Samples were taken during spring season 2021 and during late winter and spring season 2022 at two stations in Bornholm Basin, Baltic Sea: HELCOM monitoring station TF0213 (LO: 15.984, LA: 55.25) and station BB23 (Lo: 15.44, LA: 55.17; around 8 nautical miles away from TF0213, details see Supplementary Table 1). Sampling was accomplished in collaboration with regular HELCOM Baltic Monitoring Program cruises of the Leibniz-Institute for Baltic Sea Research Warnemuende (IOW), IOW project cruises (“MARNET”, “PHYTOARCHIVE”) and GEOMAR Helmholtz Centre for Ocean Research Kiel project cruises (“Baltic Sea Integrative Long-term Date Series”, “BILTS”, cruise program “Winter cod 2021-2025”). Sampling followed the guidelines and methods for phytoplankton sampling as described in the international environmental monitoring programme of HELCOM (method manual: HELCOM, 2017a). Specifically, to obtain a representative phytoplankton community sample of the whole upper surface layer, plankton samples were collected by pooling equal amounts of water collected from defined depth at 1 m, 2.5 m, 7.5 m, and 10 m using a rosette water sampler with 12 Niskin bottles. A total volume of 200 mL of these mixed-samples, respectively for phytoplankton and microzooplankton abundance, were stored in 250 mL brown-glass bottles and immediately fixed with acid Lugol’s solution, final concentration of 2%. The oceanographic parameters salinity and Sea Surface Temperature (SST) of the upper 10m water column were collected on the cruises following HELCOM guidelines and downloaded from the IOW database “ODIN” (https://odin2.io-warnemuende.de/) or specifically measured for this project. Salinity and SST data represented the average of measurements taken with the CTD sensor from the surface layer to 10 m depth. Data for dissolved inorganic nutrients (nitrate, nitrite, phosphate) was also downloaded from the IOW database “ODIN” or specifically taken and measured for this project. More precisely, 15 mL of water was filtered through acetate filters (Sartorius), immediately frozen at -20 °C and later measured colorimetrically according to Grasshoff et al. (1999) by means of a Seal Analytical QuAAtro constant flow analyzer, following HELCOM guidelines. The data sets of SST, salinity and dissolved inorganic nutrients can be found in Supplementary Table 1. Temperature data is additionally presented graphically in Supplementary Figure 1.

Trophic categorization.

As foundation for the characterization of the trophic composition of communities, we categorized all identified species/taxa and groups by their trophic mode. To do so, we used information on species trophic modes available from published literature (search engine: web of science) and from the online databases World Register of Marine Species (WoRMS, http://www.marinespecies.org/), AlgaeBase (http://www.algaebase.org/) and Nordic Microalgae (https://nordicmicroalgae.org/, for details see Table 1, Supplementary Table 2). Knowledge and literature references concerning potential mixotrophic status of Cryptophyceae and most of the Prasinophyceae and Crysophyceae are limited, species-specific or even lacking. In the most resent literature on mixotrophic species, we identified and referred to Mitra et al. (2023), Leles et al. (2019).

Depending on plankton density, 25 mL or 50 mL of fixed samples were analyzed under an inverted microscope (Nikon Eclipse TE2000-S) using the Utermöhl technique (Utermöhl, 1958). Samples were left to settle down for 24h in sedimentation chambers before counting. All protists (besides heterotrophic ciliates) were determined to the lowest possible taxonomic level and categorized in size classes in line with HELCOM guidelines (Olenina et al., 2006; HELCOM, 2017b). At least 500 individuals per sample were counted to reduce the statistical counting error to <10%, and at least 50 individuals of the most abundant species were counted. For calculating species biovolume and biomass the software “OrgaCount” (AquaEcology, Oldenburg) was used. More specifically, the biovolume of each species was assessed by taking the respective nearest geometric standard, following Hillebrand et al. (1999) and the HELCOM biovolume list PEG_BVOL2021 (http://www.ices.dk/marine-data/Documents/ENV/PEG_BVOL.zip). Afterwards, the biovolume was converted into carbon content according to Menden-Deuer and Lessard (2000) for calculating species carbon related biomass.

For heterotrophic ciliates, whenever possible the entire sedimentation chamber surface, was counted to avoid counting bias towards rare taxa. For practical purposes, for highly abundant species, this was reduced to half the chamber surface. The ciliates were identified to the lowest taxonomic level and/or body-size classes. The biovolume was calculated using geometric proxies by Hillebrand et al. (1999). Carbon biomass was calculated using the conversion factors provided by Putt and Stoecker (1989).

Changes in size composition is known to be potentially related to changes in top-down control. To account here for feeding relationships between trophic levels, total phototrophic and mixotrophic carbon biomass was separated into two size-classes respectively: biomass of small-bodied (<20 µm) prey, which are predominantly grazed by ciliates (Haraguchi et al., 2018) and large-bodied (>20 µm) prey, which are predominantly grazed by dinoflagellates (Käse et al., 2021).

Within the heterotrophic ciliate community, temporal changes in size composition have been shown to reflect seasonal succession pattern and potential relation with prey (Montagnes et al., 1988). To test for changes in size-dominance, all aloricate and loricate ciliates were divided into four size categories, according to their body-size: 10-20 µm, 20-30 µm, 30-55 µm and >55 µm, following e.g., Horn et al. (2016), Aberle et al. (2007). For loricate ciliates (genera surrounded by a cup-shell, here only tintinnids), individuals smaller 30 µm were absent in this study.

All biomass values were checked for normality by using the Shapiro-Wilk test in RStudio (R version: R4.3.1). In case of non-normal distribution, biomass values were log-transformed before statistical analyses. If biomass included zero-values, the data set was log (x+1)-transformed. However, data sets including many zero values, partly stayed non-normally distributed after transformation.

Principal component analysis (PCA) was used to visualize relationships between trophic community composition and environmental parameters (dissolved inorganic nutrients, SST), utilizing the PCA package in R. Sampling events with missing nutrient data were excluded from this analysis (2021:14 April 2021 and 21 May 2021; 2022: 22 April 2022 and 18 May 2022).

Since heterotrophic ciliates and dinoflagellates tend to prefer prey of different sizes, prey size was categorized into small-bodied (<20 µm) and large-bodied (>20 µm) prey. Associations among the biomasses of heterotrophs (total heterotrophic ciliate biomass, total heterotrophic dinoflagellate biomass, biomass of ciliates 30-55 µm and >55 µm) and the biomass within size classes of potential phototrophic and mixotrophic prey (small bodied <20 µm, large-bodied >20 µm) were assessed with Pearson’s correlation analysis, where a significant negative correlation would be consistent with the presence of a grazing relation. Additionally, Pearson’s correlation analysis was used to test potential prey association between biomass of M. rubrum and biomass of cryptophytes.

Generalized linear model (glm) analyses in RStudio were used to test for a relation between SST and the respective biomass of each specific trophic group (biomass of phototrophic phytoplankton, primarily phototrophic dinoflagellates, mixotrophic dinoflagellates, mixotrophic ciliates, mixotrophic others, heterotrophic dinoflagellates, heterotrophic ciliates, heterotrophic others). Further, separate glm model analyses were used to test for a relation between nutrient concentrations (nitrate + nitrite and PO₄) and the respective biomass of each specific trophic group. The separation of the analyses was due to temporal data gaps for some of the model parameters (see Supplementary Table 1). Accordingly, a reduced data set was used to test additionally for interaction of nutrient concentrations (nitrate + nitrite and PO₄) and SST on the respective biomass of mixotrophic dinoflagellates, mixotrophic ciliates and other mixotrophs. The combined dataset from 2021 and 2022 was tested for interaction of SST and year on the respective biomass of tintinnids and P. catenata. Associations between SST and specific species biomass (tintinnids, Mesodinium pulex) were likewise assessed with glm model analysis.

The Shannon-Weaver index was calculated using RStudio to assess taxonomic diversity, focusing here on the mixotrophic dinoflagellate community.

The protist community in Bornholm Basin exhibited a clear pattern of succession in trophic composition during spring season. By the end of the winter season in mid-February (only samples from 2022), when total microplankton protist biomass was still very low, mixotrophic ciliates and dinoflagellates accounted for half of the protist biomass, followed by phototrophic phytoplankton (predominantly diatoms) which comprised 34% (Figures 1A, B). With the beginning of the spring season in March the phototrophic community increased steadily in biomass and dominated total protist biomass between 52% (2022) and 75% (2021, Figures 1A, B). Contrasting to our expectations, the mixotrophic fraction constituted about half of the total protist biomass at total protist bloom peak (Figure 1A). From late April onward dominance shifted to the mixotrophic fraction (57-76%) in both 2021 and 2022, which retained or even increased its presence throughout May (Figures 1A, B, 2). Within this mixotrophic fraction composition changed, from predominantly ciliates to a dominance of the small-sized mixotrophs Dinobryon spp. and Prymnesiales (mixotrophic others, Figure 1B). The contribution of the heterotrophic fraction to total biomass remained minor throughout the whole spring season, reaching a maximum of 19% (2022) at the beginning of May (Figures 1 A, B).

Interannual variability between the two years was evident in the middle of the spring period (April), characterized by a significant difference in the biomass contributions of the main phototrophic groups (Figures 1B, 2). In 2021 the primarily phototrophic dinoflagellates (mainly Peridiniella catenata) accounted for 61% of total protist biomass (Figure 1B). In contrast, total protist biomass in 2022 was dominated by the phototrophic phytoplankton fraction (mainly diatoms) (Figures 1B, 2). The contribution of primarily phototrophic dinoflagellates on total protist biomass in 2022 was even below 16% throughout the entire spring period.

The mixotrophic community was to a large part of the spring period dominated by the ciliate M. rubrum. Dinoflagellates exhibited an overall diversity of species and taxa however, they were predominantly represented by a few species or groups for most of the time (see Figure 3). Other mixotrophs, predominantly Prymnesiales, occurred at the end of the spring period and dominated in mid-May 2021 even over 60% of total mixotrophic biomass (Figure 3).

More specifically, M. rubrum accounted for approximately 65% of the total mixotrophic biomass already in February 2022, reaching its peak dominance in April (81-89% on total mixotrophic biomass). During this period the dinoflagellates comprised the second largest community, predominantly consisting of Gymnodinales, but also Heterocapsa rotundata and Peridinales (Figure 3). The highest mixotrophic dinoflagellate diversity, as measured by the Shannon-Weaver index, was observed in March in both years, when several dinoflagellate species occurred in nearly equal abundances (Supplementary Table 3). Moreover, the index was lower at almost all sampling events in 2021 compared to the same periods in 2022 (Supplementary Table 3).

Interannual differences in dominance between mixotrophic ciliates and dinoflagellates were primarily observed at the beginning of the spring season in March. In 2021, mixotrophic dinoflagellates (Gymnodinales and Apocalathium CXP) dominated the mixotrophic community, accounting for 77% of total mixotrophic protist biomass, while M. rubrum entirely dominated at this time in 2022 (92%, Figure 3).

The heterotrophic protist community exhibited an overall high diversity and clear pattern of succession. At the end of the winter season in early February 2022, species and groups summarized here as “other heterotrophs” (E. tripatita, choanoflagellates, Katablepharis remiga, Leucocryptos marina, Telonema spp.) dominated the community to a large extent (Figure 4).

With ongoing spring season, both in 2021 and 2022, dominance switched to ciliates and dinoflagellates (Figure 4). In both years, in March the contribution of ciliates to total biomass increased to about 40%, consisting predominantly of Strombidium spp., and Lacrymaria spp., and (only in 2021) also tintinnids. Also in both years, dinoflagellates then simultaneously increased in biomass (up to 44%) and diversity (Figure 4), predominantly consisting of Protoperidinium species, Amphidinium spp., Gyrodinium spirale, and in 2021 also Peridiniella danica. In early to mid-April 2021, dinoflagellates dominated the heterotrophic community, accounting for 58-82%. Towards late April dominance in both years shifted back to ciliates (mainly Strombidium spp., Figure 4).

The highest interannual variability was observed in May. While in 2021 the community composition fluctuated and was composed predominantly of heterotrophic dinoflagellates and other heterotrophs, in 2022 ciliates (Strombidium spp.) accounted for over 90% of total heterotrophic biomass (Figure 4). Despite this, diversity within groups was likewise low in 2021 due to dominance of one or two species (dinoflagellates: P. danica or Katodinium glaucum; other heterotrophs: choanoflagellates, Figure 4).

Size-classes within the heterotrophic ciliate community showed clear seasonal succession from winter to spring season with interannual variability during spring period. At the beginning of February 2022 biomass was composed exclusively by the smallest size-fraction (10-20 µm), Figure 5). With start of the spring season, size-fractions shifted to predominantly 20-30 µm and 30-55 µm with an increasing amount of large-size (>55 µm) ciliates (including both aloricates and loricates, Figure 5). During April and May 2022, a high proportion (70-98%) of the biomass was composed of ciliates within the 20–30 µm and 30–55 µm size classes (Figure 5). In contrast, in 2021, the large-size-fraction (>55 µm) had completely disappeared from spring bloom after mid-April (Figure 5). Instead, the smallest size-fraction (10-20 µm) increased again and reached 70% of the biomass from mid of May 2021 (Figure 5).

The total biomass of heterotrophic dinoflagellates was found to positively correlate with total biomass of both, large-sized phototrophs and large-sized mixotrophs (size >20 µm, Table 2), as indicated by Pearsons`s correlation analyses. This pattern was predominantly driven by the 2022 data set (Table 2).

Correlations between heterotrophic ciliate biomass and the various size classes of phototrophs and mixotrophs were less well defined (Supplementary Table 5). The biomass of heterotrophic ciliates was not significantly correlated with the biomass of phototrophs <20 µm (2021 data) or mixotrophs <20 µm (2021 and 2022 data). When focusing on specific size classes, the biomass of larger-sized ciliates, i.e. ciliates 30-55 µm and ciliates >55 µm, was significantly negatively correlated with the biomass of mixotrophs <20 µm in the 2021 data set, but not with the biomass of phototrophs <20 µm, nor with the biomass of phototrophs and mixotrophs >20µm, in either the 2021 and 2022 data set.

SST steadily increased from the beginning to the end of the spring period in both years (Supplementary Figure 1). Notably, SST in February 2022 was already about 1 °C higher than SST in March 2021. This elevated temperature level remained constant throughout the entire spring period (Supplementary Figure 1), reflecting an overall warmer spring in 2022 compared to 2021.

Dissolved inorganic nutrient concentrations, specifically nitrate, nitrite and phosphate, were available in high concentrations at the end of the winter period, providing favorable starting conditions for phototrophic protists. Over the course of the spring period concentrations of nitrate and nitrite decreased strong, leading to nitrogen-limited conditions in both years (see Figure 2, Supplementary Table 1). Phosphate likewise decreased, but remained available throughout the entire spring period (see Figure 2, Supplementary Table 1).

Environmental conditions were associated with the trophic community biomass and composition during spring bloom, as indicated by PCA and glm model analyses (Table 2, Supplementary Figure 2, Supplementary Tables 4, 6). SST was positively correlated with total mixotrophic biomass (Supplementary Figure 2A) and with biomass of mixotrophic other species (2021 and 2022, Table 2) and heterotrophic ciliates (2021, Table 2). In contrast, SST was negatively correlated with biomass of primarily phototrophic dinoflagellates (Supplementary Figure 2A, Table 2), phototrophic phytoplankton, heterotrophic ciliates and total phototrophic biomass (Supplementary Figure 2). It further negatively correlated with biomass of tintinnids (Table 2). Likewise, the respective biomass of tintinnids and P. catenata was significantly negative associated with an interaction term of SST and year of sampling (Table 2).

Clear associations between community composition and dissolved inorganic nutrient concentrations were minor and predominantly observed in the 2022 data set (Supplementary Figure 2B). Such might be attributable to the inclusion of February data in 2022, a period marked by still high nutrient concentrations, low SST and low total protist biomass, thereby amplifying differences within the dataset. Heterotrophic ciliates and mixotrophic other species, for instance, were associated with low dissolved inorganic nutrient concentrations (Supplementary Figure 2B). Further, total mixotrophic biomass was negatively correlated with concentrations of nitrate and nitrite, as shown by glm analysis (Table 2). Both, biomass of mixotrophic ciliates and of mixotrophic dinoflagellates was significantly negative associated with an interaction term of SST and concentrations of nitrate, nitrite and phosphate (Table 2).

Mixotrophs fulfil several trophic positions by functioning as important primary producers, prey for other mixotrophic and/or heterotrophic protists and top-down predators. They consume a significant fraction of small-sized phytoplankton and bacterioplankton production and play an important role in remineralization of macronutrients (Rivkin et al., 1999; Calbet and Landry, 2004). Mixotrophic abundance was overall assumed to be primarily regulated by environmental conditions, which favor mixotrophs and limit phototrophs, e.g., low dissolved inorganic nutrient concentrations or light availability (Mitra et al., 2014; Dobbertin da Costa et al., 2024). However, our statistical model analyses revealed no significant associations between mixotrophic community composition and dissolved inorganic nutrient concentrations. The combination of environmental factors, i.e. nutrient concentration and SST, partly explained community composition, predominantly driven by temperature. Further, results showed various differences of the composition of the mixotrophic dinoflagellate community over time, suggesting that the occurrence of mixotrophs was not driven by a single or a few primary environmental factors. Instead, Dobbertin da Costa et al. (2024) revealed that spatial and temporal differences in mixotrophic community composition can be explained by the combination of several environmental conditions which favor specific mixotrophic species and groups and their utilization of phagotrophy. Thus, a larger data set of measured environmental factors is needed in future studies to enhance the understanding of mixotrophic species succession. Prey-availability, which is partly related to one specific prey-species, and grazing pressure additionally affect the abundance of these species.

More specifically, the mixotrophic dinoflagellate community was strongly dominated by Gymnodinales spp., followed by Heterocapsa rotundata over the entire spring period in both years. Conspicuously, mixotrophic dinoflagellate biodiversity was lower at almost all samplings in 2022 compared to 2021, indicated by the Shannon-Weaver index. The higher SST in 2022 might have been one of the explanations, as for instance cold-water associated taxa (e.g., taxa from the Apocalathium complex), typically occurring in spring communities of the northern and central Baltic Sea (Klais et al., 2013) were completely absent during that year. Contrary to our expectations, mixotrophic dinoflagellates known to graze on M. rubrum for kleptoplasts like Amylax triacantha (e.g. Park et al., 2013) and several species of Dinophysis spp. (Hansen et al., 2013), occurred only in low abundances, despite M. rubrum`s high biomass and thus potential food source. At the end of the spring bloom, bacterivorous small-sized constitutive mixotrophic species like Prymnesiales and Dinobryon spp. dominated the community making up around 50% of the total protist biomass. They prefer higher SST and benefit from an increasing bacterial community, which typically follows the primary production peak. Additionally, they might have profited from an increasing release from ciliate grazing pressure (Makareviciute-Fichtner et al., 2020), which is linked to the rising copepod abundance at that time of the year (data not shown).

The heterotrophic fraction predominantly consisted of dinoflagellates and ciliates and made up only a minor part of the total protist community (max 19%). Nevertheless, this community exhibited a succession of various species and groups throughout the spring bloom period, potentially linked to both species-specific favorable environmental conditions and prey availability. In February (only 2022 data), the community was composed of ciliates and “other small heterotrophs”, predominantly K. remiga and L. marina. By early April in both 2021 and 2022, the community included a higher proportion of larger-sized heterotrophic dinoflagellates such as Protoperidinium species, Amphidinium spp., G. spirale and K. glaucum. Experimental studies have shown that this community feeds on a variety of species and size classes, ranging from small-sized cryptophytes to larger-sized mixotrophic dinoflagellates (K. glaucum: e.g. Naustvoll, 2000), and single-cell and colonial diatoms (Yamaguchi and Horiguchi, 2008 and references therein). Such prey organisms were present in large biomass in April, offering good food conditions in both quantity (biomass) and quality (prey-size, see Section 3) for heterotrophic dinoflagellates, which potentially resulted in the observed increase in biomass.

For ciliates, knowledge on trophic relations and even seasonal succession and community composition in marine areas is limited. Despite their important trophic role within the microbial loop, they are seldom included in plankton community studies nor sampled in (inter)national monitoring programs (Mironova et al., 2014). The few studies on ciliates in the Baltic Sea show that the magnitude of ciliates in spring varies strongly between areas and study years. Our monitoring study detected maximum biomass values of around 12 µg C L^-1 in April 2022. Similar maximum biomass values for heterotrophic ciliates were found in a study by Johansson et al. (2004) at Landsort Deep (Eastern Gotland Basin) in April 1998, and in a mesocosm study from Kiel Fjord (Aberle et al., 2007). In contrast, Van Beusekom et al. (2009), observed much higher values in the Bornholm Basin in a study from 2002/2003 with a total ciliate peak of about 200 µg C L^-1 in April, which is thought to be related to an increased sampling depth up to 20 m.

The community composition of ciliates is strongly affected by water temperature (Montagnes and Weisse, 2000; Aberle et al., 2007), in addition to considerable bottom-up regulation by bacteria and phytoplankton prey and mesozooplankton top-down grazing (Mironova et al., 2012). In the Baltic Sea, composition has been shown to undergo strong changes at SST 5-12 °C, shifting during the spring period in April from a cold-season associated community to a warm-season associated community (Mironova et al., 2012). Likewise, at the start of the spring season the community in Bornholm Basin was mainly represented by species, typically associated with cold seasons (Mironova et al., 2012) such as Balanion spp., M. pulex, ciliates 10-15 µm and more seldom species like Lacrymaria spp. Glm model analysis indicated that tintinnids in the Bornholm Basin seem to be further restricted to low water temperatures. Their occurrence was limited to March/early April in 2021 and was absent in 2022, when SST in February was already about 1 °C higher than in March 2021. Future warming could, therefore, arguably lead to a reduction and less frequent occurrence of this important group of protist grazers in the Bornholm Basin. In contrast, a study of Roskilde Fjord observed that tintinnids were most abundant in late spring and summer, a period when SST is high. The two areas thus appear to harbor different tintinnid communities, supporting Mironova et al. (2014), who noted that only about 10% of all identified ciliate species are distributed across most Baltic Sea regions.

The level of dominance of dinoflagellates and ciliates within the strictly heterotrophic community has been found to be positively correlated with the availability of suitable prey sizes (Hansen, 1991; Haraguchi et al., 2018; Käse et al., 2021). For example, in Roskilde Fjord, southern Kattegatt (Haraguchi et al., 2018), during spring, ciliates dominate the heterotrophic fraction, whereas heterotrophic dinoflagellates are rare (Haraguchi et al., 2018). This corresponds with a dominant role of small-sized phototrophic organisms <20 µm, which represent the preferred prey-size of ciliates. In Bornholm Basin, ciliates likewise dominated the heterotrophic community composition over most of the spring in 2022 and part of spring in 2021. However, contrastingly to Haraguchi et al. (2018), analyses detected no clear predator-prey relations between ciliates and small-sized phototrophic prey <20 µm over time. In contrast, high biomass of heterotrophic dinoflagellates, correlated significantly with periods when larger-sized prey >20 µm was predominantly available, following at least partly Hansen (1991); Käse et al. (2021).

The absence of clear pattern is potentially related to the high abundances of mixotrophic ciliates and dinoflagellates in this area, which likely graze on the same prey-size, thus enhancing the complexity of relations between trophic levels. M. rubrum, for instance, is known to graze on cryptophytes smaller than 15 µm (Teleaulax spp.), thus potentially competing against heterotrophic ciliates. Further, Johansson et al. (2004) mentioned that larger-sized heterotrophic ciliates potentially also graze on other, smaller-sized heterotrophic ciliates. It raises the question whether such is a general relation or species related.

However, our data do not allow us to speculate further on this, but gives evidence that in areas like Bornholm Basin, size-related succession within the phototrophic community during spring period is only partly explainable by the abundance of grazers. Future size-trait studies within the Central Baltic Sea area and other regions will be needed to assess potential correlations between prey-size and ciliate abundance in more detail.

In contrast, the size-distribution within the heterotrophic ciliate community may at least partly reflect a seasonality effect related to prey-size (size concept of Montagnes et al., 1988, Kivi and Setälä, 1995; Setälä and Kivi, 2003; Yang et al., 2014). Our results showed that small-sized ciliates (10-15 µm) dominated the community in winter. These ciliates are known to graze on bacteria (no data in this study), which are available in higher abundances early in the year (Jurdzinski et al., 2024). With end of the winter/early spring season, ciliates 20-30 µm in size increased in abundance and made up a consistent part of the ciliate community during the whole spring season. These ciliates are known to graze on picoplankton <5 µm (Setälä and Kivi, 2003), which likewise increased in biomass at the end of the winter season and was availability throughout the whole spring season (February/March 2022: 0.16 µg C L^-1; April-May 2021 and 2022: 1-20 µg C L^-1). With progression of the spring season, ciliates with a size of 30-55 µm and even >55 µm occurred in higher abundances, mainly larger-sized Strombidium spp. Although these groups are thought to feed on and benefit from the increasing number of nano-sized diatoms and other algae during the spring bloom (Kivi and Setälä, 1995; Setälä and Kivi, 2003), our results did not reveal consistent patterns between the biomass of large ciliates and those of phototrophs and mixotrophs <20 µm and >20 µm. This may indicate that grazing relations between ciliates and their prey may be less generalizable and potentially more specific than currently assumed.

Sea surface temperature is known as one of the main factors shaping community composition during spring season. It typically affects community composition in spring directly via species’ differences in temperature tolerances and preferences. Thus, the end of winter/early spring protist community in Bornholm Basin consisted predominantly of cold-water tolerating or adapted species, typically for the northern and Central Baltic Sea area (e.g. Hoglander et al., 2004; Paul et al., 2023) like the phototrophic diatoms Skeletonema marinoi and Thalassiosira spp., but also the primarily phototrophic dinoflagellate P. catenana and mixotrophic dinoflagellates like Apocalathium CXP. However, the obviously prolonged seasonal dominance of phototrophic diatoms in relation to P. catenata in mid spring 2022 compared to 2021 was predominantly triggered by a diverse Chaetoceros spp. community (Ch. wighamii, Chaetoceros spp., data not shown). We assume that this was related to the earlier increase in SST in Bornholm Basin (Δ mean SST March/beginning April: 0.5 °C) and the overall slightly weaker ice-winter in 2022 (Aldenhoff, 2022). Studies have shown that increasing SST favors Chaetoceros spp. (Vrana et al., 2023), whereas high biomass of P. catenata in Bornholm Basin is significantly correlated with low minimum spring SST and ice build-up (Paul et al., 2023).

With progression of the spring period, rising SST indirectly leads to stratification of the water column, which further reduces the input and availability of new nutrients in the shallower euphotic zone. Results showed that dissolved inorganic nutrient concentrations could be associated with overall trophic community composition and succession during the spring bloom period. Specifically, phototrophs initially dominated at high nutrient concentrations, whereas mixotrophs and heterotrophs became more prevalent when nutrient availability decreased. Concentrations of dissolved inorganic nitrite and nitrate had already declined by early to mid-April. Nitrate is considered one of the main limiting factors for new primary production, and its decline has been identified as a key driver of the rapid reduction in phototrophs, especially diatoms, during spring (Lips et al., 2014). In contrast, phosphate remained at relatively low levels throughout the spring period and likely had a lesser impact on trophic community composition. In many marine areas motile mixotrophs have been observed to increase in abundance and to dominate the protist community under such nutrient-limited conditions (e.g. Hjerne et al., 2019; Makareviciute-Fichtner et al., 2020; Anschütz et al., 2024). They are more stoichiometrically stable than phototrophs under decreasing or fluctuating inorganic nutrient ratios, as they can compensate for inorganic nutrient limitation through phagotrophy (Anschütz et al., 2024). Likewise, the mixotrophic and heterotrophic community strongly dominated total biomass in April and May in both study years. These trophic groups play a significant role in improving the nutrient transfer to higher trophic levels during periods of reduced dissolved inorganic nutrient availability (Balzer et al., 2023).