Introduction

Turbulence is a ubiquitous feature of geophysical fluids, which in turn affects their physicochemical and physicobiological profiles (Margalef 1997). The importance of turbulence to the aquatic environment has been widely acknowledged. This is particularly true for phytoplankton, whose spatial distributions are to a large extent governed by physical mixing processes (Diehl et al. 2002). Turbulent mixing influences the sedimentation and most biological processes of phytoplankton, including photosynthesis, growth, nutrient dynamics and grazing (as reviewed by Fraisse et al. 2015). Mixing and mixing depth in particular, which transport populations to deeper waters and thus limits irradiance and photosynthetic activity, have been argued to affect phenology and proliferation of phytoplankton blooms (Huisman et al. 2004). Turbulence, which is the physical stressor generated during mixing, could also cause mechanical damage and induce even mortality (Thomas and Gibson 1990). Studies in vitro and in the field, however, showed that different taxa have varying levels of sensitivity and tolerance to turbulence (Sullivan and Swift 2003), and certain phytoplankton species may show specific adaptive responses to turbulence (Schapira et al. 2006). For example, Margalef’s Mandala hypothesis predicts that in high nutrient but turbulent conditions, diatoms will proliferate, while dinoflagellates favor more stratified waters (Margalef 1978). This is associated with the buoyancy and fragility of dinoflagellate cells and the high nutrient uptake capacity of diatoms (Margalef 1978). Generally, turbulent mixing can cause species replacements from dinoflagellates in oceans or cyanobacteria in lakes/rivers toward diatoms-/green algae-dominated communities (as reviewed by Huisman et al. 2004). Therefore, turbulence could select particular life-forms (Arin et al. 2002) and regulate phytoplankton community structures in aquatic ecosystems (Fraisse et al. 2015; Barton et al. 2014a).

However, studies on the role of turbulence on phytoplankton communities in lakes are still limited, and most phytoplankton competition studies reported have utilized well-mixed laboratory-based systems (Huisman et al. 1999; Fraisse et al. 2015). Few studies have focused on the wind wave turbulence in lakes, especially in large, shallow lakes which are strongly influenced by wind-driven perturbations (Zhou et al. 2015). Moreover, turbulence is also a highly variable environmental parameter, changing in both time and space over many orders of magnitude (Visser et al. 2009). Therefore, to better understand the roles of wind wave turbulence in plankton dynamics, it is necessary to know the effects of the magnitude and duration of turbulence events (Guadayol and Peters 2006). This opens a question on whether regional modulations of the turbulent kinetic energy could affect the dynamics of phytoplankton groups in eutrophic lakes (Sciascia et al. 2013).

Lake Taihu (Taihu), China’s third largest freshwater lake, is a typical large eutrophic shallow lake, with a maximum depth of 2.9 m, an average depth of 1.9 m and a water surface area of 2230 km2 (Qin et al. 2007). This lake plays a critical role in the social and economic development of the surrounding regions due to recreational opportunities and numerous natural resources (Qin et al. 2010). It is a major source of drinking water, livelihood and food supply for the surrounding communities of more than 8 million people (Wilhelm et al. 2011). Unfortunately, since the mid-1980s, severe blooms of toxin-producing Microcystis spp. have annually appeared in the lake from late spring through autumn (Qin et al. 2010). In 2007, a severe algal bloom occurred over most of Lake Taihu, which led to a drinking water crisis in Wuxi City that interrupted the drinking water supply for approximately two million inhabitants for at least a week (Qin et al. 2010). Like other shallow lakes, Taihu is strongly influenced by wind-driven wave turbulence and has been implicated in the recurrence of cyanobacterial blooms (Wu et al. 2010, 2013, 2015; Zhu et al. 2014). Zhou et al. (2015) previously showed that diatoms and green algae seem to dominate in the central part of the lake which is strongly influenced by wind, while harmful cyanobacteria dominated in the calmer Meiliang Bay. However, the effects of the magnitude and persistence of turbulence (time) on phytoplankton community structure and dynamics of harmful algal blooms in lakes are not clear and still require further study (Huisman et al. 2004; Jager et al. 2008; Zhou et al. 2015).

To address these knowledge gaps, mesocosm experiments were conducted to understand the competitive balance between phytoplankton species under different turbulence regimes in Taihu. Three levels of turbulence were tested on their effects on the composition of phytoplankton communities. Specifically, the aim of this study was to examine (1) the growth of phytoplankton groups under different turbulence regimes and (2) determine potential relationship between phytoplankton species competency and the magnitude and persistence of turbulence (time). Understanding of the interplay between the different characteristics of phytoplankton species and the physical environment will provide insights into the structure and variability of phytoplankton assemblages (Ross and Sharples 2008).

Materials and methods

Experimental setup

Mesocosm experiments were carried out in a total of twelve (12) customized tanks made of 8-mm-thick Plexiglass. The tank was right trapezoid with an external length, width and height measuring 67 (up), 53 (down) × 30 × 70 cm, which have maximal capacities of 126 L (see Fig. 1a). Each tank was filled with lake water to a height of 55 cm (about 96 L). Artificial waves were generated by frequency conversion wave-maker pumps (WP, Jebao, China; Fig. 1c), while rebounds were reduced by the slopes (5:1) on the sides (Fig. 1a). Each of the dissipation plates was also fixed on the sides to lessen the rebound effects (detailed in Fig. 1b). Potential chemical contaminants in the materials used for the tanks were removed by immersing them in water for 15 days before being used for any experiment.

Fig. 1
figure 1

Diagram of the mesocosm tank (a) used in the experiments which includes an energy dissipation plate (b) and a wave-maker pump (c). All dimensions are in centimeter

Full size image

Experimental design

The duration of the experiments was limited to 15 days to lessen potential variances caused by bottle effects. Bioassays were performed from July 7 to July 22, 2014, at Taihu Laboratory for Lake Ecosystem Research on the shores of Taihu (31°41′835″N, 120°22′044″E), in Wuxi. The pre-cleaned tanks were filled with lake water pumped from 0.2-m subsurface of the water column. Natural conditions were simulated and mimicked by floating and fixing the tanks in an outside artificial pond (10 × 10 × 2 m) filled with lake water and dug 1 m into the ground. The inside walls of each tank (except the bottom) were also gently brushed at 6:30 p.m. every day to prevent adhesion of microorganisms.

Turbulence generation

The submerged wave-maker pumps fixed under surface water by strong magnets were used to generate turbulence simulating the ones induced by natural wind waves as demonstrated in previous studies (Pekcan-Hekim et al. 2013; Harkonen et al. 2014; Zhou et al. 2015, 2016).

The pump frequency was set to 1 Hz, and the turbulence generated was monitored and measured by an acoustic Doppler velocimeter (ADV, 10 MHz ADVField; Sontek/YSI, San Diego, California, USA). Turbulent kinetic energy was measured from the middle of the tank with a 25-Hz measurement for a period of 2 min. Measurements were taken after the turbulent motion in the tank had reached a steady state after around 10 min.

The root-mean-square velocity (U rms), which helps define the characteristic speed of the turbulence, was calculated using the following formula:

$$U_{\text{rms}} = \sqrt {u_{\text{rms}}^{2} + v_{\text{rms}}^{2} + w_{\text{rms}}^{2} }$$
(1)

where

$$\begin{aligned} u_{\text{rms}} & = \sqrt {\frac{{\sum \mu_{x}^{2} - \left( {\sum \mu_{x} } \right)^{2} /n}}{n - 1}} \\ v_{\text{rms}} & = \sqrt {\frac{{\sum \mu_{y}^{2} - \left( {\sum \mu_{y} } \right)^{2} /n}}{n - 1}} \\ w_{\text{rms}} & = \sqrt {\frac{{\sum \mu_{z}^{2} - \left( {\sum \mu_{z} } \right)^{2} /n}}{n - 1}} \\ \end{aligned}$$
(2)

is the fluctuation of the flow for Cartesian vector x, y and z, and n is the number of samples per measurement. The U rms were expressed as averages for the whole tank. The energy dissipation rate (ɛ, m2 s−3), which describes the rate at which the turbulent energy decays over time, was deduced from the U rms (m s−1) following the formula described by Sanford (1997):

$$\varepsilon = A_{1} \frac{{U_{\text{rms}}^{3} }}{l}$$
(3)

where A 1 is an dimensional constant of order 1 (Moum 1996; Kundu and Cohen 2010) and l is the depth (m) of the water column describing the size of the largest vortices.

The Reynolds number (Re, the ratio of inertial forces to viscous forces) for the given turbulence levels was calculated following Peters and Redondo (1997):

$$Re = l\frac{{U_{\text{rms}} }}{v}$$
(4)

where l is the water depth (m) and v is the kinematic viscosity for water (8.5 × 10−7 m2 s−1).

The different levels of turbulence used in this study were based on actual conditions previously observed in Lake Taihu during summer 2013. The U rms which corresponded to the three turbulence levels used in the experiments were 0.85 cm s−1 (low), 2.53 cm s−1 (medium) and 4.33 cm s−1 (high). These turbulence have corresponding ɛ ranging from 1.12 × 10−6 to 1.48 × 10−4 m2 s−3 with Re between 5500 and 92,620 in Taihu (Table 1). All treatments were conducted in triplicates.

Table 1 Summary of the root-mean-square velocities (U rms), energy dissipation rates (ɛ) and Reynolds numbers (Re) of the four levels of turbulence (control or calm water, low, medium and high turbulent kinetic energy) used in the experiments and in Lake Taihu
Full size table

Measurements

Environmental parameters were measured from the start of the experiments daily until day 9, and at day 12 and 15, between 7:00 and 8:00 in the morning. Water temperature was determined using a 6600 multi-sensor sonde (Yellow Springs Instruments, San Diego, California, USA). Nutrients were also analyzed including total nitrogen (TN), total dissolved nitrogen (TDN), dissolved inorganic nitrogen [DIN, ammonium (NH4 +) + nitrate (NO3 −) + nitrite (NO2 −)], total phosphorus (TP), total dissolved phosphorus (TDP) and soluble reactive phosphorus (SRP), according to the methods described by Zhu et al. (2014). The particulate nitrogen (PN) was obtained by subtracting the TDN from the TN and particulate phosphorus (PP) as the difference between the TDP from the TP.

The chlorophyll a (Chl a) concentrations were measured by spectrophotometric method (Papista et al. 2002). Samples were first filtered through GF/F filters and frozen at −20 °C, and pigments were extracted with 90 % hot acetone.

A PHYTO-PAM™ fluorometer (Heinz Walz GmbH, Effeltrich, Germany) was used to assess the photosystem II maximum photochemical efficiency (Fv/Fm) and the maximum photosynthetic electron transport rate (ETRmax) of the phytoplankton. For that, 5 mL samples were acclimated in the dark for 10 min near in situ temperatures.

Cell counts of the autotrophic species were determined at the beginning of the experiment (day 0) and on every third day until day 15 and fixed with Lugol’s iodine solution (2 % final conc.). Samples were then allowed to settle for 48 h prior to counting with a Sedgwick–Rafter counting chamber under an upright light microscopic (Olympus, Japan) at 200–400× magnifications. Taxonomic classification followed the descriptions of Hu et al. (1980). Algal biovolumes were calculated from the cell numbers and cell size measurements by measuring at least 20 individual cells of each species in each sample (Hu et al. 1980).

Statistical analysis

To test for significant differences in physicochemical conditions among and between the various treatments, one-way analysis of variance (ANOVA) was employed, and the level of significance used was P < 0.05 for all tests. Post hoc multiple comparisons of treatment means were made by Tukey’s least significant difference procedure, and the standard deviation in the variations of the triplicates was calculated. Temporal differences in phytoplankton community structure were assessed by calculating for the Hellinger distance between each pair of samples. The Hellinger distance was chosen because it is not influenced by differences in total biomass (Legendre and Legendre 1998). A distance matrix was then obtained and was clustered by Ward method to obtain a dendrogram visualizing differences between samples according to their phytoplankton species composition. Statistical analyses were performed using the SPSS 22.0 statistical package.

Results

Environmental characteristics

Water temperature generally ranged from 25.6 to 30.2 °C in all tanks during the experiment (P > 0.05). Moreover, the TN and TP concentrations ranged from 1.24 to 1.99 mg L−1 and from 0.032 to 0.092 mg L−1, respectively. Most nutrients, like total and particulate N and P, seemed to be higher in the treatments (low, medium and high) than in the control (Fig. 2a, b, e, f). Further, they were highest in the treatments with medium turbulence (Fig. 2a, b, e, f). The TDN and DIN on the other hand slowly decreased toward the end of the experiment, except for high turbulence (Fig. 2c, d). In high turbulence, TDN increased after day 8, and DIN kept stable after day 4 (Fig. 2c, d). Moreover, fractions of the different forms of dissolved phosphate (TDP and SRP) fluctuated and were generally higher in the treatments than in the control after 5 days (Fig. 2g, h). However, no significant differences were observed in the nutrient profiles between still and turbulent treatments (P > 0.05), except for the TN and PN.

Fig. 2
figure 2

Variations of nitrogen and phosphorus fractions (a TN, b PN, c TDN, d DIN, e TP, f PP, g TDP, h SRP) in the control and three turbulence treatments (low, medium and high) during the experiment. Control (square), low (circle), medium (triangle) and high (inverted triangle). TN total nitrogen, PN particulate nitrogen, TDN total dissolved nitrogen, DIN dissolved inorganic nitrogen, TP total phosphorus, PP particulate phosphorus, TDP total dissolved phosphorus, SRP soluble active phosphorus

Full size image

The 15-day average of Fv/Fm and ETRmax was lower in the control than in the treatments (P > 0.05) and highest in medium (Table 2). Simultaneously, the average of Chl a concentration in the control was lower compared to the treatments and highest in medium, with significant difference between the control and medium (Table 2, P < 0.05).

Table 2 Mean and standard deviation of biological measurements in the control (C) and turbulence treatments (low, medium and high) during the experiments
Full size table

Phytoplankton community structure dynamics

A total of 39 phytoplankton species were observed for the entire duration of the experiment, which belonged to the four major groups: Cyanophyta, Bacillariophyta, Chlorophyta, and others. Throughout the experiment, the phytoplankton biomass was mainly composed of Cyanophyta and Bacillariophyta (71.8–97.5 %), while Chlorophyta and others accounted for only a very small proportion of the total biomass. The phytoplankton biomass increased from day 0 to day 15 in all tanks during the experiment (Fig. 3). Compared to calm waters, the phytoplankton biomass was higher in the treatments and highest in medium (Fig. 3), but there is no significant difference among the treatments (P > 0.05). Interestingly, the abundances of the four phytoplankton groups increased with the presence of turbulence on the first 6 days of the experiment, especially for Cyanophyta which dominated in all tanks before 6 days (Fig. 3).

Fig. 3
figure 3

Biomass contribution of Cyanophyta, Bacillariophyta, Chlorophyta and others in the control (a) and three turbulence treatments (b low, c medium and d high) during the experiment

Full size image

In the control, Cyanophyta remained the most dominant phyla for the entire duration of the experiments and remained on the surface waters (Fig. 3). In contrast, the biomass of Bacillariophyta dramatically increased and gradually became the dominant phyla by replacing Cyanphyta after about 8–11 days in all of the treatments (Fig. 3). The critical time for the shift in phytoplankton community structure was shortest in high (8 days), followed by low (9 days) and longest in medium (11 days). Interestingly, even though the Cyanophyta populations accounted for only a small proportion of the total biomass (11.1–19.3 %) in the treatments at the end of experiment, it was still higher than the control especially in medium (Fig. 3).

Further classification showed that pennate diatoms Fragillaria spp. accounted for 59.7 ± 34.1 % of Bacillariophyta biomass. The Cyanophyta was almost composed by buoyant Microcystis spp. (92.8 ± 7.2 %) during the experiment. Fragilaria spp. and Microcystis spp. were the two dominant species during the experiments. Like the Cyanophyta, the biomass of Microcystis spp. increased before 6 days and were mainly higher in turbulent treatments than in control especially in medium (Fig. 4, P < 0.05). In the control, the biomass of Fragilaria spp. was sparse, and Microcystis spp. was always the dominant species during the experiment (Fig. 4). However, in turbulent treatments, the biomass of Fragilaria spp. rapidly increased after 6 days and became the dominant species by replacing Microcystis spp. after 8–10 days.

Fig. 4
figure 4

Biomass variations of Fragilaria spp. and Microcystis spp. in the control (a) and three turbulence treatments (b low, c medium and d high) during the experiment

Full size image

Temporal differences in phytoplankton community structure

Hierarchical clustering showed the temporal differences in phytoplankton community structure among the four treatments during the experiment. The differences in community structures between the control and the treatments occurred after 3 days (Fig. 5). Different patterns and clustering was also observed on the succeeding days. Interestingly, control samples mainly remained similar to those in low turbulence condition (Hellinger distance <4, except L3), but it drastically changed and diversified after day 9 (Fig. 5). The community structure in medium turbulence from 6 to 12 days grouped in a distinct cluster (Fig. 5). Particularly, highest distances were obtained on the 15th day in the medium and high treatments, which were also the most different with others (Fig. 5).

Fig. 5
figure 5

Hierarchical clustering (Ward method) of the distance matrix (Hellinger distance) based on the species composition of the control and three turbulence treatments during the experiment, where M0 is the sample of M, at day 0. C, L, M and H represent the control, low, medium and high turbulence treatments, respectively

Full size image

Discussion

Most studies have focused on the effects of turbulence (or mixing) processes on phytoplankton communities in water, particularly in species interactions and shift in species composition (Huisman et al. 1999, 2004; Jager et al. 2008; Prairie et al. 2012; Barton et al. 2014b; Fraisse et al. 2015). To date, studies that have tested the effects of the magnitude and persistence of turbulence (time) in lakes are still limited (Zhou et al. 2015). In this study, our experimental setup was designed to simulate the effects of wind wave turbulent of standard kinetic energy in Lake Taihu through large volume mesocosms. Mesocosm studies allow using naturally existing communities in water bodies to be used and mimic natural conditions as close as possible, but also permitting various environmental factors such as turbulent kinetic energy (Striebel et al. 2013). The present results demonstrated that turbulence promoted the growth of phytoplankton and influenced the shift in the composition of phytoplankton species. For example, the pennate diatom Fragilaria spp. were more favored and became the dominant species by replacing buoyant cyanobacteria Microcystis spp. after a lag phase of 8–11 days in all the turbulent treatments, while buoyant Microcystis spp. seemed to be more adapted and always dominated in calm conditions (control). Interestingly, turbulence mainly promoted the growth of cyanobacteria in short-term period (6 days), which is beneficial for the harmful cyanobacterial blooms (CyanoHABs). These results are consistent with previous observations in Taihu (Zhu et al. 2014; Wu et al. 2013) and other lakes (Huisman et al. 2004; Jager et al. 2008).

Turbulence is a ubiquitous hydrodynamic feature of all surface waters and is also a highly variable environmental parameter, changing in both time and space over many orders of magnitude (Visser et al. 2009). In Lake Taihu, with 1.9 m average depth, turbulence has little space in which to dissipate. Therefore, the predicted diameter of the smallest eddies in the lake is much smaller, and the turbulent kinetic energy content and the turbulent shear forces are much higher than those in stratified deep lakes and even in the open ocean (G-Tóth et al. 2011). In Taihu, the ɛ varied between 6.01 × 10−8 and 2.39 × 10−4 m2 s−3 (Table 1), which corresponded to the range of values (from 1.07 × 10−7 to 6.67 × 10−3 m2 s−3) previously measured by G-Tóth et al. (2011 and references therein) in the large shallow Lake Balaton. Thus, the turbulent kinetic energy (ɛ, 1.12 × 10−6–1.48 × 10−4 m2 s−3) used in our experiment was commonly found in large, shallow lakes.

Turbulence can have positive effects on primary production by bringing nutrient-rich deeper waters to the oligotrophic surface water and enhancing the nutrients flux into the cells, thereby increasing nutrient available for phytoplankton use (Peters et al. 2006; Striebel et al. 2013). Our results also demonstrated that the total and particulate nutrient concentrations were larger, but dissolved nutrient concentrations were mainly smaller in the treatments (except for high turbulence) than in the control (Fig. 2). Therefore, the resulting optimum condition for photosynthesis decreased dissolved nutrient concentrations in the turbulence treatments by converting them to particulate matter (except for dissolved phosphorus as discussed in Zhou et al. 2016). Moreover, under calm conditions, gas-vacuolated cyanobacterium Microcystis concentrated in the water surface while shading themselves and other phytoplankton species (Huisman et al. 2004). So, turbulence can reduce the shading that subsidizes direct solar input which is in part responsible for the relatively high productivity under turbulence (Petersen et al. 1998). Our results also showed photosynthetic efficiency was enhanced when turbulence was present (Table 2), suggesting that it imparts a positive effect on the structure of PS II and promotes the growth rate of autotrophs (Wang et al. 2012). Furthermore, enhanced phytoplankton biomass may also be associated with resuspension (rather than sedimentation) and dilution leading to low zooplankton capture rates under turbulent conditions (Pécseli et al. 2014). Therefore, turbulence promoted the growth of all phytoplankton groups, especially for diatom and cyanobacteria (Fig. 3). However, results of this study showed that there was a lag time (around 6 days) for the dramatic increase in diatoms biomass. In short term (6 days), it was rather beneficial to the growth of Microcystis blooms, which was consistent with recent field studies demonstrating that short-term tropical cyclones stimulated buoyant Microcystis spp. blooms in Taihu (Zhu et al. 2014).

Turbulence, however, could also negatively affect phytoplankton several ways although few are well understood (Fraisse et al. 2015). Indeed, the lower Kolomogorov scale (defining a lower eddy motion) produced when there is high turbulence may be expected to result in phytoplankton mechanical stress or damage and even mortality (Thomas and Gibson 1990). That is why the phytoplankton biomass and Chl a were smaller in the high treatment compared to the medium and/or low treatments (Fig. 3; Table 2), which also resulted in higher dissolve nutrient concentrations, while particulate nutrient concentrations were lower in high turbulence (Fig. 2b, c). In addition, it should be noted that resuspension was unlikely to occur in low turbulence, but strongly manifested in high in this experiment. Studies suggested that the most likely mechanisms resulting to the observed increase in phytoplankton biomass under turbulent conditions are the increase in nutrient flux and solar input, as shown previously both in vitro and in the field (Metcalfe et al. 2004; Guadayol et al. 2009).

It is widely known that buoyant cyanobacterium Microcystis spp. have an advantage in calm waters by increasing nutrient flux and light absorption through migration. Large non-motile diatoms are favored by turbulent mixing, which are able to grow relatively quickly compared with other taxa, and have high affinities for nitrate and phosphate (Edwards et al. 2012). The available empirical evidence suggests that turbulence strongly affected functional community composition of phytoplankton (motile vs. sinking taxa; Jager et al. 2008). In this study, diatom Fragilaria spp. exhibited high growth rate and dominated the community after 8–11 days under the turbulent conditions. This may have been caused by a number of factors such as (1) the increase in nutrient influx, especially for large diatom (Prairie et al. 2012; Barton et al. 2014b); (2) high adaptability of diatoms to fluctuating light conditions (Huisman et al. 2004); (3) decreased sedimentation losses of non-motile diatoms under turbulence; and (4) the capture rates of diatoms by predators is reduced in turbulent environments (Pécseli et al. 2014; Zhou et al. 2015). The results reported in this paper are consistent with previous field studies and theoretical predictions proposed to explain shifts in the competitive balance between buoyant and sinking phytoplankton species in freshwater ecosystems (Huisman et al. 2004; Jager et al. 2008; Zhou et al. 2015). Unexpectedly, pennate diatom also dominated in the low turbulence conditions, same as those in medium and high (Fig. 4). A likely explanation is that the turbulence used in this experiment was constant and its duration was rather long, which may be actually rarely experienced for a long time in nature. Moreover, in this experiment, one surprising result was that the same diatom species dominated the community under different turbulence levels (Fig. 4). Wang et al. (2012) suggested that the dominant diatom species was determined by the coupling of hydrodynamics and nutrient concentrations, according to their habitat-dependent physiological characteristics. In this experiment, the environmental characteristics were similar among the turbulence treatments during the experiment (Fig. 2, P > 0.05). That may be the reason why the same diatom species dominated under different turbulence levels in this study.

Although the same phytoplankton species were dominating under different turbulence treatments, differences in the time needed before shift in composition manifested varied with the degree of turbulent kinetic energy (Figs. 3, 4). To have significant and quick changes in the phytoplankton community, experiments are generally conducted under constant turbulence conditions (Peters and Marrase 2000). Although it is true that relatively high turbulence levels such as those used in many laboratory experiments are common in the lake, organisms may actually rarely experience them for a long period of time in nature (Guadayol and Peters 2006). In this study, it took 8–11 days as the lag time of diatom Fragilaria spp. before it dominated in the turbulent tanks, which was consistent with previous enclosure experiments in Lake Brunnsee in which the shift took 21 days through intermittent mixing (for 5 min every 30 min; Jager et al. 2008). The community structure was first observed to shift in high (8 days), suggesting intense turbulence may accelerate consequential changes.

Conclusions

This study based on mesocosm experiment in Lake Taihu demonstrated that turbulence shifted the competitive balance from buoyant cyanobacterium Microcystis spp. to sinking diatom Fragilaria spp. after a lag phase of 8–11 days. These findings suggest that phytoplankton community structure response to turbulence seems to be independent of the turbulent kinetic energy (ɛ = 6.01 × 10−8–2.39 × 10−4 m2 s−3) used in this study. Moreover, turbulence triggered a highly significant increase in phytoplankton biomass, especially for the cyanobacteria Microcystis spp. before 6 days and then for diatom Fragilaria spp. in the remaining days. This experiment then implies that although changes in turbulence are likely to shift the species composition of the phytoplankton, short-term turbulence is rather beneficial for the cyanobacterium Microcystis blooms in eutrophic lakes. These findings may find useful application in water management, improving prediction and mitigation to prevent CyanoHABs. Our results may contribute to a better understanding of phytoplankton communities and dynamics of harmful algal blooms in lake ecosystems exposed to natural turbulent regimes.