Distribution and abundance of planktonic mollusks along a longitudinal gradient in the Southeastern Pacific off Chile

Resumen.- Los objetivos de esta investigación fueron estimar la abundancia de los principales grupos de moluscos planctónicos (larvas meroplanctónicas, gasterópodos holoplanctónicos y paralarvas de cefalópodos) y relacionarla con las propiedades fisicoquímicas del agua, a lo largo de un gradiente longitudinal entre Caldera en la costa continental de Chile y la ecorregión de Isla de Pascua (Isla de Rapa Nui e Isla Salas y Gómez), en el sureste del océano Pacífico. Se recolectaron muestras de plancton durante el Crucero CIMAR 21-Islas, en octubre y noviembre de 2015, en 33 estaciones oceanográficas con una red WP2 (180 µm de abertura Abstract.- The objectives of this research were to estimate the abundance of the main groups of planktonic mollusks (meroplanktonic larvae, holoplanktonic gastropods and cephalopod paralarvae), and relate these groups to the physical-chemical water properties along a longitudinal gradient between Caldera, on the coast of mainland Chile, and the Easter Island ecoregion (Rapa Nui Island and Salas y Gómez Island), in the Southeast Pacific Ocean. Plankton samples were collected over the course of the CIMAR 21-Islas Cruise, from October to November 2015, at 33 oceanographic stations via vertical hauls of a WP2 net (180-µm mesh size) from a maximum depth of 300 m to the sea surface. Mollusks were sorted, counted and initially assigned to Class rank, later being identified to lower taxonomic ranks. Planktonic mollusks were obtained at all stations, and were composed of 92.7% of Gastropoda and 7.3% of Bivalvia. The total abundance of mollusks varied between 55 and 4,922 individuals 100 m -3 . Euthecosomate gastropods exhibited the highest occurrence within the oceanic area. Meanwhile, no paralarvae were captured. Differences in the composition of planktonic mollusks between the continental and oceanic zones were evident. Bivalve larvae increased their abundance in warmer, salty and vertically mixed waters. These results are the first record of meroplanktonic mollusks in waters near the Chilean oceanic islands, and suggest that planktonic mollusks display spatial variation at the scale of the South Pacific Basin, which could be related to the hydrographic conditions and the water column structure.


Introduction
The Southeast Pacific region is characterized by a remarkable zonal gradient in both productivity and oceanographic conditions, from the highly productive and cold Humboldt Current ecosystem (>20 mg m -3 Chl-a) in the coastal zone to the oligotrophic/ultra-oligotrophic South Pacific central gyre (<0.02 mg m -3 Chl-a) (von Dassow & Collado-Fabbri 2014, Meerhoff et al. 2018, Espinosa-Leal et al. 2021. These contrasting water mass properties across the zonal gradient may exert a strong influence on the distributional patterns of planktonic organisms (Espinosa-Leal et al. 2021).
Planktonic mollusks play an important role in ocean trophic dynamics, by supporting secondary consumers which also inhabit the water column (Young & Chia 1987), as well as in the biodiversity conservation, since their absence, presence or abundance in a given place can be used as indicators of environmental change (Shelton 2014). In addition, the geographic dispersion of adult populations may reflect characteristics of the life history of the species (Botsford et al. 2001, Shanks 2001, Young 2002, Trakhtenbrot et al. 2005. Most benthic mollusks develop larvae in the water column as components of meroplankton, remaining as part of the planktonic community for a few hours up to months, or even years, depending on the species, environmental conditions and food availability (Scheltema 1986a, Pechenik 1987, Strathmann & Strathmann 2007. Most of the mollusk larvae are either actaeplanic or teleplanic (Strathmann & Strathmann 2007), the latter having the ability to be transported long distances (Shanks 1995). On the other hand, gastropod mollusks of the Order Pteropoda and the Superfamily Pterotracheoidea (=Heteropoda) develop their whole life cycle (larva, juvenile and adult) in the water column, forming part of the holoplankton. These gastropod groups are of particular importance because some species act as sensitive indicators of water masses (Chen & Hillman 1970) and climate change (Hughes 2000), as well as playing an important role in the marine food web (Wall-Palmer et al. 2016). Cephalopod paralarvae and juveniles also form part of the planktonic mollusk community. Along with the other planktonic mollusks, they play a key role in marine food webs and have been considered potential indicators of environmental change (García-Guillén et al. 2018).
Planktonic mollusks from the Southeast Ocean off north-central Chile have mainly been the subject of general zooplankton studies or the identification of selected species (Fagetti 1958(Fagetti , 1968Fagetti & Fischer 1964, Palma et al. 1976, Prado 1983, Campos & Ramorino 1984, Palma & Rosales 1995, Vega et al. 1999, Poulin et al. 2002aRomero & Valdebenito 2002, Labrín et al. 2015, Torreblanca et al. 2016, Carrasco et al. 2019. However, the knowledge concerning the distribution and abundance of these groups is very scarce especially in oceanic areas. It was hypothesized that the main groups of planktonic mollusks would exhibit distribution and abundance patterns along the longitudinal gradient between Caldera and the Easter Island ecoregion (Rapa Nui Island and Salas y Gómez Island) according to the environmental conditions of the water column (i.e. groups related to coastal/productive waters, oceanic/oligotrophic waters, and around islands). The objectives of this study were to determine the abundance of main planktonic mollusks groups along a longitudinal transect between Caldera and the Easter Island ecoregion, and to relate their distribution with the physical-chemical characteristics of the water column.
The continental coast of Caldera is influenced by Equatorial Subsurface Water and upwelling processes, generating a system high in nutrients. While towards Rapa Nui, the mesotrophic transition and oligotrophic zones are composed by Sub Antarctic Water and Subtropical Water, respectively, with lower nutrient concentrations compared to the coastal zone (González et al. 2020). The sea surface water temperatures and salinities, from the coast to the open ocean, range from 14.9 to 21.4 °C and from 34.7 to 35.9, respectively. A well-oxygenated surface layer (>5 mL L −1 ) has been recorded all along the latitudinal gradient, however its depth ranges from ~35 m, near the coast, to ~215 m, in the open ocean (Espinosa-Leal et al. 2021).
Mollusks were obtained from 60 plankton samples collected at 33 sampling stations ( Fig. 1). At each station, samples were obtained by means of vertical tows of a WP2 net (180-μm mesh size), equipped with a TSK flowmeter, down to a maximum depth of 300 m. The collected samples were preserved onboard with 96% ethanol. Once in the laboratory, this fixative was replaced with a solution of 96% ethanol with sodium glycerophosphate (pH> 8.5). Each mollusk was manually removed from the sample under a Leica MZ8 stereo microscope and preserved in 96% ethanol with sodium glycerophosphate. Every specimen was first assigned to one of the following Classes: Bivalvia or Gastropoda. Since gastropods have both holoplanktonic and meroplanktonic representatives, specimens of this Class were classified into three groups: 1. Meroplanktonic larvae (from benthic species), 2. Order Pteropoda and 3. Superfamily Pterotracheoidea (=Heteropoda), based on Bouchet et al. (2017).
The abundance of mollusks was standardized in number of individuals · 100 m -3 (indiv. 100 m -3 ). Salinity, water temperature and dissolved oxygen data were recorded using a Seabird SB-19 CTD (conductivity/temperature/ depth profiler) at each sampling station. The environmental conditions were also characterized by vertical temperature, salinity and oxygen gradients, ΔV= V sup -V 200m , where V sup is the surface value and V 200m is the value measured at 200 m depth for each of the corresponding variables. Prior to the correlation analyses between environmental variables and standardized mollusk abundances, the normality of the data was checked by means of the Shapiro-Wilk test. Since most of the analyzed variables did not fulfill the normality assumption (P < 0.05), the Spearman rank correlation was applied.
Adobe ® Photoshop CS6 13.0 was used for editing the digital images of the mollusks. Global Mapper (Blue Marble Geographics) 1 and Surfer ® 10 Golden Software, Llc; USA) software programs were used for graphics.
Comparisons were made between groups of stations, which were divided into oceanic waters (O), Rapa Nui Island (RN) and Salas y Gómez Island (SG). The standardized abundances of heteropods, pteropods and gastropod larvae were utilized for comparisons. A Principal Component Analysis (PCA) was applied to compare the mollusk composition among locations and was tested via a oneway PERMANOVA (Anderson 2001), using Bray-Curtis similarity index. All statistical analyses were carried out using Past 3.26 software (Hammer et.al 2001).

Distribution and abundance
Planktonic mollusks were collected at all the stations, adding to a total of 12,761 individuals.

Multivariate analyses
The PCA was explained by three components. PC1 described 74.4% of the variance; PC2 explained 16.3%, while PC3 illustrated only 9.3% (Fig. 8). PC1 was explained by the variability in the abundances of both pteropods (loading factor= 0.798) and gastropod larvae (loading factor= 0.601). PC2 showed variability in the stations around SG, and was positively correlated to larvae (loading factor= 0.661) and heteropods (loading factor= 0.580) (Fig. 8). All the stations executed around Rapa Nui (RN) showed similar compositions, e.g., all sampled stations occurred in a similar multivariate space in the PCA (Fig. 8). One-way PERMANOVA detected significant differences among locations (pseudo-F= 3.24, P = 0.005). Post-hoc analysis showed that, at a major taxonomic group scale, composition and abundance of planktonic mollusks were different between SG and O (Bonferronicorrected P = 0.011), RN and SG (P = 0.040), but not between RN and O (P = 0.743). Therefore, the planktonic mollusk community along the longitudinal gradient was heterogeneous and differentially represented.

Relationship with environmental variables
The standardized abundance of total mollusks did not significantly correlate with any of the environmental variables considered in this study (P > 0.151), nor did the gastropods (P > 0.177). However, the standardized abundance of bivalve larvae showed significant positive correlations with the mean water temperature (rs= 0.584, P < 0.001), the sea surface temperature (rs= 0.568, P = 0.001), the mean salinity (rs= 0.704, P < 0.001) and the sea surface salinity (rs= 0.626, P < 0.001). On the other hand, the standardized abundance of bivalve larvae showed a negative correlation with the vertical temperature gradient (rs= -0.541; P = 0.002) and the vertical dissolved oxygen gradient (rs= -0.542, P = 0.002). In summary, the standardized density of bivalve larvae increased in warmer, salty and vertically more mixed waters, while the hydrographic conditions and the water column structure did not have any significant effect on the abundance of total mollusks or gastropods.

Distribution and abundance
The spatial distributional patterns of planktonic mollusks clearly exhibited a longitudinal gradient along the South Pacific, from continental coastal waters to oligotrophic oceanic waters, and around the oceanic islands, indicating that both adult and larvae are able to live in different oceanographic conditions across the basin. Meroplanktonic larval distribution could be explained by oceanographic and biological processes occurring in coastal waters, the open ocean and around the oceanic islands. Larvae could be transported off the coast by mesoscale activity in the mesotrophic coastal transition zone (CTZ), which is principally characterized by northwest travelling eddies (Hormazabal et al. 2004, González et al. 2021). On the contrary, the wind-influenced coastal circulation would prevent larval transport out to the open ocean via coastal retention processes (Marín & Moreno 2002, Aiken et al. 2011. The onshore-offshore temperature gradient could generate a physical barrier for more temperature-sensitive groups of planktonic mollusks, such as bivalve larvae which was the only group that showed a significant connection with the environmental conditions and structure of the water column. Larvae found in the open ocean are more likely to be teleplanic, which have the ability to cross physicalchemical water barriers and survive even in oligotrophic waters (e.g., Teredinidae bivalve larvae; Scheltema & Williams 1983). Scheltema (1986b) reported a 77.6% occurrence of teleplanic gastropod larvae in the surface waters of the Central Pacific Ocean, compared to the 47.1% of non-teleplanic larvae. Nevertheless, the occurrence of open ocean non-teleplanic gastropod and bivalve larvae, produced by demersal mollusks inhabiting seamounts and hydrothermal vents nearby the study area, cannot be discarded (see Yáñez et al. 2009, fig. 2 for seamounts location, andFitzsimmons et al. 2014, fig. 1B for hydrothermal vents location). The absence of bivalve larvae in several oceanic stations could be explained by the short planktonic period (<4 weeks) of many coastal species (Shanks 2009), whose larvae would have completed their planktonic stage in less time than that required to be transported off the coast.
The high variability in larval spatial abundances along the longitudinal gradient could be explained by the inherently patchy distribution, produced by the limited larval duration and episodic reproduction events exhibited by benthic mollusks (Shanks 2009), along with the water mass distribution. Although increases in larval abundances associated with higher temperatures have previously been recorded in closed coastal systems of Portugal andTunisia (Chícharo &Chícharo 2000, 2001;Ziadi et al. 2015), to date this pattern had not been registered in the oceanic waters of the Pacific Ocean. The greater abundance of larvae found in waters around Rapa Nui could be due to an island-mass effect, which could cause an increase in local productivity as described by von Dassow & Collado-Fabbri (2014) and Meerhoff et al. (2018).

Mollusk groups
Among the holoplanktonic mollusks, the occurrence of the pseudothecosomatous gastropod Desmopterus pacificus Essenberg, 1919 had previously been recorded in Chilean waters, at the beginning of the 20 th century, according to Labrín et al. (2015). Campos & Ramorino (1984)  The pterotracheoid gastropod genus Atlanta, which is mainly distributed in warm waters (Moreno-Alcántara et al. 2014, 2017, 2020Wall-Palmer et al. 2016 In relation to meroplanktonic gastropod and bivalve larvae, this is almost their first record from the coast off Caldera as well as the waters adjacent to the Chilean oceanic islands. The closest record of gastropod larvae to our study area is from the coastal location 30°S, 71°21'W (Romero & Valdebenito 2002). Most of the families to which the larvae have been assigned so far in this study have previously been reported by Scheltema (1986b) for the Central Pacific islands, therefore representatives of these families seem to be common throughout the Pacific Ocean.
The absence of adult stages of some of the holoplanktonic gastropods in the samples as well as of paralarvae could have been largely due to a net effect. The WP2 net (180-200 μm mesh size) is suitable for capturing larval stages of mollusks but would not be as efficient in catching adult stages of thecosomatous gastropods, over 1 mm or cephalopod paralarvae because of their greater capacity to evade the net. Labrín et al. (2015) collected adults of several species of thecosomatous gastropods, using an IKMT net (10 mm mesh size), and Carrasco et al. (2019) captured paralarvae in waters of Rapa Nui Island using nets of 300 μm mesh size. In our research, mainly larval and juvenile stages of some of the genera collected by Labrín et al. (2015) were captured alongside a few adults of some of the species also recorded by these researchers.
The relative dominance of pteropods over the heteropods and meroplanktonic larvae in most stations should be interpreted with caution since the spatial distribution of the planktonic stages depends partly on, the developmental stage of the species, their capabilities to perform vertical migration and their kinetic behavior in response to oceanographic conditions (Young 1995, Kersten et al. 2019, as well as species-specific differences. These trends can also be affected by seasonality (e.g., the sampling period encompassed only the austral spring), mesoscale oceanographic conditions (e.g., shallow oxygen minimum zone near the continental shelf, ultra-oligotrophic conditions in the South Pacific Gyre, island-mass effect around oceanic islands) and interannual variability in environmental conditions (e.g., the presence of the strong El Niño Godzilla during late 2015-early 2016).
In summary, the pelagic mollusk community display spatial variation at the scale of the South Pacific Basin, which can be related to both the hydrographic conditions and the water column structure. A more intense spatial and temporal sampling of planktonic mollusks could help to provide advances in our knowledge of not only their probable origin, but also their contribution to the colonization and connectivity of the oceanic islands and seamounts within the Easter Island ecoregion as well as in the shallower coastal zones.