香京julia种子在线播放

The West Coast Ocean Acidification (WCOA) cruise is a periodic assessment of ocean chemistry led and conducted by NOAA’s Pacific Marine Environmental Laboratory (PMEL), and information about the 2016 WCOA cruise can be obtained from PMEL’s Carbon Program website ¹ . Some of the 2016 sites reported here were coincident with sites used in prior cruises (2007, 2011, 2012, 2013) and with California Cooperative Oceanic Fisheries Investigations (CalCOFI) sites, which were originally chosen due to their significant oceanographic features. For this study, a total of 29 nearshore stations along the northwest coast of the U.S. and Canada ( Figure 1 ) were sampled by deployment of a rosette containing a Seabird CTD and an array of Niskin bottles. Water samples were collected from Niskin bottles from vertical profiles during the northern leg of NOAA’s 2016 WCOA cruise. Temperature, salinity, pressure, and dissolved oxygen were measured during profiling with a Sea-Bird 9plus CTD. Samples for chemical analyses were processed immediately for shipboard analyses. Names of sample groups shown in Figure 1 are listed in Table 2 .

Samples for biological analyses (> 2 L) were co-collected with those used for chemical analyses. Biological sampling depths were within 3 m of the surface, within 3 m of the bottom, and two to four approximately equally spaced intervals throughout the water column, depending on maximum depth. They were briefly held on ice in the dark and processed for longer term storage within an hour of collection. Whole water flow cytometry samples were preserved in 0.2% paraformaldehyde, flash-frozen in liquid nitrogen, and stored at ≤ −80°C until analysis. Duplicate 1 L samples for microbial taxonomic analyses were vacuum-filtered through 0.2 µm polyethersulfone filters (Sterlitech Corp, Kent WA), flash-frozen in liquid nitrogen, and stored at ≤ −80°C until extraction and analysis.

Measurements for dissolved inorganic nutrients (nitrate, nitrite, ammonium, phosphate, silicate), dissolved inorganic carbon (DIC), and total alkalinity (TA) content were analyzed following the methods of Feely et al. (2016) and in detailed metadata included with the cruise dataset at the National Centers for Environmental Information (Alin et al., 2017). All measurements are reported in units of micromoles per kilogram (µmol kg⁻¹). Additional carbonate system parameters—pH (on the total scale, pH_T, unitless), carbonate ion content ([CO₃ ²⁻], µmol kg⁻¹), and carbon dioxide fugacity (fCO2, µatm)—were calculated from inputs of DIC and TA using CO2SYS (Pelletier et al., 2007). Within CO2SYS, we used the Lueker et al. (2000) method for K ₁ and K ₂ dissociation constant formulations the Dickson (1990) method for KSO₄ the Uppström (1974) method for total boron, and the seacarb option for K _f (i.e., Perez and Fraga, 1987, when temperature is 9–33°C and salinity is 10–40 and Dickson and Goyet, 1994 otherwise).

Microorganism abundances were enumerated by a BD FACSCAlibur 2-laser 4-color flow cytometer operated by the Kavanaugh Laboratory at Oregon State University. Bacteria and archaea were enumerated using the method of Longnecker et al. (2006). Coccoid cyanobacteria (Synechococcus, Prochlorococcus), eukaryotic nano- and picophytoplankton, and Cryptophytes were enumerated using the method of Sherr et al. (2005).

DNA for taxonomic analyses was extracted from frozen filters by the methods of Green and Sambrook (2018), and used to prepare amplicon libraries for Illumina sequencing. The bacterial and archaeal library was amplified using the 16S V4 primers from the Earth Microbiome Project (https://earthmicrobiome.org/protocols-and-standards/16s/):

For the eukaryotic library, the following 18S primers from Stoeck et al. (2010) were used for amplification:

Individual samples amplified with HotStarTaq Plus Master Mix (Qiagen, USA) with the following conditions: denaturation at 95°C (5 minutes); 30 amplification cycles of 95°C (5 seconds), 53°C (40 seconds), 72°C (1 minute); elongation at 72°C (10 minutes). Products were uniquely dual indexed with Nextera adapters, purified with calibrated Ampure XP beads, and size-checked and quantified in 2% agarose gel for normalization and pooling. Library preparation and sequence analysis was performed using the MiSeq reagent kit v3 (600 cycles) on a MiSeq sequencer (Illumina, San Diego CA) following manufacturer’s guidelines by Molecular Research LP (MR DNA, Shallowater, TX USA).

Dereplicated sequence reads from the bacterial and archaeal results were quality trimmed using Trimmomatic (Bolger et al., 2014), and paired ends assembled using PANDAseq (Masella et al., 2012). Additional sequence filtering removed sequences with lengths less than 250 base pairs (bp), containing homopolymers, or ambiguous bases greater than 7 bp in length. Highly similar sequences (≥ 97% identity) were grouped into operational taxonomic units (OTUs), which were treated as the highest resolution taxon for some community analyses. OTUs were identified using QIIME2 (Bolyen et al., 2019) and taxonomy determined with the SILVA SSU database (release 132; Pruesse et al., 2012; Quast et al., 2012; Yilmaz et al., 2013). Based on analysis of a mock community of equimolar amounts of genomic DNA from 14 known bacterial species, OTUs with a frequency of < 21 in any one sample were discarded as sequencing errors. After sequencing, there were a total of 52,098,424 raw sequence reads for all station and depth combinations (113 samples), with an average count of 461,048 reads per sample for the V4 region of the 16S SSU rRNA ( Supplementary Table 1 ). After trimming and joining paired reads followed by quality processing, 17,936,177 reads remained with an average count of 158,727 reads per sample.

For 18S eukaryotic sequence results, primer sequences were removed from demultiplexed paired-end reads using CUTADAPT version 3.0 (Martin, 2011). The R package DADA2 (version 1.2.0, Callahan et al., 2016) was used to quality filter, denoise, and trim reads, specifying a trimming length of 110 and a maximum number of expected errors (MaxEE) of 2. DADA2 was also used to remove chimeric sequences and export a table of amplicon sequence variants (ASVs). Taxonomic assignments were performed using the IDtaxa function in the R package DECIPHER (Murali et al., 2018) and the Protist Ribosomal Database (PR2; Guillou et al., 2012) using a confidence level of 70. Classifications were adjusted for consistency with the World Register of Marine Species (Ahyong et al., 2023). Eukaryotic phyla that did not contain marine protists were excluded from further analyses. A total of 17,936,177 reads remained after filtering, clustering and removal of chimeras, with an average of 138,727 reads per sample. For the V9 region of the 18S SSU rRNA, there were 16,132,082 reads from 112 station/depth combinations with an average of 144,045 reads per sample ( Supplementary Table 2 ). After QC, clustering and denoising, 12,709,078 sequences were kept for 6763 ASVs, with an average coverage of 113,474 reads per station/depth combination. The 18S reads were filtered further to only include protist taxa resulting in 4,425,842 reads in 1976 ASVs, with an average coverage of 39,516 reads per station/depth combination.

Data are deposited in publicly accessible repositories. Physical and chemical seawater data are archived at NOAA’s National Centers for Environmental Information ² (Alin et al., 2017). Sequencing reads are archived at the National Center for Biotechnology Information under BioProject PRJNA1018955 ³ . Nonsequence biological data are archived at NOAA’s National Centers for Environmental Information (Accession number 0265154) ⁴ .

Data exploration and statistical analyses were performed using PRIMER7 (Anderson et al., 2008; Clarke and Gorley, 2015), R statistical package 4.3.2 (R Core Team, 2021; Oksanen et al., 2022), and Stata 18 (StataCorp, 2023). When appropriate, data were transformed to satisfy parametric statistical assumptions, applying a log(1+x) transform to raw sequence reads and standardizing by sample for Bray-Curtis similarity/dissimilarity matrices. Univariate metrics (species richness [d], Shannon diversity index [H’(log _e )], Pielou’s evenness [J]) were calculated based on OTUs and ASVs (PRIMER7, Diverse; R, vegan). Principal components analysis (PCA) was performed to reduce variable datasets to influential subsets and summary indices (PRIMER7, PCA; Stata 18, Factor & principal components analysis), and canonical analysis of principal coordinates (CAP) was used to ordinate samples based on sequencing results (PRIMER7, CAP). Microbial community data were examined by permutational multifactorial and multivariate analysis of variance (PRIMER7, PERMANOVA; Stata 18, MANOVA; R, vegan). Network analysis was performed and displayed using R 4.3.2 (network, igraph, ggraph). For analyses utilizing depth intervals, they are: 1, surface to 10 m; 2, 11 to 20 m; 3, 21 to 30 m; 4, 31 to 40 m; 5, 41 m to 50 m; 6, 51 to 100 m; 7, >100 m.

The West Coast Ocean Acidification cruises are a subset of oceanographic surveys conducted since 2011 along the western coastline of the contiguous United States and portions of the Canadian west coast. Sampling for this microbial assessment was conducted in the northern portion of the cruise, ranging between latitudes 40.25°N and 52.40°N, encompassing the coastlines of northern California, Oregon, Washington, and southern British Columbia. Collecting microbial samples from the same water grabs used for analytical chemistry allows direct comparisons with microbial metabarcoding information without assumptions or adjustments for spatial or temporal differences in sampling. This direct correspondence is important for organisms with short replication periods and rapid responses to environmental conditions, such as bacteria and phytoplankton.

Overall, few physical and chemical measurements were extreme, and there was a considerable overlap in seawater profiles among locations, in spite of the large geographic range ( Figure 1 ). Samples collected from deeper than 10 m were modestly differentiated from each other, suggesting that although seasonal stratification was in early stages, vertical mixing was still prevalent at the time of the survey (late May to early June). An El Niño was in place during the survey ⁵ , which would have weakened typical upwelling along the coast. There were no extreme temperatures, with those in shallower waters (≤ 20 m) spanning 8.9–16.2°C and waters deeper than 20 m spanning 5.7–10.6°C. Only one sample was hypoxic (< 62 µmol kg⁻¹), and none were at eutrophic concentrations for dissolved inorganic nitrate (≥ 82.6 µmol kg⁻¹). For shallow samples, no phosphorous or silicate limitations for diatom or other phytoplankton growth were detected. There was no evidence of primary nitrite and ammonium maxima, which often coincide and are located at the base of the euphotic zone. Instead, dissolved nitrite and ammonium concentrations were distributed throughout the water column and increased with depth, a pattern typically observed at higher latitudes where light-limited primary production coexists with high levels of nitrification (Zakem et al., 2018). Nonetheless, there were depth-dependent variations in nutrients and carbon chemistry. Deeper samples exhibited higher dissolved inorganic nitrate, phosphate, and silicate, while shallower samples exhibited higher pH_T and dissolved oxygen ( Figure 2 ; Supplementary Figures 1 , 3 ). In general, seawater conditions across the survey sites could be considered moderate, and not under conditions associated with strong upwelling, marine heatwaves, or hypoxia.

Commonalities in bacterial and archaeal communities occurred across the geographic span of the survey. Proteobacteria were dominant at all locations and at all depths, and Bacteroidetes and Cyanobacteria were highly abundant at depths ≤ 50 m. The high abundances for Proteobacteria were due primarily to the families Rhodobacteriaceae and SAR11 clades, and for Bacteroidetes were due to the family Flavobacteriaceae. These families occur at high abundances in coastal northwest Atlantic waters, with Flavobacteriaceae associated with the particle-associated fraction and Rhodobacteriaceae and SAR11 clades associated with the free-living fraction (Zorz et al., 2019). They are also dominant in offshore northeast Pacific waters at depths < 50 m, and were identified as indicator taxa for the years preceding the marine heatwave of 2014–2015 (Traving et al., 2021). Flavobacteriaceae is considered central in maintaining multiple and variable microbial relationships as surface waters transition through seasonal temperature fluxes in northwest Pacific coastal waters (Chun et al., 2021). In our study, there was a strong association between Flavobacteriaceae and Rhodobacteriaceae in waters shallower than 50 m, with Cryomorphaceae (another Flavobacteriales family) added to the association in the midwater depths, and subsequent replacement of Flavobacteriaceae by Cryomorphaceae and Halieaceae at depths > 50 m. Cryomorphaceae are highly abundant across the seasons in temperate surface waters (< 10 m; Korlević et al., 2022), and nutrient requirements of cultured isolates suggest the genus relies on enriched organic sources, likely contributing to secondary production (Bowman, 2014). Halieaceae (also known as NOR5/OM60 clade) have a world-wide coastal distribution, are highly abundant at the surfaces of intertidal sediments (Yan et al., 2009), and based on analysis of cultured isolates, are capable of chemotrophic or heterotrophic metabolism including utilization of polysaccharides from phytoplankton blooms (Li et al., 2023).

At depths > 100 m, the second most abundant phylum was the archaeal Thaumarchaeota (also known as Marine Group I archaea), composed of the family Nitrosopumilaceae and two genera, Nitrosopumilus and Nitrosopelagicus. Both genera are characterized as chemolithoautotrophic ammonia oxidizers with the ability to fix inorganic carbon (such as bicarbonate), broad salinity tolerance, and pH optima ≤ 7.3 (Könneke et al., 2005; Santoro et al., 2015; Qin et al., 2017). Originally considered to inhabit extreme environments such as oxygen minimum zones, Nitrosopumilus and Nitrosopelagicus isolates have been cultured from coastal and open ocean habitats and from photic zone to ocean depths, indicating a more cosmopolitan distribution (Qin et al., 2014; Santoro et al., 2015; Qin et al., 2017). The high metabolic activity, flexible physiology (i.e., mixotrophy), and tolerance of low pH and low dissolved oxygen may position these archaea to be winners under conditions of ocean acidification, and we identified both genera as indicator taxa for potentially corrosive seawater conditions (pH ≤ 7.75, DIC > 2190 µmol kg⁻¹).

Among phytoplankton, Prochlorococcus cell abundance exceeded Synechococcus and eukaryotic nano- and picophytoplankton abundances by at least an order of magnitude ( Figure 4 ). This pattern was contrary to the generalized expectation of greater Prochlorococcus abundance in oligotrophic waters and greater Synechococcus abundance in higher nutrient, cooler coastal waters (Pierella Karlusich et al., 2020), and the absence of obvious upwelling conditions during the cruise may have favored the former cyanobacterium. Prochlorococcus cells were well distributed across depth and in water temperatures as low as 6°C, reminiscent of the HLI and LLI ecotypes that are adapted to high and low light levels, respectively, and that occur at higher latitudes (Partensky and Garczarek, 2010). Ecogenomic analysis of Prochlorococcus is an emerging research area with potential for discovering strains with previously undescribed physiological tolerances (Partensky and Garczarek, 2010).

Among zooplankton and non-phytoplanktonic eukaryotic microbes, several taxa were present at high relative abundances across the geographic extent of the cruise and in the depth profiles, suggesting these are important components of community structures. The taxa Chrysophyceae (golden algae), Litostomatea and Spirotrichea (ciliates), and Filosa-Thecofilosea (flagellates) span size ranges from 0.4 µm to greater than 20 µm and are dominant in Arctic melt pond communities (Xu et al., 2020). The depth associations of certain eukaryotic microbes in this survey were supported by known physiology. For example, the ciliated protozoans Trimyema (class Plagiopylea) detected in deeper water are known anaerobes with mitochondrial modifications to produce H₂ which is consumed by Trimyema’s endosymbiotic methanogens (Lewis et al., 2018). Some eukaryotic microbes with saprotrophic lifestyles, such as Labyrinthulomycetes, were abundant throughout the geographic extent of the survey. Temporal studies of this eukaryotic class found time-lagged correlations with highly abundant bacterial phylum Flavobacteriaceae, cyanobacteria Synechococcus, algae, and fungi (Xie et al., 2021), emphasizing the multiple processes they can play in moving nutrients by heterotrophy and remineralization through marine food webs. One abundant taxon of eukaryotic microbes across all survey locations was the gregarines (Gregarinomorphea; Supplementary Figure 13 ), well known apicomplexans that parasitize invertebrates. Assessments of the effects of gregarines on their host (e.g., arthropod, mollusks, worms) reveal that there is a spectrum of relationships, from beneficial to harmful, which may contribute to their widespread occurrence and high diversity (Rueckert et al., 2019).

Assessment for co-occurrences among 16S and 18S phyla identified two networks of interactions, one involving highly abundant taxa that occurred across the depth profiles and one involving less abundant taxa that were more abundant in deeper waters ( Figure 6 ). For the former network, bacteria and archaea were strong nodes for associations, with one group of phyla (Actinobacteria, Bacteroidetes, Cyanobacteria, Euryarchaeota, Verrucomicrobia) forming a tightly associated group and Proteobacteria serving as a separate strong node. A translatitudinal survey of the Atlantic Ocean using 16S and 18S metabarcoding found that Proteobacteria were a highly associated node for the free-living fraction, and Bacteroidetes and Verrucomicrobia were highly associated node for particle-associated fractions (Milici et al., 2016), indicating these taxa are central to microbial networks in both nearshore and open ocean waters. Interestingly, dominance of the nearshore network by bacteria and archaea is more similar to the Southern Ocean waters than the other open oceanographic provinces (Lima-Mendez et al., 2015), which may be a reflection of seasonal high productivity in temperate nearshore waters and in the Southern Ocean. The strongest nodes in the second network were the eukaryotic Perkinsea, Lobosa, and Streptophyta and the bacteria Acidobacteria and PAUC34f. Perkinsea include well described parasitic protists that can infect dinoflagellates, mollusks, and finfish, while Lobosa are mostly free-living amoeba that move by pseudopodial extensions. Streptophyta include some green algae and land plants, the latter presumably transported into the ocean by terrestrial runoff and rivers. The absence of an understandable functional association among these taxa suggests that co-occurrences might be driven primarily by physical factors, rather than biological interactions.

Eukaryotic phytoplankton exhibited stronger differentiation across the geographic groups than bacteria and archaea ( Figure 5 ). The patterns in eukaryotic phytoplankton community structures can result from physical factors such as wind stress and ocean currents operating in combination with mesoscale oceanographic features such as canyons and promontories (Hickey and Banas, 2003). Submarine banks and canyons contribute to the well documented retentive features of the Juan de Fuca eddy, located near the coast of Washington State and Vancouver Island, and Heceta Head, located near the central coast of Oregon State. For example, eukaryotic phytoplankton communities from Destruction Island through Barkley Sound (groups H through L, Figure 1 ) displayed similarity with each other, suggesting a retentive influence by the gently sloping coastal bathymetry and the Juan de Fuca eddy ( Figure 5B ). Terrestrial inputs such as river plumes, and especially the Columbia River, also influence phytoplankton community structure via nutrient input, hydraulically trapping cells near the coast, or acting as a conduit for community transport (Hickey et al., 2013). The potential significance of physical transport for both nutrients and particles is reinforced by the pattern of community similarity among sites proximal to each other, such as the Columbia River plume and Copalis Beach (F and G), Kildidt Sound and Hecate Strait (N and O), entrance to Queen Charlotte Strait and Johnstone Strait (M and P); Figure 5B ). Another contributing factor to the observed patterns of differentiation could be latitudinal variation in light intensity and cloud cover. The structure of ocean seascapes, which are analogous to terrestrial landscapes, can also influence the composition of microbial communities in localized areas (Kavanaugh et al., 2016).

The northern locations (groups L through P) exhibited shared microbial features, in spite of the large spatial separation among sites and geomorphic differences in adjacent coastal uplands. Bacterial and archaeal taxon richness was consistently and significantly higher at northern locations ( Supplementary Figure 6A ), and there was high similarity in bacterial and archaeal community structures ( Figure 5A ). The shortest water-based distance between Barkley Sound (L) and Hecate Strait (O) is more than 500 km, and the sites leading to and within Johnstone Strait (P) are closely bounded by land rather than exposed to open ocean locations (L through O). Although the microbial community similarity might be driven by similarity in ocean conditions, this was not reflected in the principal components analysis of the seawater parameters ( Figure 2 ), suggesting unmeasured factors (biotic and/or abiotic) may have contributed to community homogeneity. Studies of oligotrophic open ocean communities using drifters to track changes associated with moving water parcels have found community homogeneity over distances ranging from a few km up to approximately 50 km (Hewson et al., 2006). The stretch of water from Barkley Sound to Hecate Strait is east of the division of the North Pacific Current into the Alaska Current (turning north) and the California Current (turning south), potentially placing this region beyond the main influence of these large ocean currents or even the Haida Eddies that aggregate and transport coastal water and plankton (Peterson et al., 2011). The oceanography of Hecate Strait and Queen Charlotte Strait indicates complex horizontal recirculation of waters at multiple depths (Crawford et al., 1995), which could result in transport, mixing, and possible retention of water masses. Given the single timepoint nature of this survey, the observed homogeneity may have been serendipitous, and seasonal transitions are expected to induce greater microbial heterogeneity.

In contrast to shared patterns in bacterial and archaeal communities among the northern locations, there was stronger geographic differentiation for southern locations. Communities at the two southernmost locations (Cape Mendocino and Cape Ferrelo, A and B, respectively) were strongly different from each other and all of the other locations ( Figure 5 ). Both of these locations displayed narrow variations in temperature and salinity ( Figures 3B, C ), and Cape Ferrelo had low taxon diversity and evenness ( Supplementary Figures 6B, C ). Thalassotalea (Colwelliaceae) was identified as an indicator taxon for Cape Ferrelo, and the genus occurred nearly exclusively at the four southernmost locations (A–D) at depths > 10 m. Thalassotalea emerged as an important component of petroleum degradation during the 2010 Deepwater Horizon incident (Noirungsee et al., 2020), and genome analysis indicates that the genus has good capacity for metabolizing complex organic molecules, such as macroalgal-derived polysaccharides, and likely plays an important role in nutrient cycling (Kim et al., 2020).

Microbial communities from the Columbia River plume (F) and Copalis Beach (G) resembled each other but were distinctive from other locations ( Figure 5 ). These two locations are approximately 100 km from each other, and each is influenced by substantial fresh water outflows from the Columbia River (average ~7,500 m³ s⁻¹) and from the Chehalis (average ~180 m³ s⁻¹) and Copalis Rivers (average ~80 m³ s⁻¹), respectively. Historically and as recently as 2021, nearshore waters between the Columbia River and Copalis Beach are subject to seasonal hypoxia, driven by upwelling, variable water sources, and/or biochemical oxygen removal (Connolly et al., 2010; Peterson et al., 2013; Barth et al., 2024). Although their microbial communities were similar, the seawater features of these two locations were different, with Copalis Beach distinctive from all other locations ( Figure 2A ). Its samples contained high levels of dissolved nitrite and ammonium, relative to other locations ( Supplementary Figure 4 ), as well as high levels of dissolved inorganic phosphate, low pH_T, and low dissolved oxygen. The elevated concentrations of nitrite and ammonium in the presence of low dissolved oxygen suggests suitable conditions for anaerobic ammonium oxidation, or anammox, at this location (Zehr and Ward, 2002), but relative abundances of ammonia-oxidizing archaea (e.g., Thaumarchaeota) were low, and no ammonia-oxiding bacteria were identified. Elevated nitrite and ammonium concentrations can be due to local river inputs (e.g., agricultural runoff), atmospheric deposition (e.g., combustion byproduct), elevated rates of remineralization of organic matter, altered rates of nitrification and/or denitrification (e.g., due to hypoxia), or combinations of these factors (Voss et al., 2011). The adjacent uplands are forested areas with no significant agricultural activities or urbanization, and there is no obvious source or cause for the elevated ammonium and nitrite at this location.

Cape Flattery (J) was another geographic group that was distinctive: low levels of dissolved oxygen ( Figure 3D ); low pH_T/high DIC seawater ( Figure 3E ); low absolute abundances of bacteria, archaea, and phytoplankton ( Figure 4 ); and greater abundances of bacterial and archaea indicators for potentially corrosive waters ( Table 4 ). Cape Flattery is at the intersection of water flows from the Strait of Juan de Fuca and the along-coast Davidson and California currents, and it is located at the periphery of the Juan de Fuca (or Tully) eddy. Modest upwelling can occur there (Hickey and Banas, 2003) which may have contributed to the observed seawater chemistry, but the relatively low abundances of microorganisms suggest low water retention, perhaps due to rapid or frequent movement in water masses.

The survey included several sampling locations within the Columbia River estuary (E), and, not surprisingly, the community compositions were strikingly different from other locations ( Figure 5 ; Supplementary Figure 7 ). An extensive spatiotemporal and depth assessment by Fortunato et al. (2012) along the Columbia River through the estuary and plume and to the shelf edge detected distinctive spatial groups that obscured temporal variability (Fortunato et al., 2012). Our analysis found five genera that were indicators for the Columbia River estuary locations, and these taxa have a variety of geographic preferences and metabolic lifestyles. Sediminibacterium (family Chitinophagaceae), Pseudarcicella (family Spirosomaceae), and Algoriphagus (family Cyclobacteriaceae) are members of families associated with freshwater sediments and soils, and more recently, with microplastics (Miao et al., 2019; Pinnell and Turner, 2020; Rogers et al., 2020; Wright et al., 2021), and their presence is most probably linked to the estimated 5 million tons of sediment annually transported to the Pacific Ocean by the river ⁶ . The genus Solitalea (family Sphingobacteriaceae), which was detected only within the Columbia River estuary (E) and the adjacent coastal plume (F), may play a role in ammonium regulation, which is useful for wastewater treatment (Pang et al., 2022). The genus Flavobacterium (family Flavobacteriaceae) was detected only within the estuary, and potential significance is discussed in the pathogens section below.

Archaea were a prominent component of most communities, with Euryarchaeota and Thaumarchaeota each comprising ~5.5% of the overall sequence abundances, while Nanoarchaeota contributed < 0.1%. All Euryarchaeota were classified as Marine Group II, and this taxon was absent or at low relative abundance south of Newport (D) and detected at up to nearly 40% at depths < 20 m. In contrast to Thaumarchaeota, Marine Group II archaea are well distributed in photic zones and abundant in metabolically active fractions of the microbial community, strongly indicating a role in pelagic biogeochemical cycling (Zhang et al., 2015; Wemheuer et al., 2019). All Thaumarchaeota we detected were classified as Nitrosopumilaceae, including the genera Nitrosopumilus and Nitrosopelagicus, and exhibited a depth-related distribution. This family, well characterized as ammonia-oxiding archaea, is commonly associated with deeper waters and sediments, and capable of thriving under hypoxic or low pH conditions (Qin et al., 2017; Campbell et al., 2019). The prominent position of archaea in this survey without environmental extremes supports their wider role in nutrient cycle and ready availability when conditions change, such as upwelling events.

Although Flavobacteriaceae occurred at all locations and ten genera were identified, the genus Flavobacterium was detected only at locations in the Columbia River estuary (E). This genus is found in a wide variety of habitats and temperature ranges from freshwater to seawater (Loch and Faisal, 2015). Several species of Flavobacterium (F. columnare, F. psychrophilum, F. branchiophilum) are well characterized pathogens of finfish, including salmon (Starliper, 2011; Declercq et al., 2013). Over the past 20 years (2002–2022), an annual average of ~1.5 million adult salmon and ~7.9 million juvenile salmon migrate through the Columbia River estuary (Columbia River DART, 2023). At the time of our sampling (May 29, 2016), ~151,000 adult salmon and ~6.5 million juvenile salmon had passed into the estuary during that year (Columbia River DART, 2023). It is likely that the high numbers of salmon contributed to the Flavobacterium abundance at this location, and conversely, the estuary could represent a location where horizontal transmission of Flavobacterium among salmon could occur.

In contrast to Flavobacterium, the genus Tenacibaculum was detected in every sample and at every depth, with highest relative abundances at Destruction Island (H). Multiple species of Tenacibaculum cause disease (called tenacibaculosis) in a wide variety of finfish including pelagic (e.g., salmon, amberjack, sea bream) and epibenthic (e.g., sole, flounder; (Avendaño-Herrera et al., 2006; Fernández-Álvarez and Santos, 2018) fish. Although there are species of Tenacibaculum that do not have demonstrated disease-causing abilities, genomic analyses show that pathogenic and non-pathogenic strains are closely intertwined and there may be homologous recombination between species (Habib et al., 2014). For marine aquaculture, the genetic evidence for endemic colonization by Tenacibaculum, rather than introduction by long distance fish movements (Habib et al., 2014), suggests their ubiquity could pose a persistent challenge during growing operations.

The Francisella genus includes several species, including isolates from diseased marine shellfish and finfish that were originally classified as subspecies of known zoonotic agents (Brevik et al., 2011; Colquhoun and Duodu, 2011; Ramirez-Paredes et al., 2020). Although isolates from marine infections belong to the genus containing well-known human pathogens F. tularensis and F. philomiragia, the risk of zoonotic disease has been considered low (Birkbeck et al., 2011; Colquhoun and Duodu, 2011). One feature of the fish pathogen F. noatunensis is the ability to enter a low metabolic but persistent state termed viable but not culturable (VBNC; Duodu and Colquhoun, 2010). Induction of the VBNC state is usually in response to stressors such as starvation or unfavorable temperatures, and many pathogens such as Vibrio spp. adopt this strategy, allowing longer-term persistence in the environment (Oliver, 2010). Similar to Tenacibaculum, Francisella was detected at all depths and at all locations except the southernmost locations (Cape Mendocino, A; Cape Ferrelo, B) and Queen Charlotte Strait, M). The highest abundances occurred at Johnstone Strait (P), Destruction Island (H), Barkley Sound (L), and Cape Johnson (I).

In addition to fish pathogens, one potentially zoonotic genus was detected: Coxiella. The single species of Coxiella, C. burnetii, is a strictly intracellular bacterium and the causative agent of Q fever, which can result in abortions and still births in animals and humans (Eldin et al., 2016). C. burnetii infects a wide range of organisms (arthropods, reptiles, birds, and mammals including humans), and occurs in terrestrial, freshwater, and marine environments (Eldin et al., 2016). Among marine mammals, both pinnipeds and cetaceans can be infected, and population surveys found seroprevalence ranging from 24–50% from the outer coasts of central Oregon and southern Washington and the inland waters of the Salish Sea (Kersh et al., 2012), and as high as 80% among Alaskan pinnipeds (Minor et al., 2013). Infections among terrestrial agriculture animals are common (Eldin et al., 2016), and birds could serve as vectors through feeding on infected placenta (Gardner et al., 2023). In our survey, Coxiella was not detected south of Cape Johnson (I), with highest abundances in Canadian waters in Kildidt Sound (N), Hecate Strait (O), and Johnstone Strait (P). Recent efforts to employ genotyping to differentiate terrestrial from marine mammal-associated strains may enable a better understanding of the sources of Coxiella in marine environments (Gardner et al., 2022).

Microalgae capable of producing biotoxins or causing harm under bloom conditions were widely detected both geographically and across the depth profiles, and the identified taxa are well documented as problematic for seafood safety, marine organisms (including aquacultured and fisheries harvested species), and the marine ecosystem (Bates et al., 2020; Anderson et al., 2021; McKenzie et al., 2021). Alexandrium was a frequently detected HAB, with high abundances at northern Canadian locations (Kildidt Sound (N), Hecate Strait (O)), at the Columbia River plume (F), and at Heceta Head (C). As a dinoflagellate that includes a benthic resting cyst stage in its life history and vertical mobility of the vegetative stage, Alexandrium detection and distribution is likely to be greatly affected by temperature (controlling excystment and growth) and salinity (for horizontal transport; Anderson, 1997). Pseudo-nitzschia was also a dominant HAB detected in the survey, and eleven of the fifteen known toxic species are problematic along the west coast of the U.S. and Canada (Bates et al., 2020; Anderson et al., 2021; McKenzie et al., 2021). Our survey was conducted approximately eight months after the end of a severe bloom of Pseudo-nitzschia spp. extending from northern British Columbia, Canada to California in the U.S (McCabe et al., 2016). Fifteen years of intense documentation of Pseudo-nitzschia in the Southern California Bight was able to identify general oceanographic patterns favorable for bloom formation, but factors affecting distribution, magnitude, and toxin production are not yet known (Smith et al., 2018). Although the Heceta Bank eddy (C) and the Juan de Fuca eddy (K) are known “hot spots” for Pseudo-nitzschia blooms (Lewitus et al., 2012), the diatom genus was not at high relative abundance at those locations at the time of this survey ( Figure 8 ). Because Pseudo-nitzschia impacts the entire west coast of North America (Lewitus et al., 2012), targeted studies of bloom vs favorable but non-bloom conditions at such known “hot spots”, including associated eukaryotic microbial communities, is likely to yield better information about drivers of this diatom’s dynamics. Gyrodinium, a dinoflagellate capable of mass mortalities in shellfish and finfish, was a prominent HAB in a quarter of the geographic groups, around Cape Flattery and the Juan de Fuca canyon (J, K), Queen Charlotte Strait (M), and Johnstone Strait (P; Figure 8 ). Other HAB taxa with moderate abundances such as Heterocapsa, Gymnodinium, Chrysochomulina are known hazards for marine finfish and shellfish, especially those in aquaculture facilities (Karlson et al., 2021).