Biogeography and ecology of Ostracoda in the U.S. northern Bering, Chukchi, and Beaufort Seas

Ostracoda (bivalved Crustacea) comprise a significant part of the benthic meiofauna in the Pacific-Arctic region, including more than 50 species, many with identifiable ecological tolerances. These species hold potential as useful indicators of past and future ecosystem changes. In this study, we examined benthic ostracodes from nearly 300 surface sediment samples, >34,000 specimens, from three regions—the northern Bering, Chukchi and Beaufort Seas—to establish species’ ecology and distribution. Samples were collected during various sampling programs from 1970 through 2018 on the continental shelves at 20 to ~100m water depth. Ordination analyses using species’ relative frequencies identified six species, Normanicythere leioderma, Sarsicytheridea bradii, Paracyprideis pseudopunctillata, Semicytherura complanata, Schizocythere ikeyai, and Munseyella mananensis, as having diagnostic habitat ranges in bottom water temperatures, salinities, sediment substrates and/or food sources. Species relative abundances and distributions can be used to infer past bottom environmental conditions in sediment archives for paleo-reconstructions and to characterize potential changes in Pacific-Arctic ecosystems in future sampling studies. Statistical analyses further showed ostracode assemblages grouped by the summer water masses influencing the area. Offshore-to-nearshore transects of samples across different water masses showed that complex water mass characteristics, such as bottom temperature, productivity, as well as sediment texture, influenced the relative frequencies of ostracode species over small spatial scales. On the larger biogeographic scale, synoptic ordination analyses showed dominant species—N. leioderma (Bering Sea), P. pseudopunctillata (offshore Chukchi and Beaufort Seas), and S. bradii (all regions)—remained fairly constant over recent decades. However, during 2013–2018, northern Pacific species M. mananensis and S. ikeyai increased in abundance by small but significant proportions in the Chukchi Sea region compared to earlier years. It is yet unclear if these assemblage changes signify a meiofaunal response to changing water mass properties and if this trend will continue in the future. Our new ecological data on ostracode species and biogeography suggest these hypotheses can be tested with future benthic monitoring efforts.


Introduction
Biological systems of the Pacific-influenced Arctic Ocean are currently undergoing rapid climate-related transformations [1][2][3][4]. This is attributed to ocean warming in the North Pacific, Bering, and Chukchi Seas [5,6], which is accelerating sea-ice loss and extending the open water season [7,8]. Changes in sea-ice cover further alter stratification, hydrography, and circulation patterns, all of which influence primary productivity [9][10][11][12] and trophic relationships [13,14]. The quality, quantity, and timing of production reaching the sea floor ultimately affects benthic marine species abundance, distribution, and food web dynamics [2,3,15,16]. In this study, we examine a component of the benthic ecosystem: meiobenthic ostracodes and the primary factors that influence their ecology and biogeography in response to changing conditions in the Bering, Chukchi, and Beaufort Seas.
This study has three primary objectives. The first is to examine large-scale biogeographic ostracode patterns and species diversity in the BCB Seas. The second objective is to identify primary environmental factors related to ostracode faunal distributions, particularly among dominant and diagnostic species. For this, we used ordination analyses of a series of nearshore-to-offshore transects (primarily the DBO lines) to spatially assess species abundance patterns across different water masses. Lastly, we use distribution and abundance patterns of dominant and/or ecologically significant taxa as proxies for species response to environmental changes and consider whether recent temperature, sea ice, and productivity changes are affecting ostracode distributions.

Ostracodes as indicators of environmental change
Ostracodes are a bivalved group of Crustacea, ranging in size from~0.5 to 2.0 mm, that secrete a calcareous (CaCO 3 ) shell commonly preserved in sediments. Because individual marine ostracode species have ecological limits controlled by temperature, salinity, oxygen, sea ice, food, and other habitat-related factors, they are useful ecologic indicators [17][18][19][20]. Ostracodes provide ways to evaluate the effects of changing environmental conditions in different types of areas/ecotones and through geologic time. In addition to distinct species' ecology, their small size, abundant fossil record, limited stratigraphic ranges, species-level identification, and shell isotope geochemistry, ostracodes are multi-proxy tools for paleoecologic, paleoceanographic reconstructions and biostratigraphy [17][18][19][20][21].

Controls on marine ecosystems and benthic ostracode species
The distribution of biological communities-from plankton and invertebrates to epibenthic fish and sea birds-has largely been found to reflect the distribution of water masses [16,22]. While the temperature required for survival and reproduction primarily controls the distribution of cryophilic ostracode species on large biogeographic scales [23,24], other local factors also influence species spatial dynamics, such as depth, sediment properties, and primary productivity exported to the benthos [17,[25][26][27][28]. Assemblage composition in any given area is constrained by a combination of these environmental gradients. In addition to water mass properties, this study examines other parameters that affect dominant and indicator ostracode species over south to north gradients (Fig 1), as well as nearshore to offshore transects (Figs 2 and 3), to better understand the ecological preferences of important indicator species.
Here, we extend a prior study analyzing ostracode distributions in the Bering Sea [31], using additional samples collected through the Distributed Biological Observatory (DBO). This international hydrographic and biological sampling program, initiated in 2010, targets known areas of high biological activity along latitudinal transects in the Pacific-Arctic for repeated sampling and ecosystem monitoring [32]. We incorporated an ostracode species analysis into the DBO framework with samples collected from 1970 to 2018 and others in the vicinity of these locations (Fig 1) using a subset of data (S1 Table) from the Arctic Ostracode Database (AOD-2020); [33][34][35][36]. The AOD-2020 compiles data for 96 extant ostracode species from 1574 Arctic and subarctic surface (most from the upper 0-2cm) samples, with associated data that include geographic coordinates, water depth, temperature, and salinity. This allowed multivariate analyses to be done on a subset of samples in the northern Bering, Chukchi, and Beaufort (BCB) Seas over the past several decades.  Table 1.) Note that some stations along established DBO transects are not represented due to low or no ostracode abundance (i.e. DBO3-6 to 3-8 and DBO4). Smaller inset map shows context of the study area. Figure made with GeoMapApp [29] (www.geomapapp.org) / CC BY / CC BY [30]. https://doi.org/10.1371/journal.pone.0251164.g001 Strait [6], including three distinct summer water masses: Anadyr Water, Alaska Coastal Water, and Bering Sea Water (Fig 2); [37][38][39]. Anadyr Water (AW; -1.0˚<T<1.5˚, S�32.5) is relatively saline, cold, and nutrient-rich and flows along the Russian coast and the western side of the northern Bering Sea. Alaska Coastal Water (ACW; 3˚<T<9˚, S�32) in summer contains warmer and fresher water derived in part from Alaskan river runoff [38]. Bering Shelf Water (BSW; 0˚< T<3.0˚, 31.8<S<33) is an intermediate water type that is lower in macro-nutrients but with higher salinity contributions from the Bering Slope Current [37]. It runs northward and bifurcates at St. Lawrence Island to enter the Bering Strait between the westerly AW and easterly ACW.
In summer, currents flowing though the Bering Strait drive oceanographic properties in the Chukchi Sea, where water is made up of ACW and Bering Sea Water (BSW), which is a combination of both Bering Shelf Water and AW [39][40][41]. Bering Sea Water has a higher salinity and nutrient levels than ACW [41,42]. In the central Chukchi basin, BSW contains the highest primary production and chlorophyll standing stock [43]. This water mass follows bathymetric contours of the Chukchi Sea shelf, flowing northwest to Herald Canyon and northeast across the mid-shelf Central Channel to flank Hanna Shoal to the northern tip of Alaska (Barrow Canyon; Fig 2); [39,43]. Year-round, when the prevailing northeast winds are weak, outflowing water from the Chukchi Sea is advected toward the eastern Canadian Basin along the Beaufort shelfbreak jet (Fig 2); [44].
These seasonal ocean currents are key in transporting nutrients and phytodetritus into the western Arctic Ocean and maintaining high productivity and biomass [15]. In winter and early spring, the shelf areas are uniformly comprised of cold Winter Water (WW; T� -1.0˚, S>31.5); [45,46]. In summer and fall, WW remains present in the bottom waters of the northern Chukchi Sea [46].

Sample collection and processing
The AOD-2020 compiles ostracode census data collected from circum-Arctic and subarctic sites between 1970 to 2018 from a number of programs and expeditions. Geographical coordinates, depth, temperature, salinity, and original source methodology for all samples in the database are based on original published studies. Most samples were obtained from the uppermost 0-2 cm of surface sediments. For collections during 2009-2018, including the Distributed Biological Observatory (DBO) program, surface sediments were taken from the tops of multicores or Van Veen grabs prior to opening. Most sampling occurred during the months of June, July, August, or September. Corresponding bottom water temperature and salinity were collected at the time of sampling or if shipboard expedition data were not available, as was the case for some older Beaufort Sea samples from the 1970s and 1980s, measurements were derived from World Ocean Atlas 2001 [47,48] and World Ocean Database 2013 [35,49]. Salinity is given without units using the Practical Salinity Scale [50].
Sediments ranging in wet weight from 20 to 150g were washed with tap water though a 63μm sieve, oven-dried in paper filters at 40-50˚C, and then dry sieved over a mesh-size of >125μm. In most cases, all ostracode specimens present in a sample were picked. Living and dead specimens were counted together; however, the majority of samples contained well-preserved shells and carapaces with chitinous appendages that indicated specimens were alive at or near the time of collection. The number of specimens refers to valves and articulated carapaces each counted as one specimen [33].

Statistical analyses
Multivariate statistical analyses were carried out using the Paleontological Statistics (PAST) software package, version 3.24 [59]. Relative frequencies (percent abundance of total assemblage) were calculated for each ostracode species. Species relative abundance data are expressed as a percentage of an individual species relative to the total number of individuals in a sample. Our studies focused on ostracode samples yielding more than 30 total specimens. Both species (R-mode) and samples (Q-mode) were grouped by several different correspondence analyses. A principal component analysis (PCA) was used to group ostracode species by their similarities/differences in dominance based on samples designated by ecoregions. A detrended correspondence analysis (DCA) described relationships between ostracode assemblage structure and geographic DBO/other ecoregions. A canonical correspondence analysis (CCA) was used to explore multivariate relationships between the most common ostracode species and their associations with major environmental variables, and sample locations were color-coded by ecoregion. The benthic environment at each station in the northern Bering and Chukchi Seas was characterized by eight variables that could affect ostracode habitat and survival: bottom water temperature; bottom salinity; grain sizes 0-4 phi; grain size �5 phi; sediment chlorophyll-a; TOC; TON; C/N ratios. The Shannon Weaver "H" Index was used to measure ostracode diversity [60]. The values of "H" typically range from 0.0 to 5.0, and increases with increases in both richness (or the number of different species) and evenness (how close in abundance each species is in an environment). In general, values above 3.0 suggest a habitat structure that stable and balanced. Values less than 1.0 indicate a more challenging or impoverished habitat for survival and reproduction.

Results
Benthic ostracodes were abundant and diverse at continental shelf depths (20-~100m water depth) in Pacific Arctic surface sediments analyzed (ranging from 30 to 562 specimens/sample; Fig 1, Table 1). A total of 47 ostracode species and five genera were identified in 289 samples (34,369 total specimens) collected from 1970-2018 in the BCB Seas ( Table 2). Ostracode species identified represent a mixture of Arctic, subarctic, and cold-temperate taxa. At the sampling sites, summer bottom water temperatures were highly variable, ranging from 12˚C off Nome to 6-9˚C near the Alaskan coast to �0˚C in Bering and Chukchi Sea middle shelf areas (Fig 4A-4E). More specifically, ecoregion comparisons of summer near-seafloor temperatures from conductivity-temperature-depth (CTD) measurements show the northern Bering Sea, southwest of St. Lawrence Island ( [1]. Bottom water temperatures representing a similar but wider geographic area in 2015 and 2017 corroborate this finding ( Fig 4E). Salinity was relatively uniform across the sampling locations in that marine conditions prevailed and averaged 32 ±2. In the Alaskan Beaufort Sea (20-90m water depth), summer temperatures across all years of sampling averaged 0˚C ±1˚C, and salinity averaged 31±1. Coarser, sandy sediments that occur in the hydrodynamically active nearshore areas did not cause lower ostracode abundances, as might have been expected [63,64].
Diversity "H" in the Chukchi Sea ranged from 1.8 to a little more than 2 during most sample years. Samples averaged 12 (±4) different species. The Beaufort Sea samples had comparable diversity values of 1.8, which was consistent in years 1970-2018 where sampling occurred, and 11 (±3) distinct species per sample. Ostracode diversity in the northern Bering Sea was the lowest compared to these regions, and averaged 1.1-1.2 during the period 2009-2018 with an average of 6 (±2) species per sample.

Large-scale biogeography of ostracodes in the Bering, Chukchi, Beaufort Seas
Three dominant species in BCB Seas region accounted for 20% to 50% of the total population in the study area: Normanicythere leioderma, Sarsicytheridea bradii, and Paracyprideis pseudopunctillata ( Fig 5). Each of these species clearly groups with sampling regions designated by DBO ecoregions (noted by symbol on legend) in separate quadrants of the PCA (Fig 5). In some samples along the DBO3 and central Chukchi Sea (Ledyard Bay) transects (indicated by orange squares on the PCA), N. leioderma and S. bradii were co-dominant. Otherwise, all three taxa showed distinct spatial distribution patterns related to the path of the primary summer bottom water masses (and its respective properties) in the surveyed regions. Other ostracode species (labeled in blue around the center axes on the PCA) are secondary components in assemblages.
A DCA (Fig 6A) grouped ostracode samples in four distinct geographical groups (Table 3). These ostracode biofacies are associated with specific bottom water masses, sediment substrates and/or food sources: Group 1 is a subarctic assemblage dominated by N. leioderma in the vicinity of the St. Lawrence Island polynya and the western Chirikov Basin at DBO2 that is influenced in summer by Anadyr Water. Group 2 is an Arctic-subarctic assemblage dominated by S. bradii and N. leioderma found in the eastern Chirikov Basin of DBO2 and nearshore areas in the southeast Chukchi Sea. Group 2 and Group 3 is bridged by S. bradii. In sediments further offshore in the central and middle shelf areas, P. pseudopunctillata was the most abundant species in Group 3, influenced by the BSW of the northeast Chukchi Sea. This offshore sample group includes sediments of Hanna Shoal, which is characterized by stable, cold bottom water temperatures and higher salinity from brine rejection, with WW present in bottom waters year-round [46]. Group 4 are the Beaufort Sea samples where influences by river influx and sea ice are strong, and ostracodes are dominated by P. pseudopunctillata and H. sorbyana.
Environmental factors influencing 14 of the most abundant ostracode species in the northern Bering and Chukchi Seas were examined by CCA ( Fig 6B) and indicate that variables other than temperature and salinity water mass characteristics influence species distribution patterns. The environmental variables best correlated to ostracode assemblage structure at nearshore stations in the Chukchi Sea (orange squares in Fig 6B, i.e. DBO3-1 to 3-4, Ledyard Bay, and blue triangles at DBO5-1 to 5-3) are coarse, sandy sediments (0-4phi) and increased temperatures of ACW. The species most associated with these characteristics include Schizocythere ikeyai, Munseyella mananensis, Semicytherura mainensis, and Elofsonella concinna. In addition, N. leioderma and S. bradii are associated with these regions where sediments have sandy-pebbly textures and also at some locations with high sediment chlorophyll-a in the northern Bering Sea, both south of St. Lawrence Island and in the Chirikov Basin. In samples Comparison of summer bottom temperatures in studied ecoregions and months at stations nearest to shore (e.g. DBO3-1, DBO5-1) that are locations affected by Alaska Coastal Water, and more central or middle-shelf stations (DBO3-5 or DBO5-10) that are affected by Bering Sea Water (winter or summer variants). This subset of data was derived from the Pacific Marine Arctic Regional Synthesis (PacMARS) Project ( Table 1), with sites designated by DBO areas and other ecoregions (legend symbols) and major taxa (green lines with species names labeled in blue). Three species dominate the assemblage groups: N. leioderma, S. bradii, and P. pseudopunctillata. The proximity of the symbols in the plot reflects the level of similarity in taxa present at each sampling station.
https://doi.org/10.1371/journal.pone.0251164.g005 Ostracode biofacies differed among ecoregions within the study area (described in Table 3) and grouped primarily on summer water mass characteristics (temperature and salinity). b. Canonical correspondence analysis (CCA). CCA of ostracode assemblages in the northern Bering-Chukchi Sea region, with sampling ecoregion sites (noted by symbol on legend), major taxa (names are labeled in blue), and eight environmental parameters (green lines and labels in black: temperature, salinity, sediment total organic carbon [ located in the offshore northern Chukchi Sea and Hanna Shoal region (red inverted triangles and red circles respectively in Fig 6B), sustained cold temperatures (<0˚C), high salinity from brine rejection during ice formation, and finer grained sediments best align to ostracode assemblage structure and the dominant species, P. pseudopunctillata, Jonesia acuminata, Acanthocythereis dunelmensis, Robertsonites tuberculatus, Kotoracythere arctoborealis, and Heterocyprideis fascis.
The CCA yielded four axes that explained 84.4% of the variance in the relationships between ostracode assemblage structure and environmental properties. The first two axes alone explained about 52% of the variance in the data. The CCA affirms that in the northern Bering and Chukchi Sea region, continental shelf ostracode taxa are more correlated to variables associated with distance from shore and water mass than to latitude (60-71˚N). These factors include sediment substrate and food type or availability which link certain species to specific environments.
Based on ordination and multivariate results of ostracode species in the study region (Figs 5 and 6B), we consider six species, N. leioderma, S. bradii, P. pseudopunctillata, S. complanata, S. ikeyai, and M. mananensis, to be most diagnostic of specific ranges in bottom water temperatures, salinities, sediment substrates and/or food sources (Figs 5, 6A and 6B; S2 Table; S1 and S2 Figs).

Ecoregions and transects: Ecological-scale ostracode assemblages
We also incorporated DBO transect lines (Figs 2 and 3) in each ecoregion (Table 1) into our analysis of ecological preferences of dominant and ecologically significant species. Ecoregionscale plots of species abundances across seven transects over multiple sampling years help illustrate the regional variation of ostracode biofacies and their spatial composition from nearshore to offshore (Figs 7-13). Ecoregion DBO-2 in the Chirikov Basin of the northern Bering Sea (Fig 7) shows a sharp faunal boundary between N. leioderma overlain by Anadyr Water on the western side of the Basin and S. bradii in Bering Shelf Water on the eastern side. The southeast DBO-3 ecoregion is dominated by N. leioderma, S. bradii and S. ikeyai in ACW. In samples directly north of the Bering Strait in BSW, S. complanata increases in proportion among these species (Fig 8). More complex onshore-to-offshore assemblage changes are reflected in the Ledyard Bay and Icy Cape transects (Figs 9 and 10) from 2018, both in the Chukchi Sea. Ledyard Bay and Ice Cape also show clear onshore-to-offshore changes in dominant species. With the exception of a few samples, the Hanna Shoal in the northeast Chukchi Sea ecoregion ( Fig  11) is dominated by P. pseudopunctillata and Group 3 species (Table 3). The Barrow Canyon transect (Fig 12) shows perhaps the most complex faunal patterns reflecting the temporally and spatially complex and variable ocean water masses and depth changes in the Canyon. The Beaufort Sea west-to-east transect (Fig 13, sample locations on Fig 4 map) shows fairly consistent dominance of H. sorbyana, P. pseudopunctillata, S. bradii and recent contributions from N. leioderma, which is in sharp contrast to its abundant presence in samples of the western Chirikov Basin and southernmost Chukchi Sea. Biogeographic patterns of these species are further discussed below in the "Ecological transects" section.

Time-series analysis
Two time periods (1970-2012 and 2013-2018) were assessed by PCA to evaluate possible temporal changes in ostracode assemblages (Fig 14). These groupings were chosen based on the documentation of accelerated environmental changes [1,16] during the latter years, including sea ice duration [7], ocean temperature and oceanography [6,46] and primary productivity [65]. We did not find that the three most dominant species, N. leioderma (Bering Sea), P. pseudopunctillata (offshore Chukchi and Beaufort Seas) and S. bradii (all regions), changed in relative abundance. For the more recent 2013-2018 period, northern Pacific species Munseyella mananensis and Schizocythere ikeyai exhibited small but ecologically significant increases in abundance (~2%) and frequency in the Chukchi Sea region. Based on this evidence, we examined times series of relative frequencies of these five species, N. leioderma, S. bradii, P. pseudopunctillata, S. ikeyai and M. mananensis (Fig 15). The goal was to augment the synoptic time slice analyses to determine if the sampling was complete enough to detect decadal trends from interannual variability. The time series was limited to the last 18 years in Chukchi Sea (2000-2018) because too few repeat observations at similar locations to assess benthic change regionally.

Distribution and ecology of dominant and indicator taxa
Normanicythere leioderma and S. bradii are ubiquitous in this region, with very broad environmental tolerances. These species are adapted to the large seasonal variations of the inner and middle continental shelf. S. bradii is a wide-ranging, cold temperate to frigid species with temperature tolerances of -1.7 to 18˚C [23], but it demonstrates a preference for mid-shelf depths in temperatures colder than 4.5˚C and typical Arctic marine salinities on continental shelves

PLOS ONE
(31-33 is typical; [36]; S1A and S1B Fig). Likewise, N. leioderma has wide temperature tolerances (-1.7 to 19˚C; [23]) in typical Arctic shelf salinity (S1A and S1B Fig). As a dominant species in the northern Bering Sea [31], its distribution is circum-Arctic, including the Chukchi, Beaufort, and Eastern Siberian Seas [36], but is very rare in the Kara Sea and absent in the Laptev Sea [20]. Paracyprideis pseudopunctillata is a cryophilic, littoral-sublittoral species (1-50m) with a southerly range limit of~63˚N latitude, tolerant of salinities as low as 5-10 [19,23,66,67]. S. complanata is adapted to the coldest conditions on continental shelves, including where polynyas form during winter [67]. An analysis of this species in AOD-2020 reveals that its distribution is tied to near-freezing water temperatures �0˚C in sublittoral depths, even though it can withstand higher seasonal temperatures.
These four species, which dominate the study area, are ideally adapted to bottom water temperatures that normally range from −1.8˚C to~4˚C across the area (S2 Table). This temperature range coincides with the temperature window established for benthic macrofaunal animals prevalent at hotspot areas of the Pacific Arctic [15,51].
S. ikeyai and M. mananensis are members of genera endemic to the Pacific that have modern distributions off eastern Japan and the Okhotsk Sea and are adapted to a wider range of summer bottom water temperatures (0-20˚C; [27,68]) than the species discussed above. M. mananensis is also common to the cold temperate North Atlantic with bottom water temperatures from 2˚to 14˚C and water depths from 24 to 261m [69]. Siddiqui and Grigg [70] included this species in a list of sublittoral fauna from Halifax Harbour, Nova Scotia. Using the AOD-2020 to assess the subarctic-Arctic distribution of these species in samples with 10 or more specimens, M. mananensis is most commonly found in estuarine inlets, with highest abundances (2-30%) in Hudson Bay, North Star Bay (Thule, Greenland), Norton Sound (Alaska) and Chaunskaya Gulf (Eastern Siberian Sea) in bottom water temperatures ranging in summer from 0 to 7˚C and salinities from 26 to 33. S. ikeyai is rare (one sample reported) in the Beaufort Sea before 2018. It is recorded on the Chukchi Sea shelf in 8 samples (abundance of 2-38%) before 2014 in summer bottom water temperatures from 0˚C to 6˚C and salinities from 31 to 33. Single specimens occurred in three samples in the Alaskan Beaufort Sea from 1971 to 1982. In the northern Bering Sea, S. ikeyai is rare (1.2% of the cumulative assemblage in samples collected during 2010-2018), and M. mananensis is extremely rare (<1% of assemblage). Increased frequency of these species in the Arctic could reflect warming bottom water temperatures since these species reach peak abundances in lower-latitude waters that generally remain above 0˚C throughout most of the year. Overall, the abundance and distribution of these six species can be used to infer benthic environmental conditions and potential change in the Pacific Arctic in future studies and in sediment archives for paleo-reconstructions.
While temperature of the bottom water mass is a dominant control on ostracode distributions (Fig 6A), the CCA (Fig 6B) showed that some species are correlated to a particular sediment texture, and type/availability of carbon food sources. For a few species, N. leioderma and P. pseudopunctillata, these factors appear to drive their abundance, as supported by the transect analyses (see "Ecoregion transects" Discussion). For example, in the northern Bering and Chukchi Seas, N. leioderma more commonly inhabits coarse and pebbly sediments (0-4 phi) of faster moving water, as does S. bradii, S. ikeyai, and M. mananensis (Fig 6B; S1C Fig). Paracyprideis pseudopunctillata is found in greater proportions in silty, fine-grained seafloor habitats (�5 phi) as indicated by its low abundance in sediments of phi 0-4 ( Fig 6B; S1C Fig). S. complanata does not demonstrate a sediment preference (S1C Fig).
Availability of organic carbon is a top-level factor that drives species survival and reproduction [71]. N. leioderma is associated with areas of newly settled organic material (i.e. higher quality carbon), as indicated by its relative abundance in areas of higher sediment chlorophylla (>10mg/m 2 ) and lower C/N ratios (<7; Fig 6B; S2A and S2B Fig). In contrast, P. pseudopunctillata resides in areas with lower sediment chlorophyll-a and higher TOC (>0.5%; indicating a greater amount of detritus in the surface sediments; Fig 6B; S2B Fig) and C/N values (>7; indicating greater terrigenous inputs; Fig 6B; S2C Fig). This may indicate P. pseudopunctillata uses detritus or more refractory carbon sources for food. The widespread abundance of S. bradii in areas of high-and low-quality carbon further supports its eurytopic nature, and it demonstrates no clear carbon (food) preference (S2A- S2C Fig). Because S. complanata, M. mananensis and S. ikeyai are secondary components in assemblages, their proportions are lower and therefore their environmental preferences are more difficult to evaluate. M. mananensis and S. ikeyai are present in sediments with higher C/N ratios, which corresponds to their higher proportions at sampling stations nearer to shore in ACW, e.g. southern Chukchi Sea stations DBO3-1 to DBO3-4 and northern Chukchi Sea stations DBO5-1 to DBO5-3 (S2C Fig). In these nearshore sediments, coarse grains and gravel may dominant, but there is fine sediment in between the large grain size. This indicates variable settling rates over various seasons, providing complexity to nearshore statistical findings. Additional sampling of these species and corresponding sediments are required to better evaluate the role that sediment and organic carbon sources may have on their abundance.

Ecoregion transects: Spatial relationships among ostracode biofacies
Stack plots showing the primary species in sample assemblages in an ecoregion and/or along a sampling transect (Figs 7-13) provide insight into factors that control species occurrence and frequency. Generally in the study area, bottom water temperature declines with increasing latitude and distance from shore. Sediment composition also becomes finer-grained (greater silt fraction) with increasing distance from shore [15,72]. Other properties, especially in the central and northeast Chukchi Sea, involve more complex spatial patterns [51,73]. Varying seafloor bathymetry, sediment substrate, flux of organic carbon settling to the seafloor, and current velocities combine with water mass characteristics to create a complex habitat of varying ecological gradients. These transects show the variability of species occupying a sampling area of 0.1m 2 at a given location on the seafloor during a summer's day collection. At

PLOS ONE
DBO2-Chirikov Basin (Fig 7A), Ledyard Bay (Fig 9), Icy Cape (Fig 10) and DBO5-Barrow Canyon (Fig 12) transects, ostracode biofacies do not show a continuous composition but instead a faunal transition from nearshore to offshore stations. The ecoregion plots show distinct abundance changes in indicator species, which were sometimes acutely distinct, based on spatial environmental changes in water mass, sediment properties and food type. These descriptions of ostracode faunal assemblages can serve as a baseline to examine future meiofaunal changes in the BCB Seas.
Chirikov Basin-DBO2. The DBO2 transect in the Chirikov Basin (Fig 7A) is an example of ostracode indicator species clearly divided by water masses of differing productivity.

PLOS ONE
Stations DBO2-1 and 2-4 are the western-most locations, and are in the path of AW (Fig 2). Major differences between the water masses include higher average values of sediment chlorophyll-a and salinity (Fig 7B) to the west because the more eastern samples are underlain by BSW or a combination of ACW above and BSW below. Integrated primary production estimates in the Chirikov Basin range from~80 g C m -2 yr -1 on the interior shelf to up to 480 g C m -2 yr -1 in AW [11]. N. leioderma (red) is more prevalent in AW. The eastern side of the Basin (stations DBO2-2 and 2-5) is dominated by S. bradii (blue) and E. concinna (brown), with increasing numbers of S. ikeyai (light blue) in 2018 samples that may reflect slightly higher average bottom water temperatures (nearly 2˚C, Fig 4B). The abrupt change in faunas from We speculate that N. leioderma may have expanded its population perhaps due to increasing carbon export to the benthos and/or changes in current flow that affected sediment grain size. Since satellite monitoring began in 1979, the length of the ice-free season in the Chirikov Basin has increased by *25 days, stimulating greater primary productivity due to a longer growing season [74,75]. With N. leioderma as the established dominant fauna in the northern Bering Sea around the St. Lawrence Island polynya [31], additional light and productivity may be factors that contribute to its expansion. Another major feature of this area is strong currents that create coarse seafloor sediments consisting of primarily (>75%) 0-4 phisized sand, which N. leioderma favors. Macrofaunal biomass associated with finer-grained
Southeast Chukchi Sea-DBO3. At the nearest-to-shore stations, DBO3-1 to 3-4, the ostracode biofacies during 2014 to 2018 have been extremely consistent, not only in species present but species proportions as well (Fig 8). Species at DBO3-1 to 3-4 stations are in the path of ACW (temperatures 6˚-10˚C) that reach to the seafloor [76]. Higher sediment grain sizes at these near shore stations are consistent with the presence of the ACC. These nearshore
Samples just north of the Bering Strait (denoted as "N of St" on Fig 12) and DBO3-5 are offshore in a different summer water mass than DBO3 sites closer to shore. This water mass change is indicated by a different ostracode biofacies with fewer species overall and high proportions of N. leioderma and Semicytherura spp., particularly S. complanata and S. mainensis.
Ostracodes are rare or absent in sites further offshore (DBO3-6 to DBO3-8) in the path of BSW as sediment composition shifts to finer-grained sediments (slower current speeds,~50-93% �5 phi) and higher TOC (>1%) values. This pattern is consistent over multiple years of sampling. Benthic macrofaunal biomass is very high at these offshore sites, with abundant production settling to the sediments [51].
Central Chukchi Sea-Ledyard Bay 2018. While the ostracode biofacies at most sampling sites along the 2018 Ledyard Bay transect are similar to those at DBO3 sites (except S. ikeyai is lacking at Ledyard Bay [2 individuals]), the Central Channel bearing BSW north to the shelf break (Fig 2) cuts across this sampling transect and creates sediment grain size, production and temperature gradients between the sample locations that drastically change ostracode assemblage composition. The station LB-5, in the path of ACW proximal to shore, has coarse sediments (75% 0-4 phi), a summer bottom water temperature at the time of sampling of 8.7˚C, and 31 salinity. Subarctic species N. leioderma (20%), M. mananensis (6%) and Semicytherura species (32%) occupied this site (Fig 9). At the next site, LB-7, sediments change to finer silt (59% �5 phi) and P. pseudopunctillata becomes dominant with Palmenella limicola and S. complanta. N. leioderma is absent. Further offshore there is a grain size shift back to coarser sediments east of Herald Shoal (62% in the 0-4 phi category at LB-13, and of that, 25% in the 0-2 phi category) where N. leioderma, S. bradii, M. mananensis and Kotoracythere arctoborealis become predominant. In addition, the highest primary productivity rates in the Chukchi Sea are consistently observed in the central part of the shelf along the dividing line between the warm Alaskan Coastal Current to the east and in the colder nutrient-rich Bering shelf waters to the west [78]. N. leioderma's increasing abundance at LB-11 and LB-13 may reflect a fresh, more ample food source. TOC increases toward the offshore stations and sediment chlorophyll-a, although seasonally and yearly highly variable is highest at LB-13 (20 mg/m 2 ) as is N. leioderma's (30%) abundance. Along the transect, temperature declined with distance from the coast from 8.7˚C at LB-5 to 2.4˚C at LB-13, where N. leioderma, S. bradii, S. complanata, M. mananensis and K. arctoborealis persist. This transect may be an example of how subtle changes in bathymetry and current flow can result in changes in food availability and sediment texture in very localized areas that affect the spatial distribution of benthic communities [73] and, likewise, ostracode biofacies.
Northeast Chukchi Sea-Icy Cape 2018. The 2018 Icy Cape transect shows the influence of bottom water mass and sediment texture on assemblage composition. Bottom water temperature closest to shore at ICY2 was 1.4˚C, but more offshore stations were colder with temperatures of 0˚to -1˚C (Fig 10). In the northeast Chukchi Sea, sea-ice coverage usually extends

PLOS ONE
into the summer months (July-August), with areas of localized persistence possible throughout the summer. Winter water was still present on the seafloor during late summer sampling. Close to shore, grain size at ICY2 to ICY5 was coarse (45% to 76% in phi 0-4 category), where S. bradii and S. complanata were predominant and P. pseudopunctillata was absent. The presence of P. pseudopunctillata beginning at ICY6 and steadily increasing to dominate (25% to 60%) the assemblages at ICY9, ICY10 and ICY11 is indicative of a seafloor sediment transition to very fine grains, ranging from 63% to 87% �5 phi. In addition to suitable habitat, persistent cold temperature supports P. pseudopunctillata. Jonesia acuminata is correlated to near-freezing temperatures and fine sediments [31], and this species has a small but notable presence at the ICY7 to ICY11 stations.
Offshore Hanna Shoal, Northeast Chukchi Sea. Heterogeneous seafloor contours direct BSW with high nutrient concentrations to the Hanna Shoal area [79]. It is a hotspot for benthic productivity that supports marine mammals [15]. The ostracode biofacies includes P. pseudopunctillata (37% of assemblage) with secondary species S. complanata (14%), H. fascis (8%), R. tuberculatus (7%) and C. cluthae (5%; Fig 11). It is very similar to the assemblage of offshore Icy Cape sites except this biofacies includes greater proportions of A. dunelmensis (6%) and R. tuberculatus, and Argilloecia sp. (4%), which are species associated with finer-grained sediments ( Fig 6B). Normanicythere leioderma is rare (1%). Paracyprideis pseudopunctillata and H. fascis are euryhaline and can tolerate fluctuating salinities. At the time of summer collection, the average salinity at sample locations was 33, bottom temperature, -1.7˚C and water depth, 51m. This area lacks any strong gradient from currents, as sea ice tends to be trapped over Hanna Shoal, so dense bottom-confined pools formed by winter water tend to stagnate [38,39,80]). Sustained cold temperatures, fine-grained sediments, and relatively higher salinities describe this frigid-water assemblage.
Alaskan Beaufort Sea. The Alaskan Beaufort continental shelf is a highly variable environment, influenced by seasonal discharge from the Colville and Mackenzie rivers that supply fresh water and deposit terrigenous material but lacks strong gradients within the examined narrow area of 20-90m water depth (Fig 3; sample locations were limited to within 100 km of shore, and included intermittent years between 1971 and 2018). Bottom water salinity at the sampling sites during the time of collection averaged 31. In addition to river inflow, the Beaufort shelf also receives dense ''winter-transformed" Pacific water [83,84] that drains off the Chukchi shelf. This water enters the Beaufort Sea as a shelf-break current that turns eastward from Barrow (Fig 2); [85]. As an area that experiences dramatic seasonal fluctuations, particularly in salinity, two species are overwhelmingly suited to these conditions: P. pseudopunctillata and H. sorbyana, cryophilic taxa correlated with euryhaline conditions (Fig 13); [19,20]. These two species are reported by Stepanova et al. [19,20] as adapted to the inner shelf, river-affected zone of the Kara and Laptev seas where salinities range between 26 and 32. The eastern Alaskan Beaufort is influenced by the Mackenzie River, resulting in finer sediment grain sizes and better sorting in that region [86,87], which correlates to greater abundance of P. pseudopunctillata. For 2018 PRW and PRB transect samples (sites on Fig 3 map) that are most influenced by Colville River inflow, H. sorbyana is more prevalent than P. pseudopunctillata. Secondary species in the Beaufort Sea assemblage include S. bradii, S. punctillata, R. tuberculatus, H. fascis, and A. dunelmensis. S. complanata is present in low (~2%) abundance in these Beaufort Sea assemblages.
Despite limited temporal data in this area, N. leioderma and S. bradii are more prevalent in 2018 samples than in samples collected in comparable locations during the 1970s and 1980s where P. pseudopunctillata and H. sorbyana comprised the major proportion of the assemblage (S3 Fig). This may be due to recent declines in ice persistence and increases in nearshore temperatures or changes in sediment composition that may be incompatible with P. pseudopunctillata's preferences for near-freezing temperatures and finer grained sediments. However, sediment transport by ice rafting can be significant on the shelf, and can carry gravel and sand particles from the coast onto the shelf [87]. There are insufficient data available to determine if a change in sediment grain size is a factor. The Beaufort shelf, while not as productive as the Chukchi Sea, can have locally high primary production rates associated with the ice edge, reaching 200mg C m -2 d -1 [11]. The retreating sea-ice edge, in addition to upwelling, provides nutrients for phytoplankton growth. The number and strength of upwelling events in the Alaskan Beaufort Sea has increased over the past 25 years [88,89]. Greater fresh phytoplankton production may be an explanation for the increasing appearance of N. leioderma in 2018 samples (S3 Fig), but our new data must be considered cautiously given the single year of samples.

Temporal trends in BCB Sea indicator species
Criteria used to assess whether environmental factors are driving changes in benthic communities include the arrival of taxa characteristic of a temperate climate zone or shifting dominance or abundance of taxa [90]. Species can extend north and south (and east to west) to the point where certain maximum or minimum tolerable temperatures are reached [23].
Due to temporal and spatial gaps in sampling, we could confidently evaluate temporal changes only in the dominant species, N. leioderma, S. bradii and P. pseudopunctillata. Two PCA comparisons by years (1970-2012 and 2013-2018) showed that the dominant species have remained fairly consistent in the BCB Seas (Fig 14). Interannual abundances of the dominant species in the Chukchi Sea from 2000-2018 did not reveal statistically significant changes, but two indicator species, S. ikeyai and M. mananensis, show incipient increases (Fig 15).
Dispersal of meiofauna into new areas likely lags behind sediment and water column changes. Benthic marine ostracodes are transported by ocean currents [91]. Considering that the movement of Pacific water from the Aleutian Passes to Bering Strait takes more than a year [5] and an additional~4 months to transit to Barrow Canyon in the Chukchi Sea [40], benthic meiofaunal migration from the north Pacific to the Arctic could take a number of years. Over the shallow shelves of the northern Bering and Chukchi Seas, seasonal ice coverage cools the entire water column to temperatures below 0˚C. These cold temperatures limit the northern distribution of subarctic populations of groundfish [92] and may also serve as a barrier inhibiting the mixing of Arctic and N. Pacific faunas through the Bering Strait [71]. It is more likely that cold temperate species already inhabiting restricted inner bay areas like Norton Sound and coastal polynyas along Alaska will increasingly be able to extend their range and abundance as conditions become more suitable for them under changing climate scenarios.

Conclusions
We examined benthic ostracode assemblages collected during research cruises from 1970-2018 in the northern Bering, Chukchi and Alaskan Beaufort Seas. We focused on identifying indicator species and their associated ecological preferences. In particular, ostracode abundance and distribution were related to bottom temperature, salinity, organic carbon deposition (sedchla, TOC, C/N), and sediment substrate.
The variability in ostracode assemblage composition throughout the BCB Seas was linked qualitatively to localized (summer seasonal) environmental patterns. We found south-to-north (Chirikov Basin to Northern Chukchi to Beaufort) changes in dominant ostracode species and assemblage composition. Over the large biogeographic scale, four ostracode biofacies were identified in the study area primarily tied to the transit of summer water masses. Distinct changes of ostracode biofacies in nearshore to offshore transects were related to water mass properties likely in combination with food sources and sediment substrate. This study is among the first to highlight changes in ostracode abundance and species over small spatial scales in response to changes in water mass properties and productivity.
Six indicator species were identified by correspondence and multivariate analyses based upon their abundance and correlations with distinct ecological parameters by which they are defined: Normanicythere leioderma-opportunistic subarctic species correlated with ample, high-quality production (exported phytoplankton) as food source; sandy to pebbly to gravelly sediments; Paracyprideis pseudopunctillata-euryhaline cryophilic species associated with WW presence and sea ice, fine grained sediments, sustained frigid temperatures (�1˚C), TOC with high phytodetritus; Sarsicytheridea bradii-habitat generalist, eurytopic species that capitalizes in highly dynamic environments, higher frequency in sandy sediments; Semicytherura complanata-frigid temperatures (�1˚C), normal marine salinity, no discernable sediment preference; Schizocythere ikeyai-cold-temperate species, warmer temperatures, ACW, sandy sediments; Munseyella mananensis-cold-temperate species, warmer temperatures, ACW, sandy sediments.
This study found a consistent link between areas of high sediment chlorophyll, coarse sediment grain size and the abundance of N. leioderma. Despite wide tolerances, it may be a sensitive indicator of new production reaching the benthos and/or enhanced food supply. In future paleoceanographic studies using fossil ostracode faunal assemblage data, N. leioderma may signal the influence of nutrient-rich Pacific water flowing in through the Bering Strait and a food supply consisting of newly settled phytoplankton. Paracyprideis pseudopunctillata is positively correlated to very fine-grained sediment textures with high TOC values of refractory (detrital) organic carbon food sources.
During the last decade, we document an incipient increase in two secondary cold-temperate species, S. ikeyai and M. mananensis, that may be increasing. Preliminary interpretation of this finding may reflect recent increases in coastal and mid-shelf bottom water temperatures and/or carbon flux to the benthos.
Continued monitoring of temperature-sensitive ostracode species in the BCB Seas is necessary to provide information on annual and decadal variability in species distributions. This analysis of modern species ecology can help interpretation of ostracode faunal data from sediment cores in relation to past and future ocean changes. This study contributes new species ecology and further validates the sensitivity and application of ostracode fauna as biomonitors and proxies for specific environmental conditions. These results provide a baseline for assessing the effects of future water mass changes and productivity on benthic ostracode communities in the Pacific Arctic.