Dominant organisms in each site varied between ophiuroids, holothuroids, demosponges, and pycnogonids ( Figure 2B and Table 2 ). Anthozoans, ophiuroids, and ascidians were ubiquitous at all sites, but ophiuroids dominated the Montravel Rock, Seymour Island, and the Bellingshausen Sea sites. Based on the cluster and SIMPROF analysis of benthic communities, we grouped the sites into three regions: the Peninsula region, the Amundsen region, and the Ross Sea region ( Figure S1 ). There were 28 photos (16 Amundsen, 11 Ross Sea, and 1 Peninsula) that did not cluster within the same regions as the rest of the photos from their sites providing evidence of within group dissimilarity ( Figure S2 ). There were 16 photos from the Amundsen in a deeply branching cluster with 15 photos from the Ross Sea indicating the lack of similarity within this particular group indicating patchiness along the transect. These Amundsen region photos were characterized by reduced numbers of cnidarians, ophiuroids, and echinoids compared to other photos of the region. The 11 Ross Sea photos that clustered within the Amundsen Region had less annelids, holothuroids, and pycnogonids than the rest of the photos in the Ross Sea and Recovitza Island region. Generally, sites cluster by relatives location according to epifaunal community assemblages ( Figures 3A , S2 ): the Amundsen region includes the Eastern Amundsen, Bellingshausen, and Wright’s Gulf sites, the Ross region includes the Mid Ross, Near Ross Shelf, and Recovitza Island sites, and the Peninsula region includes the Seymour Island and Montravel Rock sites ( Figure 3A ). Pycnogonid sp. 1, pycnogonid sp. 2, Ophionotus victoriae, Sterechinus sp. 1, Ophioplinthus spp., Iophon sp., and the yellow branching sponge ( Figures 1A, C ) were the primary species that helped differentiate the three regions we define here ( Figures S3A, B ). Based on the cluster and SIMPROF analysis ( Figure S2 ), Recovitza Island communities were more similar to the Ross Region sites and have been included with them in the remainder of the analysis although we note some degree of within group dissimilarity ( Figure S1 ). We saw a higher prevalence of echinoids (e.g., Sterechinus spp.) and holothuroids (e.g., Rhipidothuria racowitzai and sediment dwelling holothuroid 2) in the Amundsen Region than the other two regions. The Iophon sp. (demosponge), and Ophioplinthus spp. (ophiuroid) symbiont pair dominated the Peninsula Region and was significantly less abundant in all other regions ( Table 2 ). The brittle star Ophionotus victoriae is abundant throughout the Southern Ocean, but it typically dominated softer sediments of the Peninsula region and Ross Sea region. The ophiuroid morphotype of Ophiuroid 4, characterized by small spiny pink-red arms with a darker central disc) was common in the Ross Sea region and is possibly Amphiura sp. We also saw an increase in Hexactinellida (glass sponges) in the Ross region compared to other regions ( Figure 2B ).

Our three regions had significant heterogenous dispersions with potentially significant differences between communities ( Table 3 ). Our study regions illustrate distinct community compositions ( Figure 2B ) and a latitudinal trend in abundance (again with exception of Recovitza Island; Figures 4B , S4 ). Northernmost sites portrayed higher relative abundances of demosponges and ascidians on average (12.5% and 19%, respectively) than southernmost sites (4% and 3%, respectively) which contained higher percentages echinoids and holothuroids (9% and 18%, respectively). Macrofaunal abundances in the Ross region and Recovitza Island were significantly lower than that of the northernmost Peninsula region sites. Based on a SIMPER analysis between sites ( Table 2 ), epifaunal communities in the Ross Sea and near Recovitza Island significantly differed from other communities near the peninsula as they were dominated by pycnogonids (~65%: pycnogonid 1 and pycnogonid 2 - possibly Nymphon spp.), malacostracan crustaceans (~5%: Mysid sp. 1), and anthozoan cnidarians (3%: possibly Hormosoma sp.). These taxa and those listed in Table 2 for each region are the crucial taxa that may be used to discriminate between the Ross (and Recovitza) region, the Peninsula region, and the Amundsen region ( Table S3 ).

The Peninsula region portrayed lower evenness and Simpson diversity values than the Amundsen region as a result of dominance by the yellow rope sponge, brittle stars (Ophioplinthus spp.), and their symbiotic sponge (Iophon sp. [ Table 2 : Figure 1A ]). The Amundsen Region typically had higher Simpson diversity values than the other regions, but there was no difference in this diversity index between the regions. Although the Peninsula region was dominated by a few species ( Table 2 : Ophioplinthus spp., Iophon sp., and Ophionotus victoriae), it contained the highest species richness and highest Shannon Diversity values, closely followed by the Amundsen region ( Figures 4A, C ). No significant latitudinal trends can be described here due to the inclusion of the Recovitza Island Site with the Ross region, but there is a discernable decreasing trend when looking at the sites individually ( Figure S4 ). However, species richness, abundance, and Shannon diversity tend to decrease progressing southward (Peninsula region > Amundsen region > Ross Sea region: Figures 4B–D ).

Based on abiotic data ( Table 4 ), sites generally group together based on location along our transect, similar to the assemblages mentioned above. The PCA analysis indicates the strongest driver of differences in PC1 which accounted for 46.8% of the variation ( Table S5 ) was DO content of bottom waters which was negatively correlated with temperature and maximum fluorescence, our proxy for chlorophyll ( Table S6 ). Percent sea ice cover, water depth, latitude, and salinity were the most influential factors for PC2 which accounted for 22.6% of variation in the environmental characteristics ( Figure 3B ). Phytodetritus presence and pebble substrate were associated with higher chlorophyll, temperatures, and lower oxygen levels near the benthos in the Amundsen region.

Amundsen region sediments included pebbles and cobble stones whereas other regions were primarily comprised of finer sandy sediments, and Amundsen region benthos was also characterized by the presence of algal detritus. One drop stone (>50cm) was encountered at the Seymour Island Site (denoted with an asterisk in Figure 2A ) and was classified as containing sand, pebble, and cobble. Salinity and temperature were positively correlated and displayed an inverse relationship with oxygen. For the localities we sampled, Amundsen region sites were also the warmest (average = 1.13 °CC), but there was no difference in temperature between Peninsula region and Ross Sea region sites. Peninsula region sites had the highest DO of all sites (average = 6.95 ml l^-1), while Amundsen region sites had the lowest (average = 4.26 ml l^-1).

Macrofaunal communities were significantly different based on the substrate type and presence of phytodetritus ( Table 3 and Figures S4A, B ). Sites with phytodetritus, cobbles, and pebbles in the Amundsen region were characterized by a higher abundance of echinoids and holothuroids while the sandier sediments without phytodetritus of the Peninsula region contained more ophiuroids and sponges. Sea spiders were the most abundant group on the finer sediments of the Recovitza Island site. PCA vectors illustrate a significant correlation (Spearman Rank: rho = 0.443, p = 0.001; overlain in Figure 3A ) between environmental parameters and epifaunal communities. The Amundsen region and Peninsula region are primarily differentiated based on PC1. Therefore, maximum fluorescence, DO, and temperature emerge as important factors driving community distinctness along the Western Antarctic coast ( Table S6 ). In contrast, the Ross region and Recovitza Island were differentiated from the Peninsula and Amundsen regions based on PC2 whose core factors were pH, water depth, and percent sea ice cover. Distance based linear models show the importance DO, depth, temperature, and sea ice cover in predicting community compositions on the western Antarctic shelf at these sites ( Tables S7 ).

Echinoderms are one of the most dominant fauna of the Antarctic continental shelf (Dayton et al., 1974; Clarke and Johnston, 2003; Chiantore et al., 2006; Moles et al., 2015). However, cnidarians, sponges, and pycnogonids were abundant enough to influence assemblages along the Western Antarctic Peninsula sites sampled for this investigation. Similar to data presented here from the Peninsula region ( Figure 5A ), a recent study in the Antarctic peninsula has shown high abundances of ophiuroids, ascidians, and demosponges (Gutt et al., 2019). Within the Bransfield Straight, Gutt et al. (2019) found dominance of varying phyla (e.g. Echinodermata, Hydrozoa, and Bryozoa) which further shows the patchy nature of this region on a relatively small spatial scale (100km). Previous studies on the continental shelf in the Ross Sea denoted high abundances of suspension feeding bryozoans, anthozoans, crinoids, demosponges, and ascidians (Cummings et al., 2021), which is similar to our results except that we also recorded an abundance of pycnogonids which were abundant in the Ross region and dominated Recovitza Island. Further, comparing our Amundsen sites to previous studies is difficult due to lack of data, but the abundance of holothuroids and echinoids at these sites is notable ( Figure 5C ).

Although we did not see a latitudinal trend in overall diversity across the Western Antarctic benthic communities, we did see a shift in these communities from dominance by a single taxon to a more diverse representation. Previously, higher Shannon diversity was observed on the open continental shelf compared with fjord sites (Grange and Smith, 2013). Communities at similar depths showed higher diversity and evenness on the shelf, but higher species number and abundance in the fjords. Given the proximity of these fjord and shelf sites and the lack of environmental data included, drawing comparisons between these studies is difficult. However, based on our results, predicted changes with climate warming on salinity and temperature in fjords will drastically impact the communities present there (Grange and Smith, 2013).

Our results indicate Antarctic benthic community variability resulting from abiotic differences on a local scale. Between the Amundsen region and other two regions, differences in maximum fluorescence (>2.3 mg/ml³), temperature (> 1.9 ˚C) and DO (> 1.8 ml/l), compared to salinity (0 PSU at minimum) lead us to emphasize the maximum fluorescence, temperature and DO impacts. The relationships between temperature, salinity, depth, pressure, and DO in marine environments make disentangling singular effects of one factor on a community difficult. Amundsen region sites with higher maximum fluorescence, lower DO, and higher temperature than other regions might have had more echinoids and holothuroids than other sites due to less competition in these zones from organisms that require more oxygen as a result of a higher metabolic rate, such as ophiuroids and crinoids (Woods et al., 2009; Hughes et al., 2011). These environmental characteristics are influenced by decomposition of algal particulates on the seafloor (Iken et al., 2010; Menezes et al., 2017) likely resulting in the community differences observed. These three factors in the Amundsen Sea are typically heavily impacted by algal blooms which was likely occurring during the data collection for this project (Summer 2013). Further, the abundance of holothuroids was higher in the Amundsen region than the Ross or Peninsula region where the maximum fluorescence was highest, potentially indicating a shallower calcium carbonate saturation horizon in this region (Griffiths, 2010). Phytodetritus, higher chlorophyll concentrations in the water column, and a shallower saturation horizon in the Amundsen Sea and Wright’s Gulf would help explain the greater abundance of filter-feeding holothuroids over those organisms more reliant on calcium carbonate exoskeletons, such as ophiuroids or crinoids (Griffiths, 2010). Another important factor in distinguishing communities presented here is the maximum sea ice coverage as it facilitated the differentiation between Peninsula region and Ross Sea regions. Within the Peninsula region, sea ice coverage was increasingly variable as the Montravel Rock site had less than 60% of the sea ice cover than the Seymour Island site, however, these communities present with high abundances of ophiuroids and other filter-feeders (e.g. demosponges, crinoids, and ascidians) as previously described (Sumida et al., 2008; Gutt et al., 2019). Finally, the sea ice coverage for the summer in the Southern Ocean has reached a new historic low (Charctic, 2023) which will likely impact benthic communities.

Because the evidence for algae (maximum fluorescence and phytodetritus absence) was lacking, Recovitza Island’s significantly different community from the Peninsula region sites could be a result of the slight increase in temperature (0.45-0.8 °C) or decreased oxygen (0.6-0.7 ml l^-1). Although this site lies within the Bransfield Strait, it is located at the convergence of the transitional waters from the Bellingshausen Sea and water currents from the Weddell Sea (Smith et al., 2012; Deppeler and Davidson, 2017; Tripathy and Jena, 2019). Complexity of the Bransfield water currents leads to environmental fluctuations likely resulting in the community differences presented in our study (e.g., dominance of pycnogonids at Recovitza Island and ophiuroids at Montravel Rock). This distinction in communities on either side of the Bransfield Strait is supported by the clear variation between communities on the northern (mixed assemblage) compared to the southern (more suspension feeders) side of the Bransfield Strait (Gutt et al., 2013). These hydrographic features aid in forming ecoregions in the Southern Ocean which result in these trends in community composition and diversity as shown in the present study (Deppeler and Davidson, 2017). Given impacts of these environmental fluctuations and water currents on the benthic community fluctuations at the local scale, Western Antarctica should be a priority for future climate change studies, especially the Amundsen Region which is the least studied region of all the Southern Ocean (Griffiths et al., 2011).

Although most sites presented in this study were sandy or muddy sediments by design, those that had pebble and cobble substrates in the Amundsen and Peninsula regions, portrayed comparable abundances and Shannon diversity values ( Table 1 ) to those found in other sections of the shelf (Grange and Smith, 2013), the Weddell Sea (e.g. Gutt and Starmans, 1998; Pineda-Metz et al., 2019), and Eastern Antarctica (Post et al., 2017). However, the diversity of benthic communities in the southeastern Weddell Sea were higher than the western Weddell Sea and most of the regions presented in this study (Gutt and Starmans, 1998). Further, in the southeastern Weddell Sea, a greater abundance of holothuroids, bryozoans, and ascidians were found on substrates with less pebbles and a greater abundance of sponges, anthozoans, and other taxa found on finer sandy substrates (Gutt and Starmans, 1998), similar to our Amundsen region and Peninsula region sites, respectively. An increase in arthropods (pycnogonids and others) at the finer sediment sites (Ross Sea and Recovitza region) was found in the Weddell Sea (Gutt and Starmans, 1998) and the Sabrina Coast in Eastern Antarctica (Post et al., 2017). In the Amundsen region and along the Sabrina Coast, on pebble and cobble substrates, there was an increase in brachiopods. Additionally, drop stones and iceberg scour which affect substrate were not accounted for in this study but there was notable abundance and diversity of macrofauna on our one encountered drop stone at Seymour Island ( Figure 5D ). Drop stones can act as faunal islands in Antarctic and Arctic waters and facilitate dispersal of benthic fauna by serving as steppingstones (Boucot and Carney, 1981; Dale et al., 1989), and in the Arctic, they have been shown to house communities dominated by sponges, ascidians, bryozoans, and crinoids which were found on our one drop stone (Syvitski et al., 1989).

High levels of endemism (Hempel, 1985; Griffiths, 2010; Gutt et al., 2013; Miranda et al., 2021) in the western shelf suggest research is needed to understand how diversity, community composition, and faunal richness will change with climate. Given the higher richness, abundance, and diversity of Peninsula region communities compared to the rest of Western Antarctica, further studies should be employed to better understand changes along the continental shelf and assess regions where benthic community diversity may be of interest (e.g., highly stable or highly variable regions). Diversity trends observed here are similar to those seen in meiofaunal reports from Western Antarctica (Brannock et al., 2018) which show lower richness and diversity in the center of the Bransfield Straight and higher in the Amundsen region. Curiously though, the Eastern Peninsula sites (near our Seymour Island) presented the lowest richness, abundance, and diversity for meiofaunal communities (Brannock et al., 2018) which we believe supports our assertion for localized habitats formed by water currents as seen in the Bransfield Straight. Further microbial diversity was shown to be higher in further west sites (e.g., Bellingshausen Sea and Amundsen Sea) compared to Antarctic Peninsula sites (Learman et al., 2016), however, this might be an artifact of grouping all peninsula sites together.

With predicted decreased sea ice, increased temperature, decreased DO, and algal blooms, polar regions with a dearth of data such as the Amundsen region provide target areas for future research. Studies such as this provide necessary information on benthic communities to aid in characterizing the Southern Ocean’s diverse ecoregions in this remote and isolated habitat without the need for potentially harmful trawling (Konecki and Targett, 1989). The Southern Ocean is a large and diverse habitat with varying benthic communities along the continental shelf. In order to identify economically and ecologically important ecoregions, we must continue to gather more data on biodiversity and spatial relationships in this area.