Recent observations from the Cassini spacecraft show that Saturn’s moon, Enceladus, may have an alkaline ocean beneath its icy crust(Zolotov, 2007). Chemical models suggest that Enceladus’ ocean is an alkaline Na-Cl-CO3 solution with pH in the range of 8.5 to 12, total mineral content between 10 and 40 g/L and temperature ranging from 0 to 90°C(Glein et al., 2015; Hsu et al., 2015). Geochemicaland microbiological studies of alkaline lakes on Earth may provide useful insights into the possible geochemistry and biology of Enceladus’ ocean. The majority of studied soda lakes are located in areas with arid and semi-arid, warm and temperate climates like the East African Rift Valley(Bogoria, Magadi, Natron), Egypt(Wadi Natrun), Europe(Hungary) and in North America(Oren et al., 2009; Boros et al., 2014). On the other h and Antarctic lakes, especially subglacial ones, are physically analogous to Enceladus’ ocean, but geochemically are less relevant because of their near neutral pH(~6–8; Christner et al., 2014).

A unique feature of the Mongolian plateau is its wide annual range of air temperatures, spanning from -30 to 30°C. High summer temperatures and its arid climate lead to formation of numerous alkaline soda lakes that are covered by ice during 6–7 months per year.

The Mongolian plateau(Fig. 1)includes the entire country of Mongolia, a portion of the Russian Siberian region, and part of northern China. Intensive research on these lakes started in the late 1990s after fi nding of thick(up to 2 cm)microbial mats on the bottom of lakes. A microbial mat is a multi-layered sheet of microorganisms usually dominated by cyanobacteria and anoxygenic phototrophic bacteria. Microbial mats are considered to be the predecessors of the oldest Archean stromatolites(3.5 billion years ago) and earliest forms of life on Earth for which there is good fossil evidence. During research on the lakes of the Mongolian plateau Academician Georgy Zavarzin suggested an idea that modern soda lakes could be a modern analogue of Precambrian epicontinental seas that played an important role in the origin of terrestrial biodiversity(Zavarzin et al., 1999). This research resulted in reports on the biodiversity and microbialactivity of the soda and saline lakes of the Trans- Baikal region of Russia, Inner Mongolia of China, as wellas various regions of Mongolia(Ma et al., 2004; Gorlenko et al., 2010; Namsaraev et al., 2010; Zaitseva et al., 2014). Presently, a significant body of quantitative data on the activity of microbial communities in alkaline lakes of Mongolian plateau was accumulated. These data have been obtained using radioisotopic methods that allow estimation of the rates of microbial processes in situ. Measurement of light-dependent and dark assimilation of CO ₂ allow estimates to be made of the rates of photosynthesis and non-photosynthetic CO ₂ fixation. Microbial processes of sulfate reduction and methanogenesis are the major processes of anaerobic terminal destruction, and their measurement allows the activity of anaerobic destructors to be estimated. The rates of microbial processes obtained using radioisotopic methods can be compared with each other(Sorokin, 1999), allowing wide scale quantitative comparison of the activities of biogeochemical cycles across different communities.

Fig. 1 Locations of major groups of studied lakes in northeastern part of the Mongolian plateau 1: Onon-Borzya group of lakes, Aginskii Buryatskii District (Khilganta, Gorbunka); 2: Onon-Borzya group of lakes, Onon district of Zabaykalskiy region (Dabasa Nur, Tsagan Nur, Babye, Barun Torey, Ilim Torom); 3: Selenga group of lakes, Buryat Republic (Verkhneye Beloye, Solenoye); 4: Onon-Kerulen group of lakes. Dornod aymag, Mongolia (Khotontyn Nuur, Gurvany Nuur, Dzun Davst Nuur, Barun Davst Nuur, Shara Burdiin Nuur, Tsaidam Nuur).

Figure options

The data obtained during earlier research on lakes of the Mongolian plateau show that the rates of microbial processes in different lakes varies over several orders of magnitude, but the environmental factors that explain this variability were not determined. Studies performed on single lakes and groups of lakes showed that salinity, pH and temperature were important(Sorokin et al., 2004; Medová et al., 2011; Zaitseva et al., 2012). The role of other parameters, including water chemical composition, type of sediments, etc. was not evaluated. Traditional statisticalapproaches are usually not suitable for processing of data combining both numericaland non-numerical(qualitative)variables and for analysis of nonlinear relationships between variables, even though nonlinear relationships prevail in natural ecosystems. These limitations of traditional statistics are not amenable with currently widespread methods of multidimensional data processing, including Principal Component Analysis(PCA)(Geladi, 1989). To reveal the major environmental factors that influence the activities of the microbial processes in alkaline lakes of the Mongolian plateau, we determined environmental parameters and the rates of microbial processes in 14 lakes located in different regions of Russia and Mongolia. This paper presents data collected since 1995 using a stand ard set of sampling and processing techniques and were further analyzed using a Principal Component Analysis.

The research was conducted over the period of 1995–2013 in the Buryat Republic and Transbaikalian region of Russia and in the North East region of Mongolia(Dornod aimag)(Fig. 1). Samples of sediments, microbial mats and overlying water were collected from the littoralareas of the lakes from depths not more than 0.5 m and were transferred into either sterile glass flasks(sediments and mats)or plastic bottles(water samples). Radioisotope assays were conducted immediately upon collection. All collected samples were stored at ambient temperature for one or two days until their analysis in the laboratory. Taking the highly variable nature of studied lakes with high fluctuations of chemical parameters into account, only samples for which both data on the chemical composition of the water and data on microbial processes were available were analyzed in this paper.

Water pH and salinity were measured in the fi eld upon collecting the samples by a portable pH-meter and a TDS-4(Trans Instruments, Singapore)conductivity meter. The alkalinity was determined at the same time through two-step titration of 1 to 10 mL of water with 0.1 mol/L HCl: 1)down to pH 8.0 in the presence of the indicator phenolphthalein(carbonate alkalinity) and 2)further down to pH 4.0 with methyl orange(bicarbonate alkalinity). Chloride content was determined by Moor’s argentometric method; total sulfate was determined by a gravimetric method with BaSO ₄ after acidification of the samples to pH 2 with 1 mol/L HCl(Arinushkina, 1970). Concentrations of Na ⁺, Ca ²⁺, and Mg ²⁺ were measured in a SOLAAR M6(USA)atomic absorbance spectrometer. Temperature variations in the surface layers of microbial mats, and soil were recorded over an 18 month period from 9 April 2012 to 29 September 2013 using Thermochron iButton DS1921G temperature loggers(Maxim Integrated, Inc.). The loggers were installed below 1 cm layer of microbial mats of Lake Khilganta and in soil near the lake in order to avoid direct heating by sunlight. Air temperature data for the Aginskoye meteostation were obtained from the All-Russia Research Institute of Hydrometeorological Information - World Data Centre(RIHMI-WDC). Mean monthly temperatures were calculated using MatLab 11(The MathWorks, Inc.).

Aliquots(0.1 mL)of [ ¹⁴ C]-NaHCO ₃(10 μCi; Isotop, Russia)were used to determine the rates of total photosynthesis, dark CO ₂ assimilation and methanogenesis; [ ³⁵ S]-Na ₂ SO ₄(15 μCi; Isotop, Russia)was used for the determination of the rate of sulfate reduction. Samples of the microbial mats and the uppermost layers of sediments were placed immediately after the collection in 15 mL flasks filled with lake water, which were then closed with gasproof caps. Care was taken to remove bubbles trapped within the flask before closure. The isotope solutions were injected through the caps with a tuberculine syringe. Flasks for the measurements of dark CO ₂ assimilation were wrapped with foil before the injection of radioisotope. Flasks were incubated on the surface of microbial mats or lake sediments at the depths from which the samples were collected. At the end of the incubation(2–20 hours)flasks were removed from the lake and activities of microbial processes were stopped with formalin with the final concentration of 3%–4%. The radioactivity of fixed samples was determined by means of a liquid scintillation counter(RackBeta, LKB, Sweden)using conventional techniques(Gorlenko et al., 1999, 2010; Namsaraev et al., 1999). The intensity, I, of microbial processes was calculated using the equation I= r × C /(R × t), where r is the radioactivity of the generated product, С is the natural concentration of the substrate in the sample, R is the radioactivity of the injected substrate, and t is the exposure time in days. Carbon consumption through methanogenesis was calculated using molar volume value 22.4 L/mol. Carbon consumption through sulfate reduction was calculated using balance reaction 2C _org +SO ₄ ^2ˉ →S ^2- +2CO ₂, where 24 g of organic carbon is equivalent to 32 g of sulfi de.

Principal Component Analysis(PCA)was performed using MathLab11 software(The MathWorks, Inc.). In total 11 environmental parameters(lake area, pH, salinity, the content or major ions(HCO ₃ ^ˉ, CO ₃ ^2ˉ, SO ₄ ^2ˉ, Cl ^–, Na ⁺, Mg ²⁺, Ca ²⁺), the type of sample(microbial mat or sediment) and three types of microbial processes(total photosynthesis, dark CO ₂ accumulation and sulfatereduction rates)were analyzed. The rates of methanogenesis were not included in the PCA because of the high number of missing data. The ALS algorithm estimating missing values in PCA data was used in case of missing chemical composition of lakes Dzun Davst Nuur and Barun Davst Nuur(Ilin and Raiko, 2010). Ternary diagrams of major cations and anions of all lakes except Dzun Davst Nuur and Barun Davst Nuur were computed using ProSim Ternary Diagram software(ProSim SA). Data on chemical composition of reference lakes(Big Soda Lake, Mono Lake, Sonachi, Elmenteita, Nakuru, Bogoria, Magadi, Natron) and seawater were taken from published papers(MacIntyre and Melack, 1982; Kharaka et al., 1984; Counciland Bennett, 1993; Grant, 2004). Cluster analysis was performed using MathLab11(The MathWorks, Inc.).

Most of the investigated lakes were small(0.2– 5.2 km ²)except Gurvany Nuur-2(30 km ²), Shara Burdiin Nuur(90 km ²) and Barun Torey(540 km ²). During the study period, the lakes had pH values between 8.1 to 10.4 and salinity between 1.8 and 360 g/L(Table 1). The majority of the lakes displayed salinity up to 30 g/L and pH 9.1–10.2(Barun Torei, Ilim Torom, Verkhnee Beloe, Gorbunka, Solenoye, Tsaidam Nuur). Lake Gurvany Nuur-2 was placed in the second group with salinity between 30 and 50 g/L. Babye, Shara Burdiin Nuur, Dzun Davst Nuur, Barun Davst Nuur, Khotontyn Nuur showed salinity between 85 and 360 g/L. Two lakes could not be placed in one of the above-mentioned groups due to variability of their salinity level. The salinity of Lake Khilganta changed from 30 to 150 g/L and the salinity of Lake Dabasa Nur changed from 17.4 to 200 g/L depending on the particular year in which the measurements were taken(Table 1). Due to their shallowness also, the size of lakes changed considerably. The area of lake Dabasa Nur changed from 1.4 to 0.8 km ² and that of lake Khilganta from 0.3 to 0.2 km ² .

Table1 Environmental parameters of the studied lakes

Table options

Based on their chemical composition, the lakes were grouped into two large clusters(Table 2, Fig. 2). Khilganta, Dabasa Nur, Gorbunka and Babye were included in the fi rst cluster which combined sodium chloride and sodium chloride-sulfate lakes with a low content of carbonates and bicarbonates(＜30% of the anions). Other lakes grouped in the second cluster that included lakes with a high content of carbonates and bicarbonates(>30% of anions). These lakes belonged to sodium carbonate, sodium chloridecarbonate and sodium sulfate-carbonate types of lakes. As expected the cationic composition of all lakes was dominated by sodium. Calcium and magnesium ions comprised only up to 2% and 14% of the total cations, respectively.

Table2 Chemical composition of water in the studied lakes

Table options

Fig. 2 Ternary diagrams of the major anions (a) and cations (b) in the studied and reference lakes 1: Gurvany Nuur; 2: Barun Torey; 3: Tsaidam Nuur; 4: Verkhneye Beloye; 5: Ilim Torom; 6, 7: Solenoye; 8: Khotontyn Nuur; 9: Shara Burdiin Nuur.

Figure options

Temperature loggers showed that the temperature in the microbial mats of Lake Khilganta varied from 31°C to -27°C. The mean monthly temperature in the microbial mats varied from ²1°C to -23°C. Positive temperatures in microbial mats were observed during a six-month period from mid-April to mid-October. Visual observation of the lake during winter period showed that the lakebed was dry. Ice and snow were found only in small depressions on the lakebed. The mean monthly temperature in the microbial mats during January and February was about 10°C lower than in the soil near the lake(Fig. 3). This could be explained by deposition of snow between grass and thermal isolation by snow. The lakebed was mostly free of snow, probably due to wind action.

Fig. 3 Mean monthly temperature of air (Aginskoye meteostation), the microbial mats of Lake Khilganta,and the surface layer of soil near the lake

Figure options

In total 22 samples were collected from 14 lakes. 10 samples were collected from black and blackgreen sediments with a high content of organic matter. 12 samples were collected from microbial mats. Several lakes contained very thin(up to 2-mm thick) and ephemeral microbial mats as wellas biofi lms of purple bacteria on the surface of reduced silts(Dabasa Nur, Babye, Solenoye, Verkhnee Beloe, Barun Torei, Tsaidam Nuur, Gurvany Nuur-2). Only in Lake Khilganta did we observe development of a 20-mm thick cyanobacterial mat that covered the entire bottom of the lake. Thin ephemeral microbial mats were dominated mostly by cyanobacteria belonging to the genus Geitlerinema, the thick microbial mat in Lake Khilganta was dominated by Coleofasciculus chthonoplastes(Tsyrenova et al., 2011).

Photosynthesis was the dominant process of carbon fixation. The rates of photosynthesis in microbial mats varied between 2 and 386 mg C/(L∙d)with the highest rate registered in the microbial mats of Lake Khilganta(salinity 46 g/L). The rates of total photosynthesis in sediments were comparable to those in microbial mats and varied between 0.1 and 304.9 mg C/(L∙d)with the highest rate recorded in Verkhneye Beloye Lake(salinity 9.4 g/L)(Table 3, Fig. 4). The rates of dark CO ₂ assimilation in mats varied between 0.1 and 59.8 mg C/(L∙d)with the highest rate recorded in Lake Khilganta at 150 g/L salinity. Maximal values of the dark CO ₂ assimilation in sediments were lower than in microbial mats. The values varied between 0.02 and 8.6 mg C/(L∙d)with the highest rate recorded in Lake Gorbunka at 8.3 g/L salinity. Comparison of the lakes where both samples of microbial mats and sediments were available showed that except for one sample from lake Verkhneye Beloye, the rate of total photosynthesis in microbial mats was higher than in sediments(lakes Solenoye(salinity 11.4 g/L), Tsaidam Nuur and Gurvany Nuur-2). In contrast to photosynthesis the rates of dark CO ₂ assimilation in the same lakes were higher in sediments than in microbial mats.

Table3 Activities of microbial processes in the studied lakes

Table options

Fig. 4 Activities of microbial processes in relation to salinity (a, b), temperature (c, d) and pH (e, f). Primary production processes (a, c, e), sulfate reduction (b,d, f) ●: total photosynthesis; ○: dark CO 2 assimilation; ♦: sulfate reduction.

Figure options

Sulfate-reducing bacteria play a crucial role in the final stages of organic matter decomposition in the soda lakes studied. The rates of sulfate-reduction in sediments varied between 0.4 and 60 mg S/(L∙d)with the highest rate registered in Verkhneye Beloye Lake(salinity 9.4 g/L). The rates of sulfate-reduction in the mats varied in the same range between 0.5 and 60 mg S/(L∙d)with the highest rate registered in Verkhneye Beloye and Dabasa Nur(salinity 9.4 and 200 g/L respectively). The highest carbon consumption through sulfate reduction was 45 mg C/(L∙d). In lakes Verkhneye Beloye, Solenoye(salinity 11.4 g/L), Tsaidam Nuur and Gurvany Nuur-2, where both samples of microbial mats and sediments were available, the rate of sulfate reduction was the same in mats and sediments(Table 3, Fig. 4). This observation shows that the type of sample(mat or sediment)is probably not very important for activity of sulfate reduction. The intensity of methanogenesis in mats was low, between 0.9 and 46.7 μL СН ₄ /(L∙d). In sediments the intensity of methanogenesis was significantly higher, between 0.3 and 75.6 μL СН ₄ /(L∙d). The highest rate of carbon consumption reached 302 μg С/(L∙d). The rate of organic matter consumption during methanogenesis was 3–4 orders of magnitude lower than that during sulfate reduction.

Extremely high salinities(>250 g/L), pH>10 and temperatures ＜20°C inhibited activities of microbial processes in the studied lakes, but statisticalanalysis did not show significant linear correlation between environmental parameters and the microbialactivities(correlation coefficients below 0.50). Correlation coefficients for total photosynthesis were in the range from -0.02 to 0.36, for dark CO ₂ assimilation from 0.16 to 0.40 and for sulfate reduction from -0.11 to -0.49. Only in two cases did correlation values reached a medium level. Firstly, dark CO ₂ fixation in mats correlated positively with salinity(R =0.60, P ＜0.001). This could be explained by degradation of cyanobacterial mats at high salinity during rapid fluctuations of water level in the studied lakes(Namsaraev et al., 2010). Secondly, total photosynthesis in mats correlated negatively with temperature(R =-0.54, P ＜0.001). This observation shows that salinity and temperature are important factors for activity of phototrophic bacteria in the lakes of the Mongolian plateau.

PCA of the data collected on alkaline lakes of Mongolian plateau showed that temperature determined the distribution of samples on the plot. Temperature contributed the most(-0.98)to the fi rst principal component, PC1, that explained 25.4% of the variations(Fig. 5). Salinity, dark CO ₂ fixation and major ions except carbonates were the main contributors to PC2 that determined 20.5% of the variations. PC3 explained 18.1% of variations and was determined mostly by carbonates concentrations. The type of sample, salinity and pH were less important in this component. PC4 explained 15.8% and was determined by sulfate-reduction, pH and total photosynthesis.

Our results demonstrated a relatively high productivity of the microbial mats and sediments of the alkaline lakes of Mongolian plateau. The highest rate of production(386 mg C/(L∙d)or 3.86 g C/(m ² ∙d))was recorded in a cyanobacterial mat of the Lake Khilganta. However, these values were 2–3 times lower than those obtained for the benthic cyanobacterial community of Solar Lake, Egypt, which was as high as 5–12 g C/(m ² ∙d), and for planktonic communities of African soda lakes(up to 11 g C/(m ² ∙d))(Melack and Kilham, 1974; Krumbein et al., 1977). These differences could be explained by an active process of photosynthesis in studied lakes over a relatively short, warm period that lasts only 3–4 months per year. Alternatively or in addition, a difference in cyanobacterial species composition between tropicaland Mongolian lakes may explain the difference in the rates of oxygenic photosynthesis. For example, Arthrospira plathensis(formerly known as Spirulina plathensis)is abundant in tropical soda lakes, but it was not found in Siberian and Mongolian soda lakes(Sorokin et al., 2004). The highest rates of photosynthesis were recorded at a salinity of 46 g/L. We believe that this salinity is the optimum for the development of the filamentous mat-building cyanobacteria in studied lakes.

Sulfate reduction prevailed over methanogenesis in the alkaline lakes of Mongolian plateau. The rate of organic matter consumption during sulfate reduction was 3–4 orders of magnitude higher than that of methanogenesis. Sulfate reduction values determined in studied lakes were similar to the activity of sulfate reducers in the marine shallow water sediments of the Black Sea(0.01–96 mg S/(L∙d)), saline lagoons of Sivash in the Crimea(2.4–37.6 mg S/(L∙d)) and surface sediments of the saline meromictic Lake Shira(Khakasia, Russia)(0.9 mg S/(L∙d))(Belyaev et al., 1981; Trutko et al., 2003). The activities of methane formation also were comparable with activities in marine sediments(up to 16.24 μL СН ₄ /(L∙d)) and in the bottom sediments of Lake Shira(0.015 μL СН ₄ /(L∙d))(Namsaraev, 1993; Namsaraev et al., 1995; Trutko et al., 2003). Nevertheless, they were considerably lower than those determined in freshwater lakes, where the average rates of methanogenesis reached 2.6–21 mL СН ₄ /(L∙d) and the values as high as 806 mL СН ₄ /(L∙d)were recorded in moderately eutrophic Lake Mendota(Wisconsin, USA)(Zeikus and Winfrey, 1976; Capone and Kiene, 1988).

The data presented in this paper show that the harsh continental climate of the Mongolian plateau influences the microbial processes in studied lakes. Our measurements of temperature in microbial mats of lakes showed that they experience annual temperature fluctuations in the range of about 60°C(from 31°C to -27°C). The short summer season and negative winter temperatures may explain the lower productivity of lakes of the Mongolian plateau compared to lakes of similar geochemical type, but located in regions with warmer climates. Activities of terminal destruction processes were probably less affected by regional climate. Their values and the relative roles of sulfate-reduction and methanogenesis are typical for marine and salt lake sediments with high concentrations of sulfate.

The PCA results confi rm that temperature was the most important factor for microbial processes in lakes. Salinity and pH were also important factors, but their overall importance was lower than temperature and they correlated mostly with single processes. Dark CO ₂ fixation correlated with salinity and chemical composition of lake water. Total photosynthesis and sulfate-reduction correlated with pH. These results show that the influence of temperature should be considered every time during analysis of microbial processes on alkaline lakes of the Mongolian plateau.

Fig. 5 Principal Component Analysis of the environmental factors and microbialactivity in microbial mats and sediments of investigated soda lakes ▲: microbial mats; ●: sediments; TPh: total photosynthesis; DAc: dark CO 2 assimilation; SRed: sulfate reduction; a. PC1-PC2; b. PC3-PC4.

Figure options

According to Sorokin and coauthors a study of biogeochemical cycles in soda lakes is still open(Sorokin et al., 2004). We need to use a systems biology approach combining state-of-the-art molecular techniques, cultivation, in situ analysis of the rates of microbial processes and mathematical modeling to predict the response of microbial communities to changing environmental conditions. From the point of view of future research of extraterrestrial water bodies like alkaline Enceladus’ ocean it would be interesting to see what kind of processes occur in the ice, subglacial water and sediments of alkaline lakes during winter. Especially interesting would be the winter research in the areas where permafrost, mud volcanoes and alkaline soda lakes occur in one place. One of the examples of such a place is a meromictic lake Doroninskoe located in Transbaikalian region of Russia(northern part of Mongolian plateau). Earlier we studied this lake during summer stratification, but detailed study of the lake during winter has not yet been performed(Gorlenko et al., 2010).