Species diversity and spatial distribution of pelagic amphipods in Terra Nova Bay (Ross Sea, Southern Ocean)

A greater understanding of biodiversity and the roles of various species involved in Southern Ocean pelagic food webs is needed to predict and hypothesize about responses to future scenarios in relation to climate changes. The aim of this paper was to describe for the first time the composition, relative abundance, spatial distribution and relation with water masses of pelagic amphipods in Terra Nova Bay, Ross Sea, based on stratified sampling. Zooplankton was collected by BIONESS (Bedford Institute of Oceanography Net Environmental Sampling System), during the 1987–1988 Italian Antarctic R/V Polar Queen Expedition. A total of 1331 specimens of pelagic amphipods was counted and 17 species belonging to nine families were identified. Significantly relative higher abundances were related to Modified Circumpolar Deep Water (MCDW) in the upper layer than modified Ross Sea Shelf Water (RSSW) and High Salinity Shelf Water (HSSW). Hyperiella dilatata was the most abundant species (48% of relative abundance), followed by Pseudorchomene plebs (14%), Hyperia macrocephala (8%) and Hyperiella macronyx (6%). Hyperiella dilatata was distributed widely across the study area and showed a link with Antarctic Surface Waters (AASW) and MCDW. Hyperiella macronyx and Pseudorchomene rossi were also distributed widely though were much lower in abundance. Hyperia macrocephala were high in abundance though had a narrow distribution that was linked with AASW. The vertical distributions showed variable patterns for adult and juvenile specimens. This study represents a knowledge base against which to compare more recent studies to highlight any structural changes attributable to ongoing climate change in the Terra Nova Bay and Ross Sea ecosystems.


Introduction
The Southern Ocean is characterized by low water temperatures year-round and seasonal formation of pack-ice on the surface. Sea ice surrounding the Antarctic continent varies in extent from 3.1 × 10 6 km 2 in February to a maximum of 18.5 × 10 6 km 2 in September, making it one of the largest and most dynamic ecosystems on Earth (Arrigo et al. 1997;Eayrs et al. 2019;Parkinson 2019). The processes of ice formation and melting help determine the makeup of planktonic communities (Garrison and Buck 1991;Dieckmann et al. 1998;Gleitz et al. 1998;Granata et al. 2022), particularly those processes that regulate the availability of suspended particles for planktonic and benthic suspension feeders, especially during summer Cau et al. 2021).
Terra Nova Bay is characterized by an extended polynya, which is formed by katabatic winds that blow downslope from the Trans-Antarctic Mountains, resulting in open water that extends through the winter (Kurtz and Bromwich 1985). In this region the polynya is a major driver of ice growth and decay processes, which, in turn, control the growth of coastal sea ice microalgae and the availability of suspended particles in the upper coastal waters of Terra Nova Bay Misic et al. 2002;Pane et al. 2004). The waters of the bay show peculiar physical-chemical features within the Ross Sea system, as summer heating that can cause temperature increases of the surface layers. In Terra Nova Bay, the summer meso-and macrozooplankton is mostly composed of copepods (Carli et al. 2000;Zunini Sertorio et al. 2000;Pane et al. 2004), ichthyoplankton (Granata et al. 2002), with key species including the ice krill Euphausia crystallorophias ), the notothenioid fish Pleuragramma antarcticum (Guglielmo et al 1998;Granata et al. 2009), the mysid Antarctomysis ohlini (Crescenti et al. 2000) and the pteropod Limacina helicina antarctica (Manno et al. 2010). These species represent a fundamental link for predictive distributions , food web structure and modelling (Pinkerton and Bradford-Grieve 2014;Signa et al. 2019).
Amphipods are well represented in all three macrohabitats of Terra Nova Bay: the sea floor with Gammaridea and Corophiidea, the water column with Hyperiidea and pelagic and bentho-pelagic Gammaridea, and the sea ice with epontic Gammaridea (De Broyer and Jażdżewska 2014). Amphipods of the benthic realm have been investigated, although many areas of the continental shelf remain unexplored and the deep sea has been sampled rarely (Zeidler and De Broyer 2009;De Broyer and Jażdżewska 2014;Jażdżewska and Siciński 2017). Amphipods exhibit diverse lifestyles, trophic habits, habitat and size spectra, and they constitute a significant trophic resource for Southern Ocean fishes, invertebrates, seabirds and mammals (Zeidler and De Broyer 2009). Considering the key role of this oceanic sector in the Earth system and the growing impact of global environmental change, it is fundamental to collect information on the status of Antarctic marine biodiversity. It is well-known that the majority of amphipods inhabit the benthic environment, and comparatively few species belonging to a very few families have evolved to colonize the pelagic realm (Zeidler and De Broyer 2009). Very little is known about oceanic species, in particular their vertical distribution as they have rarely been sampled with a stratified multi-net system (Zeidler and De Broyer 2009).
The aim of this paper was to describe for the first time the composition, relative abundance and spatial distribution of pelagic amphipod species in Terra Nova Bay, based on stratified sampling. An additional goal was to describe the main water masses influencing the region and, where possible, relate the species assemblages to these water masses and latitude-longitude gradient in Terra Nova Bay.

Materials and methods
Sampling procedure During the 1987During the -1988 Italian Antarctic Expedition of the R/V Polar Queen, between 5 January and 21 February 1988, zooplankton samples were collected in Terra Nova Bay, Ross Sea (Fig. 1). A total of 31 stations were investigated, where samples were collected in several depth-layers from the surface to near the seabed. All the sampling details are shown in Table 1. The study area extended from Cape Washington (74°39′00''S) to Drygalski Ice Tongue (75°29′00''S), along a longitude from 163°49′00''E to 168°04′00''E. The sampling stations were located on the continental shelf of the Ross Sea, except for two stations (1SK located offshore from the coast and ANZ near Cape Adare, see Table 1). Samples were collected with an electronic multinet BION-ESS (Bedford Institute of Oceanography Net Environmental Sampling System) (Sameoto et al. 1980), which had a mouth area of 0.25 m 2 and was equipped with 10 nets (mesh size 230-μm or 500-μm). The BIONESS continuously measured temperature, salinity and depth during tows with an Applied Microsystems digital CTD. Filtered volume was monitored using external and internal TSK flowmeters. The BION-ESS was deployed at low speed along an oblique path to the maximum selected depth to be investigated, then towed at a speed of 1.5-2 m s −1 until closing, with the simultaneous opening of a new net. The number and the thickness of the sampled strata depended on the bottom depth (Table 1). Fishing upward, the nets were opened and closed on command, at 100-to 20-m intervals in the hauls programmed in the upper 200 m according to different water masses, and between 30-and 100-m intervals below this zone to the maximum depth sampled. Due to technical reasons, it was not always possible to collect the same number of samples at all stations, whereas the lack of surface samples at some stations was our choice for 50-m intervals in the deep hauls. A total of 266 samples (for a total of 9780 m 3 of filtered water) was collected from the 31 sampling stations. The zooplankton samples collected at each station from the surface to the maximum sampled depth (800 m), during the downward deployment of the BIONESS, were useful in lab for rare species potential detection, but not used for this work. On board, all the zooplankton samples were preserved in a 4% buffered formaldehyde-seawater solution ( Table 2).

Zooplankton analysis
In laboratory, subsamples of different volumes, from 1/10 to 1/25 of the original sample, were observed under a Leica Wild M10 stereomicroscope. The whole sample was examined for the identification and enumeration of macrozooplankton and micronekton, as for amphipods. All the specimens of each taxon were counted and identified at higher taxonomic levels, while diagnosis at the species level was undertaken for the amphipods. The abundance of amphipods for each stratum was calculated by dividing the total number by the filtered water volume and expressed as individuals 100 m −3 . For the entire water column the total abundance of amphipods at each station was expressed as a weighted mean, summing all amphipods counted dividing by the total volume of seawater filtered (m 3 ), multiplied by the water column thickness and expressed as individuals m −2 . The mean total abundance of a given species for the 31 sampled stations is represented by the geometric mean and expressed as individuals m −2 , to minimize the difference in water column depth.

Statistical analysis
Temperature and salinity data collected during the BION-ESS samplings were processed with Ocean Data View (ODV version 5.5.2) software (Schlitzer 2001) to map the position of each of the 266 datapoints on the ϴ-S diagram and assign them to a specific water mass. The 'elbow' Completeness of the samples was evaluated according to the standardization approach proposed by Chao and Jost (2012). To correctly compare informative though uneven subsample aggregations, rarefaction/extrapolation curves based on the reference samples were used (Colwell et al. 2012;Chao et al. 2014). Expected species richness was calculated through the Chao1 estimator (classic formula) and first order jackknife analysis. Asymptotic species richness in subsample aggregations were obtained with the R package iNEXT (Hsieh et al. 2016). Taking into account the sampling strategy, amphipod density and diversity across relevant factors (six levels for the factor "depth range", four levels for the factor "water masses", see Table 3) were examined using R version 4.0.3 to test for significant differences between and within groups of data (ANOVA and Kruskal-Wallis non-parametric tests on medians in case of non-homogeneity of variances, coupled with post-hoc Dunn test). Indicator species analysis (IndVal) has been used to identify species that were representative of different sample aggregations (layers, water masses) and inform about their "specificity" (i.e., degree of exclusivity) and "fidelity" (i.e., probability of occurrence) as indicators of groups of samples throughout the water column (Dufrene and Legendre 1997;Podani and Csanyi 2010). The function 'multipatt' of the R package indicspecies has been used.

Environmental parameters
The ϴ-S scatter plot revealed four water masses present on the Ross Sea Shelf at depth < 700 m (Fig. 2). These were: (1) AASW with lower salinity, temperature often exceeding 0 °C and density anomaly σ0 < 27.75 (γ n ≈ 28 kg m −3 ); (2) the MCDW that includes a relatively warmer and more saline core delimited by the 28.27 kg m −3 neutral density γ n surface and, below it, (3) the RSSW modified by the interaction with coastal processes (such as e.g. upwelling due to katabatic winds) and glacier presence. In this region of the ϴ-S plane, a more saline (S > 34.75) water mass with a temperature close to the freezing temperature at the surface (ϴ ≈ − 1.85 °C) was found. The signature of a (4) slightly colder HSSW (ϴ ≈ − 1.95 °C, S > 34.62, depth 400 m) was also noted, which was formed by the deep melting of the Drygalski Ice Tongue and is a precursor of the Deep Ice Shelf Water (DISW).

Pelagic amphipod composition and abundance
Overall, about 25% samples (66 out of 266) were empty (they did not contain amphipod specimens) regardless of the filtered water volume. A total of 1331 pelagic amphipods were counted in the whole study area, representing 0.09% of A total of 20 taxa were recovered in Terra Nova Bay (Table 2). We identified 17 species belonging to nine families, two groups identified to genus (Epimeria sp. and Pseudorchomene sp.) and one taxon at family level (Acanthonotozomatidae).
Observed richness, grouping samples by depth range and by water mass, is shown in Table 3. The sampling effort was suitable to describe in detail the amphipod community in the different water masses (coverage > 99%) with only a minor reduction below 400 m depth. The reference overall sample (n = 1331 identified specimens, including 7 singleton species) exhibited a strong degree of completeness (99%) efficiently representing the amphipod community in Terra Nova Bay in summer. However, the overall species accumulation curve did not reach an asymptote within the size of the reference sample nor in the extrapolation to a 2-time size, suggesting a remarkably greater species richness for amphipods in the whole area, on the order of at least 50 species (32 ± 13 s.e. species, Chao1 estimator, classic formula). The RSSW showed the greatest richness, higher than MCDW with which there is overlap between the 95% confidence intervals (Fig. 3). The lowest diversity can be assigned to AASW, which is possibly dependent on the extremely uneven sampling in these deep waters. HSSW contained about 50% of the overall richness, with nine species (Hyperiella dilatata, Hyperiella macronyx, Pseudorchomene rossi, Pseudorchomene plebs, Pseudorchomene sp., Orchomenyx macronyx, Epimeria sp., Eusirus antarcticus, Halice tenella) recorded out of 20, none of them exclusive to this water mass, except the single observation of H. tenella. Moreover, while the extrapolated trends in richness for RSSW and MCDW did not plateau at a size near the reference sample and far beyond, HSSW appeared to have reached its asymptotic richness.

Pelagic amphipod spatial distribution
Hyperiella dilatata had the widest distribution among all the recorded species, with a percentage frequency of occurrence of 100%; i.e., it was observed at all 31 sampling stations ( Table 2). Some species, including H. macronyx and P. rossi, were present in low abundance although showed a wide horizontal distribution, while others, including H. macrocephala had high abundance but were narrowly distributed (St. 17, 35, ANZ only).
Comparison of species richness between the stations, showed that St. 34 had the highest richness, seven species, further to the presence of specimens belonging to the genus Epimeria and to the family Acanthonotozomatidae. St. 11 showed the lowest richness, with only H. dilatata and one specimen of Pseudorchomene sp. recorded.
Different patterns of vertical distribution were shown for the four most abundant species, and between adult and juvenile specimens (Fig. 6). The main distribution (74%) of H. dilatata adults occurred in the upper 50-m, with very few specimens until 700 m in depth. In contrast, H. dilatata juveniles has their centre of distribution lower in the water column, where 85% of the population occupied the 200-500 m layer, (range 50-600 m). For P. plebs, similar patterns of vertical distribution were shown by the adults and juveniles. The main part of the population (79% of adults and 93% of juveniles) occupied the same layers from the surface to 250 m. Few P. plebs adults were recorded between 300 and 600 m, while all juveniles were shallower than 350 m. The third most abundant species, H. macrocephala, was recorded above 30 m depth, except for 1 specimen occupying the 220-260 m layer m at St. ANZ. H. macronyx adults and juveniles showed different patterns in their vertical distributions: 74% of the adults occupied the 0-150 m layer, with a total range from 0 to 450 m, while all the juveniles were concentrated between 100 and 350 m.
The IndVal analysis performed on the amphipod assemblages, identified the most characteristic species of the different water masses (Table 4). In the very surface layer the clearest indicator species was H. dilatata, by far the dominant species in the area, with an occurrence of 100% in the AASW + MCDW water mass (fidelity = 1). Representative of the AASW, with a high specificity but quite low occurrence, was H. macronyx that could prefer a deeper habitat linked with AASW. In the subsurface photic layer the strongest indicator was Pseudorchomene sp. Interestingly, no species emerged as a significant indicator of either HSSW or RSSW.

Discussion
Pelagic amphipods are often not dominant in terms of abundance in zooplankton samples, but they do have a substantial role within the Antarctic food web (Pinkerton et al. 2010;Havermans et al. 2019). They are important in recycling of organic matter and are a prey item for many fish species, higher marine predators and migratory seabirds (DeWitt and Hopkins 1977;Eastman 1985a, b;Stowasser et al. 2012;Xavier et al. 2018;Waluda et al. 2012). Trematomus bernacchii, T. newnesi and T. pennelli show a wide trophic spectrum, consisting mainly of amphipods, like Pseudorchomene plebs, Debroyerella fougneri, specimens of Eusiridae, Hyperiidae and Acanthonotozomatidae families, along with gastropods and polychaetes (Richardson 1975;Vacchi et al. 1994;Vacchi and La Mesa 1995;La Mesa et al. 2004).
Recent importance was also given to amphipods in Southern Ocean for microplastic uptake phenomenon by lower trophic organisms that has been hypothesised as a conduit for microplastics up the food-chain. Evidence for this has been founded on positive identification of microplastics/ fibres in the scats of zooplanktivore higher predators (Jones-Williams et al. 2020).
In Terra Nova Bay polynya, the omnivorous ostracods and amphipods were restricted below the pycnocline and the plankton ecosystem is dominated overall by macrozooplankton food-chain (diatoms, molluscs, amphipods), and larval stages of Pleuragramma antarcticum representing the carnivorous level (Hecq et al. 2000). In the southern area of Terra Nova Bay, P. antarcticum old postlarvae and juveniles were more abundant than in the norther one and the seem to 1 3 feed in the water column where Calanoides acutus, Ctenocalanus vanus, Metridia gerlachei copepodites, pelagic amphipods and young euphausiids were abundant inside zooplankton community (Guglielmo et al. 1998). In addition, two species of amphipods (Hyperia macrocephala and Eusirus antarcticus) and eggs with embryos of P. antarcticum were found in the platelet ice (Guglielmo et al. 2007).
In the present study a high biodiversity of pelagic amphipods off Terra Nova Bay, Ross Sea, was found. All the collected species were already cited as present in the pacific Sector of the Southern Ocean, particularly in the Ross Sea, except for Cheirimedon femoratus, Orchomenyx macronyx and Debroyerella solidus. All the species were also reported as found in the Atlantic sector, Weddell Sea, except Hyperiopsis australis, Halice tenella, C. femoratus, O. macronyx, D. solidus (De Broyer et al. 2007;Zeidler and De Broyer 2009).
Hyperiidae was the most abundant family (62% of relative abundance) with the dominant species being Hyperiella dilatata. Zeidler and De Broyer (2009) indicated this species to be cosmopolitan in the Southern Ocean, widely distributed in Atlantic, Indian and Pacific Sector (Stebbing 1888;Weigmann-Haass 1989;Jażdżewski et al. 1992;Dinofrio 1997;Vinogradov 1999). This species was reported in early years as the most abundant and frequent in Ross Sea and considered an important prey for many species of fish, as above mentioned (Barnard 1930;Foster et al. 1987;Hubold 1992;Guglielmo et al. 1998Guglielmo et al. , 2011La Mesa et al. 2004). The family Hyperiidae was represented by other two species in the present study, namely Hyperia macrocephala and Hyperiella macronyx. Both these two species show wide distribution in the Southern Ocean, having been found in the Atlantic, Indian and Pacific sectors, and their presence recorded in the Ross Sea over many decades (Barnard 1930;Hempel et al. 1983;Weigmann-Haass 1989;Jażdżewski et al. 1992;Browne et al. 2007).
The family Tryphosidae was the second most abundant family and was represented by six species, with Pseudorchomene plebs the most abundant. This species is also distributed widely in the Southern Ocean and has previously been recorded in Ross Sea (d'Udekem d'Acoz and Havermans 2012;De Broyer et al. 2007). The other five species in this family were encountered only occasionally, because they are not pelagic (De Broyer et al. 2007). P. plebs has been recorded in substantial numbers at the ice-water interface, where it probably feed on under-ice algae, unlike hyperiid species that live in free waters (Arndt and Swadling 2006). P. plebs is reported to be a facultative scavenger (Stockton 1982;Arndt and Swadling 2006), as some individuals taken from beneath the Ross Ice Shelf had copepod hard parts in the guts, while stomach contents of hatchlings indicated the presence of bacterial aggregations.
Comparing species richness in the different water masses suggests that the mixed waters on the Ross Sea Shelf (RSSW) encompass the main core of the community in Terra Nova Bay (18 of the 20 species observed), though there is important input from the offshore community that is associated with the MCDW. A probable role can be assigned to HSSW that could feed the deeper basin of Terra Nova Bay with adult or juvenile forms of ice-related species from the Drygalski area, so providing a deep path of remote connectivity, especially during melting.
Hyperiella dilatata was very abundant in the sampling region; some individuals in this species exhibit an interesting behaviour that guards against predation. In situ observations have shown specimens of H. dilatata carrying a gymnosome pteropod, Clione limacina antarctica, on their dorsal side, firmly holding it with their sixth and seventh pereopods (Havermans et al. 2018). In the under-ice environment around the Antarctic continent, H. dilatata is a regular component of zooplankton communities and is highly preyed upon by fish and seabirds (Foster 1987;Vacchi and La Mesa 1995;Pakhomov et al. 1999;Havermans et al. 2018). The amphipod incurs an energetic cost of carrying the pteropod but gains the benefit of chemical protection against predators (McClintock and Janssen 1990). Clione limacina antarctica produces an efficient predator-deterrent chemical called pteroenone, a β-hydroxy ketone that has been isolated, characterized and synthesized (Bryan et al. 1995;Asao et al. 2010). The defence mechanism is more suitable for female carrying eggs, because the effort of carrying a pteropod may result in protection for both the female and the offspring  (Havermans et al. 2018). In the present study many embryonated eggs, previously attached to female specimens were detected, especially in the 270-310 m layer at St. 17, along with ovigerous females, e.g. in the 400-500 m layer at St. 5. It has been reported that abducted pteropods could have been released by the amphipods during the sampling process (Havermans et al. 2018), in which could explain the absence of tandems in our samples.
The vertical distributions of the four dominant species in our study confirm patterns that have been observed previously for the Ross Sea (De Broyer et al. 2007;Zeidler and De Broyer 2009). Hyperiella dilatata has been reported as occurring from 3700 m to the surface, although it seems to be most common in the shallower waters (0-300 m); for P. plebs the range is from the surface to 550 m; H. macrocephala is reported mainly in near-surface water; and H. macronyx from near the surface to 600 m depth. No evidence of diel vertical migration was detected with this study, and the sampling period was characterized by the constant presence of the sun. The centres of distribution for juvenile H. dilatata and H. macronyx were deeper than those observed for the adults. This might be a mechanism whereby food resources, in the form of particulate organic matter, are separated between adults and juveniles of similar species to reduce competition.
In conclusion, summer pelagic amphipods within Terra Nova Bay are represented by 17 species belonging to nine families. Only a few common species, among which just one (H. dilatata) is represented across the entire study area, particularly in surface waters and is related to AASW and MCDW water masses. Many studies conducted in recent years have confirmed that many areas of Antarctica are changing, even if ecosystems seem to respond differently to the stresses of climate change (Fraser et al. 2022;Swadling et al. 2023). A greater understanding of biodiversity and the roles of various species involved in Southern Ocean pelagic food webs is needed to predict and hypothesize about responses to future scenarios. Some important key species that are common to many Antarctic ecosystems and important in carbon cycling processes, have received very little attention, producing substantial gaps in our understanding of Antarctic ecosystem functioning. As Smith Jr et al. (2017) stated "Understanding the nature of the ecological changes that will occur in the Ross Sea in future years will be a major challenge for both oceanographers and polar ecologists".
Although our data refer to a study carried out 35 years ago we believe that, due to the scarcity of studies on pelagic amphipods, the data presented here highlight a need for further studies on the biodiversity and structure of the pelagic trophic web. These data also present a knowledge base against which to compare more recent studies to highlight any structural changes attributable to ongoing climate change in the Terra Nova Bay and Ross Sea ecosystems.
Acknowledgements Logistical and financial support for this paper was given by the Italian National Program of Research in Antarctica (PNRA). A special thank goes to prof. Giuseppe Costanzo, now retired from University of Messina, teacher and mentor, that has spent more of his life to scientific research on zooplankton taxonomy and ecology. Many thanks go to officers and crew of the research vessel Polar Queen and to technicians involved in the BIONESS sampling procedures for their excellent cooperation in field work.
Author contribution RM conceived experimental design, lab procedure, and wrote the ms; AB carried out statistical elaboration, water mass analysis, and wrote the ms; LG conceived the experimental design, wrote the ms; KS wrote the writing and done the language revision; AG and FV carried out the lab procedure; AG conceived the experimental design, supervised the lab procedure, wrote the ms. All authors read and approved the manuscript.

Data availability
The data presented in this study are available by Roberta Minutoli, rminutoli@unime.it.

Declarations
Conflict of interest There are not conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.