Trophic state (TSI_SD) and mixing type significantly influence pelagic zooplankton biodiversity in temperate lakes (NW Poland)

Introduction

The mixing regime and thermal structure of lakes have a profound effect on ecosystem functioning because they strongly affect the availability of nutrients, light, and oxygen (Hutchinson, 1966; Wilhelm & Adrian, 2008; Gauthier, Prairie & Beisner, 2014). Polymictic lakes are shallow and mix to the bottom intermittently during the thawing period (Hutchinson, 1966; Shatwell, Adrian & Kirillin, 2016). In shallow lakes, regular mixing guarantees a stable light-dark cycle as well as higher heterotrophic activity, which stimulates primary production (Nixdorf & Deneke, 1997; Eleveld, 2012). Dimictic lakes are typically deep, mix only in spring and autumn, and stratify continuously over the warmer months (Kirillin & Shatwell, 2016). As was shown using a modeling approach, in dimictic lakes, a deep thermocline could decrease photoautotrophic production because the phytoplankton spends more time in darker waters, thus altering the productivity of aquatic organisms (Berger, Diehl & Kunz, 2006). In contrast, a greater thermocline depth could potentially boost the release of nutrients stored in the sediment, and in turn increase primary production (Scheffer & Van Nes, 2007). The diversity of mixing type regimes has implications for the biodiversity and composition of phytoplankton, macrophytes, macrobenthos, and ichthyofauna (Timms, 1982; Nixdorf, Mischke & Rücker, 2003; Grzybowski, 2014; Napiórkowska-Krzebietke & Hutorowicz, 2013; Czerniawski et al., 2015). Therefore, summer thermal stratification is also believed to be one of the key physical factors structuring the zooplankton communities of also for northern temperate lakes which was demonstrated in a whole lake experiment (Gauthier, Prairie & Beisner, 2014). Moreover, some authors suggest that changes in plankton communities may lead to mixing regime shifts in temperate lakes, which in turn can reveal the complexity between mixing regimes and aquatic organisms (Hambright, 1994; Shatwell, Adrian & Kirillin, 2016).

In both shallow and deep lakes, there are various mechanisms that can affect zooplankton biodiversity. Shallow lakes may be transparent and covered with macrophytes or turbid and dominated by phytoplankton (Scheffer & Van Nes, 2007). Shallow polymictic lakes usually present more diversity because of the array of microhabitats and refuge against predators (Manatunge, Asaeda & Priyadarshana, 2000; Meerhoff et al., 2007), as well as resources (Rautio & Vincent, 2006) offered by macrophytes. Phytoplankton is not the only source of food for the zooplankton. The proximity of the lake bottom (benthic resources) and mixing events increase the availability of dissolved organic matter, allochthonous materials, and bacteria for microinvertebrates (Rautio & Vincent, 2006; Mariash et al., 2014). Thus, there is expected to be a great diversity of Rotifera and Crustacea in polymictic lakes (Kuczyńska-Kippen, 2007; Magalhães Braghin, Simões & Bonecker, 2016). However, it would seem that polymictic lakes are unsuitable for certain zooplankton taxa if the lake conditions are not suitable for individual species (e.g., stenotherm species such as Cyclops abyssorum, Daphnia hyalina, Eurytemora lacustris, Filinia terminalis, Heterocope appendiculata, Notholca squamula or pH-sensitive species such as Ascomorpha ovalis (Radwan, 2004; Błędzki & Rybak, 2016). In dimictic lakes, a deep thermocline could decrease photosynthetic activities, altering the productivity of various zooplankton taxa or of the entire community (Berger, Diehl & Kunz, 2006; Gauthier, Prairie & Beisner, 2014). Less resource availability for various zooplankton taxa could lead to less diversity in the epilimnion, nonetheless many taxa could inhabit the deeper layers of lakes. However, in dimictic lakes with distorted stratification, the loss of a hypolimnetic refuge for large crustaceans is expected to occur (Maier et al., 2011; Gauthier, Prairie & Beisner, 2014).

Water transparency also has an impact on lake processes and the behavior of freshwater organisms (Fee et al., 1996; Berger, Diehl & Kunz, 2006). Transparency is considered a trophic state proxy (TSI_SD) in water management and especially for ecological assessments (Carlson, 1977; Egan et al., 2009; Pyhälä, Fleming-Lehtinen & Laamanen, 2014; Stock, 2015; Binding et al., 2015; Heddam, 2016; Alikas & Kratzer, 2017). However, transparency is more than just an indicator. Transparent lakes are also an important supplier of viable water. Thus, transparency could affect zooplankton biodiversity. The variables that could cause a shift in zooplankton diversity (e.g., decreasing richness of sensitive species) in the deep layers of dimictic lakes, are anoxic conditions below the thermocline (Maier et al., 2011) or the presence of a large community of planktivorous fish (Gliwicz, 1986). These factors escalate the eutrophication process and decrease water transparency (Jeppesen et al., 2011). Rapid eutrophication often negatively affects the ecological balance of aquatic ecosystems and may lead to a decline in biodiversity (Dunne, Williams & Martinez, 2002; Dudgeon et al., 2006; Cardinale et al., 2012; Scherer & Pfister, 2016). Rotifer and crustacean communities are also affected by eutrophication, and therefore serve as indicators of changing ecological status (Karabin, 1985; Jeppesen et al., 2011; Ejsmont-Karabin, 2012; Ejsmont-Karabin & Karabin, 2013; Ochocka & Pasztaleniec, 2016). According to Ejsmont-Karabin (2012), relationships between zooplankton and trophic state are different between polymictic and dimictic lakes, and in both lake types rotifers are more responsive to trophic changes than crustaceans (Karabin, 1985; Ejsmont-Karabin & Karabin, 2013). When examining lake zooplankton biodiversity, authors often find a unimodal peak in species richness at intermediate primary productivity levels (Dodson, Arnott & Cottingham, 2000; Waide et al., 1999; Barnett & Beisner, 2007). However, Jeppesen et al. (2000) showed that zooplankton species richness declines along the eutrophication gradient of Danish lakes. That is, due to lake eutrophication, it is difficult to find oligotrophic lakes; therefore, it is difficult to observe a normal distribution of zooplankton biodiversity along a trophic gradient.

The zooplankton response to the depth and mixing regime is complex due to the interplay of environmental processes (such as eutrophication) and certain traits of micro-invertebrates. Thus, the present study focused on how depth and mixing regime affect zooplankton diversity in lakes. The authors also investigated the vertical distribution of diversity across a transparency gradient (used as a trophy proxy) among the lakes. We hypothesized that the epilimnion of dimictic lakes is less diverse than that of polymictic lakes. Furthermore, we hypothesized that when the transparency decreases, the biodiversity of zooplankton decreases and that the zooplankton composition would be more similar among lakes of the same mixing type.

Methods

A set of zooplankton samples was collected once for each lake during the day in the summer, from 15 July 2011 to 15 August A total of 329 zooplankton samples were collected from 79 lakes in northwestern Poland (Fig. 1). Lakes were chosen according to depth gradient (Choiński, 2006). All sampled lakes are affected by commercial fishing and none of the lakes are temporary. Completely stratified lakes and lakes that were deep enough to partially stratify (lakes with a thermocline) were classified dimictic (deep lakes). Lakes without a fully developed thermocline were classified polymictic (shallow lakes). According to those criteria, we classified 36 polymictic and 43 dimictic lakes (Table 1). Sampling stations were set up at the deepest point in the lake (Jańczak, 1996). If the lake was morphologically diverse, several sampling stations were established. Individual thermal layers were determined for each sampling station in each lake, based on temperature measurements at 1-m intervals. In the case of shallow lakes, the sampling site was located at the furthest distance from plants and the littoral zone. In thermally stratified lakes, samples were taken from three zones: the epilimnion, the metalimnion, and the hypolimnion, whereas in shallow lakes, the samples were collected only from the epilimnion (Lewis Jr, 1983).

The summer period was chosen based on the methodology used by other authors in zooplankton research (Karabin, 1985; Dodson et al., 2009). Summer stagnation results in the accumulation of factors (e.g., color, chlorophyll and phosphorus concentration) affecting the trophic state of lakes (Karabin, 1985). During the stagnation period, the abiotic and biotic factors fluctuate very little, compared to the spring and autumn periods, when intensive water mixing occurs. Table 2 shows the ranges of temperature, dissolved oxygen, pH and conductance that were measured with a Hydrolab DS5 sensor (USA).

For water samples, a Van Dorn 3-liter sampler was used. 50 liters of water was taken at each sampling site and from each layer and passed through a plankton mesh of 30 µm. The samples from each zone were not combined with other thermal zones. Concentrated samples were poured into a 110-ml tube and fixed in a 4% formalin solution. Zooplankton was analyzed in four subsamples in 2-ml plankton chambers using a Nikon Eclipse 50i microscope and a Zeiss Primo Vert reverse microscope. Zooplankton samples were identified using taxonomic keys (e.g., Einsle, 1996; Hołyńska et al., 2003; Radwan, 2004; Benzie, 2005; Błędzki & Rybak, 2016).

Zooplankton biodiversity was calculated with two commonly used indicators: species richness SR (number of species) and the Shannon index SI (log base e) (Shannon & Weaver, 1949). Species richness is the simplest and fundamental measurement of community. It is also the basis of many ecological models of community structure (May, 1988). Maintaining biodiversity is also a central objective for various monitoring and management projects (Williams & Gaston, 1994; Green et al., 2005; Pereira & Cooper, 2006). Another commonly used biodiversity metric is the Shannon index. Entropy is a reasonable index of diversity, which takes SR and species evenness into consideration (Shannon & Weaver, 1949; Jost, 2006).

An index based on the visibility of the Secchi disc (SD) was used to determine the trophic state of the lakes (TSI_SD) (Carlson, 1977). The TSI_SD was calculated using the 60-14.41 ln(SD) formula, where SD is transparency in meters. The TSI_SD value referred to all thermal zones in a given lake. The SD is affected by various water quality parameters, such as chlorophyll, total phosphorus concentration, and color in addition to light scattering and light absorption (Carlson, 1977; Fee et al., 1996; Brezonik et al., 2015). Despite the numerous variables affecting transparency, the TSI_SD is commonly used as a key eutrophication indicator of different water body types (Karabin, 1985; Pyhälä, Fleming-Lehtinen & Laamanen, 2014; Stock, 2015; Binding et al., 2015; Heddam, 2016; Alikas & Kratzer, 2017). Therefore, the Secchi depth (SD) is an important visual indicator of water clarity that can be used for the water quality index calculation (Carlson, 1977; Alexakis et al., 2016; Heddam, 2016; Ochocka & Pasztaleniec, 2016).

To determine the correlations and significant differences in zooplankton biodiversity, we used the mean value for SR and SI for lakes and thermal layers (Statistica 12 software; StatSoft, Tulsa, OK, USA). To find the best predictors for zooplankton biodiversity a multiple stepwise regression was used (P < 0.05) (Sokal & Rohlf, 1995). The relationship between TSI_SD and depth was verified by the Pearson correlation (P < 0.05). The parametric one-way ANOVA and the post hoc Duncan test were used to determine significant differences in zooplankton biodiversity between mictic lake types and thermal layers (P < 0.05). The rarefaction method and the “EstimateS” software were used to calculate the species number (SR) and the Shannon index (SI) values. These tools were used to test differences among lakes types and thermal layers. In an effort to avoid bias for lakes with larger sample numbers, we followed the rarefaction methods, and used all samples for each layer. Recent examples emphasize the importance of quantifying SR using taxon sampling curves (Allen et al., 1999; Gotelli & Colwell, 2001; Chao et al., 2014). Therefore, we based our calculations on the number of samples. Based on Sørensen similarity analyses (MVSP 3.22 software), we developed four levels of taxonomic similarity (<0.5—low degree of similarity; 0.5–0.65—moderate degree of similarity; 0.8–0.65—high degree of similarity; >0.8—very high degree of similarity) and compared similarities between pairs (i.e., between two lakes) for polymictic and dimictic lakes, and their percentage contribution to the similarity classification. Nonmetric multidimensional scaling (NMDS) (Bray–Curtis distance metric, a square root transformation, and Wisconsin double standardization) describe the similarity patterns in species composition among lakes in terms of mictic type, thermal layer and the TSI_SD value.

Results

The ranges of environmental variables in the 36 polymictic and 43 dimictic lakes are shown in Tables 1 and 2. Lake depth was significantly correlated with the TSI_SD value (R =  − 0.44; p < 0.001). We identified a total of 151 taxa in the sampled lakes (Table 3). The most frequent taxa (with the frequency of occurrence of more than 75% among lakes) were: Keratella cochlearis 100%, Polyarthra vulgaris 94%, Mesocyclops leuckarti 88%, Trichocerca similis 82%, Daphnia cucullata 84%, Diaphanosoma brachyurum 78%, K. cochlearis v. tecta 77%, and Thermocyclops oithonoides 75%.

Multiple regression revealed that the TSI_SD and depth significantly affected the Rotifera SR, Copepoda SR, and Copepoda SI. However, the TSI_SD significantly affected the total zooplankton SR (P < 0.05) (Table 4). The analysis (R value) explained 25%–38% of the variability in the SR of Rotifera, Copepoda and total zooplankton and the SI of Copepoda (P < 0.05). The Duncan test results show that the mictic type had a significant effect on the zooplankton biodiversity (Fig. 2). Rotifera SR was significantly higher in polymictic lakes than in dimictic lakes and significantly affected by thermal zone (p < 0.05). The deepest layers of dimictic lakes presented the lowest levels of Cladocera SR and SI (p < 0.05). The Copepoda SR and SI values in the metalimnion of dimictic lakes were significantly higher than in the epilimnion of both shallow and deep lakes (p < 0.05).

The SR rarefaction curve of zooplankton showed differences between the groups of polymictic vs. dimictic lakes (Fig. 3), and between lakes with a different trophic state (Fig. 4). In general, the SR of zooplankton in the epilimnion showed a higher value in lakes with a low trophic state. The SR of Rotifera and Copepoda was the highest in the mesotrophic lakes, and the lowest in the polytrophic lakes. In contrast, in the above results, the highest Cladocera SR was found in the eutrophic lakes, whereas fewer species were present in the polytrophic lakes. In all cases, polytrophic waters represented the lowest SR of the organisms studied. In the metalimnion, the SR of Rotifera and especially of Cladocera, was the highest in the mesotrophic lakes, and lowest in the eutrophic lakes (Fig. 5). Furthermore, the highest Copepoda SR was found in the mesoeutrophic lakes, whereas fewer species were present in the eutrophic lakes. In all cases, eutrophic waters represented the lowest SR. In the hypolimnion, the Crustacea SR was the highest in the mesotrophic lakes and the lowest in the eutrophic lakes (Fig. 6). The highest Rotifera SR value was recorded in the mesoeutrophic lakes, whereas fewer species were present in the eutrophic lakes. Eutrophic waters in the hypolimnion demonstrated the lowest SR. The SI rarefaction curve reached a higher value for dimictic than polymictic lakes and is highest for the lowest trophic status lakes in all thermal zones. This pattern was most pronounced in the epilimnion (Fig. 7).

In the dimictic lakes studied, a group of taxa (mainly Crustacea) that typically occur in deep lakes was found, some of which are not common in East-Central Europe (e.g., Bythotrephes longimanus, Daphnia hyaline, Daphnia longiremis, Eubosmina longicornis, Heterocope appendiculata), or even globally (the rare glacial relict Eurytemora lacustris) (Table 3). In contrast, a few taxa considered typical for the littoral zone were found only in the polymictic lakes (e.g., Brachionus quadridentatus, Lepadella quadricarinata, Simocephalus serrulatus). We determined that polymictic lakes are more heterogeneous than dimictic lakes (which were homogenous) in terms of zooplankton composition. A high degree of similarity (0.8–0.65) was observed in 21% of polymictic lake pairs and 50% of dimictic lake pairs (Fig. 8). A low degree of similarity (<0.5) was observed in 34% of polymictic lake pairs and only in 5% of dimictic lake pairs.

Ordination by NMDS showed clustering of the different mictic types, thermal layers, and the change in composition throughout the transparency profile (Fig. 9). The zooplankton communities in the metalimnion and the hypolimnion (of dimictic lakes) appeared to be very distinct from those in the epilimnion of shallow lakes. However, in many cases, the community of zooplankton in the epilimnion of dimictic lakes was more similar to those found of shallow lakes, than those found in deeper layers of dimictic lakes. The TSI_SD also had an influence on the zooplankton communities, which is demonstrated by the least-transparent lakes being clustered. Moreover, communities of several deep eutrophic lakes were similar to communities found in lower trophic level lakes.

Discussion

Conclusion

Polymictic lakes are characterized by having a higher average species richness than dimictic lakes, which is attributed to a large number of Rotifera species in shallow waters. Shallow lakes are usually more eutrophic, yet as they lose their thermal niches (metalimnion and hypolimnion), they gain macrophyte niches or gain proximity to the substrate, which together causes a decrease in Copepoda biodiversity and increase in Rotifera biodiversity. However, at a regional scale (using rarefaction), the highest species richness and Shannon biodiversity values are found in Cladocera, Copepoda and Rotifera in non-eutrophic waters. This finding indicates a large number of zooplankton species in the low trophic waters of the studied region. Moreover, polymictic lakes are characterized by higher variability of zooplankton composition than dimictic lakes.

Supplemental Information