Introduction

Small-scale turbulence is an abiotic factor that contributes to the feeding of aquatic animals by affecting encounter rates between predators and prey and thereby influencing ingestion rates of predators (Rothschild & Osborn, 1988; Marrasé et al., 1990; Härkönen et al., 2014a). The effects of turbulence depend on the size and swimming speed of the organisms. Large and strong swimmers such as adult fish are usually unaffected by turbulence due to their high maneuverability (Kiørboe & Saiz, 1995). On the other hand, very small organisms may also be immune to turbulence (Peters & Marrasé, 2000; Zhou et al., 2016). The most pronounced effects of turbulence have been described for organisms of intermediate sizes, such as crustacean zooplankton, macroinvertebrates, and larval fish (Rothschild & Osborn, 1988; MacKenzie et al., 1994; Kiørboe & Saiz, 1995).

Studies on the effects of turbulence on rotifers, which are usually clearly smaller than crustacean zooplankton and in size range of 100–700 µm, have been relatively sparse and conclusions have been variable. Both positive and negative responses of rotifers to increasing turbulence have been reported (Miquelis et al., 1998; Sluss et al., 2008; G.-Tóth et al., 2011; Zhou et al. 2018). Clarification is needed because rotifers can make significant contributions to zooplankton abundance and productivity in numerous lakes (Mackarewicz & Likens, 1979; Pace & Orcutt, 1981). Due to their high abundance in some ecosystems, rotifer community grazing rates may match, or even exceed those of some crustaceans both on a daily and annual basis (Bogdan & Gilbert, 1982). Additionally, rotifers form a substantial food source for many planktivorous organisms, including cladocerans, copepods, chaoborids, and fish larvae (Williamson, 1983; Moore & Gilbert, 1987; Nagata et al., 2006). Therefore, any disturbance affecting rotifer communities may influence the whole aquatic food web, and turbulence levels in many aquatic ecosystems are probably increasing due to increasing wind speeds and decreasing water levels, and these changes will influence the planktonic communities (Pryor et al., 2005; G.-Tóth et al., 2011).

In an outdoor pond experiment, Härkönen et al. (2014b) showed that intermediate turbulence affected crustacean zooplankton communities via changes in the feeding efficiency and prey selection of predators. The predation pressure exerted by invertebrate predators (larvae of the phantom midge Chaoborus flavicans (Meigen)) on crustacean zooplankton, especially the cladoceran genus Bosmina sp., increased when turbulence was introduced to the system. The study focused on crustacean zooplankton, and for rotifers, only total biomasses were reported. Therefore, in this study, the effects of turbulence on the rotifer community in the experiment were explored in detail. It was expected, that turbulence has only minor effects on rotifer communities due to their small individual size. Additionally, it was hypothesized that inclusion of chaoborids in the system would enhance the abundance of rotifers (e.g., Yan et al., 1991).

Materials and methods

Experimental set-up

The study was conducted between 23 July and 31 August in summer 2012 in outdoor ponds filled with humic water (area 8.1 m2, volume 3200 l, mean depth 40 cm, maximum depth 60 cm) in the Evo district (61°13′N, 25°12′E), southern Finland. The ponds were rectangular, had a sand-gravel bottom covered with a 0.5–1 cm layer of organic debris and had no macrophyte vegetation, but only natural communities of phytoplankton. The experiments included two different predation regimes: no predation (CTRL), and invertebrate predation (IP; 3th and 4th instar larvae of the phantom midge C. flavicans) and two different turbulence conditions: no induced turbulence (CALM), and intermediate turbulence (TURB; RMS velocity of 1.4 cm s−1, dissipation rate 5.6 × 10−6 m2 s−3). In the CALM-treatments, the natural background RMS velocity was 0.3 ± 0.1 cm s−1 (dissipation rate 4.6 × 10−8 m−2 s−3). In lakes, the dissipation rate in the surface mixed layer during calm conditions often varies between 10−9 and 10−8 m2 s−3, and can rise to 10−6–10−5 m2 s−3 during wind forcing (Saggio & Imberger, 1998; MacIntyre et al., 1999). Thus, the turbulence conditions in the experimental ponds corresponded to natural lake circumstances. Based on the dissipation rate and kinematic viscosity of the water (Tennekes & Lumley, 1972; Zhou et al., 2016), the smallest eddy diameter in our experiment was c. 0.7 mm. Each experiment of the four combinations of turbulence and predation (CALM-CTRL, CALM-IP, TURB-CTRL, TURB-IP) included three replicate ponds. The different treatments and replicates were situated randomly in the experimental pond area. C. flavicans larvae were used as they are important predators of rotifer communities and are often abundant in humic lakes common also in lakes in the Evo district (Liljendahl-Nurminen et al., 2003; Wissel et al., 2003).

The ponds were filled with water from the nearby dystrophic River Majajoki (coming from Lake Majajärvi) after filtration through a 50 µm net. The watercolor in the ponds was 140 mg Pt l−1. Turbulence was generated with computer-controlled submersible pumps (Tunze Turbelle nanostream 6055; Tunze Aquarientechnik GmbH, Penzberg, Germany) and measured with an acoustic Doppler velocimeter (ADV, 10-MHz ADVField; Sontek/YSI, San Diego, CA, USA). The natural zooplankton community was collected from the dystrophic Lake Majajärvi, and equal aliquots of zooplankton were added to each pond before the experiments. The zooplankton samples were taken with a 153 µm plankton net in the daytime by horizontal hauls in the uppermost 2 m layer, which prevented chaoborids larvae of being included in the samples. During daytime in Lake Majajärvi (max depth 12 m), third and fourth instars of Chaoborus larvae inhabit water layers below 4 m (Horppila et al., 2018). The relatively large mesh size was chosen for zooplankton collection because Lake Majajärvi is inhabited by the flagellate Gonyostomum semen (Ehrenberg) Diesing, which makes the use of smaller mesh sizes impossible due to clogging (Cronberg et al., 1988; Estlander et al., 2009). Third and fourth instars of Chaoborus larvae were collected from Lake Majajärvi with vertical hauls of a plankton net (mesh size 180 µm). Each experimental pond with IP - treatment received 960 larvae, leading to an initial density of 0.3 ind. l−1 that corresponds to densities found in numerous lakes (Wissel et al., 2003; Estlander et al., 2009).

Sampling and analyses

The ponds were sampled at 4-d intervals for 5 weeks. Zooplankton samples were taken with a tube sampler (diameter 5.4 cm, length 50 cm). From each pond during each sampling, five samples from random places were lifted and combined, resulting in a 6 l sample volume per pond. The water was filtered through a 50 µm plankton net and the samples were preserved in 4% formaldehyde. Rotifers were analyzed by inverted microscopy (Olympus CK40; 125 × magnification; Olympus Corporation, Tokyo, Japan), measured for length and identified to species or genus level. The biomasses were calculated from individual lengths, using length–weight regressions (Dumont et al., 1975; Bottrell et al., 1976). Together with chaoborids, the biota of Lake Majajärvi includes also other invertebrate predators preying on rotifers; cyclopoid copepods and the cladoceran Polyphemus pediculus (L.). As the filtration did not exclude these taxa, their abundance in the different treatments was also examined. Because Bosmina is usually the main crustacean prey for chaoborids (Stenson, 1990; Lüning-Krizan, 1997; Liljendahl-Nurminen et al., 2003), results on Bosmina biomass are also shown to explain the turbulence-mediated variations in the feeding preferences of chaoborids.

Water temperature, dissolved oxygen and pH in each pool on each sampling date were measured with a YSI 6600V2 sonde (YSI Inc. Yellow Springs, OH, USA). Total phosphorus and total nitrogen samples were taken with the tube sampler and analyzed with Lachat Autoanalyzer [Quickchem Series 8000, Lachat Instruments, (Hach Company)] according to Koroleff (1979). Chlorophyll a samples were taken with a tube sampler, filtered with Whatman GF/C filters, and analyzed spectrophotometrically (Shimadzu UV-260, UV–Visible Recording Spectrophotometer, Shimadzu Corporation, Tokyo, Japan) after extraction with ethanol according (Finnish Standard Association, 1993).

Statistical analyses

The effect of the different treatments on the density and biomass of various rotifer taxa were analyzed with analysis of variance for repeated measurements (ANOVAR), accounting for the temporal autocorrelation of the sequential samples. The data were ln(x + 1) transformed before the analyses. The analyses were conducted on genus level on those rotifer genera exceeding the average density of 5 ind. l−1 in at least one treatment. Other rotifer taxa were grouped, except Asplanchna spp. Asplanchna species are predaceous rotifers and can regulate other rotifer species (Gilbert, 1980; Conde-Porcuna & DeClerck, 1998). Therefore its abundance in the different treatments was tested separately. Additionally, because variable responses of different species within genera might be masked if only genera were analyzed, the density and biomass fluctuations of six dominant species [Chromogaster ovalis (Bergendal), Conochilus unicornis (Rousselet), Keratella cochlearis (Gosse), Polyarthra remata Skorikov, Trichocerca porcellus (Gosse), Asplanchna priodonta (Gosse)] were analyzed with ANOVAR. Moreover, the between-treatment differences in the biomass of cyclopoids, P. pediculus, and Bosmina sp. were studied with ANOVAR. Pairwise comparisons were performed with Bonferroni t-tests and tests with P values < 0.05 were judged significant. The possible difference in the initial rotifer biomass between treatments was analyzed with analysis of variance (ANOVA, ln ( x+ 1)—transformed data). Differences in the average individual length of rotifers were analyzed with ANOVA (all taxa combined and different taxa separately). Additionally, as rotifer abundance often increases with increasing chlorophyll a concentration via increasing food abundance (Nagata et al., 2006), the variations in chlorophyll a concentration were analyzed with ANOVAR.

Results

The physical and chemical water quality variations between the treatments were very small. Water temperature fluctuated between 18 and 23°C, and between-treatment differences remained below 0.5°C. The concentration of dissolved oxygen was between 8 and 9 mg l−1 and water pH between 6.8 and 6.9 in all the treatments. The average total phosphorus concentration varied between 16 and 21 µg l−1 and total nitrogen concentration between 800 and 850 µg l−1. Chlorophyll a concentration fluctuated between 13 and 19 µg l−1, with no differences between the treatments (P = 0.244).

Altogether 27 rotifer species were discovered from the ponds during the study period. In the initial densities or biomasses of the different rotifer taxa, no between-treatment differences were observed (P > 0.05), except in Polyarthra, which was more abundant in the TURB treatments than in CALM- treatments (density P = 0.0256, biomass P = 0.0497). The total density of rotifers varied between 200 and 500 ind. l−1 at the beginning of the experiment, decreased during July in all treatments and remained below 100 ind. l−1 until the end of the study (Fig. 1). The total density in TURB-IP (average 131.6 ind. l−1) was significantly higher than in TURB-CTRL (106.5 ind. l−1) and in CALM-IP (64.2 ind. l−1) (Fig. 1, Table 1). The total biomass of rotifers was 5–15 μg C l−1 at the beginning of the experiments and dropped below 2 μg C l−1 during the study (Fig. 2). No significant differences in total biomass of rotifers were observed between treatments (Table 2). In CALM-IP, a biomass peak of 15 μg C l−1 was detected on 30 July (Fig. 2). It was caused by a temporary increase in abundance of the large-sized Asplanchna priodonta and Asplanchna herricki (Guerne).

Fig. 1
figure 1

Total rotifer density and biomass during the experiment in the different treatments

Full size image
Table 1 Pairwise, between-treatment comparisons (ANOVAR, Bonferroni t-tests) for the differences in the total density and density of the dominant rotifer genera and dominant species (Chromogaster ovalis, Conochilus unicornis, Keratella cochlearis, Polyarthra remata, Trichocerca porcellus) under calm (CALM) and turbulent (TURB) water
Full size table
Fig. 2
figure 2

Density of different rotifer genera during the experiment in the different treatments

Full size image
Table 2 Pairwise, between-treatment comparisons (ANOVAR, Bonferroni t-tests) for the differences in the biomass of different rotifer taxa, Polyphemus pediculus, cyclopoida, and Bosmina sp. under calm (CALM) and turbulent (TURB) water
Full size table

The most abundant rotifer genera in all treatments during the study period both regarding density and biomass were Keratella (mainly K. cochlearis), Synchaeta spp., Polyarthra (P. remata, P. vulgaris Carlin, P. euryptera (Wierzejski)), and Trichocerca (T. porcellus, T. cylindrica Imhof). Synchaeta dominated especially during the first week of the experiment with a density of 100–250 ind. l−1 (biomass 3–7 μg C l−1) (Figs. 2, 3). In August, Synchaeta was almost absent while the density Polyarthra and Trichocerca varied mostly between 2 and 30 ind. l−1 (biomass 0.05–2 μg C l−1). Towards the end of the experiment, especially the abundance of Conochilus (mainly C. unicornis), and Chromogaster ovalis increased (Figs. 2, 3).

Fig. 3
figure 3

Biomass of different rotifer genera during the experiment in the different treatments

Full size image

Rotifers were not affected by turbulence alone (Tables 1, 2). Introduction of chaoborids into the calm water had no effect either, as no differences between CALM-CTRL and CALM-IP treatments were detected in density or biomass. In turbulent water, however, chaoborids affected the rotifer community. In TURB-IP, the density of all studied genera was higher than in CALM-IP (Table 1). For most taxa, the density in TURB-IP also exceeded their density in TURB-CTRL. In biomass, fewer significant differences were detected, but the biomass of Chromogaster, Keratella, Polyarthra, and Trichocerca was significantly higher in TURB-IP than in CALM-IP (Table 2). Trichocerca and Chromogaster were also more abundant in TURB-IP than in TURB-CTRL. In the statistical analyses on the dominant species, fewer significant differences were found than at the genus level, but the most significant differences were again found between TURB-IP and CALM-IP and between TURB-CTRL and TURB-IP (Table 3).

Table 3 Pairwise, between-treatment comparisons (ANOVAR, Bonferroni t-tests) for the differences in the density and biomass of six dominant rotifer species (Chromogaster ovalis, Conochilus unicornis, Keratella cochlearis, Polyarthra remata, Trichocerca porcellus, Asplanchna priodonta) under calm (CALM) and turbulent (TURB) water
Full size table

The average body length of rotifers varied mostly between 110 µm and 140 µm (Table 4). When all sampling days were pooled, the average individual size was 125.4 µm in CALM-CTRL, 132.9 µm in CALM-IP, 132.3 µm in TURB-CTRL and 135.3 µm in TURB-IP. When the large-sized Asplanchna were excluded, the average sizes were 120.1, 123.4, 126.5 and 129.1 µm, respectively. In all treatments, the average individual body size of rotifers decreased during the experiment (Table 4). No differences in individual size were detected in any of the genera (P > 0.05 in all comparisons between treatments) (Fig. 4). Thus, turbulence had no effect on the individual size of rotifers. The minimum individual size of rotifers varied mostly between 30 µm and 80 µm, and the maximum individual size between 250 µm and 800 µm (Table 4).

Table 4 The average, minimum and maximum individual size of rotifers in the different treatments during the experiment
Full size table
Fig. 4
figure 4

The average individual size of the dominant rotifer genera in the different treatments

Full size image

The biomass of cyclopoid copepods (mainly Mesocyclops spp.) varied mostly between 2 and 10 μg C l−1 (1–10 ind. l−1) during the first week of the experiment and decreased after that (Fig. 5). No differences between the treatments were observed (Tables 1, 2). For Polyphemus, both turbulence and invertebrate predation affected. In TURB-CTRL, the density (max. 26 ind. l−1) and biomass (max. 15 μg C l−1) were significantly lower than in CALM-CRTL (max. 47 ind. l−1, 30 μg C l−1) (Fig. 5, Tables 1, 2). Additionally, in TURB-IP, Polyphemus was less abundant than in TURB-CTRL and CALM-IP. The biomass of Bosmina sp. was lower in TURB-IP than in TURB-CTRL or in CALM-IP (Table 2).

Fig. 5
figure 5

Biomass of Bosmina sp., Polyphemus pediculus, and cyclopoid copepods during the experiment

Full size image

Discussion

Both positive and negative responses of rotifers to increasing turbulence have been previously reported (Miquelis et al., 1998; Sluss et al., 2008; G.-Tóth et al., 2011). The present study suggested that intermediate turbulence alone does not have a direct influence on rotifers. This was expected because the effect of turbulence on planktonic organisms depends on the size of the organisms in relation to the size of the turbulent eddies. Especially organisms larger than the diameter of the smallest eddies may be directly affected by turbulent forces (Peters & Marrasé, 2000; Zhou et al., 2016). In the rotifer community of the experimental ponds, Asplanchna was the only taxon, where the individual size occasionally reached the size of the smallest eddies. The individual size of all other taxa was considerably lower explaining why the direct effects of turbulence on rotifers were weak. The size range of rotifers confirmed that the large mesh size used in plankton sampling did not bias the community structure because the minimum size of all rotifer taxa was considerably beyond 100 µm. This was explained by reproduction and by rotifers introduced from River Majajoki with water used to fill the ponds. The blocking effect of Gonyostomum could also decrease the realized mesh size of the 153 µm net. The species composition and diversity of the rotifer communities in the ponds were typical of humic lakes in the temperate zone (Havens, 1991; Sarvala et al., 1999). The overall decrease of rotifers in all treatments during the experiment was due to natural succession. In many lakes, including those in the Evo area, the abundance of rotifers decreases in late summer (Haberman, 1983; Salonen et al., 1991; Habidja et al., 1993).

The experiment demonstrated on the other hand that indirect, predation-mediated effects of turbulence on rotifer communities could be considerable. In treatments, where chaoborids were included, turbulence resulted in the increased abundance of most rotifer taxa. This was most likely a consequence of increased predation by chaoborids on crustacean zooplankton. Predators can affect their prey populations also through non-consumptive effects via influence on reproduction (Heuschele et al., 2014). However, no effect of chaoborids on the abundances of cladoceran embryos and copepod nauplii, their ratios to adult crustaceans (Härkönen et al., 2014b), or in the size structure of the rotifer communities was detected during the experiment. Thus, predation of chaoborids on crustacean zooplankton and its indirect effect on rotifers was the explanation for the observed increase of rotifers in the TURB-IP treatment. Turbulence enhanced the predation by chaoborids on cladocerans and the biomass of the preferred cladoceran prey, Bosmina sp. decreased in TURB-IP compared with CALM-IP. Accordingly, previous studies have shown that rotifers can dominate lake zooplankton when the abundance of chaoborids is high because chaoborid predation is concentrated on crustaceans (Hanazato & Yasuno, 1989; Havens, 1991; Yan et al., 1991). Larvae of C. flavicans are ambush predators that do not actively search for prey, but attack prey entering their strike area (Swift & Fedorenko, 1975). Therefore, moderate turbulence such as used in the present experiment can enhance their prey capture rate (MacKenzie & Kiørboe, 1995; Härkönen et al., 2014b), which is reflected as an increased predation rate especially on their preferred prey items.

The increased chaoborid predation on cladocerans could promote the increase of rotifers by two mechanisms. Firstly, predation on rotifers was relaxed. This had a strong effect on the rotifer community because most dominant rotifer taxa in the ponds belong to the diets of chaoborids (Elser et al., 1987; Moore & Gilbert, 1987; Vanni, 1988; Liljendahl-Nurminen et al., 2003). Secondly, the reduction of cladocerans increased rotifer abundance through reduced competition between cladocerans and rotifers (MacIsaac & Gilbert, 1989; Lampert & Rothhaupt, 1991). Rotifer density was enhanced more strongly than their biomass, which was due to a negative dependence of individual size and density, a common phenomenon in rotifer and other zooplankton communities (Stemberger & Gilbert, 1985; Pace, 1986; Cyr et al., 1997).

Another explanation for the increase of rotifers in turbulent conditions could be the decreased abundance of the predatory cladoceran Polyphemus in TURB ponds. Polyphemus feeds for instance on Keratella, Polyarthra, and Conochilus (Packard, 2001) and a reduction in its abundance could be reflected in rotifers. This explanation can, however, be rejected, because Polyphemus was also depressed in TURB-CTRL, which showed no difference in rotifer abundance compared with CALM-CTRL with high Polyphemus biomass. The decrease of Polyphemus in turbulent conditions could be expected, since it is vulnerable to environmental disturbances and usually exists only in very sheltered habitats (Butorina, 1986). Cyclopoid copepods can regulate rotifers (Lapesa et al., 2002; Sluss et al., 2008), but no between-treatment differences in cyclopoid abundance were observed. Turbulence can enhance the prey encounter rate of cyclopoids copepods (Kiørboe & Saiz, 1995), but the minor differences between TURB-CTRL and CALM-CTRL confirmed that copepods did not regulate rotifers even in the turbulent treatments.

A factor that could also potentially affect the abundance of rotifers in the experiment was the impact of turbulence on rotifers attached on substrata. Most of the dominant rotifer taxa in the experiment can be periphytic (Duggan, 2001). Compared with the calm ponds, water movements in the turbulent ponds probably snatched rotifers from the periphytic mode on bottom substratum to the limnetic mode (Horner & Welch, 1981; Schmid-Araya, 1993). This was not, however, a major factor behind between-treatment differences in rotifer abundance, because no differences between TURB-CTRL and CALM-CTRL were observed. The impact of turbulence on periphytic rotifers probably affected the results in the first sampling day, when all dominant taxa were somewhat more abundant in the turbulent than in the calm ponds. The effect of turbulence on the attachment of periphyton to substrata was not an experimental artifact, but the phenomenon is also present in natural shallow aquatic systems (Weisner et al., 1997).

The experiment suggested that the previously very variable results on the effects of turbulence on rotifers may have depended on the differences in the predation regimes between the studied systems. For instance, (G.-Tóth et al., 2011) reported negative effects of increased turbulence on Keratella, while Sluss et al. (2008) suggested that Keratella could be favored by turbulence. The former study was conducted in field conditions including invertebrate predators and the latter with an experimental setup lacking large invertebrate predators. The present study suggested that the effects of turbulence on rotifers can depend on the abundance of their invertebrate predators. This study was conducted in humic water, but invertebrate predators are important predators also in other types of waters. Chaoborus flavicans can for instance form high densities also in many eutrophic, clay-turbid, as well as in acidified waters, in both fishless and fish-rich conditions (Yan et al., 1991; Xie et al., 1998; Dawidowicz et al., 2002; Liljendahl-Nurminen et al., 2003). Due to their ability to burrow into the sediment, chaoborids may prevail both in shallow and in deep stratifying lakes (Parma, 1971; Liljendahl-Nurminen et al., 2002). Bosminids are an important group in the herbivorous zooplankton of many types of lakes (Lieder, 1983; Beaver et al., 2018). Thus, the present results on the indirect effect of turbulence on rotifers via relaxed chaoborid predation are probably applicable to a large number of water bodies. The study also has implications for the future. Because climate models have predicted increasing wind velocities and wind is a major force inducing turbulence in lake ecosystems (Imboden & Wüest, 1995; Giorgi et al., 2004) climate change can lead to an increasing abundance of rotifers in lakes with abundant populations of invertebrate predators.