A growing interest in the carbon cycle of subarctic and arctic regions is caused primarily by climate change, which is most pronounced in high latitudes, where it leads to increased emissions of greenhouse gases carbon dioxide (CO2) and methane (CH4) from the water surfaces, which potentially enhances the effect of global warming. In this regard, over the past decades, a priority for basic research has led to a complex ecosystem study of the northern lakes. However, study of the status and the change of lake ecosystems in the context of climate change requires consideration of the influence of anthropogenic factors.
Annual trends in the physical, chemical, and biological parameters of boreal lakes became a topic of high interest within the context of climate warming and the response of biogeochemistry of carbon in high-latitude aquatic ecosystems, with increasing ground and water temperatures being among the key issues of the prediction of ecosystem evolution under on-going environmental changes [1
]. This prediction, however, is significantly hampered by the lack of knowledge of aquatic ecosystem response to one of the main consequences of climate warming—the rising of surface water temperatures. One of the most efficient traditional methods of assessing the effect of long-term environmental change is continuous monitoring of whole lake ecosystems, which includes a suite of parameters, from hydrochemical to fish population, routinely conducted in various boreal and temperate regions (see refs. [5
] and [6
] for a recent review).
In Russia, such long-term monitoring is known for large lakes [7
], with almost no observations of small lakes. Yet, small lakes are likely to be the most sensitive indicators of on-going environmental changes, being at the same time highly vulnerable to both local anthropogenic pressure and global climate change. In this regard, small boreal humic lakes of European Russia deserve special attention. In the Arkhangelsk region (NW Russia), there are over 200,000 lakes. Most of them are located to the west of the Onega River, between the middle and upper reaches of the Onega River and the southwestern boundary of the area. This region is dominated by lakes having surface areas of ≤50 km2
, conventionally classified as small lakes with moderate or significant anthropogenic impact. The global trends of water ecosystem evolution can be, however, strongly biased by local effluent discharge, such as from the dairy industry [8
]. Thus, phosphate is widely recognized as a very efficient tracer of urban wastewater efflux to the environment [9
Detailed study of organic carbon and trace metal speciation in the water column of the Svyatoe Lake—a representative humic stratified lake of the Russian boreal zone—allowed the creation of a comprehensive picture of organic and mineral colloids formation, biodegradation, and exchange between surface and deep waters and sediments. In particular, we demonstrated the impact of the anomalously hot summer in 2010 on the increase of low molecular weight (potentially bioavailable) forms of dissolved organic carbon (DOC) and metals, primarily linked to enhanced photo- and bio-degradation of large size allochthonous soil dissolved organic matter (DOM) coupled with production of phytoplankton exometabolites [10
]. In addition, observations of the diel photosynthetic cycle of the shallow zone of this lake during the hot summer’s cyanobacterial bloom revealed the existence of coupled DOC-Fe-Mn speciation cycle in the surface water layer linked to alternative photosynthetic activity and heterotrophic aerobic plankton respiration [11
]. In contrast to rather detailed understanding of lake water hydrochemistry and mechanisms regulating the biogeochemical cycle of organic carbon during recent years, the evolution of the lake ecosystem over past decades remains poorly known.
This work presents an overview of results of multidisciplinary study of the Lake Svyatoe, conducted since 2001 to the present time. We demonstrate the synchronous behavior of macro- and micro-nutrients in the deep and shallow zones of the lake and assess the degree of local anthropogenic pollution versus global climate trends, as well as catastrophic weather events. We conclude on the necessity of complex, multidisciplinary studies of small humic lakes in the boreal zone, because these “model” lakes can serve as important proxies of on-going environmental changes in highly abundant aquatic ecosystems of poorly accessible regions of NW Russia.
2. Study Site Description, Sampling, Methods, and Analysis
Lake Svyatoe is located in the Arkhangelsk region (northern Europe) within the watershed of the Onega River near the Rotkovetz Scientific Monitoring Station (60°51’ N; 39°32’ E, Figure 1
). The watershed’s lithology is represented by glacial moraine (products of granite-gneisses erosion) and carbonate rocks. The lake has an overall surface area of 2.11 km2
, its volume is 0.00749 km3
, the maximum depth is 16 m, and the average depth is 3.6 m. There are several inlets to this lake and one outlet river, the Svyataya. The lake is strongly influenced by Dissolved Organic Matter (DOM) delivered from adjacent bogs which contain significant peat deposits. The lake exhibits a deep zone in its southern (relatively pristine) part, with minimal inlets having anthropogenic influence. Little human activity has been present near this southern part of the lake over the past 50 years, so the studied part of the lake can be considered almost pristine, similar to other lakes in the region [12
]. In contrast, the northern-most bay (Maslozavod) is subjected to significant anthropogenic pollution from both local industry (milk and butter factory) and local inhabitants of approximately 30 houses, since there is no sewage system in the Rotkovetz village.
The mean summer air temperature and precipitation of the region from 2001 to 2015 are illustrated in Figure 2
A. Similar to other boreal and subarctic lakes, the lake exhibits two main periods of pronounced stratification (November to April and June to September) and two periods of lake overturn (October and May). The typical ice thickness at the end of winter, based on our 2 year period of observation, is 50 ± 10 cm. Maximal winter stratification is in March, and the highest water temperature typically occurs in July. The length of the ice-covered period of the lakes is between 195 and 201 days; the average freeze over date is 25–30th October and the average ice out date is 12–14th May [12
]. Mean seasonal water temperatures over the full period of observations on the deepest stratified station of the lake are illustrated in Figure 2
The water samples were taken using a PVC boat in the summer time and from ice cover during the winter time from the two selected stations of the lake (Figure 1
) using a pre-cleaned polycarbonate horizontal water sampler (Aquatic Research Instruments Inc., Hope, ID, USA); they were immediately filtered through sterile, single use Minisart®
filter units (acetate cellulose filter, Sartorius®, Göttingen, Germany) with a pore size of 0.45 µm.
Filtration was performed directly on the boat in summer or on the ice in winter. Dissolved O2 level and temperature were measured in situ using the Oxi 197i oximeter (WTW®, Weilheim, Germany) with a Cellox 325 submersible sensor (WTW®, Weilheim, Germany; ±0.5% and ±0.2 °C uncertainty). The Winkler titration method was also used for O2 determination. The conductivity was measured on-site using a Hanna HI991300 conductivity meter (Hanna® Instruments, Inc., Ann-Arbor, MI, USA).
pH was measured in the field using a combined electrode calibrated against NIST (National Institute of Standards and Technology, Gaithersburg, MD, USA) buffer solutions (pH = 4.00 and 6.86 at 25 °C). The accuracy of pH measurements was ±0.02 pH units. Filtered water samples for nutrient analyses were frozen (−20 °C) within 1 to 3 h after collection and analyzed within 1 week after sampling.
DOC and dissolved inorganic carbon (DIC) concentrations were measured using methods routinely used in the Geosciences and Environment Toulouse (GET) laboratory to analyze boreal water samples [10
]. The nutrient determinations were based on colorimetric assays [13
]. Total dissolved organic nitrogen (DON) was measured from the difference between the total dissolved nitrogen (persulfate oxidation) and the total dissolved inorganic nitrogen (DIN, or the sum of NH4+
, and NO3−
). Uncertainties of DON and DIN analyses were between 10% and 20%, and detection limits were between 10 and 50 µg/L.
Major and trace elements, including Fe, were measured by Inductively Coupled Plasma Mass Spectrometry ICP-MS (7500 ce, Agilent Technologies, Santa Clara, CA, USA) without pre-concentration, an uncertainty of 5%, and a detection limit of 0.02 ppb. Indium and rhenium were used as internal standards. The international geostandards SLRS-4 and SLRS-5 (Riverine Water Reference Material for Trace Metals certified by the National Research Council of Canada) were used to check the validity and reproducibility of each analysis (see refs. [10
] for analytical details). There was good agreement between our replicated measurements of SLRS-5 and the certified values (relative difference <15%).
The biological availability of trace metal is known to be fully controlled by their speciation in aqueous solution. For example, permeation of free ions through the biological membrane is determined by ion transport channels having approx. 600 Da molecular size. Conventional filtration through a 0.45 µm membrane cannot, therefore, account for the bioavailability of metals, and thus a fine size separation procedure via ultrafiltration/dialysis is necessary. In order to better characterize trace element speciation in the studied lake, we performed in situ dialysis over the full depth of the water column. These dialysis experiments were performed using 20 to 50 mL pre-cleaned dialysis bags placed at various depths during the different seasons. The duration of the dialysis procedure was between 72 and 96 h, following kinetic experiments of dialysis equilibrium attainment for DOC, Si, and trace metals [10
]. For dialysis experiments, EDTA-cleaned trace-metal pure SpectraPor 7®
dialysis membranes made of regenerated cellulose with pore sizes of 1 kDa were thoroughly washed in 0.1 M double-distilled HNO3
and ultrapure water, and then filled with ultrapure MilliQ deionized water and placed into the lake at the desired depth. The interior compartment of the dialysis bag was retrieved after 3–4 days of exposure and analyzed for organic carbon and trace metals.
Samples for microbiological analyses were collected in sterile 250-mL flacons and stored for less than 2 h in a portable cooler at 5 °C. Active bacteria number count (Colony forming units, CFU/mL) was done using Petri dishes inoculation (0.1 to 1.0 mL of lake water in three replicates) performed in a specially organized microbiological field laboratory (i.e., refs. [15
]). All manipulations were performed in the vicinity of open flame on thoroughly alcohol- and detergent-treated and UV-sterilized microbiological table. Samples were inoculated on Nutrient Agar (5 g/L beef extract, 5 g/L gelatine peptone, 15 g/L bacteriological agar, pH = 6.8 ± 0.2 at 25 °C) to determine the total number of heterotrophic bacteria. Difco™ agar (granulated powder) inoculation was used to assess the number of oligotrophic bacteria. Inoculation of blanks was routinely performed to assure the absence of contamination from external environments. Two duplicates of different volumes of lake water (normally 0.5, 1, 5 mL) were inoculated on three types of agar nutrient media (nutrient Bacto agar, 1:10 diluted nutrient agar, and Difco nutrient-poor agar, for enumeration of eutrophic (E), facultatively oligotrophic (FO) and oligotrophic (O) bacteria, respectively). The number of colony-forming units (CFU) was evaluated by visual counting of colonies on the agar plate after 5, 10, and 30 days of incubation at 25 °C in the dark for E, FO, and O-type bacteria, respectively.
In situ measurement of primary production and mineralization were performed via incubation of bottles with oxygen detection, either by the WTW oxymeter with dissolved oxygen (DO) polarographic sonde or by Winkler titration method. On each selected horizon, two transparent polycarbonate and two dark 0.5-L bottles were filled by filtered and non-filtered lake water and closed without bubbles. Exposure time varied from 24 to 72 h. The oxygen analyses were performed in duplicate with an uncertainty of 2%, a resolution of 0.1 mg/L, and a detection limit of 0.2 mg/L.
The chlorophyll a
was measured by spectrophotometry at six different wavelengths after extraction from 0.45 µm filter using 90% acetone. For this, from 1 to 1.5 L of the lake water was processed. Typical analytical uncertainty of Chl a
determination is ±10% in the concentration range 0.1 to 10 µg/L. For phytoplankton analyses, the lake water samples were stored in the refrigerator, concentrated by sedimentation, and counted in a Najotte chamber of 0.05 mL volume [18
]. The biomass was determined as algal volume for each lake and converted to wet weight assuming a density of 1 g·cm−3
after approximating the geometry of the cell [20
]. The microscopes AxioLab А1 and Axiovert CFL-40 (Carl Zeiss, GmbH, Oberkochen, Germany) were used to assess the cell number and morphology. The phytoplankton diversity was evaluated using the Shannon–Weaver’s index (H
) and Pielou’s index based on the biomass [21
]. The saprobity state of the water bodies was evaluated using the modified indices of Pantle and Bukk [22
]. Ichthyological material for the study of fish population in Lake Svyatoe was conducted in 2004 to 2006. Sampling was carried out by net with mesh size of 18 to 60 mm.
The lake water components were analyzed using best fit functions based on the least squares method, Pearson correlation, and one way ANOVA with STATISTICA version 10 (StatSoft Inc., Tulsa, OK, USA). The regressions of concentration versus year of observation were calculated using mean average of several surface (oxygenated) horizons, provided that no trend of concentration with depth was observed. For constructing the general annual trends, we mainly used the weighted least squares regressions. The statistical test was aimed to assess the difference in element concentration between two sampling sites for the whole period of observation, from 2007 to 2010 during summer and winter periods. It included the Mann–Whitney U Test and allowed the estimation of the difference between two independent sets of data (shallow and deep site) based on one given parameter (nutrient or metal concentration), at the significance level of p < 0.1.
For ten years, we have been studying the hydrochemistry and trophic status of Lake Svyatoe, a heterogeneous seasonally stratified lake with varying degrees of human impact on different parts of this area. Assessment of the environmental status of Lake Svyatoe is particularly relevant at present, because it can serve as a model of water bodies used by the local population for economic, household, and recreational purposes, while being isolated from intensive anthropogenic and industrial influence. Such rural facilities dominate in the Nordic countries and in Western Europe and will inevitably become a point of “attraction” of population growth, provided adequate infrastructure development of the road network in the northwest region of Russia. In the course of many years of complex work, the results obtained on hydrological and hydro-chemical, hydro-biological and ichthyological studies of the lake allow us to describe the current state and functioning of aquatic communities, and to evaluate the self-purification potential of the ecosystem. The distribution tendencies of ecosystem components at different time scales allowed us to assess the daily dynamics and seasonal and interannual variabilities. Based on a comprehensive analysis of the ecological status of the lake, we performed field experiments to assess the potentially “bioavailable” DOC and to model microbial degradation of autochthonous and allochthonous DOC.
Measurements of the relative proportion of colloids as a function of depth during different seasons allowed for an assessment of the autochthonous and allochthonous pools of colloidal material as a function of the (1) biological processes of phytoplankton production leading to the presence of mainly organic low molecular weight colloids; (2) input of allochthonous soil and bog OM present in the form of high molecular weight organo-mineral colloids; and (3) diffuse flux of Fe(II), certain trace metals, and organic matter from the sediment porewater to the bottom horizons. Results of the present work allow the identification of the redox processes in the hypolimnion and the organic detritus degradation leading to the chemical stratification in small boreal lakes. Assessing these processes over several years of observation provided the necessary database for subsequent quantitative modeling.
The ratio of the contribution of organic and inorganic components to the overall carbon balance depends on the season, as follows from the measurements of intensity of primary production–heterotrophic mineralization processes. Over more than a decade of monitoring of the lake ecosystem in the summer period, we demonstrate that the trophic status of the water body, the gas exchange with the atmosphere, and the exchange of dissolved components with the sediments are largely determined by the concentration of autochthonous and allochthonous DOC.