al. 2004;Holst and Laakmann 2014;Gómez Daglio and Dawson 2017). It is necessary to describe the Discomedusae diversity in such important hotspot and fishery area. Herein, we aim to describe the Discomedusae diversity on an unexplored area of Indonesia (Java Sea: Strait of Surabaya), employing an integrative approach (genetic and morphologic data). In addition, we include some notes about its records along the coastline.

Sample collection
We carried out two major collection projects. The first one consisted of a survey the east coast of Surabaya at 22 stations located in two transects parallel to the shore (0.5 km and 1.5 km from the coastline, Fig. 1) between September and December 2010 every 2 weeks. At each station, we measured five physico-chemical parameters (temperature, salinity, pH, transparency, and flow velocity). Surface temperature was measured using a mercury thermometer with a precision of ± 1 °C. Salinity was measured using a hand refractometer (ATAGO ATC FG-217 with a precision of ± 1%). pH was measured using testing pH paper (Hydrion, range 0.0 to 13.0 pH). Transparency levels were measured using a Secchi disk following the protocol of English et al (1994). Flow velocity was estimated using buoys, which were held together with 2 m of rope and floated at each sampling point then timed with a chronometer.
Discomedusae were collected during mornings with Gillnets (Ø = 1.5 cm) and Push-nets (Ø = 3 cm) at 2 m deep, with immersion time ranging from 10 to 15 min over the 5-h period. In addition, NORPAC nets (Ø = 150 μm) were laid horizontally on the surface at 2-min intervals. A piece of oral arm, tentacle, or umbrella margin were clipped and preserved in ethanol 90% for subsequent DNA analyses; the whole medusa was preserved in 10% formalin. To complete the taxonomic sampling, we carried out a different sampling; here we collected jellies from shrimp nets set by local fishermen in Kenjeran in the strait of Surabaya during the months of September through November 2010. All the samples were labeled and separated by morphotypes groups at the taxonomic family level (Pelagiidae, Catostylidae, Mastigiidae, Cyaneidae, Lychnorhizidae, and Rhizostomatidae sensu (Kramp 1961). In addition, on each site, we recorded the number of Discomedusae captured per station.

Morphological identification
Morphological characterization and identification were performed in the Ecology Department Laboratory of Biological Science ITS Surabaya (Indonesia). Each medusa was photographed and examined to record morphological and meristic characters that the literature suggests are helpful for Discomedusae identification (Mayer 1910;Light 1914;Stiasny 1920Stiasny , 1921Stiasny , 1922Stiasny , 1932Marques and Collins 2004;Morandini and Marques 2010;Gómez Daglio and Dawson 2017). For example, macro-morphological characters, coloration patterns; bell shape, distribution, position, number, and shape of oral arms; shape and number of terminal clubs; position, insertion, and number of tentacles; detailed characters such as filaments, scapula, and exumbrella papillae; and rhopalia insertion, shape and position, manubrium, position, number and shape of the oral pillars, shape and size of the oral disk, and shape and number of velar lappets, were recorded and photographed using a Stereo Microscope Olympus SZ51 (list of measured characters is available in Gómez Daglio and Dawson (2017)). Specimens revised were deposited in the Museum of Comparative Zoology, Harvard University (MCZ Cat. No.153057-153060;153062-153066;153068-153072).

DNA extraction, amplification, and sequencing
Total genomic DNA was extracted from a subset from the morphotypes' tissue samples (12 individuals) using a modified CTAB phenol-chloroform protocol (Dawson and Jacobs 2001). We amplified the mitochondrial locus cytochrome c oxidase subunit I (COI). COI is widely use within the Discomedusae to identify species (Ortman et al. 2010;Gómez Daglio and Dawson 2017

Molecular identification
All the sequences were assembled into contigs, primers trimmed, and inspected manually to check base calls and open reading frames using Sequencher v. 5.2.4 (Gene Codes Corp. Ann Arbor). Sequences were deposited in GenBank (Accession number MN395673-MN395694). We aligned congeneric sequences using ClustalX v. 2.0 (Larkin et al. 2007). Each alignment was visualized using Jalview (Waterhouse et al. 2009); we estimated the sequences similarity and export the consensus sequence. Each consensus sequence was compared by BLASTn searching GenBank (Benson et al. 2012), accessed February (2018), to affirm the correct morphological identification. In addition, a subset of the samples was barcoded under a global context by Abboud et al. (2018) and in extent under a phylogenetic context by Gómez Daglio and Dawson (2017) to corroborate the species identification.

Environmental parameter analyses
We analyzed the environmental parameters (Supplementary  Material Table S2) recorded at each station. Each parameter data was tested for normality and homoscedasticity in R v. 3.2.4. (R Development Core Team 2020). All the data sets met the assumption of normality and homogeneity of variance. We performed a two-tailed t-test implemented in R v. 3.2.4. (R Development Core Team 2020) to evaluate the differences between the stations sampled close to the coastline (transect 1, 0.5 km) and the stations sampled at 1.5 km (transect 2) from the coast line.

Results
Using an integrative taxonomic approach, we record seven species of Discomedusae jellyfish in the eastern waters of Surabaya (Table 1). Four belong to the Order Rhizostomeae, represented by four families, Mastigiidae, Rhizostomatidae, Lychnorhizidae, and Catostylidae; two species are classified within the order Semaeostomeae, represented by two families, Pelagiidae and Cyaneidae.

Order Semaeostomeae
Family Pelagiidae: Chrysaora chinensis Vanhöffen, 1888. Exumbrella surface fine granulated; brown-yellowish stripes radiating from the apical central up to the umbrella margin. Small (< 3 mm in diameter) flecks that covers the exumbrella (Fig. 2a). Four simple curtain-like oral arms attach to a manubrium. Bell margin with four semi-oval velar lappets per octant, three primary tentacles are inserted at the bell margin alternating between each velar lappet. Two rhopalia per quadrat, each rhopalium with two semi-oval lappets (Fig. 2b). Central stomach circular limited by the origin of the radial septae. Radial septae (four per quadrant) are bent proximally to the bell margin (Fig. 2b). Gonads are paired on the interadial axis. Subgenital ostia rounded (Fig. 2c). Family Cyaneidae: Cyanea sp. Medusae reddish to brownish color (Fig. 2d). Bell with 16 velar lappets (eight primary and eight secondary clefts). Four perradial curtainlike oral arms. Tentacle clusters are located in the adradial axis; the eight clusters have a horseshoe shaped (Fig. 2e). Eight rhopalia are embedded and located ¼ proximally from the lobes (Fig. 2f). Gonads are in out-folded pockets. The specimens were damaged, and features such as the number of coronal and radial muscles and pits were not accounted.

Order Rhizostomeae
Family Catostylidae: Acromitus flagellatus (Maas, 1903) and Catostylus townsendi Mayer, 1915. A. flagellatus is identified by the umbrella dome shape with fine granular protuberances that cover the entire exumbrella. Eight threewinged oral arms covered with short filaments (Fig. 3a). Each quadrant has two rophalia, each rhopalium with two pointed rhopaliar lappets (Fig. 3b) and eight rounded velar lappets per quadrant. Four subgenital ostia, each one with a heart-shaped prominent papillae arranged on the perradial axis (Fig. 3c). The canal system presents a ring canal which connects with one perradial, two interradial, and 12 adradial canal per octant. The perradial and interradial canals extend up to the stomach; adradial canals covers ½ of the surface but do not connect with the stomach. Anastomoses extend up to the bell margin and rhopalia (Fig. 3c). C. townsendi presents a fine granular exumbrella with yellowish-brownish pigmented flecks (Fig. 3d). Oral arms are three-winged with no filaments or terminal clubs. Rhopalium (two per quadrant) is embed on a furrow, with two rounded rhopaliar lappets (Fig. 3e) and 8-12 rounded and elongated velar lappets, some bifurcated. Subgenital ostia has a bone-shaped papillae. The canal system has one perradial, three interradial, and 8-10 adradial canals per octant. Adradial-adradial anastomoses do not reach the central stomach and present sinuses. The canal system extends up to the bell margin and rhopalia (Fig. 3f).
Family Lychniorhizidae: Lychnorhiza malayensis Stiasny, 1920. Medusae with eight three-winged oral arms, without terminal clubs or filaments. Bell margin with eight bifurcated velar lappets and one rhopalium per octant. Rhopalium with two tapering lappets (Fig. 3g). Cruciform gonadal cavity (Fig. 3h). Subgenital ostia with leaf-shape papillae. Canal system with one perradial, one interradial, and two adradial canals that connects to the gastrovascular cavity per quadrant; eight adradial canals that originates in the ring canal but do not reach the gastrovascular cavity (Fig. 3i).
Family Rhizostomatidae: Rhopilema hispidum (Vanhöffen, 1888). Exumbrella with dark red spots radiating from the center up to the bell margin; pointy conical protuberances cover the entire exumbrella (Fig. 4e).  Kramp (1961 Bell margin with eight semi-oval velar lappets and one rhopalia per octant, each rhopalia with two tapering lappets (Fig. 4f). Eight oral arms with multiple short spatulate shape terminal clubs, two times larger attached at the distal portion of the oral arm. Each oral arm has two scimitar-shaped attached to the smoot portion (Fig. 4e). Scapulae with filaments. Subgenital ostia with a flattened oval papillae (Fig. 4g). Canal system with one perradial, three interradial, and two adradial canals that join the gastrovascular cavity per quadrat. Adradial anastomoses of Cyanea sp.; e tentacle horse-shoe shape clusters; f close-up of the rhopalium embedded at ¼ distance from the bell margin. Scale bar = 10 mm. Abbreviation: cm, coronal muscles; g, gonads; m, septae; l, lobes; r, rhopalium; rl, rhopaliar lappets; rm, radial muscles; t, tentacles; vl, velar lappets make a complex mesh that includes multiple (> 179) sinuses per quadrant (Fig. 4h).

Molecular characterization
We amplified 23 COI sequences. The resultant alignments per genus did not show sequence difference larger than 0.5%. The BlastN search result for the consensus sequences per genus confirms the morphological identification at the genus level (Table 2).

Environmental parameters
The analyses of the environmental parameters between the two transects (T-1, 0.5 km and T-2, 1.5 km) show that pH, salinity, flow velocity, and lucidity are statistically different between both transects (Table 4). There is no statistical difference between the temperatures between both transects (p = 0.19).
Discomedusae identification can be problematic due the high incidence of cryptic species (Dawson 2003(Dawson , 2005bGómez Daglio and Dawson 2017). For example, within the order Semaeostomeae, the genera Chrysaora (Morandini and Marques 2010; Bayha et al. 2017) and Cyanea are cryptic species (Dawson 2005b;Holst and Laakmann 2014). We identify Chrysaora chinensis. C. chinensis is the only pelagiid reported for Indonesia waters (Stiasny 1921: originally named C. quinquecirrha Desor, 1848 (Morandini and Marques 2010)). Other records are from China, Vietnam, and the Philippines (Mayer 1915(Mayer , 1917Stiasny 1940). The individuals collected in this study present some morphological variations with respect to the redescription of C. chinensis published by Morandini and Marques (2010), such as gonads' shape (rounded pockets, Fig. 2c), gastric filaments (absent), and coloration of exumbrella (yellow and brownish, Fig. 2a). The variation in these characters might be the results of the high intraspecific morphological variation of Chrysaora species, as it is demonstrated by other studies in other regions of the world and species (e.g., C. lactea (Morandini et al. 2006) and C. quinquecirrha (Bayha et al. 2017)). The identification of individuals for the family Cyaneidae is done to the genus level. The individuals collected were broken or in poor conditions, which make difficult to account for a reliable morphological characterization. Cyanea buitendijki Stiasny, 1919 is the only species described for the area (Malayan Archipelago; Stiasny and Van der Maaden 1943). The poor conditions of the samples and the lack of molecular data (our study or published) did not give enough characters to provide an accurate identification for this species.
We identify five rhizostome medusae. The specimens of C. townsendi show the diagnostic characters (Fig. 3d-c) described by Mayer (1915) and Stiasny (1921). This species has been reported for Malayan Archipelago, South East Asian ocean, and Indonesia (Stiasny 1921;Stiasny 1922;Ranson 1949;Rizman-Idid et al. 2016). The alignment from BlastN (Table 2) and the phylogenetic analyses made by Gómez Daglio and Dawson (Fig. 2 pag. 644; corroborates the identification as a Catostylus. A. flagellatus have been reported previously for the Malayan Archipelago and Java Sea (Stiasny 1921(Stiasny , 1922Rizman-Idid et al. 2016). The individuals show the diagnostic characteristic of the species: intermediate filaments in the oral arms with a pointy terminal club (Fig. 3a) and canal system with a heart shape papillae on each perradial canal (Fig. 3c). The alignment comparison (Table 2) shows a ~ 90% identity from other individuals collected from China and Malayan Archipelago. That might suggest that the individuals collected in Surabaya might represent a different species from the individuals collected by Rizman-Idid et al. (2016) in the Malayan Archipelago. A description of specimens from the Malayan Archipelago and the recorded distribution of the species (China, Taiwan, Japan, Indian-Ocean, and Borneo (Kramp 1961)) are needed to clarify the identification. Table 4 t-test results comparing the environmental parameters for transects 1 and 2 (see Fig. 1 and Supplementary Material Table S2). For all the analyses N = 11, * = p statistically significant. T-1 transect 1 (0.5 km from the coastline); T-2 transect 2 (1.5 km from the coastline) The identification of Lychnorhiza malayensis relies on the comparison of the morphological data with the original description (Stiasny 1920(Stiasny , 1932. However, some morphological variations were noticed with respect reports from different geographic areas (India and Pakistan (Menon 1930;Gul and Osmany 2017). According to Abboud et al. (2018), there is a high likelihood that different species of Lychnorhiza might occur on small geographic ranges. To define the intraspecific variation and corroborate the identification with molecular data, we cannot discard the possibility of the presence of multiple species in the coast of Surabaya.
Phyllorhiza individuals were identified as P. cf. pacifica. The alignment of COI (Table 2) shows a high similarity with P. punctata von Lendenfeld, 1884 reported for the Strait of Malacca (Rizman-Idid et al. 2016). However, the morphological characters (terminal clubs and lappets number and shape, Fig. 4a-c) do not resemble the original description of P. punctata. In addition, these samples were analyzed under a phylogenetic context; the individuals from Surabaya belong to a different clade with respect to P. punctata (see Gómez Daglio and Dawson 2017: Fig. 2: pp 644).
Rhopilema hispidum is morphologically similar to those recorded in Pakistan (Gul and Morandini 2015) and the strait of Malacca (Rizman-Idid et al. 2016). The BlastN alignment shows a ~ 94% of similarity with the documented sequences from Malayan Archipelago. R. hispidum presents a broad distribution in the western Pacific (Japan, China, Philippines, Indonesia, Indian Ocean, and the Red Sea (Kitamura and Omori 2010;Gul and Morandini 2015;Rizman-Idid et al. 2016)). There are no molecular data available to compare our samples with individuals from other geographic locations.

Distribution and seasonality
The distribution and seasonality of jellyfish have been attributed to changes in environmental conditions, anthropogenic activities, and intrinsic factors (e.g., life cycle, reproduction; Suchman et al. 2012;Pikesley et al. 2014;Ceh et al. 2015;Quiñones et al. 2015). Our surveys from November to December (2010) show a higher incidence of jellyfishes close to the shoreline (Transect 1, Table 3). Transect 1 stations (0.5 km from the coast) are exposed to tidal and current conditions, whereas Transect 2 stations (1.5 km) are not highly influenced by those parameters. The environmental parameters in transect 1 are statistically different to transect 2 (offshore), except for the temperature. That suggest that the environmental conditions limit the distribution of jellyfishes close to the shore, and rhizostomes (A. flagellatus, L. malayensis, and R. hispidum) are present in transect 1 and 2 (Table 2); meanwhile the semaeostomes (C. chinensis and Cyanea sp.) only in transect 2. The correlation between the presence of jellyfish and environmental parameters shows a high correlation (Purcell et al. 2000;Lee et al. 2013;Quiñones et al. 2015;Kienberger and Prieto 2017), where the salinity, temperature, and currents are the main drivers of its presence.
The environmental factors solely cannot be the only explanation for the high incidence of rhizostomes close to the shore, and the biology of the species (feeding mechanisms and behavior, diet, life cycle, and reproduction) plays an important role in the distribution and presence of jellyfishes. Rhizostomes are shallow-water, coastal, and tropical-temperate species; their differences in lifestyle with respect to semaeostomes, such as the mutualistic association with symbiotic zooxanthella (e.g., Phyllorhiza spp. and Catostylus spp. (Dawson and Hamner 2009), their feeding behaviors (filter feeders) (Arai 1997)), could explain their presence and high occurrence close to the shore. On the other hand, semaeostome jellyfish are voracious predators on zooplankton (Purcell 2003), and their life cycle does not always follow the metagenesis-alternation between sexual and asexual generations (Ceh et al. 2015, but see Morandini et al. 2016). For the species recorded here, there are no information about the biology of the species and their seasonality. The information available about the Discomedusae distribution and abundance is scarce, and the information comes from the fisheries (landing values). However, these data are unreliable, due to the high incidence of misidentification of the species (Gul and Morandini 2015;Rizman-Idid et al. 2016).

Conclusion
Herein we identify seven species of Discomedusae and provide the first records for R. hispidum, A. flagellatus, C. townsendi, L. malayensis, and P. cf. pacifica along the coast of Surabaya. The taxonomic knowledge for this area is still patchy, and misidentifications occur oftentimes. We highlight the importance to document the taxonomic diversity in the area, using an integrative approach (e.g., morphological and molecular data), that proves to solve part of the taxonomic uncertainty. Indonesia, as a country with a high fishery activity, requires a good documentation of its unique biodiversity, tools that help to identify the species, and a continuous monitoring of the targeted species. In addition, to understand the seasonality and distribution of the jellyfish, it is necessary to grow the body knowledge of the biology of the species that should include the reproduction, feeding behavior, diet, and life cycle. This knowledge will help to document and understand the ecological patterns of these economic important species.