Water chemistry parameters (SPM, POC, both expressed in g m^-³; nutrients in mmol m^-³) were analyzed following protocols previously described in McGovern et al. (2020). Briefly, SPM and POC were determined gravimetrically and via acid-fumigation of GF/F filters, respectively. Dissolved inorganic nitrogen (DIN: NO₂ ^- + NO₃ ^- + NH₄ ⁺), phosphate (PO₄³^-), and silica (SiO₂) were analyzed at NIVA (Oslo) using a Flow Solution IV Analyzer and accredited protocols (Kaste et al., 2018). Chlorophyll a concentrations were determined fluorometrically on a Turner 10-AU fluorometer after methanol extraction (Parsons, 2013).

Qualitative and quantitative analyses of planktonic protists were performed at the Marine Ecology Department of the Institute of Oceanology, Polish Academy of Sciences (IO PAN), following established protocols (Utermöhl, 1958; Edler, 1979). Subsamples of 10–50 mL were transferred to sedimentation chambers and allowed to settle for 24 hours prior to examination. Protists were identified and enumerated using an inverted microscope (Nikon Eclipse TE-300) equipped with phase-contrast and interference optics. Due to high concentrations of suspended solids that hindered protist identification, eight samples from August and one from June, collected near river inlets and at the glacier front, were excluded from the analysis (details in Supplementary Table 1 ). Microplankton (>20 µm) were counted across the entire chamber surface at 100× magnification, while nanoplankton (3–20 µm) were enumerated at 400× magnification by scanning three transverse transects. For the most abundant taxa, counts were performed until at least 50 individuals were recorded, with the number of fields of view adjusted as necessary. Taxa were identified to the lowest possible taxonomic level and assigned to major taxonomic groups (class or phylum) according to the World Register of Marine Species (WoRMS; http://www.marinespecies.org). Cells were further classified by size into three categories: small nanoplankton (≤10 µm), large nanoplankton (10–20 µm), and microplankton (>20 µm). Indeterminate flagellates ≤13 µm were categorized as mono- or biflagellates and assigned to Eukaryota incertae sedis, following the nomenclature used in the European Arctic long-term monitoring program (IO PAN and UiT-The Arctic University of Norway). Additionally, two distinct unidentified taxa were recognized: “Indeterminate 1” and “Indeterminate 2”, corresponding to the IO PAN-UiT database entries Eukaryota incertae sedis (“ring”) and “cf. Raphidophyceae/Heterosigma”, respectively. These forms were distinguished by (1) unique morphological traits, (2) their notable contribution to the protist assemblages, and (3) potential for future taxonomic resolution. “Indeterminate 1” resembles a representative of the genus Telonema (class Telonemea), and its size places it near the upper limit of currently described members of this group. “Indeterminate 2” is most likely a member of Raphidophyceae, with at least some observed cells potentially belonging to Heterosigma, based on their numerous peripheral golden-brown discoid chloroplasts and a blackberry-like appearance – a response typical of this genus after fixation. Additional support for the possible link to Heterosigma comes from its bloom typically occurring in mid-summer in cold-temperate coastal waters, as well as from previous records in the Svalbard region. However, without molecular confirmation, both assignments remain tentative and based solely on morphological traits observed under light microscopy. Due to the limitations of Lugol’s fixative, including masking of chlorophyll fluorescence (Stoecker et al., 1994), trophic classification was not attempted.

To minimize the influence of highly localized conditions and to better capture the dominant environmental gradients across Isfjorden, sampling stations were grouped into four habitat categories ( Figure 1 ). These were defined as follows: river estuary (RE), located near river mouths; inner fjord (IF), situated in the innermost parts of the sub-fjords, either adjacent to glaciers or close to the shoreline; outer fjord (OF), positioned at the mouths of the sub-fjords and farther offshore than the inner stations; and the main basin (MB), representing the marine endpoints of the fjord system, with minimal terrestrial influence. This habitat classification was consistent with that used in the complementary zooplankton study by Vereide (2019), facilitating direct comparisons between the two datasets.

Water masses were classified using Ocean Data View (version 5.6.2), following definitions previously identified for Isfjorden (Nilsen et al., 2008): surface waters (SW) = Sal < 34, T > 1 °C, intermediate waters (IW) = 34 < Sal < 34.7, T > 1 °C, Atlantic waters (AW) = Sal > 34.9, T > 3 °C), transformed Atlantic water (TAW) = Sal > 34.7, T > 1 °C, Arctic water (ArW) = 34.4 < Sal < 34.8, -1.5 > T < 1 °C, winter cooled water (WCW) = Sal > 34.74, T < -0.5 °C) and local water (LW) = T < 1 °C. Freshwater content (FWC) within the upper 10 m of the water column – representing freshwater influence in Isfjorden – was calculated from CTD profiles at all stations using the method described by Proshutinsky et al. (2009). A reference salinity of 34.7 was applied, corresponding to the boundary between surface and advected waters in Isfjorden (Nilsen et al., 2008). In addition, the salinity difference between the surface and 10 m depth (dS) was calculated as an index of water column stratification at the time of sampling.

Subsamples were grouped by month and habitat to investigate temporal and spatial dynamics in protistan communities across the melt season and associated environmental gradients. This approach follows the methodology applied by Delpech et al. (2021) for bacterial and archaeal communities.

Analyses focused on overall protist abundance, with particular attention given to main groups – defined as those contributing at least 5% to total abundance – and to dominant taxa within those groups, as in the year-round Isfjorden study by Kubiszyn et al. (2017). In addition, total biomass (mg m^-3) was estimated using taxon-specific average carbon content, based on size classes and taxonomic identity. Biomass conversion factors were derived from the HELCOM Phytoplankton Expert Group (Olenina, 2006) and IO PAN internal datasets. These values were also applied in the food-web modeling study based on the same field campaign (McGovern et al., in preparation). Due to methodological constraints, ciliates were excluded from biomass estimates. Accurate quantification of ciliate biomass requires specialized fixation and preparation (Quantitative Protargol Staining, QPS; Montagnes and Lynn, 1993) to avoid significant shrinkage and morphological distortion (Stoecker et al., 1994). Furthermore, the dark coloration of Lugol’s fixative complicates ciliate identification under optical microscopy. Although ciliates contributed less than 1% to total abundance, their exclusion likely results in a slight underestimation of total biomass. Accordingly, statistical analyses were performed exclusively on abundance data. Supplementary Table 1 contains total abundance and biomass data for protist classes recorded during this study.

Associations between environmental variables and protistan community metrics were evaluated using Spearman’s rank-order correlation, performed in R version 4.4.0 (R Core Team, 2024) with the vegan package version 2.5-6 (Oksanen et al., 2017). Correlation matrices were visualized using the corrplot package (Wei and Simko, 2017).

Additional statistical analyses were conducted in PRIMER 7 (PRIMER-E, Plymouth, UK) with the PERMANOVA+ add-on (Anderson, 2008). Except for diversity estimates, all statistical analyses were based on abundance data from the main protist groups.

Community richness, diversity, and evenness were estimated using Hill’s numbers: N0, N1, and N2 (Hill, 1973; Chao et al., 2014), based on protists identified at the species level to minimize bias from potential overestimation of taxa assigned to higher taxonomic ranks. Specifically, N0 represented species richness, treating rare and abundant taxa equally; N1 indicated diversity, with moderate emphasis on common taxa; and N2 emphasized dominant taxa, reflecting the number of equally abundant species. To assess community evenness, the ratio of N2/N1 was calculated, representing the proportion of very abundant species relative to common species (Birks et al., 2016). A ratio of 1 indicated complete evenness across taxa.

To test for differences in community composition among months and habitats, and their interaction, permutational analysis of variance (PERMANOVA) with Monte Carlo sampling was applied. Pairwise post hoc comparisons were conducted using Monte Carlo sampling as well. Community structure variability was further explored using principal coordinates analysis (PCO) based on log-transformed abundance data and Bray-Curtis dissimilarities.

To identify the environmental variables most strongly influencing protistan community structure throughout the melt season, distance-based linear models (DistLM) were constructed and visualized with distance-based redundancy analysis (dbRDA), both based on Bray-Curtis distances. Marginal tests were initially performed to evaluate the effect of each explanatory variable individually. Subsequently, a stepwise forward selection procedure, based on the adjusted R² criterion (Legendre and Anderson, 1999), was used to build the final model, accounting for the combined influence of multiple environmental drivers.

Seasonal associations of protist taxa were illustrated using a heatmap, showing the contributions of taxa (representing >0.01% of total abundance) to overall community composition (cells m^-3). Community groupings were further examined using Similarity Profile (SIMPROF) tests and hierarchical clustering of taxa based on Bray-Curtis dissimilarity.

All boxplots were generated in OriginPro 2021 (version 9.8.0.200, OriginLab Corporation, Northampton, MA, USA) and finalized in CorelDRAW 2018 (Corel Corporation, Ottawa, ON, Canada).

A total of 107 taxa were identified at the species (s) level and 64 at the genus (g) level, spanning 18 classes. The identified taxa included: Bacillariophyceae (31 s, 22 g); Chlorophyceae (1 g); Choanoflagellatea (2 s); Chrysophyceae (2 s, 1 g); Ciliophora: Gymnostomatea (1 g), Hypotrichea (1 g), Litostomatea (1 s), Oligotrichea (16 s, 7 g), Prostomatea (1 s); Cryptophyceae (2 s, 2 g), Cryptophyta incertae sedis (1 s); Dictyochophyceae (3 s); Dinophyceae (40 s, 25 g); Euglenoidea (1 g); Imbricatea (1 s); Prasinophyceae (3 s, 2 g); Prymnesiophyceae (3 s, 1 g); Telonemea (1 s). Due to the absence of distinct morphological features, several individuals were identified only to higher taxonomic levels. These included: Bacillariophyceae (Pennales indet. 5-140 µm); Ciliophora indet. 20-160 µm and Oligohymenophorea indet.; Cryptophyceae indet.; Dinophyceae (Gymnodiniales 20-60 µm, Peridiniales 5-60 µm); Euglenoidea indet. Additionally, non-identified mono- and biflagellates (3–13 µm), classified as Eukaryota incertae sedis, were present in the samples. Two morphologically distinct yet taxonomically unresolved taxa were also recorded, categorized as “Indeterminate 1” and “Indeterminate 2.” These are illustrated in Figure 5 .

Alpha diversity exhibited clear seasonal and spatial variability ( Supplementary Figure 1 ). Species richness (N0) peaked in May, particularly at the outer stations (MB, OF), while nearshore river-influenced sites (RE) consistently displayed the lowest N0 values across all months, with a general increasing trend toward the main basin. Diversity indices N1 (expected number of equally common taxa) and N2 (expected number of equally abundant taxa) showed broadly similar patterns, both reaching notably higher values in June when comparing habitats across months. Across all months, N1 and N2 values generally increased along the gradient from the innermost fjord regions toward the open waters. An exception to this trend was observed in the IF stations during May and June, where both the median and variability of the diversity indices were substantially higher than in adjacent habitats. In May, the N2/N1 ratio – an indicator of community evenness – suggested greater equitability in less terrestrially influenced locations (MB, OF). By contrast, in June and August, the N2/N1 ratio revealed the highest evenness in nearshore habitats, particularly within the inner fjords (IF), with increasing dominance (lower evenness) observed toward the outer fjord areas (OF, MB).

Protistan abundance and size structure exhibited substantial spatial and temporal variability throughout the study period ( Figure 6 ). Seasonal patterns in total abundance closely mirrored biomass trends, which, in turn, corresponded well with variations in chlorophyll a concentrations – a proxy for primary producer biomass. Notably, these patterns differed markedly between the inner fjord areas (RE, IF) and the outer regions (OF, MB).

A limited number of protist groups, hereafter referred to as the main groups, accounted for the majority of the community composition, each contributing at least 5% to the total abundance ( Figure 7 ). These dominant groups included Bacillariophyceae, Cryptophyceae, Dinophyceae, Eukaryota incertae sedis (non-identified mono- and biflagellates, predominantly 3–7 µm in size), “Indeterminate 1”, “Indeterminate 2”, and Prymnesiophyceae. Collectively, these groups represented over 97% of total protist abundance.

Subsequent analyses based on these main groups revealed significant temporal and spatial variation in community composition. PERMANOVA confirmed significant differences between months (Pseudo-F = 24.975, p(perm) = 0.001), as well as significant interactions between month and habitat (Pseudo-F = 2.3994, p(perm) = 0.007), indicating considerable intra-seasonal and spatial heterogeneity in protistan communities.

In May, overall protist abundance was high, typical of the spring bloom period ( Figure 6a ), with a maximum of 17.84 × 10⁸ cells m^-3 recorded in surface waters at station IsG. This peak was primarily driven by massive colonies of the prymnesiophyte Phaeocystis pouchetii (≤10 µm), likely in its vegetative (mixotrophic) stage, which accounted for approximately 51% of total abundance across the study area ( Figures 7 , 8 ). Microplanktonic Bacillariophyceae, mainly Chaetoceros furcillatus and Thalassiosira nordenskioeldii, contributed around 36%, while mono- and biflagellates (approximately 4.7% of total abundance), potentially including flagellated forms of P. pouchetii, dominated in outer waters (OF, MB).

Spatially, the community structure exhibited a clear dichotomy: higher abundances were observed in the outer fjord (MB: 53.2%, OF: 30.2%), while inner habitats (RE: 10.8%, IF: 5.8%) supported lower proportions of total abundance, a pattern also reflected in community composition differences ( Table 1 ). In inner habitats, flagellates – particularly nano- and microplanktonic dinoflagellates of the order Gymnodiniales (Gymnodinium and Gyrodinium) – were the dominant contributors.

Biomass distribution diverged from abundance patterns. The highest biomass was recorded in MB, with a median of 91.66 mg C m^-3, compared to lower values across other habitats (median 24.44 mg C m^-3). This was accompanied by elevated chlorophyll a concentrations in MB (median 3.85 mg m^-3) relative to other areas (0.30–0.65 mg m^-3). Dinophyceae accounted for the largest share of total biomass (49.9%, mainly in IF and RE), followed by Bacillariophyceae (30.8%, primarily in MB), and Prymnesiophyceae (18.0%, mostly in MB and OF).

By June, the community had shifted toward a dominance of nanoplanktonic flagellates spanning the full class size range (3–20 µm; Figure 6b ). Numerically, the community was dominated by Chrysochromulina (Prymnesiophyceae), along with two unidentified taxa (“Indeterminate 1” and “Indeterminate 2”; Figures 7 , 8 ). Chrysochromulina and “Indeterminate 1” were widespread, with Chrysochromulina peak abundances at stations B-Outer (OF: 59.07 × 10⁸ cells m^-3) and B-RE (RE: 57.61 × 10⁸ cells m^-3), and with “Indeterminate 1” peak abundance at B-Outer (23.81 × 10⁸ cells m^-3).

In contrast, “Indeterminate 2” was limited to the stations in Adventfjorden, with the highest abundance in RE (station A-F1, 7.79 × 10⁸ cells m^-3), slightly lower densities in IF at 15 m depth (4.94–6.79 × 10⁸ cells m^-3), and a secondary peak in OF (IsA: 10.49 × 10⁸ cells m^-3). Overall, community abundance in June was exceptionally high for the post-spring-bloom period and showed minimal spatial differences ( Table 1 ). Compared to May, total protist abundance increased markedly across most habitats. Respective totals (× 10⁸ cells m^-3) for May and June were: RE (13.21; 98.58), IF (7.07; 65.22), OF (36.80; 135.93), and MB (64.99; 55.57). Biomass during this transitional period also remained high, surpassing spring values in the OF area (median June: 80.91 mg C m^-3 vs. May: 27.51 mg C m^-3). Biomass variability increased, particularly in inner habitats (RE, IF). Besides the dominant flagellates, unidentified biflagellates, Cryptophyceae, and Gymnodiniales (Dinophyceae) contributed substantially to the communities across Isfjorden.

By August, protistan communities were substantially depleted in both abundance and biomass, representing a taxonomic legacy of the early-summer assemblage ( Figure 6a ). The community, dominated by large nanoplankton (10–20 µm; Figure 6b ), varied significantly along the side fjords (RE vs. IF) and between extreme habitats (RE vs. MB; Table 1 ). The lowest protist abundances were recorded in the RE area, increasing gradually along the fjord axis toward the MB region. Station IsG, located at the outermost part of Isfjorden, recorded the highest protist abundance (5.89 × 10⁸ cells m^-3) and biomass (26.48 mg C m^-3).

In addition to previously observed flagellates (Cryptophyceae and naked dinoflagellates), the community included thecate Dinophyceae, such as autotrophic Prorocentrum cordatum, Azadinium spinosum, Gonyaulax gracilis, and the heterotrophic Protoperidinium bipes ( Figure 8 ). Bacillariophyceae were largely restricted to the MB area and occurred at low abundances (maximum 0.18 × 10⁸ cells m^-3 at IsG). The assemblage was composed primarily of Thalassiosira spp. (20–30 µm), Lennoxia faveolata, Chaetoceros furcillatus, Thalassiosira nordenskioeldii, Licmophora gracilis, and unidentified pennate diatoms (20–30 µm).

Spring blooms in the West Spitsbergen fjords typically occur between April and June, triggered by water column stabilization and increasing solar irradiance, and sustained by nutrient replenishment from winter convection (Hegseth and Tverberg, 2013; Hoppe et al., 2024). Their timing and species composition are shaped by a complex interplay of physical (light availability modulated by snow and ice cover, AW advection, wind-driven mixing), chemical (nutrient concentrations), and biological factors (micro- and mesozooplankton grazing, as well as “seed” populations from the water column, sea ice, or surface sediments) (Hegseth and Tverberg, 2013; Hegseth et al., 2019; Assmy et al., 2023). In highly advective systems such as Isfjorden, blooms may arise from locally germinating resting stages or be advected from distant open ocean regions, including the Fram Strait (Lampe et al., 2021).

Daily surface chlorophyll a concentrations derived from satellite data (CMEMS biogeochemical products, 4 km resolution, April 1 – May 11 2018) for the Nordic Seas (10°W–30°E, 70–82°N) revealed a notable increase in phytoplankton biomass in the second half of April on the West Spitsbergen Shelf, progressively advancing towards Isfjorden, where peak values were detected in early May. This timing places the bloom approximately two weeks earlier than reported by Lampe et al. (2021), who described two distinct blooms in the Atlantic sector of the Fram Strait in 2018: the first in mid-May, followed by a stronger bloom in mid-June. These discrepancies likely stem from methodological differences in data collection. Lampe et al. (2021) employed a 15-day moving average over a broader spatial domain (southeastern Fram Strait), whereas our study focused on daily chlorophyll a values in the waters adjacent to the Svalbard shelf. Given this localized approach, we propose that AW inflows and the associated offshore bloom could have influenced fjord bloom development. Supporting this, mooring data from Skogseth and Ellingsen (2019), deployed at the mouth of Isfjorden (78°03.653′N, 13°31.346′E; 30 m depth) between October 2017 and August 2018, recorded substantial inflows of TAW in mid-April (10–23 April), just before the satellite-detected phytoplankton peak in Isfjorden.

Further insights from high-resolution in situ data collected at the “Isfjorden Adventfjorden” (IsA) time series station (UNIS, 2011–2019; Vader et al., 2020, 2024) suggest that the 2018 bloom was relatively early, with lower chlorophyll a concentrations compared to 2012–2016 but similar to 2017–2019. The bloom at IsA peaked around 1 May (≥5 mg m^-3), consistent with satellite observations and supported by Nyeggen (2019), who reported similar values (3.59 mg m^-3) for the same date. By the time of our sampling (10–11 May), chlorophyll a concentrations at IsA had declined to approximately 0.5 mg m^-3 in surface waters, while UNIS data for the same site indicated subsurface maxima at ~75 m depth (IsA station depth: 86 m), reaching ~3 mg m^-3 – half the value of the early bloom peak. In comparison, chlorophyll a concentrations at our main basin (MB) stations ranged from 0.7 to 5.5 mg m^-3 at the surface (median: 2.8 mg m^-3) and from 1.2 to 5.4 mg m^-3 at 15 m depth (median: 5.0 mg m^-3), with the highest concentrations observed at stations IsK and ME-3. These findings suggest that, at IsA, the bloom had shifted to deeper layers, likely driven by nutrient depletion in surface waters and increased turbidity from Adventelva runoff, fostering subsurface chlorophyll a maxima, as observed previously by Meire et al. (2016). Additionally, given IsA’s location along the southern inflow pathway into Isfjorden, it is plausible that phytoplankton populations there were advected into the fjord alongside intruding oceanic waters, representing remnants of the offshore bloom.

Collectively, the comparison of our data with high-resolution seasonal UNIS observations, satellite records, and mooring data supports the conclusion that our sampling captured the late phase of the spring bloom (approximately one week after the peak), at chlorophyll a concentrations consistent with bloom thresholds (Nöthig et al., 2015). Furthermore, low nutrient concentrations at inner stations suggest that the bloom was already in decline in these areas, in contrast to the outer stations, where nutrient levels remained higher. Notably, the satellite-detected peak bloom in Isfjorden during 2018 coincided with bloom timing in 2013 and 2015, as reported by Chitkara et al. (2024). Across 2012–2018, these years exhibited similar patterns of sea ice cover in Isfjorden from late April to mid-May (primarily in Billefjorden and Tempelfjorden; https://cryo.met.no/), accompanied by nearly parallel chlorophyll a dynamics for both >10 µm and GFF fractions (Vader et al., 2020, 2024, 2025). Thus, similar spring bloom dynamics appear to emerge in years characterized by comparable sea ice conditions, as also demonstrated by Meire et al. (2016) in Godthåbsfjord (SW Greenland). However, the lower phytoplankton biomass recorded in spring 2018 relative to 2013 and 2015 indicates that additional factors beyond sea ice extent play a decisive role in controlling bloom intensity.

During our May sampling, protist communities reflected a typical late-bloom scenario, dominated by spore-forming diatoms (Chaetoceros, Thalassiosira) and the colony-forming prymnesiophyte Phaeocystis pouchetii, mirroring the floristic composition observed in Isfjorden during the springs of 2012–2014 (Kubiszyn et al., 2017; Chitkara et al., 2024). For most diatom taxa, we recorded a clear gradient along the fjord axis, with increasing cell numbers and community diversity towards the outer fjord. Only Chaetoceros furcillatus colonies and long chains of Thalassiosira nordenskioeldii were relatively widespread throughout the fjord, yet they still followed this spatial pattern.

P. pouchetii, commonly found in springtime assemblages across Svalbard fjords (e.g., Kongsfjorden: Piquet et al., 2014; Hornsund: Smoła et al., 2017; Van Mijenfjorden: Kvernvik et al., 2021), emerged as the overwhelmingly dominant species in May. This prymnesiophyte, often regarded as a biological indicator of Atlantification (Hegseth and Tverberg, 2013; Kubiszyn et al., 2017; Bischof et al., 2019; Orkney et al., 2020), typically succeeds diatoms in Arctic bloom succession, particularly under silicate-depleted conditions or elevated nitrate-to-silicate ratios – both characteristic of the late spring bloom (Ardyna and Arrigo, 2020). The species appears to rely on diatom cells or colonies in spring to initiate its own colony formation, a potential dependency that could help explain its variable expression under differing hydrographic conditions (Hegseth et al., 2019). Under low-light or deeply mixed conditions, environments less favorable for diatoms, it often occurs as single cells or small colonies, forming subsurface chlorophyll maxima (Sakshaug and Skjoldal, 1989; Rokkan Iversen and Seuthe, 2011; Lacour et al., 2017; Assmy et al., 2023; Hoppe et al., 2024). Beyond its capacity to persist across variable hydrographic settings, P. pouchetii may have further benefited from limited top-down control, which appears to have played little role in constraining bloom development. This reduced grazing pressure likely reflects the fact that P. pouchetii colonies are less readily consumed than diatoms (Ray et al., 2016). High zooplankton densities observed in May, particularly meroplanktonic cirriped nauplii (~20,000 ind. m^-3 in the RE and ~12,000 ind. m^-3 in the OF; Vereide, 2019), support the idea that colony formation provides an effective size-based refuge from predation (Nejstgaard et al., 2007; Grattepanche et al., 2011; Ryderheim et al., 2022). Altogether, the combination of favorable physical structure, physiological traits, and reduced grazing pressure likely contributed to the persistence of the May bloom. The abundance of herbivorous zooplankton further supports the notion of an early bloom onset, likely initiated in the innermost fjord regions (Vereide, 2019).

Our observations of P. pouchetii dominance in May align with broader Arctic trends that suggest an increasing prevalence of this taxon under ongoing climate change. Shifts favoring Phaeocystis over diatoms are expected to increase the proportion of organic carbon that is remineralized within the water column, thereby weakening the coupling between primary producers and traditional pelagic–benthic food webs (Rousseau et al., 2000). Changes in phytoplankton composition may also influence summer succession, as shown by the late-summer diatom bloom in Kongsfjorden in 2019, which capitalized on silicic acid remaining unutilized during the Phaeocystis-dominated spring bloom (Assmy et al., 2023). The rising prominence of P. pouchetii in the European Arctic therefore has significant implications for carbon cycling and trophic transfer efficiency, underscoring the need for sustained, long-term studies to understand the ecological interplay between Phaeocystis, diatoms, and changing environmental conditions.

The Phaeocystis–diatom assemblage was accompanied by a diverse but relatively low-biomass community of dinoflagellates and small flagellates, predominantly ≤10 µm in size, which were evenly distributed across Isfjorden. Among the dinoflagellates, unarmored (naked) taxa were more prevalent than thecate (armored) forms. Members of the order Gymnodiniales, particularly Gymnodinium and Gyrodinium species, were especially numerous. Due to fragile cell structures and limited distinguishing morphological features, many Gymnodinium and Gyrodinium taxa were identified only to the genus level with associated size classes – an approach commonly used in studies of this group (Kubiszyn and Wiktor, 2016). Improved resolution of these taxa will require integrative methodologies, including scanning electron microscopy (SEM), live-cell observations, or combined morphological and molecular analyses of micromanipulated specimens. Nonetheless, identifiable species within Gymnodinium included G. arcticum, G. gracilentum, G. galeatum, and G. wulffii, while the Gyrodinium assemblage was largely composed of G. fusiforme, G. grave, and G. spirale. These species are well-documented constituents of dinoflagellate communities in West Spitsbergen waters (Kubiszyn and Wiktor, 2016) and form persistent elements of the protistan plankton throughout much of the year (Błachowiak-Samołyk et al., 2015; Marquardt et al., 2016; Hegseth et al., 2019; Meshram et al., 2017; Dąbrowska et al., 2021). Additionally, unidentified nanoflagellates (≤10 µm), particularly biflagellates (3–7 µm), were detected primarily in the outer fjord areas (OF, MB), and likely represented the flagellated stages of P. pouchetii.

Marine-derived particulate organic matter predominated in May, likely sustained by early-season protists; however, near river inlets (RE), early signs of the approaching freshet – such as reduced salinity, elevated turbidity, higher NH₄ ⁺ concentrations, and the presence of potentially ice-associated diatoms (Pennales 20–30 µm, Fragilariopsis cylindrus, Navicula transitans) – foreshadowed the transition to a regime increasingly shaped by terrestrial inputs (McGovern et al., 2020). Increased NH₄ ⁺ concentrations, a typical regeneration product from bacterial activity, suggest rapid remineralization and entrainment from shallow benthic sources following sediment resuspension. This additional nitrogen input may play a pivotal role in shaping fjord-scale productivity, particularly in shallow sub-fjord regions where NH₄ ⁺ can be released from sediments or subglacial upwelling (Halbach et al., 2019; Hopwood et al., 2020). Moreover, the significant negative correlation observed between NH₄ ⁺ availability and both total chlorophyll a concentrations and the abundance of certain nanoflagellates (“Indeterminate 1”, Cryptophyceae, Prymnesiophyceae) highlights the potential importance of NH₄ ⁺ in structuring protistan communities and influencing their development in subsequent seasons.