Layout 1 ARTICLE Introduction Phytoplankton are microscopic algae and cyanobacteria, but despite being tiny, their intra- and interspecific range in size is huge. Interspecific linear cell sizes span 5 orders of magnitude, and cell volumes more than 8 orders of magnitude (Finkel et al., 2010; Ryabov et al., 2021). The amazing morphological phytoplankton diversity in terms of cell size and shape represents a successful adaptive strategy addressed at facing pelagic environment variability (Naselli- Flores, 2021). Specifically, cell size is the dominant trait affecting the physiology and ecological function of phytoplankton, including metabolic rate (growth, photosynthesis, respiration), light acquisition, nutrient uptake, sinking rate from the illuminated Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon Silvia Pulina1,2, Cecilia Teodora Satta1,2*, Bachisio Mario Padedda1,2, Jacopo Culurgioni3, Riccardo Diciotti3, Nicola Fois3, Antonella Lugliè1,2 1University of Sassari, Department of Architecture, Design and Urban Planning, Sassari; 2National Biodiversity Future Center (NBFC), Palermo; 3AGRIS, Agricultural Research Agency of Sardinia, Sassari, Italy ABSTRACT The predator-prey relationship is generally size-specific in the pelagic food webs. Phytoplankton cell size structure can provide information on the successive levels of consumers and therefore on the energy that can flow towards the top consumers. This work focuses on phytoplankton cell size structure in a coastal lagoon (Cabras Lagoon, Italy) considered one of the most important for fishing productivity in the Mediterranean. The inter-annual and seasonal dynamics of picophytoplankton (Pico, cell size <3 μm) and Utermöhl Fraction of Phytoplankton (UFP, cell size >3 μm) were considered during almost three years in relation to the temporal dynamics of selected environmental variables and zooplankton. Small-sized cells with a mean linear cell size <10 µm and a mean cell volume <103 µm3 mainly represented UFP along the entire study period. This size class contributed the most to total phytoplankton biomass (up to 86%) and density (up to 99%) during the first part of the investigation period. A compositional change was detected: smaller species of Chlorophyceae, Bacillariophyceae, filamentous Cyanophyceae, and autotrophic nanoflagellates thrived in the second part of the study, replacing larger Mediophyceae that dominated UFP at the beginning. Picocyanobacteria rich in phycocyanin were the dominant taxa of Pico along the entire investigation period and this size class contributed the most to total phytoplankton biomass (up to 30%) and density (up to 96%) at the end of the study. The observed shift towards different and even smaller UFP and Pico in the second part of the study was most probably due to complex interactions between top-down and bottom-up effects. Indeed, an increased temperature, a decreased salinity and decreased concentrations of nutrients (mainly ammonium and orthophosphate), as well as an increased grazing pressure of rotifers on the larger Mediophyceae were simultaneous with the changes detected in phytoplankton. The obtained results highlight a longer planktonic trophic web in Cabras Lagoon that includes small phytoplankton at the base, ciliates, rotifers, and copepods. This suggests low energy availability for planktivorous fish, with possible future relevant consequences for fishing activities in this coastal lagoon. Corresponding author: Cecilia Teodora Satta, University of Sassari, Department of Architecture, Design and Urban Planning, Via Pian- danna 4, 07100 Sassari, Italy; NBFC, National Biodiversity Future Center, Palermo 90133, Italy. E-mail: ctsatta@uniss.it Key words: cell size; pelagic food-webs; coastal lagoons; phyto- plankton; zooplankton; Mediterranean. Authors’ contributions: all the authors made a substantive intellec- tual contribution. All the authors have read and approved the final version of the manuscript and agreed to be held accountable for all aspects of the work. Conflict of interest: the authors declare no potential conflict of in- terest. Funding: none. Availability of data and materials: all data generated or analyzed during this study are included in this published article. Acknowledgments: the authors thank Dr Bastianina Manca and Dr Pasqualina Farina for analyses of algal nutrients, and fishermen’s consortium of Cabras Lagoon for logistic support during sampling activities. Received: 22 December 2022. Accepted: 14 April 2023. Publisher’s note: all claims expressed in this article are solely those of the authors and do not necessarily represent those of their affil- iated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher. ©Copyright: the Author(s), 2023 Licensee PAGEPress, Italy Advances in Oceanography and Limnology, 2023; 14:11112 DOI: 10.4081/aiol.2023.11112 This work is licensed under a Creative Commons Attribution- NonCommercial 4.0 International License (CC BY-NC 4.0). No n- co mm er cia l u se on ly S. Pulina et al.20 surface layer, and predatory grazing by zooplankton (Litchman et al., 2008). For these reasons, Litchman et al. (2010) have suggested considering cell size a master trait helpful in the analysis of phytoplankton responses to environmental conditions and their change. For example, being small is a particularly good strategy to deal with nutrient depletion because of the high surface-to-volume ratio that makes phytoplankton cells less diffusion-limited (Litchmann et al., 2007). Consequently, smaller- sized phytoplankton generally dominate phytoplankton in the open ocean where nutrients are depleted and primary production is low, while larger-sized phytoplankton are generally more dominant in nutrient richer coastal waters where primary production is higher (Cabré et al., 2016; Marañón et al., 2007). As primary producers at the base of aquatic food-webs, phytoplankton size structure can impact on grazers and, consequently, on the functioning of aquatic ecosystems and their relative ecosystem goods and services (Como et al., 2012, 2018). However, the functioning of plankton communities is complex, and in addition to trophic interactions, various kinds of less studied biological links, such as mutualistic symbiosis, commensalism, allelopathy and competitions, may be relevant (Kéfi et al., 2015; Pilosop et al., 2017; Russo et al., 2022). A size- based predation has been reported for plankton food chains and it is generally assumed that larger prey is eaten by larger predators: heterotrophic nanoflagellates (cell size <20 μm) feed on picoplankton (cell size <2 μm), microzooplankton (20–200 μm) feed on heterotrophic nanoflagellates and nanoplanktonic phytoplankton, and mesozooplankton (200–2000 μm) feed on microplankton (Peter and Sommer, 2012; Sommer et al., 2017a). This means that the energy flow in marine pelagic food webs is largely based on size specific predator–prey relationships (Boyce et al., 2015): a shift to smaller size phytoplankton at the base of the food web results in a lower quality food for mesozooplankton, making the carbon flow between primary producers and mesozooplankton longer, through heterotrophic nanoflagellates and microzooplankton, with relative losses of energy, nutrients and carbon which can reduce productivity of higher trophic levels as described by Berglund et al. (2007). Actually, the trophic diversity of plankton is huge, the functioning of marine pelagic food webs is complex and the efficiency of energy transfer is more variable than previously thought (D’Alelio et al., 2016a; Eddy et al., 2021). For example, there are pelagic tunicates centimeters in length that filter particles, including phytoplankton, four to five orders of magnitude smaller than themselves (Sutherland et al., 2010). Both planktonic metazoan and protozoan consumers may change their feeding behaviors based on the available resources, modifying the overall functioning of the plankton food web in different seasons (D’Alelio et al., 2016b). Seven-fold decrease in phytoplankton biomass translated into only a two-fold decrease in potential planktivorous fish biomass in a plankton community in the Gulf of Naples (D’Alelio et al., 2016a). Large temporal fluctuations in the dominance of primary producers (i.e. macroalgae vs. phytoplankton) and, thus, in the available food items, resulted in major changes in the diet and stable isotopic signatures of several primary consumers in the hypertrophic Yundang Lagoon (Xiamen, China; Zheng et al., 2020). The economic consequences of plankton composition changes (and food web alterations) are especially important in ecosystems that are strongly exploited for fishing, such as Mediterranean coastal lagoons. In general, coastal lagoons are transitional shallow ecosystems highly sensitive to natural and anthropogenic environmental changes, such as climate crises, whose impacts occur at global and local scales (Barbosa et al., 2010; Newton et al., 2014). Coastal lagoons function as tightly linked benthic-pelagic systems and, in addition to the environmental stressors and the pelagic biological interactions, phytoplankton structure and dynamics can be strongly influenced by the benthic environment in these ecosystems. Benthos can be a strong sink for phytoplankton biomass (for example, grazing by benthic size-specific filter feeders), or a source for meroplanktonic species from germination of resting stages in sediments (Cloern and Jassby, 2010) and/or for tychoplanktonic species removed from substrates by turbulence (Cahoon, 2016). Therefore, decoding phytoplankton spatial and seasonal dynamics is challenging in heterogeneous and dynamic ecosystems. In this regard, the analyses of multiannual data turned out to be a useful instrument for disentangling the intrinsic complexity of phytoplankton dynamics, at least at a local level (Morabito et al., 2018; Zingone et al., 2010). The scarcity of detailed observations, and the high variability within lagoons, seriously limited the establishment of general models of planktonic succession in lagoons as has been done for other marine and freshwater environments (Gilabert, 2001). To date, the environmental and biological factors affecting the entire plankton community composition and size structure in coastal lagoons have been little studied, preventing a deep understanding of plankton food web functioning in these ecosystems (Trombetta et al., 2021). Mediterranean lagoons are considered particularly vulnerable environments that deserve attention as the Mediterranean region is deemed as one of the most sensitive areas regarding on-going global warming and increased extreme climate events (Ferrarin et al., 2014; IPCC, 2021). Temperature rise is expected to favour the selection of smaller-sized phytoplankton with profound consequences for the aquatic food web structure and efficiency from the very basis to the upper trophic levels (Boyce and Worm, 2015; Polovina et al., 2012; Sommer et al., 2017b). However, local dynamics are strongly site-specific, and processes within each region can modulate the overall patterns observed at a global level (Chust et al., 2014; Lomas et al., 2022). A significant temporal variation in the trophic levels of dominant fish strongly related to fish size has been demonstrated in Cabras Lagoon, the largest coastal lagoon in Sardinia (Italy, Mediterranean Sea) and one of the most important for fishing productivity in the Mediterranean basin (Como et al., 2018). Cabras Lagoon is known to be a highly complex eutrophic ecosystem (Magni et al., 2022). Consistently, very high δ15N values in the small size seston (<55 µm), likely related to a high input of anthropogenic nitrogen from a large catchment area, resulted in a 15N enrichment maintained from primary producers up to the predators at higher trophic levels (Como et al., 2012). In addition, studies on ecological data collected since 1999 showed the affirmation of very small-sized phytoplankton (linear cell size <10 µm) in Cabras Lagoon (Padedda et al., 2012; Pulina et al., 2011, 2012). To verify if it was only an occasional feature of phytoplankton in this ecosystem, the present work focused on the size structure of phytoplankton from Cabras Lagoon. Two different phytoplankton size classes, the picophytoplankton (cell size <3 μm; Pico) and the Utermöhl No n- co mm er cia l u se on ly Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon 21 fraction of phytoplankton (cell size >3 μm; Utermöhl Fraction of Phytoplankton, UFP), were studied over three additional years (2017, 2018, 2019) not analyzed in previous studies. Inter- annual and seasonal phytoplankton dynamics were related to the temporal variation of environmental conditions in the lagoon and, for the first time in this ecosystem, to the temporal dynamics of zooplankton. We aimed: i) to confirm the dominance of small-sized phytoplankton in the lagoon also during the latest investigation period, suggesting that it is an intrinsic aspect characterizing phytoplankton of this ecosystem that reveals a low energy-efficient trophic web; ii) to detect which environmental variables affected significantly the temporal dynamics of the two phytoplankton size classes (Pico and UFP); iii) to assess if and how zooplankton grazing activity may have affected the temporal dynamics of the two size classes in the considered years. Materials and Methods Study area Cabras Lagoon (western Mediterranean Sea; 39°56′37″ N, 08°28′43″ E; Figure S1) is a eutrophic lagoon with a mean depth of 1.5 m and about 23.8 km2 large, where phytoplankton are the main primary producers. Freshwater originates from a large catchment of about 430 km2 and mostly arrives in the northern part of Cabras Lagoon, carried by the river Rio Mare’e Foghe. The communication with the adjacent Gulf of Oristano is on the southern opposite side of the lagoon and it happens mainly via narrow creeks that flow into a southernmost canal. Because of an ecological dataset available since 1999 (Padedda et al., 2010, 2012; Pulina et al., 2011, 2012, 2016, 2020; Satta et al., 2014), Cabras Lagoon has been included in the “Marine ecosystems of Sardinia” site of the Italian Long Term Ecological Research network since 2006 (https ://deims .org/d5071 d21-9c8f-47ccb534-1b116 2a5e0 9c; Morabito et al., 2018; Pugnetti et al., 2013). Samplings and chemical-physical analyses Data on phytoplankton and environmental variables derive from monthly samplings carried out over three years, from April 2017 to October 2019. Water samples were collected at three sampling stations, located along the natural gradient of salinity: station 1, near the lagoon’s connection with the sea; station 2, in the center; station 3, near the main input of freshwater (Figure S1). In situ, water transparency (Tra) was measured with a Secchi disk, and water temperature (Tem) and salinity (Sal) were detected using a multiparameter probe (YSI 6600V2). Water samples for phytoplankton and nutrients analyses were collected at about 20 cm depth: 250 mL samples for studying UFP (cell size >3 μm) were immediately fixed in 2% acid Lugol’s solution, and 100 mL samples for studying Pico (cell size <3 μm) and autotrophic nanoflagellates (cell size >3 μm) were immediately fixed with formaldehyde (2%). In the laboratory, orthophosphate (P-PO4), ammonium (N-NH4), nitrate (N-NO3), nitrite (N-NO2), silicate (Si-SiO4), total phosphorus (TP) and total nitrogen (TN) concentrations were obtained according to the methods of Strickland and Parsons (1972). Zooplankton samples were collected seasonally (in summer and autumn 2017, in winter, spring and summer 2018, in winter 2019) at station 2, with a hand-held conic net (mouth diameter 31 cm, mesh size 10 μm), towed for 30 m at 50 cm depth with speed of 0.8 - 1 m s-1 and immediately fixed in 70% ethanol. Phytoplankton analyses Five mL of fixed water samples were analyzed to estimate UFP cell density using Utermöhl’s technique (Utermöhl, 1958), with an inverted microscope (Axiovert 25, Zeiss, Oberkochen, Germany). Cell counts were made at magnifications of 200 X and 400 X for more easily identifiable species, and at magnifications of 1000 X for smaller cells from at least 10% of the total bottom area of the settling chamber. Non-fixed samples were also observed to facilitate the identification of species. The species were identified according to the taxonomic literature listed in Pulina et al. (2012) and grouped at the class level for data analysis. Cell sizes of 20 randomly selected individuals of each taxon were measured in each sample. Cell volumes were calculated approximating the shape of each taxon to a geometric shape following Hillebrand et al. (1999), Vadrucci et al. (2013) and “Atlas of Shapes” powered by LifeWatch Italy (http://phytobioimaging.unisalento.it/Products/AtlasOfShapes.a spx?ID_Tipo=0). For each taxon, the cell carbon content was obtained by applying the conversion formulas suggested by Menden-Deuer and Lessard (2000) to the mean cell volume. The biomass of each taxon, in terms of carbon content, was calculated by multiplying cell carbon content by the corresponding cell density. For Pico cell counts, from 2 to 5 mL (depending on cell density) of fixed water samples were filtered onto 0.2-μm black- stained polycarbonate membranes (Nucleopore). Duplicate slides were prepared and observed using a microscope (Axiovert 100, Zeiss) equipped with green (BP520–560 nm/FT580 nm/LP590 nm) and blue (BP450–490 nm/FT510 nm/LP520 nm) filter sets, at 1000 X magnification. At least 20 random fields of view and a minimum of 100 randomly selected cells of each taxon (picocyanobacteria rich in phycoerythrin, P-PE; picocyanobacteria rich in phycocyanin, P-PC; picoeukaryotes, P-Euk) were counted for each slide (MacIsaac and Stockner, 1993). When it was not possible to recognize a specific taxon, cells were counted as picophytoplankton undetermined (Pico- und). Total picophytoplankton (Pico) counts were the sum of P-PE, P-PC, P-Euk, and Pico-und. Cell sizes of about 200 randomly selected individuals from each Pico group were measured on each slide. The cell volume of each group was calculated assuming that the shape of the cell was spherical or cylindrical with hemispheric ends and using the Bratbak formulas (1985). To calculate cell carbon content, the conversion factors of 250 fg C μm-3 and of 220 fg C μm-3 were applied to the mean volume of picocyanobacteria and P-Euk, respectively (Tamigneaux et al., 1995). The biomass of each Pico group was obtained by multiplying the cell carbon content by the corresponding cell density. Autotrophic nanoflagellates (cell size 3-20 µm) were also counted during Pico cell counts. At least 200 randomly selected flagellate cells were counted and measured for each slide. Their cell volume and biomass were calculated as described above for UFP. For each sample, total UFP and Pico density and biomass No n- co mm er cia l u se on ly S. Pulina et al.22 were calculated as the sum of the cell density and biomass, respectively, of all taxa and groups observed, including autotrophic nanoflagellates into UFP. Since the arithmetic average of individual morphological traits does not properly represent the community structure because common and rare taxa receive the same weight when calculating the average, weighted average of volume was also estimated for the whole UFP community according to Kruk et al. (2015). Zooplankton analyses In the laboratory, fixed zooplankton samples were fractionated into mesozooplankton (Meso) and microzooplankton (Micro), using sieves of mesh size 200 μm and 20 μm, respectively, and preserved in fresh 70% ethanol. Each subsample was brought to the volume of 10 mL or 5 mL (depending on the richness of the subsample) by removal of the supernatant and analysed under a stereoscope (Leica M80, Wetzlar, Germany) at 20-50 X magnification to perform a qualitative-quantitative analysis by direct enumeration of the organisms in 1 mL aliquots. For each subsample, counted specimens were identified according to available literature (Avancini et al., 2006; Cicero et al., 2016). The individuals were grouped at different levels: phylum (Nematoda, Rotifera), subphylum (Naupli), class (Copepoda, Ostracoda, Bivalvia, Gastropoda, Polychaeta, Appendicularia), subclass (Teleostei), order (Decapoda). Taxa that never contributed at least 2% to total zooplankton density were not included into the data analyses. Metazoan eggs were not considered either, as they are not primary consumers. Data analyses Monthly data from the three sampling stations of the entire study period were included in all statistical analyses. In addition, seasons were considered as: winter = January – March, spring = April – June, summer = July – September, autumn = October – December. Autumn 2019 was only represented by the month October. To ascertain annual, seasonal, and spatial differences in Pico density and biomass, in UFP density and biomass, and selected environmental variables (Tra, Tem, Sal, N-NH4, N-NO2, N-NO3, Si-SiO4, P-PO4), three-way analysis of variance (ANOVA) was performed considering three fixed factors (years, seasons and sampling stations), individually and combined. Since autumn 2019 included only the month “October”, one-way ANOVA was also performed to assess significant differences in the same aforementioned biological and environmental variables among October 2017, October 2018, and October 2019, to confirm (or not) what we detected for the season “autumn”. Prior to ANOVA analyses, the two assumptions were confirmed for all pooled data: normal distribution (Kolmogorov-Smirnov test) and homogeneity of variance (Bartlett test). All data required logarithmic [ln(x)] transformation to meet ANOVA assumptions. When significant differences in the dependent variables based on factors were observed, post hoc Tukey’s pairwise comparisons test was performed. The Generalized Linear Mixed Model (GLMM) was used (R package MASS) to investigate the influence of selected environmental variables related to the total biomass and total density of Pico and UFP. This kind of model is an extension of the Generalized Linear Model (GLM) and provides a more flexible approach to analyse non-normal data when random effects are present (Bolker et al., 2008). Two GLMM were applied: one model included Tem, Tra, Sal, P-PO4, N-NO3, N- NO2, N-NH4 and P-PO4 as fixed terms (predictor variables) and the seasonality as the random variable (random term; Pinheiro et al., 2012); the other model included the same selected environmental variables and the months as the random term. All the statistical analyses were performed using R 2.15.2 software (R Core Team, 2020), and an a priori level of significance was established at p<0.05. Results Pico and UFP inter-annual and seasonal dynamics Monthly values of Pico and UFP biomass and density along the study period were reported in Table S1. Considering Pico inter-annual dynamics, after an initial peak in spring 2017 and another peak in spring 2018, maximum values of both biomass and density were detected at the end of the study period, from spring to autumn 2019 (Figures S2, S3). Pico biomass and density observed in 2019 resulted significantly different from the values observed in 2017 and 2018 (Figures S2, S3; Table 1; Tukey’s test: p<0.001). Specifically, spring density in 2019 was higher than spring density values in 2018 Table 1. Results of the three-way ANOVA (F-test and P-value) to assess significant differences in cell biomass and density of picophy- toplankton (Pico) and Utermöhl fraction of phytoplankton (UFP) among years, seasons and sampling stations. Interaction between factors (years x seasons, years x stations, years x seasons x stations) are also shown. Significant results (p<0.05) are in bold. Years Seasons Stations Years x Years x Seasons x Years x seasons stations stations seasons x stations F p F p F p F p F p F p F p Cell biomass Pico 10.30 <0.001 4.64 0.01 0.01 0.99 3.47 0.01 0.03 0.99 0.11 0.99 0.33 0.97 UFP 2.21 0.12 3.90 0.01 0.18 0.83 3.44 0.01 0.61 0.65 0.64 0.70 0.36 0.96 Cell density Pico 21.31 <0.001 6.84 <0.001 0.07 0.93 6.66 <0.001 0.15 0.96 0.22 0.97 0.41 0.93 UFP 16.80 <0.001 1.98 0.13 0.02 0.98 2.40 0.04 0.68 0.61 0.42 0.86 0.35 0.96 No n- co mm er cia l u se on ly Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon 23 (Tukey’s test: p=0.040), and summer and autumn values of both biomass and density observed in 2019 were higher than the values observed in summer and autumn in 2017 and 2018 (Tukey’s test: p<0.001). Even Pico biomass (ANOVA: F=26.21, p=0.002) and density (ANOVA: F=33.48, p=0.001) detected in October 2019 were significantly higher than the biomass (Tukey’s test: p=0.025) and density (Tukey’s test: p=0.019) observed in October 2018 (Tukey’s test: p=0.020) and higher than the biomass (Tukey’s test: p<0.001) and density (Tukey’s test: p<0.001) observed in October 2017. Pico biomass and density varied significantly among seasons of the same year as well (Figures S2, S3; Table 1). Autumn Pico biomass and density were higher than winter biomass (Tukey’s test: p=0.020) and density (Tukey’s test: p<0.001) in 2019, and summer Pico density was also higher than winter density in the same year (Tukey’s test: p=0.002). Spring Pico biomass and density were higher than summer biomass (Tukey’s test: p=0.030) and density (Tukey’s test: p=0.020) in 2017. Considering UFP inter-annual dynamics, maximum values of both biomass and density were detected at the beginning of the study period, from summer 2017 to winter 2018 (Figures S2, S3). UFP density in summer 2017 was significantly higher than the values observed in summer 2019, and UFP density in autumn 2017 was higher than the values observed in autumn 2018 (Figures S2, S3; Table 1; Tukey’s test: p=0.004 and Tukey’s test: p=0.031, respectively). Even UFP density in October 2017 was higher than in October 2018 and 2019 with a significant difference only with October 2018 (ANOVA: F=26.39, p=0.002; Tukey’s test: p=0.003). UFP biomass detected in winter 2018 was significantly higher than the values of biomass observed in winter 2019 (Figures S2, S3; Table 1; Tukey’s test: p=0.022). In addition, while no significant differences in UFP biomass among autumns of the different years were observed, UFP biomass in October 2017 was significantly higher than the biomass detected in October 2018 and in October 2019 (ANOVA: F=18.39, p=0.005; Tukey’s test: p=0.004 and Tukey’s test: p=0.03, respectively). UFP biomass and density also varied among seasons of the same year (Figures S2, S3; Table 1): autumn UFP biomass and density were higher than spring biomass (Tukey’s test: p=0.029) and density (Tukey’s test: p=0.031) in 2017, and winter UFP biomass was higher than spring biomass in 2018 (Tukey’s test: p=0.022). Excluding spring 2017, Pico contribution to total phytoplankton biomass and density increased throughout the study period, while UFP contribution decreased (Figure S4). The highest Pico contribution to total phytoplankton biomass (up to 30%) and density (up to 96%) was observed in spring, summer, and autumn 2019. On the contrary, the highest UFP contribution to total phytoplankton biomass (up to 86%) and density (up to 99%) was observed from summer 2017 to winter 2018. Significant differences in Pico and UFP biomass and density were not detected among sampling stations in Cabras Lagoon during the investigated years (Table 1). Pico and UFP composition Considering Pico composition, P-PE (mean linear cell size 1.16±0.27 µm and mean volume 0.86±0.43 µm3) and P-PC (mean linear cell size 1.23±0.40 µm and mean volume 0.96±0.55 µm3) were observed throughout the investigated period with similar contributions to total biomass and density (Figures S5, S6). P-PC dominated in all seasons in 2017 and 2019 with contributions always >60% to Pico total biomass and density. P-PE maximum contributions to Pico total biomass and density were observed in autumn 2017 and in winter 2018 (>20%). P-Euk (mean linear cell size 1.50±0.01 µm and mean volume 1.77±0.01 µm3) were observed only in spring 2018, when dominated Pico total biomass and density (85% and 71%, respectively). Considering UFP composition (Figures S5, S6), Treboux- iophyceae (mainly Chlorella sp., mean linear cell size 3.84±0.69 µm and mean volume 8.05±0.30 µm3) contributed the most to total UFP biomass in spring 2017 (43%). This class was even more important in terms of density, with a contribu- tion of at least 13% up to a maximum of 41% from spring 2017 to winter 2018. Mediophyceae (mainly Cyclotella/Thalas- siosira sp., mean linear cell size 9.74±6.73 µm and mean vol- ume 869.01±1680.44 µm3) dominated the total UFP biomass, especially at the beginning of the study period, from summer 2017 to spring 2018 and in autumn 2018 with a contribution >60%. The contribution of Mediophyceae to total phytoplank- ton biomass and density was lower in the second part of the in- vestigation period, especially from winter 2019 onwards, when a significant presence affirmation of other groups was also de- tected. Autotrophic nanoflagellates (mainly Pyramimonado- phyceae and Cryptophyceae undetermined, mean linear cell size 7.33±0.25 µm and 9.39±3.54 µm, respectively, and mean volume 37.57±8.50 µm3 and 130.68±115.65 µm3, respectively) contributed the most from spring 2018 to the end of the study period, dominating the total UFP biomass in winter (61%) and autumn 2019 (57%). The contribution of Bacillariophyceae (mainly Pennales undetermined, mean linear cell size 9.01±2.12 µm and mean volume 15.31±2.21 µm3) and of Cyanobacteria (mainly Pseudanabaena catenata Lauterborn, mean linear cell size 4.09±1.00 µm and mean volume 6.82±2.30 µm3) to total phytoplankton biomass was major from summer 2018 onwards (up to 23% in autumn 2019 and up to 18% in summer 2018, re- spectively), and the contribution of Chrysophyceae (mainly Ke- phyrion spirale (Lackey) Conrad, mean linear cell size 4.33±0.81 µm and mean volume 27.16±12.68 µm3) to total phy- toplankton biomass was up to 18% in summer 2019. Chloro- phyceae (mainly Monoraphidium minutum (Nägeli) Komárková-Legnerová, mean linear cell size 3.42±0.66 µm and mean volume 1.82±1.49 µm3) contributed especially in terms of density. Contribution was always >20% from autumn 2017 to spring 2019 and this class dominated the total UFP density in spring 2017 (57%), in autumn 2018 (62%), and in autumn 2019 (78%). As described above, the most abundant UFP taxa detected in the Cabras Lagoon throughout the investigated period were small in linear cell size (<10 µm) and in mean cell volume (<103 µm3; Figure S7). Considering all UFP taxa observed during the entire study period, those proliferating with cell densities >107 cell L-1 were those with a mean cell volume <1000 µm3, and taxa proliferating with cell densities >108 cell L-1 were those with a mean cell volume <100 µm3 (Figure S8). In addition, the observed change in UFP composition resulted in a reduction in the UFP community volume in the second part of the investigation period (Figure S9). No n- co mm er cia l u se on ly S. Pulina et al.24 Inter-annual and seasonal environmental conditions Monthly values of the selected environmental variables along the study period were reported in Table S2. According to the GLMM with seasonality as a random term, among the selected environmental variables, Temp, Sal, N-NH4, and P-PO4 significantly influenced the seasonal dynamic of both Pico and UFP throughout the multiannual period in Cabras Lagoon, but exactly in an opposite way: the environmental variables that positively influenced one phytoplankton size class, negatively influenced the other, and vice-versa (Table 2). The same findings came out from the GLMM with monthly as a random term, but in this case, N-NH4 affected significantly only UFP dynamics (Table S3). Considering only the environmental variables affecting significantly phytoplankton throughout the multiannual period in Cabras Lagoon, the values of Sal and P-PO4 differed significantly among sampling stations (Table 3). These differences never occurred in the same season of the same year, according to Tukey’s test. Looking at the inter-annual and seasonal dynamics, summer and autumn Temp values increased during the years, and autumn 2019 was statistically warmer than autumn 2017 (Figure S10; Table 3; Tukey’s test: p=0.002). October 2019 was significantly warmer than October 2017 and October 2018 (ANOVA: F=303.3, p<0.001; Tukey’s test: p<0.001). Spring and summer Temp were significantly higher than autumn in 2017 (Tukey’s test: p<0.001; winter 2017 data are not available) and winter in 2018 (Tukey’s test: p<0.001). Instead, summer and autumn were significantly warmer than winter in 2019 (Tukey’s test: p<0.001). Sal in 2017 was significantly higher than Sal detected in 2018 and 2019 (Figure S10; Table 3; Tukey’s test: p<0.001). Specifically, Sal observed in spring, summer and autumn 2017 was higher than the values observed in the same seasons of 2018 (Tukey’s test: p<0.001). Sal in October 2017 was significantly higher than Sal in October 2018 and in October 2019 (ANOVA: F=31.5, p=0.001; Tukey’s test: p=0.001 and Tukey’s test: Table 2. Results (t-value and p-value) of the generalized linear mixed-effect model (GLMM) applied on the Picophytoplankton (Pico) and Utermöhl fraction of phytoplankton (UFP) total cell density and on the Pico and UFP total biomass along the study period (2017, 2018, 2019) in Cabras Lagoon. The selected environmental variables (Secchi disk transparency, Tra; water temperature, Temp; salinity, Sal; ammonium, N-NH4; nitrate, N-NO3; nitrite, N-NO2; orthophosphate, P-PO4; reactive silica, Si-SiO4) were used as fixed terms and the seasonality was used as random term. Significant effects (p<0.05) are in bold. Cell biomass Cell density Pico UFP Pico UFP t p t p t p t p Tra -0.365 0.716 0.911 0.366 1.003 0.320 -0.825 0.412 Temp 0.849 0.399 -3.692 0.000 1.820 0.044 -3.971 0.000 Sal -2.452 0.017 6.842 0.000 -1.159 0.251 6.677 0.000 N-NH4 2.418 0.019 -2.995 0.004 1.045 0.300 -1.589 0.117 N-NO3 0.906 0.368 -1.027 0.308 -0.038 0.970 -0.883 0.381 N-NO2 -1.820 0.074 1.942 0.057 -1.161 0.250 1.510 0.136 P-PO4 -2.845 0.006 1.425 0.159 -2.852 0.006 2.024 0.047 Si-SiO4 0.182 0.856 0.841 0.404 1.237 0.221 -0.700 0.487 Table 3. Results of the three-way ANOVA (F-test and P-value) to assess significant differences in the selected environmental variables (Secchi disk transparency, Tra; water temperature, Temp; salinity, Sal; ammonium, N-NH4; nitrite, N-NO2; nitrate, N-NO3; orthophosphate, P-PO4; reactive silica, Si-SiO4) among years, seasons and sampling stations. Interaction between factors (years x seasons, years x stations, seasons x stations, years x seasons x stations) are also shown. Significant results (p<0.05) are in bold. Years Seasons Stations Years x Years x Seasons x Years x seasons stations stations seasons x stations F p F p F p F p F p F p F p Tra 6.03 0.00 9.41 <0.001 0.18 0.83 3.42 0.01 1.73 0.16 1.46 0.22 1.12 0.38 Temp 1.16 0.32 54.03 <0.001 0.19 0.83 4.52 0.00 0.05 0.99 0.09 0.99 0.13 0.99 Sal 67.83 <0.001 6.09 <0.001 7.14 0.00 10.14 <0.001 0.73 0.58 0.25 0.96 0.36 0.96 N-NH4 23.09 <0.001 3.12 0.03 1.99 0.15 2.32 0.06 0.30 0.87 0.31 0.93 0.23 0.99 N-NO2 4.91 0.01 6.04 0.00 1.68 0.20 5.69 <0.001 0.26 0.90 0.87 0.52 0.37 0.95 N-NO3 4.51 0.02 16.97 <0.001 6.65 0.00 4.84 0.00 1.16 0.34 6.08 <0.001 1.06 0.41 P-PO4 2.64 0.06 0.92 0.40 7.48 0.00 0.15 0.99 0.05 0.99 2.40 0.05 0.05 0.99 Si-SiO4 12.45 <0.001 19.85 <0.001 10.56 <0.001 0.74 0.62 4.86 0.00 0.11 0.98 0.54 0.85 No n- co mm er cia l u se on ly Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon 25 p=0.007, respectively). Sal in spring 2017 was also higher than Sal in spring 2019 (Tukey’s test: p=0.002). Sal in winter 2018 was higher than Sal in winter 2019 (Tukey’s test: p<0.001). Considering seasonal intra-annual differences, Sal in spring and summer was higher than Sal in winter in 2019 (Tukey’s test: p<0.001); Sal detected in autumn was significantly higher than in winter (Tukey’s test: p<0.001) and spring (Tukey’s test: p=0.03) in 2019. Regarding algal nutrients, N-NH4 values in 2017 were significantly higher than values observed in 2018 (Figure S10; Table 3; Tukey’s test: p<0.001) and 2019 (Tukey’s test: p<0.001). Specifically, values observed in spring 2017 were significantly higher than values observed in spring 2018 (Tukey’s test: p<0.001) and 2019 (Tukey’s test: p<0.001); values observed in autumn 2017 were significantly higher than values observed in autumn 2018 (Tukey’s test: p=0.01). Even the concentrations detected in October 2017 were only significantly higher than the concentrations detected in October 2018 (ANOVA: F=29.02, p=0.002; Tukey’s test: p=0.001). N-NH4 values varied significantly among seasons only in 2017, with spring values higher than summer ones (Tukey’s test: p=0.04). P-PO4 values detected in 2017 were significantly higher than the values observed in 2018 and 2019 (Figure S10; Table 3; Tukey’s test: p<0.001). Specifically, P-PO4 values observed in summer and spring 2017 were significantly higher than values observed in summer (Tukey’s test: p<0.001) and spring 2019 (Tukey’s test: p=0.04). Also, P-PO4 in summer 2018 was significantly higher than P-PO4 in summer 2019 (Tukey’s test: p<0.001). P-PO4 values varied significantly among seasons only in 2018, with winter values lower than summer values (Tukey’s test: p<0.001). Zooplankton temporal dynamics and composition Zooplankton density was <1 ind L-1 until winter 2018 and increased strongly in spring 2018 (maximum peak of 2.59 ind L-1) and in summer 2018 (Figure S11). The Meso fraction dominated the total zooplankton density from autumn 2017 (60%) to spring 2018 (100%), while the Micro fraction dominated in summer 2017 (70%), in summer 2018 (90%), and in winter 2019 (60%) (Figure S11). Regarding zooplankton composition, Polychaeta larvae represented zooplankton in both Meso and Micro fractions in summer 2017 (100% to the total zooplankton density) and contributed 63% to the total Micro density in autumn 2017 as well (Figure S12; Table S4). Copepods (mainly adults and copepodite stages) dominated total Meso density from autumn 2017 (91%) to spring 2018 (85%) and contributed up to 50% (mainly as copepodites and naupli) to the total Micro density in winter 2018, showing the same contribution of Crustacea naupli. Rotifera contributed 44% to the total Meso density in winter 2018 and dominated both Meso and Micro in summer 2018 (88% and 98%, respectively) and in winter 2019 (56% and 89%, respectively). Nematoda contributed 4% to the total Micro density in autumn 2017. Discussion In this work, the temporal dynamics of two phytoplankton size classes, Pico (cell size <3 µm) and UFP (cell size >3 µm), were investigated throughout a multiannual period of almost three study years, from April 2017 to October 2019, in a eutrophic Mediterranean lagoon. Results showed that the size structure of phytoplankton changed strongly, alternating periods of prevalence of UFP on Pico and vice versa: UFP dominated at the beginning of the study, especially from summer 2017 to winter 2018; Pico dominated the total phytoplankton density in spring 2017 and from summer 2018 to the end of the study period. Pico maximum cell density peak of 1.87x109 cell L-1 observed in summer 2019 doubled the UFP maximum peak of 0.95x109 cells L-1 observed in autumn 2017. The contribution of Pico to total phytoplankton biomass was clearly lower, up to 30% in spring 2017 and in autumn 2019 when maximum peaks of about 300 µg C L-1 were detected. The Pico community was mainly composed of picocyanobacteria, with a higher contribution of P-PC for a longer period compared to the contribution of P-PE. This is in line with the known preference of P-PC for salinity <25 (Bec et al., 2011; Xia et al., 2017), normally measured in Cabras Lagoon (Pulina et al., 2012). Such a relevant presence of Pico in Cabras Lagoon agrees with other studies that have indicated increased Pico presence in other Mediterranean eutrophic coastal lagoons (Bec et al., 2011; Collos et al., 2009; Pulina et al., 2018; Sorokin et al., 2004). Pico are well known to be the main contributors of biomass and primary productivity in oligotrophic regions, such as oceanic ecosystems (Wang et al., 2022), whereas they have been little investigated in coastal lagoons so far. For this reason, our findings add new ecological information on this component, underlining the necessity of considering them even in studies on phytoplankton from eutrophic transitional environments. The Pico cell density detected in Cabras Lagoon was higher than the values detected previously in other eutrophic Sardinian Lagoons (up to 108 cells L-1; Pulina et al., 2017 and 2018), and closer to values detected in other Mediterranean coastal lagoons (Venice Lagoon, Italy; Sorokin et al., 2004; Bagnas Lagoon, France; Bec et al., 2011), in an Australian coastal lagoon (Coorong Lagoon, Shapira et al., 2010) and other brackish waters worldwide (Baltic Sea; Mazur-Marzec et al., 2013; Neuse River Estuary, USA, Paerl et al., 2020). Comparison of biomass values is harder since Pico biomass data are scarce in the literature. Pico biomass values observed in Cabras Lagoon in the current work were much higher than the values detected previously in the other Sardinian lagoons (up to 60 µg C L-1; Pulina et al., 2017 and 2018) and much higher than the few values available for coastal waters (about 10 µg C L-1 in Chesapeake Bay, Atlantic Ocean; Cai et al., 2010, and in Ionian Sea; Caroppo et al., 2014) and for brackish waters (up to 180 µg C L-1 in Baltic Sea; Mazur- Marzec et al., 2013). In the present study, small-sized cells mainly represented UFP along the entire study period: they were <10 µm in mean linear cell size and <103 µm3 in mean cell volume. An UFP compositional change was also observed from spring 2018 onwards: the increased abundance of Pico at the end of the study was accompanied by the increased abundance of smaller species of Chlorophyceae (Monoraphidium minutum) and Bacillariophyceae (Pennales undetermined), smaller autotrophic nanoflagellates, and a smaller filamentous species of cyanobacteria (Pseudanabaena catenata), which replaced the larger Mediophyceae (Cyclotella/Thalassiosira sp.) which dominated at the beginning of the study when the maximum UFP abundances were observed. Already published results on No n- co mm er cia l u se on ly S. Pulina et al.26 ecological data collected previously reported small-sized phytoplankton in Cabras Lagoon (Padedda et al., 2012; Pulina et al., 2011, 2012, 2016). Specifically, cyanobacteria of picoplanktonic cell size (about 3 µm) were the most abundant during the ecological studies performed in the periods 2000- 2002 and 2007-2009 in this ecosystem, together with other small species (cell size 3-10 µm) of Cyclotella, Fragilaria, Thalassiosira, Chlorella and with Monoraphidium minutum among the most abundant taxa. The size and taxonomic structure of phytoplankton detected in previous studies were confirmed by data reported in the present work obtained with a more appropriate method of analysis applied to investigate Pico, the smallest fraction of phytoplankton. Hillebrand et al. (2022) reported that phytoplankton cell volume around 100 µm3, such as the main representative mean cell volume of the size class UFP in Cabras Lagoon, is a very interesting ecological trait: carbon fixation was, in fact, reported to be maximum at cell sizes around 100 μm3, leading to the lowest respiration to photosynthesis ratio at this value of cell volume. In the present work, inter-annual and seasonal dynamics of Pico and UFP were related to the temporal dynamics of selected environmental variables, and of micro- and meso-zooplankton for the first time in Cabras Lagoon. October 2019 (autumn 2019 in this study) was significantly warmer than October 2017 and October 2018 in Cabras Lagoon and it was characterised by temperatures very close to the summer values. In addition, with the narrowest range of temperature variation, from 24.3 °C to 27.9 °C, the summer 2019 plus October 2019 was the longest warm period in the present study. Changes in species composition with a shift towards smaller-sized organisms, as well as decrements in the average cell size of individual species, have been predicted as a response to ocean warming (Sommer et al., 2017b). A strong positive correlation between temperature and small phytoplankton, even among temperature and picocyanobacteria, has been extensively reported in the literature in relation to a direct effect of temperature or due to an indirect effect mediated by a lower nutrient availability in the water and/or a more intense zooplankton grazing activity (Collos et al., 2009; D’Alelio et al., 2020; Peter and Sommer, 2012, 2013). In an experimental activity, a plankton community from Cabras Lagoon was incubated in a laboratory excluding mesozooplankton and it was exposed to heating (Pulina et al., 2020). A decreased abundance of Pico was detected under warming together with a significant change in UFP taxonomic composition, consisting of the increase of the smaller Chlorella sp., which replaced the larger Cyclotella sp. and Thalassiosira sp. It was related to a direct and positive effect of heating on ciliate abundance in the absence of their predators (mesozooplankton), as well as a taxonomic composition shift of ciliates with the affirmation of ciliates Scuticociliatida that preyed upon nanoflagellates and Pico. Indeed, studies on eutrophic areas show grazing as one of the main drivers behind picocyanobacteria seasonal dynamics, surpassing the role of temperature and nutrient availability in structuring communities (Caroppo, 2015). Interpreting phytoplankton dynamics in relation to environmental changes in field-based study is tricky, because a lot of different variables including both environmental variables (at local and global scale) and biological relationships can interact and overlap with each other in nature. In the present work, the period summer- autumn 2019 (maximum Pico contribution) was the longest warm period, and values of salinity and of N-NH4 and P-PO4 were lower than the values observed in the summer 2017-winter 2018 period (maximum UFP contribution). According to statistics, smaller UFP species and Pico benefited from these conditions of warmer water and lower salinity and N-NH4 and P-PO4 concentrations at the end of the study period. Small-cell sized phytoplankton are expected to be favoured under lower nutrient concentration because of a higher surface area to volume ratio, a smaller diffusion boundary layer that enables rapid nutrient exchange through the cell surface, and a better competitive ability in nutrient utilisation compared to large-cell sized phytoplankton (Harris, 1986; Litchman and Klausmeier, 2008; Van de Whaal et al., 2020). Temperature, salinity and N- NH4 concentrations, strictly related to the freshwater flows through the lagoon, were identified as the main drivers of plankton community dynamics investigated in the hypersaline Coorong Lagoon (South Australia): during high freshwater flow, phytoplankton–zooplankton interactions dominated the food web, while at low flow, bacteria, viruses, and nano/picoplankton interactions were more dominant (Hemraj et al., 2017). Mesozooplankton (mainly copepods) are well known to feed on microzooplankton and phytoplankton >10 µm in linear cell size (especially diatoms large 20–200 µm) and >500 µm3 in cell volume (Lewandowska and Sommer, 2010; Lewandowska et al., 2014; Peter and Sommer, 2012). Pico, nanoflagellates, flagellates >100 µm, and filamentous cyanobacteria are instead well-known to be inedible for copepods (Paul et al., 2021). Among smaller zooplankton, rotifers are generally under-represented in zooplankton feeding studies, they are more abundant in freshwaters, and they prey upon bacteria, heterotrophic flagellates, small ciliates and phytoplankton (Bonecker and Aoyagui, 2005; Sweeney et al., 2022). Rotifers may be eaten by copepods but can also compete with copepods for the same prey (Fussman, 1996). Instead, ciliate feeding habits were studied in depth: they are recognized to prey mainly upon pico- and nanoplanktonic cell size, but they are also able to eat larger phytoplankton, such as many diatoms (Johansson et al., 2004; Strom et al., 2001). Therefore, according to the literature, the cell size and taxa composition structure of phytoplankton observed in the present work in Cabras Lagoon was not palatable to copepods, and they may have been more likely eaten by microzooplankton. Unfortunately, apart from rotifers, naupli and small copepode stages, the available data-set does not provide deeper insights into microzooplankton, and data on ciliates and heterotrophic phytoplankton are not available in the current work. A significant role of ciliates can only be hypothesized in Cabras Lagoon also according to the results described by Pulina et al. (2020) which demonstrated a significant grazing activity of ciliates on phytoplankton in this ecosystem. Looking at the temporal dynamic of phytoplankton and zooplankton in the present study, the strong decrement of UFP abundance and the shift in UFP composition towards smaller UFP taxa started in spring 2018, just when the maximum peak of mesozooplankton copepods was detected in the lagoon. Copepods may have prayed upon microzooplankton since microzooplankton was not detected at this time in the lagoon. No n- co mm er cia l u se on ly Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon 27 In turn, the increased abundance of the smallest phytoplankton may have been determined by the reduced grazing activity of microzoopankton on them. Rotifers appeared in Cabras Lagoon in winter 2018, just when salinity started to decrease, confirming the strong physiological constraints of saltwater on most rotifers, and that they are particularly sensitive to salinity changes, as highlighted by Suikkanen et al. (2013). In a study where plankton composition was investigated in three different North African coastal lagoons, rotifers dominated the zooplankton biomass in the less salty lagoon (Lake Manzala, Egypt, mean salinity of 10.85; Ramdani et al., 2009). The peaks of maximum abundance of rotifers observed in summer 2018 and in winter 2019 in Cabras Lagoon coincided with a strong decreased abundance of Mediophyceae, suggesting an intense rotifer grazing pressure on this phytoplankton group, while smaller phytoplankton thrived. This suggests a preferred feeding of rotifers on Cyclotella/Thalassiosira sp., but not on the smaller phytoplankton such as Monoraphidium minutum, small Pennales undetermined, autotrophic nanoflagellates and the filamentous Pseudanabaena catenata. Since the abundance of these smaller phytoplankton started to increase from spring 2018 onwards, just when the highest abundances of copepods first and rotifers then were detected, we can hypothesize that both copepods and rotifers preyed upon ciliates as well, reducing their grazing pressure on the smaller phytoplankton. In Mar Menor Lagoon (Spain), a Mediterranean coastal lagoon where the composition and temporal dynamics of both phytoplankton and zooplankton were investigated, copepods abundance decreased in late summer when a maximum peak of ciliates and of larger phytoplankton were also observed. On the contrary, the abundance of autotrophic nanoflagellates decreased in this period, suggesting a preferred feeding for copepods on large phytoplankton and ciliates, and for ciliates on the autotrophic nanoflagellates (Gilabert, 2001). In Cabras Lagoon, the longer warm period in 2019 may have kept high the grazing rates of copepods and/or rotifers on ciliates whose decreased grazing pressure on Pico may explain the strong affirmation of Pico from spring 2019 onwards, together with the decrement in salinity and concentrations of N-NH4 and P-PO4. Indeed, the literature reported extensively a direct positive effect of experimental warming on the planktonic top- predators (mesozooplankton or microzooplankton when mesozooplankton are absent) and indirect cascade effects in the trophic web on phytoplankton (Lewandowska and Sommer, 2010; Lewandowska et al., 2014; Sommer et al., 2007; Vidussi et al., 2011). Conclusions The dominance of phytoplankton large <10 µm in mean linear cell size and <103 µm3 in mean cell volume in Cabras Lagoon was highlighted in this work. According to our findings, we retain that both environmental variations and changes in zooplankton composition affected synergistically phytoplankton in Cabras Lagoon during the study period 2017- 2019. Specifically, higher temperature and lower salinity and N-NH4 and P-PO4 values, together with a higher grazing pressure of rotifers on the larger UFP species and very likely of copepods and rotifers on ciliates determined a shift towards the affirmation of even smaller UFP taxa and of Pico. This framework suggests a longer planktonic trophic web in Cabras Lagoon, including copepods, rotifers, ciliates, and small phytoplankton that may translate into a low availability of energy for planktivorous fish. References Avancini M, Cicero AM, Di Girolamo I, et al., 2006. Guida al Riconoscimento del Plancton dei Mari Italiani. Volume II: Zooplancton Neritico-Testo. ICRAM, 198 pp. Available from: https://www.isprambiente.gov.it/it/pubblicazioni/manuali-e- linee-guida/guida-al-riconoscimento-del-plancton-dei-mari Barbosa AB, Domingues RB, Galvão HM, 2010. Environmental forcing of phytoplankton in a Mediterranean Estuary (Guadiana Estuary, South-western Iberia): A decadal study of anthropogenic and climatic influences. Estuaries Coast. 33:324–41. Bec B, Collos Y, Souchu P, et al., 2011. Distribution of picophytoplankton and nanophytoplankton along an anthropogenic eutrophication gradient in French Mediter- ranean coastal lagoons. Aquat. Microb. Ecol. 63:29–45. Berglund J, Müren U, Båmstedt U, Andersson A, 2007. Efficiency of a phytoplankton-based and a bacterial-based food web in a pelagic marine system. Limnol. Oceanogr. 52:121–31. Bolker BM, Brooks ME, Clark CJ, et al., 2008. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24:127–35. Bonecker CC, Aoyagui ASM, 2005. Relationships between rotifers, phytoplankton and bacterioplankton in the Corumbá reservoir, Goiás State, Brazil. Hydrobiologia 546:415–21. Boyce DG, Frank KT, Leggett WC, 2015. From mice to elephants: overturning the ‘one size fits all’ paradigm in marine plankton food chains. Ecol. Lett. 18:504–15. Boyce DG, Worm B, 2015. Patterns and ecological implications of historical marine phytoplankton change. Mar. Ecol. Prog. Ser. 534:251–72. Bratbak G, 1985. Bacterial biovolume and biomass estimation. Appl. Environ. Microbiol. 49:1488–1493. Cabré A, Shields D, Marinov I, Kostadinov TS, 2016. Phenology of size-partitioned phytoplankton carbon-biomass from ocean color remote sensing and CMIP5 models. Front. Mar. Sci. 3:39. Cahoon L, 2016. Tychoplankton. In: M.J. Kennish (eds.), Encyclopedia of Estuaries. Encyclopedia of Earth Sciences Series. Springer, New York, USA. 778 pp. Cai H, Wang K, Huang S, et al., 2010. Distinct patterns of picocyanobacterial communities in winter and summer in the Chesapeake Bay. Appl. Environ. Microbiol. 76:2955–60. Caroppo C, 2015. Ecology and biodiversity of picoplanktonic cyanobacteria in coastal and brackish environments. Biodivers. Conserv. 24:949–971. Caroppo C, Musco L, Stabili L, 2014. Planktonic assemblages in a coastal mediterranean area subjected to anthropogenic pressure. J. Geogr. Nat. Disasters 4:121. Chust G, Allen JI, Bopp L, et al., 2014. Biomass changes and trophic amplification of plankton in a warmer ocean. Glob. Chang. Biol 20:2124–39. Cicero F, Marino G, Nasta E, et al., 2016. Biodiversità, habitat No n- co mm er cia l u se on ly S. Pulina et al.28 pelagico e specie non indigene – mesozooplancton. ARPA Sicilia ST3 Area Mare. EU Marine Strategy Framework Directive 2008/56/EC, 198 pp. Cloern JE, Jassby DJ, 2010. Patterns and scales of phytoplankton variability in estuarine–coastal ecosystems. Estuaries Coast. 33:230–41. Collos Y, Bec B, Jauzein C, et al., 2009. Oligotrophication and emergence of picocyanobacteria and a toxic dinoflagellate in Thau lagoon, southern France. J. Sea Res. 61:68–75. Como S, Magni P, Van Der Velde G, et al., 2012. Spatial variations in δ13C and δ15N values of primary consumers in a coastal lagoon. Estuar. Coast. Shelf Sci. 115:300–8. Como S, Van Der Velde G, Magni P, 2018. Temporal variation in the trophic levels of secondary consumers in a Mediterranean coastal lagoon (Cabras lagoon, Italy). Estuaries Coast. 41:218–32. D’Alelio D, Libralato S, Wyatt T, Ribera d’Alcalà M, 2016a. Ecological-network models link diversity, structure and function in the plankton food-web. Sci. Rep. 6:1–13. D’Alelio D, Montresor M, Mazzocchi MG, et al., 2016b. Plankton food-webs: to what extent can they be simplified? Adv. Oceanogr. Limnol. 7:67–92. D’Alelio D, Rampone S, Cusano LM, et al., 2020. Machine learning identifies a strong association between warming and reduced primary productivity in an oligotrophic ocean gyre. Sci. Rep. 10:3287. Eddy TD, Bernhardt JR, Blanchard JL, et al., 2021. Energy flow through marine ecosystems: Confronting transfer efficiency. Trends Ecol. Evol. 36:76–86. Ferrarin C, Bajo M, Bellafiore D, et al., 2014. Toward homogenization of Mediterranean lagoons and their loss of hydrodiversity. Geophys. Res. Lett. 41:5935–41. Finkel ZV, Beardall J, Flynn KJ, et al., 2010. Phytoplankton in a changing world: Cell size and elemental stoichiometry. J. Plankton Res. 32:119–37. Fussmann G, 1996. The importance of crustacean zooplankton in structuring rotifer and phytoplankton communities; an enclosure study. J. Plankton Res. 18:1897–915. Gilabert J, 2001. Seasonal plankton dynamics in a Mediterranean hypersaline coastal lagoon: the Mar Menor. J. Plankton Res. 23:207–17. Harris GP, 1986. Phytoplankton Ecology: Structure, Function and Fluctuation. Chapman and Hall, London: 384 pp. Hemraj DA; Hossain A, Ye Q, et al., 2017. Anthropogenic shift of planktonic food web structure in a coastal lagoon by freshwater flow regulation. Sci. Rep. 7:1–10. Hillebrand H, Acevedo-Trejos E, Moorthi SD, et al., 2022. Cell size as driver and sentinel of phytoplankton community structure and functioning. Funct. Ecol. 36:276–93. Hillebrand H, Dürselen CD, Kirschtel D, et al., 1999. Biovolume calculation for pelagic and benthic microalgae. J. Phycol. 35:403–24. IPCC 2021. Sixth Assessment Report WGII. Cross-Chapter Paper 4: Mediterranean Region. Johansson M, Gorokhova E, Larsson U, 2004. Annual variability in ciliate community structure potential prey and predators in the open northern Baltic Sea proper. J. Plankton Res. 26:67–80. Kéfi S, Berlow EL, Wieters EA, et al., 2015. Network structure beyond food webs: mapping non-trophic and trophic interactions on Chilean rocky shores. Ecology 96:291–303. Kruk C, Martínez A, Nogueira L, et al., 2015. Morphological traits variability reflects light limitation of phytoplankton production in a highly productive subtropical estuary (Río de la Plata, South America). Mar. Biol. 162:331–41. Lewandowska AM, Hillebrand H, Lengfellner K, Sommer U, 2014. Temperature effects on phytoplankton diversity – The zooplankton link. J. Sea Res. 85:359–64. Lewandowska AM, Sommer U, 2010. Climate change and the spring bloom: a mesocosm study on the influence of light and temperature on phytoplankton and mesozooplankton. Mar. Ecol. Prog. Ser. 405:101–11. Litchman E, Klausmeier CA, 2008. Trait-based community ecology of phytoplankton. Annu Rev Ecol. Evol. Syst. 39:615–39. Litchman, E, de Tezanos Pinto P, Klausmeier CA, et al., 2010. Linking traits to species diversity and community structure in phytoplankton. Hydrobiologia 653:15–28. Litchman E, Klausmeier CA, Schofield OM, Falkowski PG, 2007. The role of functional traits and trade-offs in structuring phytoplankton communities: scaling from cellular to ecosystem level. Ecol. Lett. 10:1170–81. Lomas MW, Bates NR, Johnson RJ, et al., 2022. Adaptive carbon export response to warming in the Sargasso Sea. Nat. Commun. 13:1211. MacIsaac EA, Stockner JG, 1993. Enumeration of phototrophic picoplankton by autofluorescence microscopy, p. 187-197. In: P.F. Kemp, B.F. Sherr, E.B. Sherr and J.J. Cole (eds.), Handbook of methods in aquatic microbial ecology. Lewis Publisher. 800 pp. Magni P, Semprucci F, Gravina MF, 2022. Joint analysis of macrofaunal and meiofaunal assemblages improves the assessment of lagoonal environmental heterogeneity. Estuar. Coast. Shelf Sci. 266:107740. Marañón E, Cermeño, Jodríguez J, et al., 2007. Scaling of phytoplankton photosynthesis and cell size in the ocean. Limnol. Oceanogr. 52:2190–8. Mazur-Marzec H, Sutryk K, Kobos J, et al., 2013. Occurrence of cyanobacteria and cyanotoxin in the Southern Baltic proper. Filamentous cyanobacteria versus single-celled picocyanobacteria. Hydrobiologia. 701:235–52. Menden-Deuer S, Lessard EJ, 2000. Carbon to volume relationships for dinoflagellates, diatoms, and other protist plankton. Limnol. Oceanogr. 45:569–79. Morabito G, Mazzocchi MG, Salmaso N, et al., 2018. Plankton dynamics across the freshwater transitional and marine research sites of the LTER-Italy Network Patterns fluctuations drivers. Sci. Total. Environ. 627:373–87. Naselli Flores L, Zohary T, Padisák J, 2021. Life in suspension and its impact on phytoplankton morphology: an homage to Colin S. Reynolds. Hydrobiologia 848:7–30. Newton A, Icely J, Cristina S, et al., 2014. An overview of ecological status vulnerability and future perspectives of European large shallow semi-enclosed coastal systems lagoons and transitional waters. Estuar. Coast. Shelf Sci. 140:95–122. Padedda BM, Lugliè A, Ceccherelli G, et al., 2010. Nutrient-flux evaluation by the LOICZ biogeochemical model in mediterranean lagoons: the case of Cabras Lagoon (central- western Sardinia). Chem. Ecol. 26:147–62. No n- co mm er cia l u se on ly Dominance of small-sized phytoplankton in a Mediterranean eutrophic coastal lagoon 29 Paerl RW, Rebecca E, Venezia RE, et al., 2020. Picophytoplankton dynamics in a large temperate estuary and impacts of extreme storm events. Sci. Rep. 10:22026. Paul C, Sommer U, Matthiessen B, 2021. Composition and dominance of edible and inedible phytoplankton predict responses of Baltic Sea summer communities to elevated temperature and CO2. Microorganisms 9:2294. Peter KH, Sommer U, 2012. Phytoplankton cell size, inter- and intraspecific effects of warming and grazing. PLoS ONE 7:e49632. Peter KH, Sommer U, 2013. Phytoplankton cell size reduction in response to warming mediated by nutrient limitation. PLoS ONE 8:e71528. Pilosof S, Porter MA, Pascual M, Kéfi S, 2017. The multilayer nature of ecological networks. Nat. Ecol. Evol. 1:1–9. Pinheiro J, Bates D, DebRoy S, Sarkar D, 2012. R development core team, nlme: linear and nonlinear mixed effects models. R Package Version 3:1–105. Polovina JJ, Woodworth PA, 2012. Declines in phytoplankton cell size in the subtropical oceans estimated from satellite remotely-sensed temperature and chlorophyll, 1998–2007. Deep Sea Res. Part II: 77–80, 82–88. Pugnetti A, Acri F, Bernardi Aubry F, et al., 2013. The Italian Long-Term Ecosystem Research (LTER-Italy) network: results, opportunities, and challenges for coastal transitional ecosystems. Transit. Waters Bull. 7:43–63. Pulina S, Padedda BM, Satta CT, et al., 2012. Long-term phytoplankton dynamics in a Mediterranean eutrophic lagoon (Cabras Lagoon, Italy). Plant Biosyst. 146:259–72. Pulina S, Padedda BM, Sechi N, Lugliè A, 2011. The dominance of cyanobacteria in Mediterranean hypereutrophic lagoons: a case study of Cabras Lagoon (Sardinia Italy). Sci. Mar. 75:111–20. Pulina S, Satta CT, Padedda BM, et al., 2017. Picophytoplankton seasonal dynamics and interactions with environmental variables in three Mediterranean coastal lagoons. Estuaries Coast. 40:469–78. Pulina S, Satta CT, Padedda BM, et al., 2018. Seasonal variations of phytoplankton size structure in relation to environmental variables in three Mediterranean shallow coastal lagoons. Estuar. Coast. Shelf Sci. 212:95–104. Pulina S, Suikkanen S, Satta CT, et al., 2016. Multiannual phytoplankton trends in relation to environmental changes across aquatic domains: a case study from Sardinia (Mediterranean Sea). Plant Biosyst. 150:660–70. Pulina S, Suikkanen S, Padedda BM, et al., 2020. Responses of a Mediterranean coastal lagoon plankton community to experimental warming. Mar. Biol. 167:22. R Core Team, 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.r-project.org/index.html. Ramdani M, Elkhiati N, Flower RJ, et al., 2009. Environmental influences on the qualitative and quantitative composition of phytoplankton and zooplankton in North African coastal lagoons. Hydrobiologia 622:113–31 Russo L, Casella V, Marabotti A, et al., 2022. Trophic hierarchy in a marine community revealed by network analysis on co- occurrence data. Food Webs 32:e00246. Satta CT, Anglès S, Garcés E, et al., 2014. Dinoflagellate cyst assemblages in surface sediments from three shallow Mediterranean lagoons (Sardinia, north western Mediterranean Sea. Estuaries Coast. 37:646–663. Schapira M, Buscot MJ, Pollet T, et al., 2010. Distribution of picophytoplankton communities from brackish to hypersaline waters in a south Australian coastal lagoon. Saline Syst. 6:2. Sommer U, Aberle N, Engel A, et al., 2007. An indoor mesocosm system to study the effect of climate change on the late winter and spring succession of Baltic Sea phyto- and zooplankton. Oecologia 150:655–67. Sommer U, Charalampous E, Genitsaris S, Moustaka-Gouni M, 2017a. Benefits, costs and taxonomic distribution of marine phytoplankton body size. J. Plankton Res. 39:494–508. Sommer U, Peter KH, Genitsaris S, Moustaka-Gouni M, 2017b. Do marine phytoplankton follow Bergmann’s rule sensu lato? Biol. Rev. 92:1011–26. Sorokin PYu, Sorokin YuI, Boscolo R, Giovanardi O, 2004. Bloom of picocyanobacteria in the Venice lagoon during summer–autumn 2001: ecological sequences. Hydrobiologia 523:71–85. Strickland JDH, Parsons TR, 1972. A practical handbook of seawater analysis. Fisheries Research Board of Canada, Ottawa. 167 pp. Strom SL, Brainard MA, Holms JL, Olson MB, 2001. Phytoplankton blooms are strongly impacted by micro- zooplankton grazing in coastal North Pacific Waters. Mar. Biol. 138:355–68. Suikkanen S, Pulina S, Engström-Öst J, et al., 2013. Climate change and eutrophication induced shifts in northern summer plankton communities. PLos ONE 8:e66475. Sutherland KR, Madin LP, Stocker R, 2010. Filtration of submicrometer particles by pelagic tunicates. PNAS 107:15129–34. Sweeney K, Rollwagen�Bollens G, Hampton SE, 2022. Grazing impacts of rotifer zooplankton on a cyanobacteria bloom in a shallow temperate lake (Vancouver Lake, WA, USA). Hydrobiologia 849:2683–703. Tamigneaux E, Vasquez E, Mingelbier M, et al., 1995. Environmental control of phytoplankton assemblages in nearshore waters, with special emphasis on phototrophic ultraplankton. J. Plankton Res. 17:1421–47. Trombetta T, Vidussi F, Roques C, et al., 2021. Co-occurrence networks reveal the central role of temperature in structuring the plankton community of the Thau Lagoon. Sci. Rep. 11:17675. Utermöhl H, 1958. Zur vervollkommung der quantitativen phytolankton-methodik. Mitt d Internat Vereinig f. Limnologie 9:1–39. Vadrucci MR, Mazziotti C, Fiocca A, 2013. Cell biovolume and surface area in phytoplankton of Mediterranean transitional water ecosystems: methodological aspects. Transit. Waters Bull. 7:100–23. Van de Waal DB, Litchman E, 2020. Multiple global change stressor effects on phytoplankton nutrient acquisition in a future ocean. Phil. Trans. R. Soc. B. 375:20190706. Vidussi F, Mostajir B, Fouilland E, et al., 2011. Effects of experimental warming and increased ultraviolet B radiation on the Mediterranean plankton food web. Limnol. Oceanogr. 56:206–18. Wang F, Wei Y, Zhang G, et al., 2022. Picophytoplankton in the West Pacific Ocean: a snapshot. Front. Microbiol. 13:811227. No n- co mm er cia l u se on ly S. Pulina et al.30 Xia X, Guo W, Tan S, Liu H, 2017. Synechococcus assemblages across the salinity gradient in a salt Wedge Estuary. Front. Microbiol. 8:1254. Zheng X, Como S, Huang L, Magni P, 2020. Temporal changes of a food web structure driven by different primary producers in a subtropical eutrophic lagoon. Mar. Environ. Res. 161:105128. Zingone A, Phlips EJ, Harrison PJ, 2010. Multiscale variability of twenty-two coastal phytoplankton time series: A global scale comparison. Estuaries Coast. 33:224–9. Online supplementary material: Figure S1. Study site and location of sampling stations. Figure S2. Inter-annual and seasonal dynamics of (A) Picophytoplankton (Pico) and (B) Utermöhl Fraction of Phytoplankton (UFP) total biomass in Cabras Lagoon throughout the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S3. Inter-annual and seasonal dynamics of (A) Picophytoplankton (Pico) and (B) Utermöhl Fraction of Phytoplankton (UFP) total density in Cabras Lagoon throughout the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S4. Picophytoplankton (Pico) and Utermöhl Fraction of Phytoplankton (UFP) contribution to (A) total phytoplankton biomass and (B) density in Cabras Lagoon throughout the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S5. Contribution of (A) each Picophytoplankton (Pico) group to total picophytoplankton (Pico) biomass (P-PE, picocyanobacteria rich in phycoerythrin; P-PC, picocyanobacteria rich in phycocyanin; Pico-Euk, picoeukaryotes; Pico-und, picophytoplankton undetermined), and contribution of (B) each Utermöhl Fraction of Phytoplankton (UFP) group to total UFP biomass (Bac, Bacillariophyceae; Chl, Chlorophyceae; Chr, Chrysophyceae; Cya, Cyanophyceae; Dino, Dinophyceae; Fla, autotrophic nanoflagellates; Med, Mediophyceae; Tre, Trebouxiophyceae) in Cabras Lagoon along the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S6. Contribution of (A) each Picophytoplankton (Pico) group to total Picophytoplankton (Pico) density (P-PE, picocyanobacteria rich in phycoerythrin; P-PC, picocyanobacteria rich in phycocyanin; Pico-Euk, picoeukaryotes; Pico-und, picophytoplankton undetermined), and contribution of (B) each Utermöhl Fraction of Phytoplankton (UFP) group to total UFP density (Bac, Bacillariophyceae; Chl, Clorophyceae; Chr, Chrysophyceae; Cya, Cyanophyceae; Dino, Dinophyceae; Fla, autotrophic nanoflagellates; Med, Mediophyceae; Tre, Trebouxiophyceae) in Cabras Lagoon along the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S7. Inter-annual and seasonal dynamics of mean cell volume of all UFP taxa observed in Cabras Lagoon during the study period (2017, 2018, 2019). Taxa abundance is not included. *Autumn 2019 was represented only by October. Figure S8. Scatterplot of mean cell volume (x axis) versus cell density (y axis) of each Utermöhl Fraction of Phytoplankton taxon recorded in Cabras Lagoon along the study period (2017, 2018, 2019). Figure S9. Inter-annual and seasonal dynamics of UFP community volume in Cabras Lagoon during the study period (2017, 2018, 2019). Weighted- average of volume of the entire UFP community is showed. *Autumn 2019 was represented only by October. Figure S10. Inter-annual and seasonal dynamics of selected environmental variables (A, temperature; B, salinity; C, ammonium, N-NH4; D, orthophosphates, P-PO4) in Cabras Lagoon throughout the study period (2017, 2018, 2019). *Autumn 2019 was represented only by October. Figure S11. Inter-annual and seasonal dynamics of (A) Mesozooplankton (Meso) and (B) Microzooplankton (Micro) total density and their contribution to total zooplankton density in Cabras Lagoon throughout the study period (2017, 2018, 2019). Figure S12. Contribution of (A) each taxon of Mesozooplankton (Meso) and of (B) Microzooplankton (Micro) to total zooplankton density (Biv, Bivalvia veliger; Cop, Copepoda; Nau, Crustacea Naupli; Pol, Polychaeta larvae; Rot, Rotifera; Nem, Nematoda) observed in Cabras Lagoon throughout the study period (2017, 2018, 2019). Table S1. Monthly mean and Standard Deviation (SD) of Picophytoplankton (Pico) and Utermöhl Fraction of Phytoplankton (UFP) cell density and biomass along the study period. Table S2. Monthly mean and Standard Deviation (SD) of the selected environmental variables (Tra, Secchi disk transparency; Temp, water temperature; Sal, salinity; N-NH4, ammonium; N-NO3, nitrate; N-NO2, nitrite; P-PO4, orthophosphate; Si-SiO4, reactive silica) along the study period. Table S3. Results (t-value and p-value) of the Generalized Linear Mixed-Effect Model (GLMM) applied on the Picophytoplankton (Pico) and Utermöhl Fraction of Phytoplankton (UFP) total cell density and on the Pico and UFP total biomass along the study period (2017, 2018, 2019) in Cabras Lagoon. The selected environmental variables (Tra, Secchi disk transparency; Temp, water temperature; Sal, salinity; N-NH4, ammonium; N-NO3, nitrate; N-NO2, nitrite; P-PO4, orthophosphate; Si-SiO4, reactive silica) were used as fixed terms and months were used as random term. Significant effects (p<0.05) are in bold. Table S4. Total density (ind L-1) and relative abundance (%) of main zooplankton groups detected in Cabras Lagoon. No n- co mm er cia l u se on ly