TY - JOUR
AU - Boros, Gergely
AB - Abstract A crucial point in current research on plankton ecology is how global warming will change community functioning, which has led to numerous studies addressing the question with a variety of conclusions. We previously observed a long-term increase in the compositional variability of phytoplankton in a large shallow lake with a concurrent rise in mean temperature, and we conducted an experimental test of this hypothetical link in a mesocosm system. Following predictions of water temperature for the end of the century, 12 mesocosms were filled with prefiltered and sterilized lake water with six of the tanks kept 3°C above ambient levels. Phytoplankton colonization and subsequent changes in its composition were monitored using microscopic analysis and flow cytometry. Chlorophytes were the most successful colonizers, with no treatment-specific effect on dominant taxa. However, heated mesocosms showed higher variation in community structure (distance-based beta diversity), due to higher variability in subdominant species, a considerable portion of which were flagellated taxa. Our hypothesis of temperature-driven compositional variation was supported by both cytometric and taxonomic data, implying that higher spring temperatures can enhance variability in phytoplankton, which might increase the chance of alternate pathways during succession and reduce the predictability of its annual cycle. INTRODUCTION Understanding how global warming modifies the dynamics of various communities is a pivotal point in the effort to assess climate-change-driven effects on ecosystems, which is particularly true for the rapidly reacting planktonic communities. Lake surface water temperatures have been found to exhibit a general increase during the last few decades (O’Reilly et al., 2015), which inevitably urges more in-depth research on how this trend can alter planktonic community composition and functioning. A key element of this issue is phytoplankton and the repercussions their response can have in the aquatic environment. Warming-related effects on algal communities have already been demonstrated in a number of cases. Long-term increasing trend in temperature was found to influence phytoplankton succession (Wiltshire and Manly, 2004), induce dramatic compositional shifts (Hsieh et al., 2010) and is likely to promote cyanobacterial dominance (Paerl et al., 2011). The observed shifts in natural communities have also stimulated intensive experimental work in controlled environments to test hypotheses with regard to the causalities behind in situ observations. Studies conducted in mesocosms are becoming increasingly widespread, however, these systems are quite diverse, and various set-ups can lead to different, sometimes contrasting conclusions. Thus, although an experimental warming of 3°C was found to alter phytoplankton composition without significantly changing its total biomass in one instance (Moss et al., 2003), an additional increase in phytoplankton diversity, biomass and primary production occurred in another study conducted at 4°C above ambient levels (Yvon-Durocher et al., 2015). To make things more complicated, other authors have found just the opposite, with decreasing diversity under different levels of warming (1–5°C, Urrutia-Cordero et al., 2017). Regardless of the outcome, a common observation of these studies is that temperature-induced changes in composition can also imply alterations in community functioning, e.g. through the dominance of small-sized, fast-growing species (Rasconi et al., 2015) or mixotrophic flagellates (Urrutia-Cordero et al., 2017) under elevated temperatures. In addition to differences in the level of warming among various studies, there is a myriad of other direct or indirect factors with potential influence on a mesocosm scale, making the assessment of temperature-related impacts more problematic. Many indirect effects are related to trophic links (Lewandowska and Sommer, 2010; He et al., 2018), and although each species of algae might respond differently to elevated temperatures (Huertas et al., 2011), compositional changes in the phytoplankton can also be elicited by the impact of warming on the grazer community (Strecker et al., 2004), overriding what could be expected from mere taxon-specific responses. Fortunately, there are a number of options at our disposal to enhance the reliability of a mesocosm study. One important aspect is the temporal resolution of the experiment. If we want to understand how community functioning changes under various temperature scenarios, the focus on compositional changes should be extended to the dynamics of those changes as well. As algal generation times are on the order of days, ideally, sampling frequency should be as close to this as possible. This notion and recent technological advancements have led to pioneer studies dealing with natural phytoplankton communities on a considerably finer temporal scale, such as those using automated flow cytometry (FC) (Sosik et al., 2010; Hunter-Cevera et al., 2016; Thomas et al., 2018). Although current high-frequency sampling and analysis efforts come at a cost with a considerably lower taxonomic resolution, increasing temporal resolution is an important key to a better understanding of the drivers and consequences of phytoplankton functioning. An often neglected factor affecting the outcome of any experimental approach is the ecological history of the tested communities, which can affect present and future dynamics as a result of ecological memory (Padisák, 1992) or legacy effects (Dal Bello et al., 2017). This impact is only absent in the case of pioneer communities colonizing newly formed habitats, such as artificial ponds, free or deprived of prior conditions or effects. On the other hand, our knowledge about the aerial dispersal and subsequent colonization capabilities of algae is very limited, with relatively few studies reporting on the subject (Genitsaris et al., 2011; Incagnone et al., 2015). The aim of the present study originates from an observation detailed in a previous paper of the first author (Pálffy and Vörös, 2019). Using data on nearby Lake Balaton, that study found a long-term increase in the compositional variation (distance-based beta diversity) of phytoplankton along a rise in annual average water temperature. This link between warming and variability can have important ecological implications, raising questions about future impacts on phytoplankton functioning, however, it remains highly hypothetical without experimental evidence. The study described here is our first step in the direction of testing this hypothesis. Considering all the aforementioned aspects of the issue, we investigated the effect of elevated temperature on the dynamics of pioneer phytoplankton communities during and after colonization in a mesocosm system. In order to identify fine-scale differences in dynamics, traditional microscopic analysis was complemented with a computational flow cytometric approach, and our initial hypothesis was that communities emerging at higher temperatures show higher compositional variability. METHOD The experiment was conducted in an outdoor mesocosm system consisting of 12 cylindrical plastic tanks situated at the Balaton Limnological Institute, Centre for Ecological Research, Tihany, Hungary, near the shore of Lake Balaton (46°54′45″N, 17°53′36″E). The tanks have an insulated wall, a diameter of 2 m and a depth of 1.5 m. Each tank was filled up with filtered and sterilized freshwater to a depth of 1.2 m. Water originated from Lake Balaton, pumped from a fixed point 100 m offshore, and was filtered through gravel- and sand filters and sterilized with a germicidal ultraviolet lamp before entering the tanks. Before filling, the mesocosms were washed with high-pressure tap water three times, including hot (90°C) water rinsing. During the experiment, six randomly selected mesocosms were not heated (control treatment), whereas water in the remaining six mesocosms was heated to a temperature constantly adjusted to 3°C above the average value of the control tanks (heating treatment). The rate of temperature increase in the heating treatment was intended to simulate future climate conditions, and was based on the global mean increase of 0.34°C decade−1 between 1985 and 2009 reported in the comprehensive study of O’Reilly et al. (2015). In the continental climate of Central Europe, such a rate of change may lead to a 3°C rise in water temperature in general by the end of the century. At the start of the experiment all tanks were supplemented with K2HPO4 and NaNO3 to reach an initial molar concentration of 1.053 and 4.838 μM−1 for PO43− and NO3−, respectively. Filling of the mesocosms took place on 8 April 2019, after which phytoplankton sampling was performed every third day until 20 May 2019. An airlift mixing system (with 0.6 m3 h−1 carrying capacity) circulated the water in the tanks during the course of the experiment to induce water currents and to ensure constant mixing of the water column. The tanks were also covered with a 1 mm mesh size net to avoid litter fall. Water temperature, dissolved oxygen and pH were monitored in the mesocosms with sensors deployed at mid-depth of the water column. A programmable logic controller (PLC) system recorded all data measured by the sensors, and at the same time controlled water temperature in the heated tanks. Chlorophyll a concentration of the samples was determined spectrophotometrically after hot methanol extraction using the absorption coefficients determined by Iwamura et al., (1970). Phytoplankton composition was monitored using FC and traditional microscopy. Although FC analysis was performed on each sampling day (every third day of the experiment), microscopic analysis was narrowed down to samples collected on every second sampling day (every sixth day of the experiment). For the latter, 50 mL aliquots of the collected samples were fixed with 50 μL Lugol’s iodine and stored at 5°C until subsequent analysis. Species composition and abundance of nano- and microplankton were determined with an inverted microscope (Utermöhl, 1958) at 600× magnification. Cytometric analysis was performed using a Sysmex CyFlow Space flow cytometer equipped with a blue (488 nm emission) and a red (638 nm emission) laser immediately after each sampling. The freshly collected samples were thoroughly mixed directly prior to FC, after which subsamples of 830 μL were injected into the 200 μm capillary of the instrument at a flow rate of 1 μL s−1 using sheath fluid provided by the manufacturer (Sysmex Europe GmbH). All samples were analyzed using the same instrument settings (identical subsample volume, flow rate and detector gain values). Algae were detected through the plots of forward scatter light (FSC) versus blue laser-dependent red fluorescence (FL3), FSC correlating with cell size and FL3 with cellular chlorophyll a content (Marie et al., 2005). Fluorescence channels of the red laser were not evaluated due to the absence of cyanobacteria during the experiment (no signals were detected). Flow cytometric scatter plots of the samples collected on 5 May are shown in Fig. 1 as an example, also representing the arrangement of the mesocosm tanks. Fig. 1 Open in new tabDownload slide Gated flow cytometric plots (FSC versus FL3) of phytoplankton collected from each mesocosm using data from 5 May as an example. The arrangement of the plots represents the actual arrangement of the mesocosms. Plots with blue background represent mesocosms kept at elevated temperature. Fig. 1 Open in new tabDownload slide Gated flow cytometric plots (FSC versus FL3) of phytoplankton collected from each mesocosm using data from 5 May as an example. The arrangement of the plots represents the actual arrangement of the mesocosms. Plots with blue background represent mesocosms kept at elevated temperature. All analyses were performed in the run-time environment R version 3.6.0 (R Core Team, 2019), the entire script is publicly available on GitHub (https://github.com/k-palffy/colonisation-experiment.git). Flow cytometric data were analyzed in the framework provided by the R packages flowCore (Ellis et al., 2019) and flowWorkspace (Finak and Jiang, 2020). First, raw FC files were converted into GatingSet objects to facilitate analysis in R. After the data were log-transformed, samples were gated on the basis of the FL3 signals in order to remove background noise. Analysis was performed based on the approach devised by Li (1997) and further elaborated by others (García et al., 2015; Props et al., 2016). The FCS and FL3 data range was divided into 64 categories, resulting in a 64×64 matrix with 4096 compartments (bins) in the FCS-FL3 space, assigning each cell to a specific bin according to its FSC and FL3 signals (Fig 2). Next, we created a vector of the counts in each bin for every sample, which can be interpreted as a FC fingerprint of the phytoplankton. These sample-specific fingerprints were subsequently used for a comparative analysis of the treatments. Shannon diversity (H) and evenness (J) of the bin counts (Magurran and McGill, 2011), and the mean and mode of FSC and FL3 values, were determined for each sample. Modes (highest peaks in the histograms representing the most frequent values) were determined using the R package flowDensity (Taghiyar, 2020). Within-treatment compositional variability was determined as multivariate dispersion for each date using the method of Anderson et al. (2006), which takes the pairwise dissimilarities between samples on the basis of community abundance/biomass data and calculates the mean distance of the samples to their centroid in the multivariate space of a principal coordinates ordination. Mean distances were determined for each set of samples within a treatment for each date. This mean distance is interpreted, in a broader sense, as a type of beta diversity capturing variation in phytoplankton community structure (Anderson et al., 2006). The Bray–Curtis index was applied to calculate dissimilarity between the samples, and centroid distances were determined with the betadisper function of the R package vegan (Oksanen et al., 2019). Prior to determining dissimilarity, counts per bin were log (x + 1)-transformed to incorporate abundance information without order-of-magnitude differences in dominant bins among the samples, which can mask the contribution of less populated bins. Fig. 2 Open in new tabDownload slide Schematic representation of the binning process applied on flow cytometric data (left figure) and the creation of a sample-specific fingerprint (right figure). Bin indices were created starting from the upper left corner in a row-by row manner. Bins empty in all the samples of a particular date were omitted. Fig. 2 Open in new tabDownload slide Schematic representation of the binning process applied on flow cytometric data (left figure) and the creation of a sample-specific fingerprint (right figure). Bin indices were created starting from the upper left corner in a row-by row manner. Bins empty in all the samples of a particular date were omitted. Taxonomic composition of the samples was compared in two different ways. We used correspondence analysis (CA) to highlight taxon-specific differences among the samples and to determine which taxa had a high contribution to the observed compositional variability. In order to obtain a better representation of data along the first two dimensions of the CA output, rare genera of common ancestry and corresponding functional traits were merged into larger taxonomic groups (e.g. dinophytes, chrysophytes), whereas species found only in a maximum of three samples were omitted from the analysis. CA was performed using the R packages FactoMineR (Le et al., 2008) and factoextra (Kassambara and Mundt, 2019), of which the former also facilitates the determination of taxon (column) contributions to the CA dimensions (as described in Husson et al., 2017). Taxonomic diversity, eveness and compositional variability of the samples were determined with the same methods and indices as those applied on FC data. Significant differences in taxonomic variability, flow cytometric variability, diversity measures, FSC and FL3 values and chlorophyll a between the treatments were identified with Dunn’s test (Dunn, 1961). At the end of the experiment, zooplankton samples were collected with a 50-cm-high, 34 L Schindler–Patalas sampler equipped with a 60 μm mesh net. Two samples were taken from each tank, one from the upper and one from the lower half of the water column. The net was washed down after every sampling. Each pair of samples representing the different depths of the same mesocosm were integrated and concentrated to 200 mL. Zooplankton was preserved with PATOSOLV (77–82% ethanol, 16–22% isopropyl alcohol, 0.5–1% tert-Butyl alcohol) in 70% final concentration. Samples were identified on the level of species, sex and developmental stage. Common species were counted in 10 mL subsamples with Olympus IX73 inverted microscope. The number of rare species was determined with a Leica WILD M8 stereomicroscope. Despite the artificial mixing, a certain amount of sedimentation did occur in the mesocosms, which eventually resulted in the development of benthic algal mats on the bottom of the tanks. The growing phytobenthos can act as a nutrient sink with consequent indirect influence on phytoplankton biomass, thus, we also took phytobenthos samples for biomass estimation after draining the tanks at the end of the experiment. Each tank was sampled on three representative spots using a 15-cm-diameter plastic tube segment. After placing the tube segment on the bottom, the benthic mat inside the segment was scraped off and resuspended in distilled water. Chlorophyll a of the resuspended phytobenthos samples was measured with the same method as the one used for phytoplankton. The putative inverse relationship between phytobenthos and phytoplankton chlorophyll a was tested with linear regression. RESULTS Among the measured background variables, temperature fluctuated roughly between 9°C and 19°C, with an anticipated 3°C difference between the control and the heated mesocosms (Fig. 3). The values of pH showed a slight increase during the study, rising from 8 (characteristic of Lake Balaton) to 8.75, probably due to the photosynthetic activity of the emerging phytoplankton communities. Dissolved oxygen, besides a clear daily fluctuation, changed in accordance with the amount of chlorophyll a, which showed an increase around halfway through the experiment and a decline in its second half, although with relatively high variability on certain dates. Although by the end of the experiment chlorophyll a dropped to below 10 μg L−1 in all the 12 mesocosms, the heated tanks showed significantly lower values than the control tanks on the last sampling day (Dunn’s test, P < 0.05). Moreover, a regression model on phytoplankton chlorophyll a as the dependent variable showed a significant negative effect of both benthic chlorophyll a and elevated temperature (R2 = 0.62, P < 0.01, Supplementary Fig. S1, Supplementary Table SII). Fig. 3 Open in new tabDownload slide Changes in temperature, dissolved oxygen, pH and chlorophyll a during the colonization experiment in mesocosms kept at ambient and elevated (+3°C) temperatures. Asterisk at chlorophyll a denotes significant difference at P < 0.05 according to Dunn’s test. Fig. 3 Open in new tabDownload slide Changes in temperature, dissolved oxygen, pH and chlorophyll a during the colonization experiment in mesocosms kept at ambient and elevated (+3°C) temperatures. Asterisk at chlorophyll a denotes significant difference at P < 0.05 according to Dunn’s test. The flow cytometric analysis revealed certain changes through time and treatment-specific differences as well. Mean FSC signals showed an increase during the experiment, with no significant difference between the treatments, whereas mean FL3 values decreased after a marked initial increase (Fig. 4). Apart from one occasion, there were no significant differences in FL3 either, however, there was generally higher standard deviation in both FSC and FL3 in the samples from the heated mesocosms. Similar differences and temporal patterns can be seen in the mode of the FSC and FL3 signals (Fig. 4). On the first three sampling days, FL3 means and modes were slightly higher in the control mesocosms, after that the values were generally higher in the heated tanks. Both cytometric and taxonomic diversity (Hc, Ht) and evenness (Jc, Jt) showed a general decline in the course of the experiment, without any significant treatment-specific differences (Fig. 5). It is, however, important to notice that there was higher standard deviation in both diversity and evenness in the phytoplankton communities growing at elevated temperature. On the basis of the cytometric fingerprints, phytoplankton communities in the heating treatment showed higher variation in community structure (distance-based beta diversity), with significant differences throughout the second half of the experiment (Dunn’s test, P < 0.05, Fig. 6). After an initial decrease, cytometric variation remained relatively constant in both treatments. Fig. 4 Open in new tabDownload slide Mean and mode of flow cytometric FSC and FL3 signals of phytoplankton samples collected from mesocosms kept at ambient and elevated (+3°C) temperature in the course of the colonization experiment. *: significant difference at P < 0.05; **: significant difference at P < 0.01 according to Dunn’s test. Error bars represent standard deviations. Fig. 4 Open in new tabDownload slide Mean and mode of flow cytometric FSC and FL3 signals of phytoplankton samples collected from mesocosms kept at ambient and elevated (+3°C) temperature in the course of the colonization experiment. *: significant difference at P < 0.05; **: significant difference at P < 0.01 according to Dunn’s test. Error bars represent standard deviations. Fig. 5 Open in new tabDownload slide Flow cytometric and taxonomic Shannon diversity (Hc, Ht) and evenness (Jc, Jt) of phytoplankton communities in mesocosms kept at ambient and elevated (+3°C) temperature in the course of the colonization experiment. Error bars represent standard deviations. The Dunn’s test found no statistically significant differences in any of the indices. Fig. 5 Open in new tabDownload slide Flow cytometric and taxonomic Shannon diversity (Hc, Ht) and evenness (Jc, Jt) of phytoplankton communities in mesocosms kept at ambient and elevated (+3°C) temperature in the course of the colonization experiment. Error bars represent standard deviations. The Dunn’s test found no statistically significant differences in any of the indices. Fig. 6 Open in new tabDownload slide Cytometric and taxonomic variation (measured as distance-based beta diversity) in phytoplankton community structure at ambient and elevated (+3°C) temperatures in the course of the colonization experiment. *: significant difference at P < 0.05; **: significant difference at P < 0.01 according to Dunn’s test. Error bars represent standard deviations. Fig. 6 Open in new tabDownload slide Cytometric and taxonomic variation (measured as distance-based beta diversity) in phytoplankton community structure at ambient and elevated (+3°C) temperatures in the course of the colonization experiment. *: significant difference at P < 0.05; **: significant difference at P < 0.01 according to Dunn’s test. Error bars represent standard deviations. As for taxonomic composition, green algae were by far the most successful colonizers during the experiment, gaining dominance in all the mesocosm tanks (Supplementary Table SI). Fig. 7 shows the result of the CA performed on the taxonomic data of the phytoplankton samples collected during the experiment. The biplot, with the first two dimensions representing 54.7% of total variance, shows that the highest contribution to the dimensions was attributable to Koliella longiseta, a dominant chlorophyte. Other important species included Chlorella vulgaris, the most abundant chlorophyte during the experiment, Chloroidium ellipsoideum, another frequently occurring green alga, and Haematococcus sp., which tended to occur in tanks kept at higher temperature. The biplot also indicates two important findings. First, while the first dimension (representing 37.4% of total variance) was defined mostly by the dominant Chlorella vulgaris and Koliella longiseta, the actual difference between the treatments, manifested along the second dimension, was defined by other, less abundant green algal taxa (Haematococcus, Chloroidium ellipsoideum, Scenedesmus, Lagerheimia) and by dinophytes to some extent. Second, although most of the samples from the control treatment are spread along the first dimension, the spread of points representing the heated mesocosms covers a larger area of the principal coordinate space, indicating higher compositional variability compared to the control in the course of the experiment. Fig. 7 Open in new tabDownload slide Symmetric biplot of correspondence analysis on the abundance of phytoplankton taxa in the mesocosms. Points represent phytoplankton samples, names with larger font size correspond to higher contribution of the respective taxon to the dimensions. Fig. 7 Open in new tabDownload slide Symmetric biplot of correspondence analysis on the abundance of phytoplankton taxa in the mesocosms. Points represent phytoplankton samples, names with larger font size correspond to higher contribution of the respective taxon to the dimensions. Taxonomic variation based on the Bray–Curtis dissimilarity of the samples indicated differences between the treatments that were similar to those found during the analysis of the flow cytometric data (Fig. 6). The mean distance between phytoplankton samples and their respective group centroid was higher in the heating treatment than in the control, which means that beta diversity increased due to the increased level of temperature, although this difference proved to be statistically significant only on three of the six investigated days (Dunn’s test, P < 0.05). As expected, we found very low abundances of zooplankton (<1 ind. L−1) in the mesocosms, with rotifers as the dominant group and with no significant difference between the treatments (Fig. 8). Rotifers were the only group present in all the 12 mesocosms, but we did not find any significant correlation between their abundance and the abundance of dominant algal taxa. Fig. 8 Open in new tabDownload slide Zooplankton abundance at the end of the colonization experiment in mesocosms kept at ambient and elevated temperatures. Fig. 8 Open in new tabDownload slide Zooplankton abundance at the end of the colonization experiment in mesocosms kept at ambient and elevated temperatures. DISCUSSION Our study has demonstrated that a 3°C increase in temperature can enhance compositional variability in the phytoplankton also without substantial changes in species dominance. Both FC and traditional taxonomic analysis showed that there was higher compositional variability among the communities kept at elevated temperature as compared to the control mesocosms with ambient temperatures. The change in phytoplankton biomass over the course of our study was similar to those demonstrated by Lewandowska and Sommer (2010), who also found a persistent decline during the second half of the experiment and lower biomass in heated tanks. They also found a warming induced shift toward smaller cell sizes, which did not manifest in our case. A possible reason for this could be either the shorter duration or the low representation of various taxonomic groups and consequently low diversity in the newly established communities as compared to the already diverse phytoplankton studied by Lewandowska and Sommer (2010). Our assumption about the role of the growing phytobenthos as a nutrient sink is supported by the significant inverse relationship of planktonic and benthic chlorophyll a, however, we also found a significant treatment effect at the end of the experiment (Supplementary Fig. S1, Supplementary Table SII). Nevertheless, with all the tanks shifting into oligotrophic conditions, this difference was considerably small, making it difficult to find an unambiguous explanation. Due to the lack of effect on zooplankton abundance, we can rule out top–down control as a possible cause of lower phytoplankton biomass in the heated tanks. On the other hand, we did not investigate the abundance of protozoa, similarly important consumers of phytoplankton (Tillmann, 2004), and since warming can induce a stronger match between protist grazers and algal prey (Aberle et al., 2012), higher protozoan grazing could have caused reduced phytoplankton biomass in the heating treatment. Then again, the difference might as well have arisen due to bottom-up effects. Nutrient addition took place only at the start of the experiment, and the decline of biomass in the second half means that phytoplankton was way past its stationary phase by the end of the study, having used up most of the available nitrogen and phosphorus. This suggests that if the effect of elevated temperature manifested through bottom-up processes, any difference in biomass must have mostly depended on the rate of nutrient recirculation. Bacterial mineralization is highly temperature-dependent (Li et al., 2014), which could have indirectly affected phytoplankton biomass in the second half of the experiment, although bacterial activity was not the focus of our study. The same is true for denitrification, the rate of which increases with rising temperature and can be further intensified as a result of inherently lower DO levels (Veraart et al., 2011). In addition to differences in temperature, DO was also slightly lower in the heated mesocosms (Fig. 2), which could thus have led to a reduction in total nitrogen, but we do not have relevant data to verify this assumption. However, whatever the cause, the effect of elevated temperature on phytoplankton biomass toward the end of the experiment cannot be linked to the observed differences in composition and its variation, which occurred shortly after phytoplankton started to grow in the mesocosms. A comparison of phytoplankton chlorophyll a and variability (Supplementary Fig. S2) also confirms that although there was substantial overlap in biomass, the variation in community structure showed a consistent difference between the treatments shortly after the colonizing communities started to proliferate. Little is known about the aerial dispersal and subsequent colonization capabilities of algae, but chlorophytes have so far been found to be effective airborne settlers (Genitsaris et al., 2011), which is in absolute agreement with our findings. All the mesocosms were dominated by green algae, although with varying relative abundance of the dominant species (Supplementary Table SI, Fig. 4). According to a review paper on airborne algae (Tesson et al., 2016), the chlorophyte genus Chlorella is a very common air-dispersed taxon, which was the most abundant genus colonizing the tanks in our study (the similarly characteristic Chloroidium ellipsoideum formerly also belonged to the Chlorella genus). We could not find any information about the dispersal abilities of the other prevalent species, Koliella longiseta, although it has a functional implication worth mentioning. According to the functional classification of phytoplankton created by Reynolds and his coauthors (Reynolds et al., 2002; Padisák et al., 2009), most of the species identified in the tanks belong to associations characteristic of shallow environments with clear mixed layers (codons X1–X3), which the mesocosm system was intended to represent. Although Koliella sp. is regarded as a representative of oligotrophic water bodies in Reynolds’s system (codon X3), Chlorella sp. are mostly categorized as members of group X1, indicating eutrophic habitats. This co-occurrence of species indicating various trophic states was likely caused by the fact that although our initial nitrogen and phosphorus supplement provided meso-eutrophic conditions, the lack of repeated nutrient replenishment and a certain degree of sedimentation shifted trophic state slightly toward the oligotrophic end of the spectrum. Further considering the ecological aspects of Reynolds’ classification, our results might also serve as a typical case of environmental filtering. Almost all of the observed taxa can also be found in the nearby Lake Balaton, however, the chlorophyte species dominating the mesocosms are by far not the predominant components in the lake, particularly not in spring. At this time of the year, phytoplankton is usually dominated by diatoms (Hajnal and Padisák, 2007; Pálffy et al., 2013), although we assume that their aerial dispersal is less effective owing to their heavy silica frustule, and the frequently wind-mixed water of the lake may be a more favorable habitat for their proliferation. The only nonchlorophyte species that reached a relatively high abundance during our study and is also a characteristic species of the lake was the haptophyte Chrysochromulina parva, whereas other typical and functionally similar species, such as those belonging to Cryptophyta, showed very sporadic occurrences in the mesocosms. These observations imply that besides their excellent dispersal ability, chlorophytes are also successful colonizers with an ability to outcompete other newcomers, at least in shallow, clear-water ponds. The effect of temperature was not discernible in terms of the dominant taxa, which is not a surprise per se. The ecological repercussions of the current long-term change in mean temperatures are gradually evolving phenomena often too subtle to detect during a shorter period of time. What temperature did affect in our case was the relative occurrence of seemingly lesser, subdominant species, a number of which, interestingly, were flagellated taxa. Explaining this tendency is a hard nut to crack. The study of Urrutia-Cordero et al. (2017) connected the dominance of mixotrophic flagellates to an increase in bacterial production at higher temperatures, however, that might not have been the case in our experiment, where the trophic state of the mesocosms was lower and the most characteristic flagellates were green algae (Chlamydomonas and Haematococcus), none of which are known to possess any phagotrophic potential. This is in contrast with the other successful flagellate, Chrysochromulina parva, which has a mixotrophic lifestyle, but its abundance did not exhibit any treatment-specific response. The differences with regard to subdominant species also reflect how seemingly minor changes in community structure can lead to an increase in variation. The flow cytometric analysis has demonstrated that there was a certain amount of fluctuation in mean cellular chlorophyll content in the first half of the experiment, and was followed by a moderate increase in mean cell size in the second half (Fig. 4). The similar temporal pattern of the modes suggests that these shifts in cytometric signals are linked to the increasing frequency of the dominant taxa, which also led to a decline in Shannon diversity and evenness. These temporal changes were similar in the two treatments, however, we also found higher variation (standard deviation) in both the cytometric signals and the diversity indices of the communities growing at elevated temperature. The variation of all these variables is also a manifestation of the aforementioned difference in compositional variability (distance-based beta diversity), and can only be explained by the varying contribution of subdominant species in the heated mesocosms. Since zooplankton abundance was very low during the experiment, the influence of top–down control on compositional variability, just as in the case of phytoplankton biomass, can be ruled out, although we did not extend our investigation to ciliates and other important protozoans potentially feeding on algae. Zoochory is a common way of passive dispersal for zooplankton, and we wanted to minimize this probability in order to avoid unwanted and uncontrollable variability within the treatments. However, transport of eggs or resting structures by wind is also feasible over short distances (see Incagnone et al., 2015 and references therein), which likely caused the emergence of mesozooplankton during the study. Given the longer life cycle of these organisms, the duration of the experiment prevented the development of a thriving zooplankton community. The most conspicuous and far-reaching finding of our study bears similarity to those demonstrated with regard to decades-long changes in the phytoplankton of Lake Balaton (Pálffy and Vörös, 2019). Considering the apparent difference in temporal scale, this similarity might sound as an overstatement without an elaborate synthesis. Nevertheless, seasonal succession is itself a chain of short-term events shaping phytoplankton structure, and plankton research has already found the means to find mechanistic links between phenological shifts in the annual cycle and species-specific changes detected at a considerably finer temporal resolution (Hunter-Cevera et al., 2016). Both the present study and the long-term observations have found that higher temperatures lead to higher compositional variability in the phytoplankton. There was a crucial distinction in terms of the nature of this variability: the long-term data suggested an increase in seasonal (temporal) variability with a simultaneous rise in annual average temperature at one specific site within the lake, whereas the mesocosm experiment has yielded higher spatial variability (beta diversity) at elevated temperatures. Global warming concurrently means an expansion of the vegetation period in the temperate zone, which can affect plankton dynamics (Klausmeier and Litchman, 2012) and hence has the potential to increase annual variability. The putative link with the results of our experiment is that the observed higher beta diversity at elevated temperatures increases the chance for alternative routes in phytoplankton succession, ultimately enhancing the variability and reducing the predictability of the entire annual cycle. This assumption could lead to and definitely needs further, more in-depth studies in the future. From a practical point of view, the presented results also support the view that an increased level of temporal resolution can lead to novel findings about phytoplankton dynamics. Traditional microscopic analysis has always represented a bottleneck of compositional analysis, on the other hand, it is still indispensable for high taxonomic resolution, a requirement FC was originally not designed to meet. However, recent technological developments linking FC with other techniques have the potential to narrow this gap in the future. Such promising alternatives include the combinations of FC with microscopic imaging (Dashkova et al., 2017) and high-throughput sequencing (Fiedler et al., 2018). CONCLUSIONS The widely observed long-term rise in temperature can affect phytoplankton community structure and functioning in numerous direct and indirect ways. Our study has experimentally demonstrated that elevated temperature can significantly increase compositional variability in spring, which presumably leads to repercussions in the seasonal succession of the phytoplankton. This temperature-driven effect is also in agreement with our previous paper on long-term changes in phytoplankton variability. Our finding was corroborated by both microscopic and flow cytometric analysis, also implying that certain aspects of phytoplankton dynamics can be effectively elucidated with a highly automated approach without performing an otherwise essential yet time-consuming taxonomic analysis. Green algae were the most successful colonizers and there was no significant difference between treatments with regard to the dominant taxa. The observed higher variation in community structure (beta diversity) in the heated mesocosms was mainly due to higher variability in the less abundant, subdominant species, flagellated taxa in particular. This increased variability suggests that higher mean temperature can increase the chance of alternate pathways in phytoplankton succession, which thus might reduce the reliability of predictions about its ecological consequences. DATA ARCHIVING The basic data that support the findings of this study are available from the corresponding author upon reasonable request and are partially presented in the electronic supplementary material. All data are archived at a local server at the Centre for Ecological Research (Tihany, Hungary). Raw flow cytometric data are available online at www.flowrepository.org through the following link: https://flowrepository.org/id/FR-FCM-Z35Z. The R script written for flow cytometric and taxonomic analysis was uploaded to GitHub and is available at https://github.com/k-palffy/colonisation-experiment.git. ACKNOWLEDGEMENTS Special thanks to Máté Burányi for his help during sampling and sample processing. FUNDING National Research, Development, and Innovation Office of Hungary (grant number: GINOP-2.3.2-15-2016-00057). G. Boros was supported by the Bolyai János Research Fellowship from the Hungarian Academy of Sciences. References Aberle , N. , Bauer , B., Lewandowska , A., Gaedke , U. and Sommer , U. ( 2012 ) Warming induces shifts in microzooplankton phenology and reduces time-lags between phytoplankton and protozoan production . Mar. Biol. , 159 , 2441 – 2453 . Google Scholar Crossref Search ADS WorldCat Anderson , M. J. , Ellingsen , K. E. and McArdle , B. H. ( 2006 ) Multivariate dispersion as a measure of beta diversity . Ecol. Lett. , 9 , 683 – 693 . Google Scholar Crossref Search ADS PubMed WorldCat Dal Bello , M. , Rindi , L. and Benedetti-Cecchi , L. ( 2017 ) Legacy effects and memory loss: how contingencies moderate the response of rocky intertidal biofilms to present and past extreme events . Glob. Chang. Biol. , 23 , 3259 – 3268 . Google Scholar Crossref Search ADS PubMed WorldCat Dashkova , V. , Malashenkov , D., Poulton , N., Vorobjev , I. and Barteneva , N. S. ( 2017 ) Imaging flow cytometry for phytoplankton analysis . Methods , 112 , 188 – 200 . Google Scholar Crossref Search ADS PubMed WorldCat Dunn , O. J. ( 1961 ) Multiple comparisons among means . J. Am. Stat. Assoc. , 56 , 52 – 64 . Google Scholar Crossref Search ADS WorldCat Ellis , B. , Haaland , P., Hahne , F., Le Meur , N., Gopalakrishnan , N., Spidlen , J., Jiang , M. and Finak , G. ( 2019 ) flowCore: Basic structures for flow cytometry data. R Package Version 1.50.0. https://bioconductor.org/packages/flowCore/. Fiedler , C. J. , Schönher , C., Proksch , P., Kerschbaumer , D. J., Mayr , E., Zunabovic-Pichler , M., Domig , K. J. and Perfler , R. ( 2018 ) Assessment of microbial community dynamics in river bank filtrate using high-throughput sequencing and flow cytometry . Front. Microbiol. , 9 , 2887 . Google Scholar Crossref Search ADS PubMed WorldCat Finak , G. and Jiang , M. ( 2020 ) flowWorkspace: Infrastructure for representing and interacting with gated and ungated cytometry data sets. R package version 4.2.0. https://www.bioconductor.org/packages/flowWorkspace/. García , F. C. , Alonso-Sáez , L., Morán , X. A. G. and López-Urrutia , Á. ( 2015 ) Seasonality in molecular and cytometric diversity of marine bacterioplankton: the re-shuffling of bacterial taxa by vertical mixing . Environ. Microbiol. , 17 , 4133 – 4142 . Google Scholar Crossref Search ADS PubMed WorldCat Genitsaris , S. , Moustaka-Gouni , M. and Kormas , K. ( 2011 ) Airborne microeukaryote colonists in experimental water containers: diversity, succession, life histories and established food webs . Aquat. Microbiol. Ecol. , 62 , 139 – 152 . Google Scholar Crossref Search ADS WorldCat Hajnal , É. and Padisák , J. ( 2007 ) Analysis of long-term ecological status of Lake Balaton based on the ALMOBAL phytoplankton database. In Nõges , T. et al. (eds.), European Large Lakes Ecosystem Changes and Their Ecological and Socioeconomic Impacts. Developments in Hydrobiology book series , Vol. 199 , Springer , Dordrecht , pp. 227 – 237 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC He , H. , Jin , H., Jeppesen , E., Li , K., Liu , Z. and Zhang , Y. ( 2018 ) Fish-mediated plankton responses to increased temperature in subtropical aquatic mesocosm ecosystems: implications for lake management . Water Res. , 144 , 304 – 311 . Google Scholar Crossref Search ADS PubMed WorldCat Hsieh , C. H. , Ishikawa , K., Sakai , Y., Ishikawa , T., Ichise , S., Yamamoto , Y., Kuo , T. C., Park , H. D. et al. ( 2010 ) Phytoplankton community reorganization driven by eutrophication and warming in Lake Biwa . Aquat. Sci. , 72 , 467 – 483 . Google Scholar Crossref Search ADS WorldCat Huertas , I. E. , Rouco , M., Lopez-Rodas , V. and Costas , E. ( 2011 ) Warming will affect phytoplankton differently: evidence through a mechanistic approach . Proc. R. Soc. B , 278 , 3534 – 3543 . Google Scholar Crossref Search ADS WorldCat Hunter-Cevera , K. R. , Neubert , M. G., Olson , R. J., Solow , A. R., Shalapyonok , A. and Sosik , H. M. ( 2016 ) Physiological and ecological drivers of early spring blooms of a coastal phytoplankter . Science , 354 , 326 – 329 . Google Scholar Crossref Search ADS PubMed WorldCat Husson , F. , Lê , S. and Pagès , J. ( 2017 ) Exploratory Multivariate Analysis by Example Using R , 2nd edn, Chapman and Hall/CRC Press , New York . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Incagnone , G. , Marrone , F., Barone , R., Robba , L. and Naselli-Flores , L. ( 2015 ) How do freshwater organisms cross the “dry ocean”? A review on passive dispersal and colonization processes with a special focus on temporary ponds . Hydrobiologia , 750 , 103 – 123 . Google Scholar Crossref Search ADS WorldCat Iwamura , T. , Nagai , H. and Ichimura , S.-E. ( 1970 ) Improved methods for determining contents of chlorophyll, protein, ribonucleic acid, and deoxyribonucleic acid in planktonic populations . Int. Rev. Ges. Hydrobiol. Hydrogr. , 55 , 131 – 147 . Google Scholar Crossref Search ADS WorldCat Kassambara , A. and Mundt , F. ( 2019 ) factoextra: Extract and visualize the results of multivariate data analyses. R Package Version 1.0.6. https://cran.r-project.org/package=factoextra. Klausmeier , C. A. and Litchman , E. ( 2012 ) Successional dynamics in the seasonally forced diamond food web . Am. Nat. , 180 , 1 – 16 . Google Scholar Crossref Search ADS PubMed WorldCat Le , S. , Josse , J. and Husson , F. ( 2008 ) Facto mine R: an R package for multivariate analysis . J. Stat. Softw. , 25 , 1 – 18 . Google Scholar Crossref Search ADS WorldCat Lewandowska , A. and Sommer , U. ( 2010 ) Climate change and the spring bloom: a mesocosm study on the influence of light and temperature on phytoplankton and meso-zooplankton . Mar. Ecol. Prog. Ser. , 405 , 101 – 111 . Google Scholar Crossref Search ADS WorldCat Li , W. K. W. ( 1997 ) Cytometric diversity in marine ultraphytoplankton . Limnol. Oceanogr. , 42 , 874 – 880 . Google Scholar Crossref Search ADS WorldCat Li , Y. , Gal , G., Makler-Pick , V., Waite , A. M., Bruce , L. C. and Hipsey , M. R. ( 2014 ) Examination of the role of the microbial loop in regulating lake nutrient stoichiometry and phytoplankton dynamics . Biogeosciences , 11 , 2939 – 2960 . Google Scholar Crossref Search ADS WorldCat Marie , D. , Simon , N. and Vaulot , D. ( 2005 ) Phytoplankton cell counting by flow cytometry. In Andersen , R. A. (ed.), Algal Culturing Techniques , Elsevier , London , pp. 253 – 267 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Magurran , A. E. and McGill , B. J. ( 2011 ) Biological Diversity , Oxford University Press Inc. , New York . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Moss , B. , McKee , D., Atkinson , D., Collings , S. E., Eaton , J. W., Gill , A. B., Harvey , I., Hatton , K. et al. ( 2003 ) How important is climate? Effects of warming, nutrient addition and fish on phytoplankton in shallow lake microcosms . J. Appl. Ecol. , 40 , 782 – 792 . Google Scholar Crossref Search ADS WorldCat Oksanen , J. , Blanchet , F. G., Friendly , M., Kindt , R., Legendre , P., McGlinn , D., Minchin , P. R., O'Hara , R. B. et al. ( 2019 ) Vegan: community ecology package . R Package Version , 2 , 5 – 6 . https://CRAN.R-project.org/package=vegan. Google Scholar OpenURL Placeholder Text WorldCat O’Reilly , C. M. , Sharma , S., Gray , D. K., Hampton , S. E., Read , J. S., Rowley , R. J., Schneider , P., Lenters , J. D. et al. ( 2015 ) Rapid and highly variable warming of lake surface waters around the globe . Geophys. Res. Lett. , 42 , 10773 – 10781 . Google Scholar OpenURL Placeholder Text WorldCat Padisák , J. ( 1992 ) Seasonal succession of phytoplankton in a large shallow lake (Balaton, Hungary) – a dynamic approach to ecological memory, its possible role and mechanisms . J. Ecol. , 80 , 217 – 230 . Google Scholar Crossref Search ADS WorldCat Padisák , J. , Crossetti , L. O. and Naselli-Flores , L. ( 2009 ) Use and misuse in the application of the phytoplankton functional classification: a critical review with updates . Hydrobiologia , 621 , 1 – 19 . Google Scholar Crossref Search ADS WorldCat Paerl , H. W. , Hall , N. S. and Calandrino , E. S. ( 2011 ) Controlling harmful cyanobacterial blooms in a world experiencing anthropogenic and climatic-induced change . Sci. Total Environ. , 409 , 1739 – 1745 . Google Scholar Crossref Search ADS PubMed WorldCat Pálffy , K. , Présing , M. and Vörös , L. ( 2013 ) Diversity patterns of trait-based phytoplankton functional groups in two basins of a large, shallow lake (Lake Balaton, Hungary) with different trophic state . Aquat. Ecol. , 47 , 195 – 210 . Google Scholar Crossref Search ADS WorldCat Pálffy , K. and Vörös , L. ( 2019 ) Phytoplankton functional composition shows higher seasonal variability in a large shallow lake after a eutrophic past . 10.1002/ecs2.2684 . OpenURL Placeholder Text WorldCat Crossref Props , R. , Monsieurs , P., Mysara , M., Clement , L. and Boon , N. ( 2016 ) Measuring the biodiversity of microbial communities by flow cytometry . Methods Ecol. Evol , 7 , 1376 – 1385 . Google Scholar Crossref Search ADS WorldCat Core Team , R. ( 2019 ) R: A Language and Environment for Statistical Computing , R Foundation for Statistical Computing , Vienna, Austria . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rasconi , S. , Gall , A., Winter , K. and Kainz , M. J. ( 2015 ) Increasing water temperature triggers dominance of small freshwater plankton . PLoS One , 10 , e0140449 . Google Scholar Crossref Search ADS PubMed WorldCat Reynolds , C. S. , Huszar , V., Kruk , C., Naselli-Flores , L. and Melo , S. ( 2002 ) Towards a functional classification of the freshwater phytoplankton . J. Plankton Res. , 24 , 417 – 428 . Google Scholar Crossref Search ADS WorldCat Sosik , H. M. , Olson , R. J. and Armbrust , E. V. ( 2010 ) Flow Cytometry in Phytoplankton Research. Chlorophyll a Fluorescence in Aquatic Sciences: Methods and AApplications , Springer , Dordrecht , pp. 171 – 185 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Strecker , A. L. , Cobb , T. P. and Vinebrooke , R. D. ( 2004 ) Effects of experimental greenhouse warming on phytoplankton and zooplankton communities in fishless alpine ponds . Limnol. Oceanogr. , 49 , 1182 – 1190 . Google Scholar Crossref Search ADS WorldCat Taghiyar , M. M. J. ( 2020 ) flowDensity: sequential flow cytometry data gating. R package version 1.24.0. https://bioconductor.org/packages/flowDensity/. Tesson , S. V. M. , Skjøth , C. A., Šantl-Temkiv , T. and Löndahl , J. ( 2016 ) Airborne microalgae: insights, opportunities, and challenges . Appl. Environ. Microbiol. , 82 , 1978 – 1991 . Google Scholar Crossref Search ADS PubMed WorldCat Thomas , M. K. , Fontana , S., Reyes , M., Kehoe , M. and Pomati , F. ( 2018 ) The predictability of a lake phytoplankton community, over time-scales of hours to years . Ecol. Lett. , 21 , 619 – 628 . Google Scholar Crossref Search ADS PubMed WorldCat Tillmann , U. ( 2004 ) Interactions between planktonic microalgae and protozoan grazers . J. Eukaryot. Microbiol. , 51 , 156 – 168 . Google Scholar Crossref Search ADS PubMed WorldCat Urrutia-Cordero , P. , Ekvall , M. K., Ratcovich , J., Soares , M., Wilken , S., Zhang , H. and Hansson , L.-A. ( 2017 ) Phytoplankton diversity loss along a gradient of future warming and brownification in freshwater mesocosms . Freshw. Biol. , 62 , 1869 – 1878 . Google Scholar OpenURL Placeholder Text WorldCat Utermöhl , H. ( 1958 ) Zur Vervollkommnung der quantitativen Phytoplankton-Methodik . Int. Ver. Theor. Ang. Limnol. Mit. , 9 , 1 – 38 . Google Scholar OpenURL Placeholder Text WorldCat Veraart , A. J. , de Klein , J. J. and Scheffer , M. ( 2011 ) Warming can boost denitrification disproportionately due to altered oxygen dynamics . PLoS One , 6 , e18508 . Google Scholar Crossref Search ADS PubMed WorldCat Wiltshire , K. H. and Manly , B. F. ( 2004 ) The warming trend at Helgoland Roads, North Sea: phytoplankton response . Helgol. Mar. Res. , 58 , 269 – 273 . Google Scholar Crossref Search ADS WorldCat Yvon-Durocher , G. , Allen , A. P., Cellamare , M., Dossena , M., Gaston , K. J., Leitao , M., Montoya , J. M., Reuman , D. C. et al. ( 2015 ) Five years of experimental warming increases the biodiversity and productivity of phytoplankton . PLoS Biol. , 13 , e1002324 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author(s) 2021. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permission@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Elevated temperature results in higher compositional variability of pioneer phytoplankton communities in a mesocosm system
JF - Journal of Plankton Research
DO - 10.1093/plankt/fbab013
DA - 2021-03-23
UR - https://www.deepdyve.com/lp/oxford-university-press/elevated-temperature-results-in-higher-compositional-variability-of-VScD7mdFg8
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -