1 INTRODUCTION

Lake Shira is a meromictic saline terminal lake (90.11°E, 54.30°N) with a surface area of 35.9 km² and a maximum depth of 24 m (Degermendzhy et al., 2010). The dominant anions in lake water are sulfate and chloride, and cations are sodium and magnesium (Kalacheva et al., 2002). The dissolved salt concentration in the mixolimnion varies with depth between 11.5 and 14.5 g/L, decreasing to 9.5 g/L close to the surface after the ice melt in May. The small river Son empties into the Lake. The trophic status of Lake Shira is defined as mesotrophic (Kopylov et al., 2002). The Lake is covered with ice about 6 months a year from November to late April or beginning of May. From the depth of 13–16 m and downward, there is a steady anaerobic zone, with dissolved hydrogen sulfide, which is detected down to the deepest point of the Lake (Rogozin et al., 2017b). The concentrations of sulfates, sulfides, phosphates, and ammonium in the monimolimnion exceed those in the mixolimnion (Zadereev et al., 2014). The mixolimnion is thermally stratified during summer. The deep-water chlorophyll a maximum and the highest abundance of phytoplankton are located in the metalimnion just below the thermocline (Degermendzhy et al., 2010).

The data of the year-round monitoring conducted since 2002 suggest that the lake had been meromictic until 2015 (Rogozin et al., 2017a). During the winter of 2014–2015, deep mixing occurred and stable stratification was disturbed (Belolipetskii et al., 2017; Rogozin et al., 2017b). In May 2015, no hydrogen sulfide was detected in the water column, and there was no chemocline. In June 2015, hydrogen sulfide was again detected in the water column, but the redox zone did not rise above the depth of 20 m. In August, the redox zone rose to the depth of 17–18 m, but the total amount of hydrogen sulfide was the lowest during the monitoring period. In 2016, similar patterns of the chemocline depths and hydrogen sulfide concentrations were observed. Since early 2017, the redox zone has not descended below 18 m, suggesting that the Lake has returned to meromixis (Rogozin et al., 2017b).

Usually, the breakdown of meromixis releases nutrients from the monimolimnion, which leads to outbreaks of phytoplankton blooms and a change in the species composition of planktonic organisms (Melack and Jellison, 1998). Such an increase in biomass and primary production of phytoplankton after the breackdown of meromixis was observed in Mono Lake (U. S. A.), Lake Is eo (Italy), and Lake Lugano (Switzerland and Italy) when they changed their circulation regime (Simona, 2003; Leoni et al., 2014; Melack et al., 2017). Changes in zooplankton biomass have also been observed. In Lake Shira, during 2015–2016, there was also a change in the biomass of almost all components of the ecosystem, including zooplankton (Rogozin et al., 2017b).

Water temperature is one of the most important factors for the development of zooplankton, and in general, the effect of the temperature on the zooplankton abundance is positive (Heinle, 1969; Kalikhman et al., 1992; Haberman and Haldna, 2017). Although temperatures above 25 ℃ may suppress the zooplankton development (Moore et al., 1996), such temperatures are not typical for the pelagic zone of Lake Shira.

In the total composition of the zooplankton community in Lake Shira, 42 taxonomic units were identified, 7 of which were Cladocera, 12 were Copepoda, and 23 were Rotifera (Anufrieva et al., 2011). The majority of these species were observed in the littoral zone. The dominant zooplankton species in the pelagic part of Lake Shira are the copepod Arctodiaptomus salinus (Daday, 1885) and two rotifer species—Brachionus plicatilis (Müller, 1786) (currently considered as a species complex (Mills et al., 2017)) and Hexarthra sp. (Degermendzhy et al., 2010). These species are found in lakes of different salinity in this region (Zadereev et al., 2022). Copepods are present in the lake in all seasons, and rotifers are the most abundant in the second half of summer, which is characterized by high air and water temperatures in this region.

The aim of this work was to study time series of zooplankton abundance and biomass in Lake Shira over more than twenty years. As a hypothesis to test, we assumed that the change in the circulation regime of the lake (breakdown and reestablishment of the meromixis), as well as the air temperature, should have a significant effect on the abundance and biomass of zooplankton.

2 MATERIAL AND METHOD

Zooplankton samples were collected in the daytime in the central part of Lake Shira (above the maximal depth), in the mixolimnion, with an interval of one meter, during 1996–2021. The frequency of sample collection varied from year to year: between 0 and 7 times a year. Most of the samples were taken during the growing season from late May to mid-September, usually at the end of May, end of June, and end of July to the beginning of August. In 1996–2011 and partially in 2015–2016, 3 or 5 L were taken from each depth. Samples were collected with a discrete point water sampler (Molchanov bathometer) or were pumped using a submersible hose with a funnel at the end (funnel diameter 10 cm). Three liters of water were taken to determine the zooplankton abundance at each depth. The entire volume was filtered through a 75-μm plankton net and concentrated to a volume of 10 mL. The zooplankton were preserved in 70% ethanol and later counted under a light microscope (×32). Rotifers, Brachionus plicatilis and Hexarthra sp., and nauplii, copepodites of young stages C1–C3, copepodites of later stages C4–C5, males and females of Arctodiaptomus salinus were counted. In 2012–2014, 2017–2021, and partially in 2015–2016, zooplankton samples were taken in a volume of 1 L in the central part of the lake with a hose sampler or a Rutner bathometer from different depths with an interval of 1 m from the surface to the lower boundary of the oxygen layer. Zooplankton samples were fixed using Kuzmin's solution (10-g potassium iodide, 50-mL distilled water, 5-g iodine, 5-mL chromic acid (1%), 10-mL acetic acid, 80-mL formalin (1% final concentration)) (Kuzmin, 1975). Samples were fixed immediately after collection. Before counting, the samples were concentrated by sedimentation (Hasle, 1978) from 1 000 mL to about 50 mL. The abundance of zooplankton (inds./L) was determined by counting in the Bogorov chamber all animals from the entire sample volume. The abundance of zooplankton in the water column was defined in terms of numbers below a square meter (inds./m²) and calculated as the total sum of the abundances (with interpolation where necessary) in the vertical sampling multiplied by 10³. The calculation of the zooplankton biomass was carried out according to the equation of the dependence of the body weight of an individual on its length (Tevyashova, 2009) according to the average size of the organism in the population of Lake Shira (which was obtained by the authors at the beginning of the observation period). The average lengths of zooplankters in the lake were as follows: A. salinus (male) (1.15×10^-3 m), A. salinus (female) (1.25×10^-3 m), A. salinus (copepodid C1–C3) (0.62×10^-3 m), A. salinus (copepodid C4–C6) (0.9×10^-3 m), A. salinus (nauplius) (0.32×10^-3 m), B. plicatilis (0.5×10^-3 m), Hexarthra sp. (0.25×10^-3 m).

The relationship between the number of zooplankton groups and the two-week average air temperature preceding the date of sampling was assessed using linear regression. Temperature data for the Shira weather station, Russia, WMO_ID= 29756, were taken from online open sources (https://rp5.ru). Data on the abundance of zooplankton groups were taken for 2012–2020, since for earlier dates, neither complete sets of data on abundance nor data from a meteorological station were available.

To determine significant differences between the vertical distributions of A. salinus at different times of sampling, the paired Wilcoxon test was used (Wilcoxon, 1945). The numbers of A. salinus during the period of meromixis breakdown (2015–2016) were compared with their numbers in years of stable stratification (1996–2014 and 2017–2021) of Lake Shira. Comparisons were carried out both for individual months (May, June, July, August) and on the basis of average numbers for each spring-summer period from May to August. The two-year data for the period of meromixis breakdown were averaged over the corresponding months. In each comparison, a pair of vertical distributions of animals was used. If several verticals were selected in a certain month, then such verticals were averaged over the corresponding depths. Pairs for evaluation by the Wilcoxon criterion were the abundance of A. salinus at the corresponding depths. If there were no data for some depths in at least one vertical, then such a depth was not taken into account. If samples were taken on less than three dates during the season from May to August, such years were excluded from the analysis. All analyses were performed in R (R Core Team, 2021).

3 RESULT

We analyzed more than 70 vertical zooplankton sample datasets in the pelagic part of Lake Shira from 1996 to 2021. The average annual abundance in the water column of the copepod Arctodiaptomus salinus (with separation of the nauplii) and the rotifers Brachionus plicatilis and Hexarthra sp. did not have a unidirectional tendency to change over the entire observation period (Figs. 1–2). Largely, this applies to the copepod A. salinus, whose population numbers fluctuated from 2×10⁵ to 8×10⁵ inds./m². The numbers of rotifers varied in a wider range–the peak abundances of B. plicatilis exceeded 2×106 inds./m². The high abundance of this species was usually associated with the season of high temperatures―the second half of summer. At other times, the numbers of B. plicatilis were much lower or animals of this species were completely absent. For A. salinus, an increase in abundances in 2015–2016 (the period of temporary breakdown of the meromixis) and a decline during the restoration of the meromixis were observed. No similar tendency was observed for rotifers. On the contrary, one of the peaks of the maximum numbers of rotifers was detected in summer 2014, during the period of stable stratification. These maximum numbers were most likely caused by a long period of high temperatures. With the exception of this date, the numbers of rotifers also increased in 2015–2016 and then decreased.

Our hypothesis about the positive effect of temperature on the abundance of zooplankton was confirmed. We observed a significant positive correlation between the two-week average air temperature near Lake Shira, preceding the date of sampling and the abundance of zooplankton (Fig. 3). The coefficient of determination between air temperature and the abundance of rotifer B. plicatilis was higher than for other zooplankton groups. This confirms our assumption that a long period of high temperatures can cause a considerable increase in B. plicatilis abundance. The strong air temperature dependence of the growth of B. plicatilis may be associated with the characteristic of the vertical distribution of this species and its high growth rate. The density of B. plicatilis in natural water bodies can reach 6×10³ inds./L (Shiel and Koste, 1986). During the period of summer temperature stratification of the water column, the air temperature primarily affects the temperature of the epilimnion. B. plicatilis in Lake Shira has a near-surface abundance peak (Zadereev and Tolomeyev, 2007). Other dominant zooplankton species, along with epilimnetic biomass peak, also have deep ones. Prolonged high air temperatures lead to the warming of the epilimnion, which creates favorable conditions for the rapid development of rotifers B. plicatilis, which reproduces in summer by parthenogenesis. For other species, local warming of the epilimnion has a less pronounced effect on the abundance and biomass due to their slower growth rates and smaller contribution of their epilimnetic abundance to the total abundance in the water column.

Analysis of the vertical distribution of the abundance of copepod A. salinus (excluding nauplii) and rotifers B. plicatilis and Hexarthra sp. in the pelagic zone of Lake Shira during two years before the change in circulation regime(2012, 2014) and two subsequent years of complete mixing (2015– 2016) showed that at the beginning of the growing season (Figs. 4–5, left row), the rotifer B. plicatilis was practically absent in the lake, and copepods dominated zooplankton. During full mixing A. salinus tends to accumulate in the upper layers of the water column (above 10 m). This is especially evident in the spring of 2016, when the peak of the copepod abundance is located at the surface layer. For the rotifer Hexarthra sp. there is an increase in abundance during the years of complete mixing in the spring (Fig. 5, left row), but the shape of the vertical distribution remains the same.

In the second half of summer (Figs. 4–5, right row), both rotifer species demonstrate vertical abundance peaks separated in depth under any circulation regime. For B. plicatilis these are depths of 3–5 m and for Hexarthra sp. are 12–15 m. However, Hexarthra sp. is present even in the surface layer. At the end of July 2014, B. plicatilis shows a sharp increase in abundance, which makes it difficult to estimate the effect of a change in the circulation regime on the change in the abundance of this species. The abundance of Hexarthra sp. in late July–early August, as well as in spring, in the years of breakdown of meromixis is significantly higher than in previous years. The copepod A. salinus also increased its abundance during the years of meromixis breakdown at the end of summer (75–130 inds./L at the peak of abundance).

If we consider the abundance of rotifers in the water column in the years before and after the breakdown of meromixis (Fig. 2), then for the spring time (late May) we observe a remarkable increase in Hexarthra sp. (17×10³–2×10⁵ inds./m²) during the period of meromixis breakdown. At the end of summer (August) (compared with 2012–2013 rather than with 2014, which was anomalous for rotifers), we observed a significant increase in the abundance of both Hexarthra sp. (1.8×10⁵–5.8×10⁵ inds./m²) and B. plicatilis (1×10⁵–5.5×10⁵ inds./m²).

During the complete mixing of the lake, an increase in the abundance of A. salinus was not observed in May, and it was weakly expressed in August. In early and mid-July, the numbers of copepods in the water column increased by a factor of more than two (from 3×10⁵ to 8×10⁵ inds./m²).

In May–August, the layer-averaged vertical distribution of A. salinus abundance in the years of meromixis breakdown (2015–2016) and stable stratification (1996–2014 and 2017–2021) was significantly different for 10 years out of 12 available for pairwise comparison (83.3%) (Table 1). When compared in the vertical distributions by months, the smallest number of years that differed between each other was observed in May (2 years out of 9; 22.2%) and the largest in August (9 years out of 12; 81.8%). In addition, for the years immediately preceding the breakdown of meromixis (2010–2014), as well as for the years following this period (2017–2019), all compared pairs of abundances in different months (except May) were statistically different.

Thus, we clearly demonstrated that the breakdown of meromixis led to a significant increase in the abundance of the copepod A. salinus in Lake Shira both during the entire growing season and in individual summer months: June, July, and August. In May 2015–2016, the abundance of A. salinus did not change significantly.

Dynamics of the abundance of copepod A. salinus during the growing season for all years of observation clearly demonstrates the actual difference between years with and without meromixis (Fig. 6). Only at the beginning of the season (end of May to the beginning of June), abundances in the years with and without meromixis are comparable. At the beginning of the growing season, the abundances of copepods are most probably determined by other factors.

4 DISCUSSION

The abundance of zooplankton in Lake Shira in years without meromixis increased approximately two fold. In terms of biomass (wet weight), this increase was from 8–12 g/m² in 2007–2014 to 15–18 g/m² in 2015–2016 for A. salinus; and from 2–5 g/m² in 2012–2014 to 4–9 g/m² in 2015–2016 for rotifers (excluding the extreme outbreak in July 2014), which corresponded to changes in the biomass of other major components of the lake ecosystem (Rogozin et al., 2017b; Khromechek et al., 2021). When assessing the dynamics of zooplankton biomass, as well as other components of the ecosystem, we cannot ignore the expansion of the oxygenated mixolimnion due to the breakdown of meromixis. It could increase the habitat available for zooplankton and consequently the zooplankton biomass in the water column. However, the data show (Figs. 3 & 4) that zooplankton in Lake Shira at depths below 15 m during the period of meromixis breakdown were practically absent. The only exception was the rotifer Hexarthra sp.

Another obvious factor of the increase in zooplankton abundance is the bottom-up effect, which manifests itself as the release of nutrients from the monimolimnion and the subsequent increase in phytoplankton biomass (Jellison and Melack, 1993; Melack et al., 2017). In Lake Shira, an increase in the mass of organic carbon in seston was also observed in 2015 (up to 50 g/m²) compared to previous years (20–25 g/m²) (Rogozin et al., 2017b).

An increase in the abundance of zooplankton with a change of the mixing regime was already recorded in meromictic lakes. For example, the average annual biomass of brine shrimp (Artemia monica) in Mono Lake during the breakdown of meromixis increased from 8 to 17 g/m² (dry weight) (Melack et al., 2017). Daphnia longispina and other cladocerans increased their abundances by a factor of two or three during the complete vertical mixing (2005 and 2006) of Lake Iseo (Leoni et al., 2014). During the deep mixing events in 2005 and 2006 in Lake Lugano the biomass of grazers increased from 5 to 7.5 g/m² (Lepori and Roberts, 2017).

After the reestablishment of meromixis in Lake Shira in 2017, the abundance of dominant zooplankton groups declined to the values of previous years. However, not all groups of organisms in the lake responded to the reestablishment of meromixis in this way. For instance, the biomass of ciliates in subsequent years remained at the level observed in 2015–2016 (Khromechek et al., 2021), which was apparently associated with an increase in the number of ciliates species in lake. Thus, intra-ecosystem interactions can also be an important factor determining the abundance of a species in a water body. There are possible complex interactions between different zooplanktors (e.g., competition, allelopathy). Moreover, different species can have complex vertical distribution. It was demonstrated earlier (Zadereev and Tolomeyev, 2007) (and you may see it here on Fig. 5, right row) that rotifers B. plicatilis and Hexarthra sp. in Lake Shira are partially separated in depth with Hexarthra sp. more abundant in deep waters and B. plicatilis—close to the surface. Here we aggregated data over the entire water column to follow the general changes in abundances. There are possible effects that is more complex but they will require more precise and focused studies.

5 CONCLUSION

The present study demonstrated that all dominant species of zooplankton in the pelagic zone of Lake Shira responded to the elevated air temperatures by an increase in abundance and biomass. At the same time, the greatest air temperature dependence was observed for the rotifer B. plicatilis, with the epilimnetic abundance peak and high reproduction rate. In addition, we observed that during the breakdown of the meromixis in Lake Shira (2015– 2016), the abundance of zooplankton increased on average by a factor of two and amounted to 9×10⁵ inds./m² for copepods. This growth correlated with the increase in other components of the lake's mixolimnion ecosystem, which changed their biomass approximately twofold during this period. After the reestablishment of the meromixis, the abundance of zooplankton decreased to previous values. Thus, the abundance of zooplankton is largely determined by weather and climatic factors, either directly (the effect of air temperature) or indirectly (the effect of weather conditions on the mixing regime).

6 DATA AVAILABILITY STATEMENT

All data generated or analyzed during this study are included in this published article or available from the corresponding author on reasonable request.

7 ACKNOWLEDGMENT

The authors are grateful to F. F. Kozlov, D. Y. Rogozin, V. V. Zykov (Krasnoyarsk Research Center, Institute of Biophysics, Siberian Branch of Russian Academy of Sciences) for their assistance during field studies. We are grateful to professional English translator Elena Krasova for linguistic improvements. The research was supported by the State Assignment of the Ministry of Science and Higher Education of the RF (No. 0287-2021-0019).