INTRODUCTION
Many recent studies considering the impact of climate change within the circumpolar North have confirmed its polar amplification, so even small climatic fluctuations cause more significant changes in the northern regions than in other parts of the world (Pörtner et al., Reference Pörtner, Roberts, Tignor, Poloczanska, Mintenbeck, Alegría and Craig2022). This is due to the positive feedback associated with sea ice and snow cover, anthropogenic pollution of the Arctic environment, and an increase in cloud cover and water vapor (Kirtman et al., Reference Kirtman, Power, Adedoyin, Boer, Bojariu, Camilloni, Doblas-Reyes, Stocker, Qin, Plattner, Tignor, Allen, Boschung and Nauels2013). Under modern global warming, these effects can lead to significant changes in ecosystems, and the forecast of these changes is especially relevant for the Arctic and adjacent regions (Self et al., Reference Self, Jones and Brooks2015). In turn, to justify these forecasts, one needs information on climate-induced changes that occurred in the past.
The part of Central Siberia locked between the Yenisei River and the western border of the Republic of Sakha (Yakutia) that includes Evenkia (Evenki District of Krasnoyarsk Krai, Russia) is currently much less explored than the neighboring regions of Siberia, the main reason being the inaccessibility of this region and its remoteness from more developed areas.
The Central Tunguska Plateau, located in the middle of the Central Siberia Plateau, is almost unexplored in terms of its paleoenvironment (Fig. 1). The paleolimnological reconstructions of the Holocene in Siberia published so far have been based on data collected from the neighboring territories that belong to other climatic zones (i.e., Blyakharchuk et al., Reference Blyakharchuk, Wright, Borodavko, van der Knaap and Ammann2007; Müller et al., Reference Müller, Tarasov, Andreev and Diekmann2009; Fedotov et al., Reference Fedotov, Vorobyeva, Vershinin, Nurgaliev, Enushchenko, Krapivina, Tarakanova, Ziborova, Yassonov and Borissov2012, Reference Fedotov, Vorobyeva, Vershinin, Nurgaliev, Enushchenko, Krapivina, Tarakanova, Ziborova, Yassonov and Borissov2013; Krivonogov et al., Reference Krivonogov, Takahara, Yamamuro, Preis, Khazina, Khazin, Kuzmin, Safonova and Ignatova2012, Reference Krivonogov, Zhdanova, Solotchin, Kazansky, Chegis, Liu and Song2023; Rudaya et al., Reference Rudaya, Nazarova, Nourgaliev, Palagushkina, Papin and Frolova2012, Reference Rudaya, Nazarova, Novenko, Andreev, Babich, Kalugin, Daryin, Li and Shilov2016, Reference Rudaya, Krivonogov, Słowinski, Cao and Zhilich2020; Nazarova et al., Reference Nazarova, Lüpfert, Subetto, Pestryakova and Diekmann2013; Frolova et al., Reference Frolova, Ibragimova, Ulrich and Wetterich2017; Subetto et al., Reference Subetto, Nazarova, Pestryakova, Syrykh, Andronikov, Biskaborn, Diekmann, Kuznetsov, Sapelko and Grekov2017; Zhilich et al., Reference Zhilich, Rudaya, Krivonogov, Nazarova and Pozdnyakov2017; Wetterich et al., Reference Wetterich, Schirrmeister and Nazarova2018; Glückler et al., Reference Glückler, Geng, Grimm, Baisheva, Herzschuh, Stoof-Leichsenring, Kruse, Andreev, Pestryakova and Dietze2022; Kobe et al., Reference Kobe, Hoelzmann, Gliwa, Olschewski, Peskov, Shchetnikov and Danukalova2022; Karachurina et al., Reference Karachurina, Rudaya, Frolova, Kuzmina, Cao, Chepinoga and Stoof-Leichsenring2023, etc.). Of the lakes that have been investigated so far, the closest in terms of distance and climatic conditions is Lake Khamra (Khamra), located 600 km to the east in Yakutia (Baisheva et al., Reference Baisheva, Biskaborn, Stoof-Leichsenring, Andreev, Heim, Meucci and Ushnitskaya2024) (Fig. 1).
Overview map of published paleolimnological studies in Central Siberia.
The Evenkia region is interesting not only because it is poorly studied compared to neighboring territories, but also because the influence of moisture transfer from both the Atlantic and Pacific oceans is weaker here than in neighboring regions of Siberia (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003). Additionally, the Central Tunguska Plateau is located at the border of the permafrost zone, which, in view of global warming, is of particular interest (Palagushkina et al., Reference Palagushkina, Wetterich, Schirrmeister and Nazarova2017; Biskaborn et al., Reference Biskaborn, Nazarova, Pestryakova, Syrykh, Funck, Meyer and Chapligin2019; Wetterich et al., Reference Wetterich, Rudaya, Nazarova, Syrykh, Pavlova, Palagushkina and Kizyakov2021).
This territory is also known worldwide for the so-called 1908 Tunguska impact event (herein Tunguska event or Event), which occurred here on June 30, 1908. The Event was a unique, powerful atmospheric explosion of unknown nature, its estimated epicenter located at 60°54′11″N and 101°54′35″E (Farinella et al., Reference Farinella, Foschini, Froeschlé, Gonczi, Jopek, Longo and Michel2001; Jenniskens, Reference Jenniskens2019; Gladysheva, Reference Gladysheva2020), that remains the most significant impact in the Earth's recorded history, though much larger impacts occurred in prehistoric times (Longo, Reference Longo, Bobrowsky and Rickman2007). The explosion over the sparsely populated east Siberian taiga flattened over 80 million trees over an area of 2200 km2 and was registered at seismic stations across Eurasia. The air waves from the blast were detected as far away as Jakarta, Indonesia (then known as Batavia, Dutch East Indies), and Washington, DC (Whipple, Reference Whipple1934). The Event is generally attributed to the atmospheric explosion of a stony asteroid about 50–60 m in size (Longo et al., Reference Longo, Serra, Cecchini and Galli1994; Hou et al., Reference Hou, Ma and Kolesnikov1998, Reference Hou, Kolesnikov, Xie and Kolesnikova2004; Kolesnikov et al., Reference Kolesnikov, Boettger and Kolesnikova1999, Reference Kolesnikov, Longo, Boettger, Kolesnikova, Gioacchini, Forlani, Giampieri and Serra2003; Morrison, Reference Morrison2018), but several other scenarios have been proposed to explain it (Longo, Reference Birks, Gordon, Birks and Gordon2007; Khrennikov et al., Reference Khrennikov, Titov, Ershov, Pariev and Karpov2020).
The possible impact of the Event on the surrounding ecosystems is still poorly studied, and nearly no data are available about the influence it had on the aquatic ecosystems situated in close proximity to the explosion. Until now, no clear traces of the Event have been found in lake sediments (Rogozin et al., Reference Rogozin, Darin, Kalugin, Melgunov, Meydus and Degermendzhi2017).
We investigated Lake Zapovednoye, a small deep lake of unknown origin situated about 70 km to the south of the Tunguska event epicenter (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003; Gladysheva, Reference Gladysheva2020). The main aim of our study was to reconstruct the paleoclimate and paleoenvironment of the Central Tunguska Plateau (Central Siberia), one of the least-studied regions of Eurasia, using a multiproxy approach that considers the data accumulated on the geochemistry, pollen, chironomids, Cladocera, and diatoms of the lake. Special attention was given to possible traces of the Tunguska event, since their identification in the bottom sediments could increase our knowledge about this still incomprehensible phenomenon and support the dating of the upper layers of the studied sediment core.
Study area
The Central Tunguska Plateau is a relatively low tuffogenic volcanic mountain area in Central Siberia, Russia, and is a part of the Central Siberian Plateau situated between the Yenisei and Lena rivers (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003; Fig. 1). Deeply incised river valleys dissect the plateau. The average elevations are 200–300 meters above sea level (m asl), and maximum heights are slightly over 500 m asl (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003).
The climate of the region is sharply continental with large amplitudes in daily and seasonal temperatures. The annual precipitation is 380–420 mm, the maximum reached in July–August. The mean annual air temperature is −6°С. The mean July temperature is +17°С; however, it often rises above +30°С during the warmest month. The average January temperature is −30°C and can drop to −55 or even −59°C. In summer, a low-pressure zone with weak winds prevails here (Vasiliev and Lvov, Reference Vasiliev, Lvov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003). The rivers are fed mainly by snow (70% of the annual runoff) (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003).
The Central Tunguska Plateau is located at the southern border of the permafrost distribution zone and is characterized by active cryogenic processes: accumulation and degradation of permafrost (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003). The permafrost boundary runs along the Podkamennaya Tunguska River. Vegetation in this area occupies an intermediate position between the pine southern taiga and northern larch taiga forests. Mixed Larix–Pinus and Betula–Pinus–Larix forests with a well-defined shrub layer and a weakly expressed grass cover prevail (Shumilova, Reference Shumilova1962; Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003).
Lake Zapovednoye
Lake Zapovednoye (60°31.685′N, 101°43.710′E) is located on the Central Tunguska Plateau, 60 km northwest of the nearest settlement, the village of Vanavara, and approximately 70 km south of the estimated epicenter of the Tunguska event. The lake is part of the small Verkhnyaya Lakura River basin, a tributary of a larger river, Podkamennaya Tunguska, which flows into the Yenisei River (Fig. 1).
The lake is almost round and about 500 m in diameter. The maximum depth is 60.3 m at 60°31.688′N, 101°43.740′E, located near the geometric center of the lake, slightly closer to the eastern shore. The lake bottom is shaped like a regular conical funnel and has a shallow littoral zone. Seismoacoustic studies of the bottom showed that the thickness of the lake sediments is about 4 m, with hard rocks lying below (Rogozin et al., Reference Rogozin, Krylov, Dautov, Darin, Kalugin, Meydus and Degermendzhy2023). The Verkhnyaya Lakura River flows into the lake at its northwestern corner and flows out of the lake in the south (Fig. 2). There is no information regarding the origin of the lake.
Geographic position and morphological characteristics of Lake Zapovednoye. The dashed line indicates the estimated boundary of the 1908 Tunguska fire with the triangle to mark its core.
METHODS
Sampling
A sediment core of 124 cm in length was retrieved on March 23, 2015, from the deepest part of the lake (Fig. 2; 60°31.699′N, 101°43.648′E). The core was taken from the ice surface, using a UWITEC gravity sampler (Austria) and removable plastic tubes of 90 mm in diameter. A transparent tube was used for visual control of the sediments. The sediment–water interface of the core was clearly visible and did not indicate loss or destruction of the upper sediment layers.
The core was split in half lengthwise. One half was cut into 1 cm samples. The samples were placed in sealed polyethylene bags, stored in darkness at −20°C, and used for 14C, 210Pb, and 137Cs dating and proxy analyses. The other half was used to prepare solid blocks for X-ray fluorescence (XRF) element analysis. To do so, 170 × 15 × 6 mm sediment blocks were extracted from the wet half core, freeze-dried, impregnated with epoxy resin, and polished following Boës and Fagel (Reference Boës and Fagel2005).
Detection of chironomids, Cladocera, pollen, and spores was performed at a 5 cm resolution between 0 and 10 cm, then at a 1 cm resolution in the layers that were presumed to correspond to the Tunguska event (10 to 17 cm), and then at a 10 cm resolution between 17 and 124 cm. Diatom analysis was done at a 1 cm resolution over the entire depth of the sediments. Water content was determined in subsamples every 1 cm, and loss-on-ignition at 550°C (LOI550) was determined in subsamples every 5 cm.
Dating
Twenty-three uppermost sediment samples (1 cm thick) were taken for 137Cs, 210Pb, and 226Ra measurements. The activity of 137Cs, 210Pb, and 226Ra was measured using previously described methods (Gavshin et al., Reference Gavshin, Sukhorukov, Bobrov, Melgunov, Miroshnichenko, Kovalev, Romashkin and Klerkx2004) of semiconductor low-background gamma spectrometry, on a coaxial Ge detector with a low-background cryostat EGPC-192-P21 connected to a spectrometer with a FP-6300B processor (EURISYS MESURES).
The modern sedimentation rate was estimated based on an assumption that the 137Cs maximum in the lake sediments marks the global fallout after nuclear tests at the Novaya Zemlya site in 1961 (Krishnaswami and Lal, Reference Krishnaswami, Lal and Lerman1978). The 226Ra activity values were subtracted from the 210Pb activity values to obtain the atmospheric 210Pb (210Pb ex) according to generally accepted methods (Melgunov et al., Reference Melgunov, Gavshin, Sukhorukov, Kalugin, Bobrov and Klerkx2003).
Two samples were taken from the core for the 14C dating. Radiocarbon analysis of bulk organic matter was carried out by the accelerator mass spectrometry (AMS) method in the NTUAMS laboratory of the National Taiwan University. The procedure for AMS 14C dating in the NTUAMS lab is described in Li et al. (Reference Li, Chang, Berelson, Zhao, Misra and Shen2022). The program OxCal 4.4 was used for radiocarbon age analyses. The measured 14C ages were calibrated with the IntCal 20 curve (Reimer et al., Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey and Butzin2020) (Table 1).
Radiocarbon dates from the Lake Zapovednoye bottom sediments.
Elemental composition
Elemental composition was measured in the solid blocks at a resolution of 0.02 to 0.1 mm by applying synchrotron radiation micro-XRF analysis (XRF-SR) at the Institute of Nuclear Physics (Novosibirsk, Russia). The sediments in the lowest 10 cm were poorly preserved and excluded from the XRF analysis; thus, elemental analysis was performed at depths from 0 to 113 cm.
Pollen
The laboratory treatment for pollen analysis was carried out according to standard methods (Faegri and Iversen, Reference Faegri and Iversen1989). One tablet of Lycopodium marker spores (batch 483216) was added to each sample for calculating the total pollen and spore concentrations as per Stockmarr (Reference Stockmarr1971).
Counting was performed with a Zeiss AxioImager D2 light microscope at 400× magnification. Keys and atlases were used to determine the taxonomic identification of pollen and spores (Kupriyanova and Alyoshina, Reference Kupriyanova and Alyoshina1978; Moore et al., Reference Moore, Webb and Collison1991). Each sample contained a minimum of 300 pollen grains and spores. The spore–pollen diagram was produced in the TILIA software; zoning was performed after cluster analysis in the CONISS software (Grimm, Reference Grimm2004).
Chironomids
Treatment of sediment samples for chironomid analysis followed standard techniques described in Brooks et al. (Reference Brooks, Langdon and Heiri2007). At least 50 head capsules were extracted from each sample to ascertain the diversity of the chironomid populations, sufficient for accurate estimation of the inferred temperature (Heiri and Lotter Reference Heiri and Lotter2001; Quinlan and Smol, Reference Quinlan and Smol2001). Chironomids were identified to the highest taxonomic resolution possible with reference to Wiederholm (Reference Wiederholm1983) and Brooks et al. (Reference Brooks, Langdon and Heiri2007). Information on the ecology of the chironomid taxa was taken from Brooks et al. (Reference Brooks, Langdon and Heiri2007), Moller-Pilot (Reference Moller Pillot2009, Reference Moller Pillot2013), and Nazarova et al. (Reference Nazarova, Pestryakova, Ushinckaja and Hubberten2008, Reference Nazarova, Self, Brooks, Solovieva, Syrykh and Dauvalter2017, Reference Nazarova, Syrykh, Grekov, Sapelko, Krasheninnikov and Solovieva2023). Percentage stratigraphic diagrams were built in C2 version 1.7.7 (Juggins, Reference Juggins2007). Zonation of the stratigraphy was accomplished using the optimal sum-of-squares partitioning method (Birks and Gordon, Reference Birks, Gordon, Birks and Gordon1985) in the ZONE program (Lotter and Juggins, Reference Lotter and Juggins1991). The number of significant zones was assessed using the broken stick model (Bennett, Reference Bennett1996) in the BSTICK program (Birks and Line, unpublished).
Cladocera
To analyze the subfossil chitinous remains of Cladocera from the bottom sediments of Lake Zapovednoye, we used the technique described by Korhola and Rautio (Reference Korhola and Rautio2001). Subsamples of dry sediments were deflocculated in 10% KOH and heated to 75°С for 30 minutes. The suspended matter was then rinsed through a 50 μm sieve. All samples were treated with ethanol to prevent decay, stained with safranin, and examined under an AxioLab A1 light microscope at 100–400× magnification. The remains were identified using specialized keys for both subfossil (Szeroczyńska and Sarmaja-Korjonen, Reference Szeroczyńska and Sarmaja-Korjonen2007; Frolova, Reference Frolova and Nazarova2012) and modern Cladocera (Smirnov, Reference Smirnov1971; Flößner, Reference Flößner2000; Kotov et al., Reference Kotov, Sinev, Glagolev and Smirnov2010). A total of 4899 Cladocera remains were found. In each sample, between 217 and 437 chitinized remains of Cladocera were detected, of which 112 to 310 specimens per sample were identified. The number of specimens was calculated by using the largest number of parts of the organisms. The stratigraphic diagram was built in the TILIA program; zoning was performed using cluster analysis in the CONISS application (Grimm, Reference Grimm2004).
Diatoms
Samples of the investigated bottom sediments were dried for 24 hours at a temperature of 70°C. An aliquot of approximately 0.05 g was taken from each sample. Processing for diatom analysis followed the modified water bath method (Battarbee, Reference Battarbee and Berglund1986). The samples were treated with a 30% hydrogen peroxide solution by heating on a solid-state thermostat at a temperature of 90°C for 4 hours and constantly adding peroxide (Bolobanshchikova et al., Reference Bolobanshchikova, Rogozin, Firsova, Radionova, Degermendzhy and Shabanov2015). After cooling, the samples were washed from peroxide by centrifugation in distilled water and diluted with distilled water to a final volume of 1.5 mL. Permanent slides were prepared using the Naphrax high refractive index resin. Diatoms were counted along parallel transects of up to 300 valves per sample using light microscopy (Axoiscope 40, Zeiss) with immersion oil. Diatoms were identified as per Krammer and Lange-Bertalot (Reference Krammer and Lange-Bertalot1986, Reference Krammer and Lange-Bertalot1988, Reference Krammer and Lange-Bertalot1991) and confirmed with modern taxonomy as given in the AlgaeBase database (Guiry and Guiry, Reference Guiry and Guiry2023).
Numerical methods and mean July air temperature reconstruction
Effective occurrence numbers for biological communities were estimated using the N2 index (Hill, Reference Hill1973). Principal Component Analysis (PCA) was used to explore the elemental core composition and main taxonomic variation patterns to compare the patterns against the hydrobiological data accumulated throughout the core (ter Braak and Prentice, Reference ter Braak and Prentice1988). Cluster analysis was performed in the PAST software (Hammer et al., Reference Hammer, Harper and Raan2001).
The quantitative reconstruction of the mean July air temperatures (T July) was performed using the North Russian (NR) chironomid-based temperature inference model (weighted averaging partial least squares [WA-PLS], two component; r2 boot = 0.81; root mean square error of prediction [RMSEP] boot = 1.43°C) based on a modern calibration data set of 193 lakes and 162 taxa from northern Russia (61–75°N, 50–140°E, T July range 1.8–18.8°С) (Nazarova et al., Reference Nazarova, Self, Brooks, Hardenbroek, Herzschuh and Diekmann2015). Mean July air temperature for the lakes in the calibration data set was derived from New et al. (Reference New, Lister, Hulme and Makin2002). The T July NR model had been previously applied for paleoclimatic inferences in Siberia and northern Eurasia and has demonstrated the high reliability of the reconstructed parameters (Meyer et al., Reference Meyer, Chapligin, Hoff, Nazarova and Diekmann2015; Plikk et al., Reference Plikk, Engels, Luoto, Nazarova, Salonen and Helmens2019; Nazarova et al., Reference Nazarova, Syrykh, Mayfield, Frolova, Ibragimova, Grekov and Subetto2020).
The chironomid-inferred T July were corrected to 0 m asl using a modern July air temperature lapse rate of 6°C/km (Livingstone et al., Reference Livingstone, Lotter and Walker1999; Renssen et al., Reference Renssen, Seppä, Heiri, Roche, Goosse and Fichefet2009; Heiri et al., Reference Heiri, Brooks, Renssen, Bedford, Hazekamp, Ilyashuk and Jeffers2014). According to New et al. (Reference New, Lister, Hulme and Makin2002), mean July air temperature at the sampling site was 16.1°C. Taking into account the altitude of the lake at 329 m asl, the reconstructed T July was corrected by adding 1.97°C to the reconstructed values. Chironomid-based reconstructions were performed in C2 version 1.5 (Juggins, Reference Juggins2007). The chironomid data were square-root transformed to minimize species variance.
To assess the reliability of the chironomid-inferred T July reconstruction, we calculated the abundances of the rare (Hill's N2 < 5) or absent fossil chironomids in the modern calibration data set. To assess how similar the fossil samples were in respect to the temperature-based training data set, we used the goodness-of-fit (GoF) statistic derived from a canonical correspondence analysis (CCA) of the modern calibration data with the passively fitted fossil samples, using T July as the sole constraining variable (Birks et al., Reference Birks, Line, Juggins, Stevenson and ter Braak1990).
Optima of the taxa that were rare in the modern data set were likely to be unreliably estimated (Brooks and Birks, Reference Brooks and Birks2001). To assess how unusual the fossil assemblages were with respect to the training set samples along the temperature gradient, we used the GoF statistics derived from a CCA of the modern calibration data and down-core passive samples with T July as the sole constraining variable (Birks et al., Reference Birks, Line, Juggins, Stevenson and ter Braak1990). Fossil samples with a residual distance to the first CCA axis larger than the 90th and 95th percentile of the residual distances of all the modern samples were identified as samples with a “poor fit” and a “very poor fit” with the reconstructed variable (T July) (Birks et al., Reference Birks, Line, Juggins, Stevenson and ter Braak1990).
PCA and CCA were performed using CANOCO 4.5 (ter Braak and Šmilauer, Reference ter Braak and Šmilauer2002). Chironomid percentage data were square-root transformed and rare taxa were downweighted. PCA for XRF elemental composition was performed in the R software environment (https://www.r-project.org; R Development Core Team, 2018). XRF data were previously (log+1) transformed. Zonation was carried out using cluster analysis and confirmed by PCA.
RESULTS AND INTERPRETATIONS
Core lithology and chronology
The sediments of Lake Zapovednoye consist of dark brown and black ferruginous clay with a high content of water and organic matter and have a layered structure. Below 107 cm the sediments became more compact with more-pronounced layering. The vertical 137Cs distribution showed a distinct maximum between 8 and 9 cm (Fig. 3A); we interpreted this interval to be 1964, the year of technogenic radionuclide deposition after ground-based nuclear tests in 1961 (Walker, Reference Walker2006). The 210Pb profile is well approximated by an exponential function (Fig. 3), indicating constant sedimentation conditions and no redeposition of sedimentary material.
(A) Vertical distributions of 137Cs and 210Pb activity in the bottom sediments of Lake Zapovednoye. (B) Age model and calibrated radiocarbon dates for the bottom sediments of Lake Zapovednoye.
The uniform structure of the sediments along the entire core indicated their relatively constant accumulation rate. The approximation of radiocarbon dates to the sediment surface gave a value close to zero (Fig. 3, Table 1), so the reservoir effect for this lake was considered negligible and taken as zero.
Elemental composition
The XRF-SR scanning of the core produced uniform profiles of the vertical Cr, Ga, Ni, V, Pb, U, and Th distributions, which meant they were not informative for our study. PCA distinguished two main groups of elements (Fig. 4): Group I included K, Ca, Ti, Rb, Sr, Zr, Y, and Nb and is related to the PCA 1 axis (46.8%); Group II included Mo, Br, Zn, Cu, and As (Fig. 4). Fe gravitated towards Group I, whereas Mn showed the greatest difference in behavior from all the other elements (Fig. 4).
PCA biplot illustrating the distributions of chemical elements in the sediment core of Lake Zapovednoye. The black dots indicate samples, arrows chemical elements.
Selected profiles shown in Figure 5 provide information on shifts in the allochthonous input (Ti), organics (Br), and oxic/anoxic conditions (Fe, Mo, Mn).
Vertical distributions of Ti and selected element ratios obtained from the X-ray fluorescence (XRF) measurements of the bottom sediments of Lake Zapovednoye. The gray lines present XRF-SR scan data; the black ones are fitted by LOESS smoothing with span 0.5. The horizontal pink line shows the sediment layer of the 1908 Tunguska impact event.
Br and Fe data were normalized to Ti because they could be incorporated into the lake from the surroundings, whereas Ti remained stable once deposited in the lake bottom and served as an unambiguous indicator of allochthonous inputs (Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018). In our case, Ti was used as a common reference, while Fe/Ti, Mn/Fe, and Mo/Mn were used as well-known indicators of redox conditions.
Ti is strongly correlated with Rb, Sr, Y, Zr, Cr, Y, and Nb, which are also known as indicators of surrounding-rock weathering, associated mainly with chemogenic and terrigenous-clastic processes and indicative of allochthonous clastic input (Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018; Sorrel et al., Reference Sorrel, Jacq, Van Exem, Escarguel, Dietred, Debret, McGowan, Ducept, Gauthier and Oberhänsli2021). Fe was also used as an indicator of allochthonous input (Sorrel et al., Reference Sorrel, Jacq, Van Exem, Escarguel, Dietred, Debret, McGowan, Ducept, Gauthier and Oberhänsli2021). Moreover, Mo may have been associated with organics as well.
Based on the distributions of all the elements, the sediment core was divided into seven zones based on the cluster analysis confirmed by PCA (Fig. 5).
In zone G I (1900–1600 BP), corresponding to the denser sediments, the Fe/Ti and Mn/Fe ratios were the highest in the core, whereas that of Mo/Mn was the lowest (Fig. 5). Other elements had no variations in G I, so this zone corresponded to oxidation conditions, a probable low water level, and a relatively low primary production.
In zone G II (1600–1200 BP), the Fe/Ti and Mn/Fe ratios significantly decreased, while that of Mo/Mn increased, indicating more reduced conditions. The sediments became less compact.
In zone G III (1200–600 BP), Br/Ti increased along with fluctuations in Ti and Br/Ti, while the ratios of Mo/Mn, Fe/Ti, and Mn/Fe remained unchanged. The amplitude and frequency of variations in K, Ca, Ti, Rb, Sr, Zr, Y, and Fe increased.
In zone G IV (600–200 BP), all the profiles remained almost constant, and a small trend towards a Mn/Fe decrease and Mo/Mn increase was observed. The amplitude of Br/Ti decreased.
In zone G V (200–100 BP), a trend towards a Mn/Fe increase was identified; all other profiles remained constant.
In zone G VI (100 to −50 BP), the Mn/Fe concentrations continued increasing and that of Mo/Mn began to increase. The abrupt changes registered at a depth of 16 cm corresponded to about 1909 CE. There was a visually discernible light gray interlayer about 3 mm thick, also marked in Figure 5 by a pink band. In this layer, K, Ti, Rb, Sr, Zr, Y, and Nb demonstrate synchronous peaks (Fig. 6). We interpret this layer as corresponding to the Tunguska event (Fig. 6). Above this layer, Mo/Mn increased, while Mn/Fe reduced abruptly along with Br/Ti.
Photo, water content (WC, %) and XRF core-scan (ppm) of the upper section of core ZAP-1. The light gray layer corresponds to the date of the 1908 Tunguska impact event.
In zone G VII (−50 BP until present), Br/Ti, Fe/Ti, and Mn/Fe increased while Ti decreased.
Pollen
The pollen diagram of the bottom sediments of Lake Zapovednoye was divided into three pollen zones (Fig. 7).
Pollen diagram for Lake Zapovednoye. Horizontal red line shows the sediment layer of the 1908 Tunguska impact event. AP, arboreal pollen; NAP, non-arboreal pollen; Sp, spores.
PZ I (2200–1600 BP) was dominated by Pinus subgenus Haploxylon and Betula pollen. The abundance of Abies pollen was also the highest compared to the other zones. However, the percentage of Cyperaceae pollen was lower than in the other zones.
In PZ II (1600–350 BP), Pinus and Betula pollen dominated while the amounts of Abies pollen decreased. The percentage of herbaceous plants increased slightly, e.g., Artemisia and Poaceae. Notable was the appearance of Larix pollen.
PZ III (350 to −64 BP) was characterized by an increased abundance of Pinus and Betula; the proportion of herbaceous plant pollen was similar to that in PZ II. The percentage of Picea pollen in this zone decreased. Abies pollen disappeared in the middle of the zone, but then its amount increased considerably.
At a depth of 16 cm (ca. 1908 CE), pollen concentration increased.
Chironomids
In total, 84 chironomid taxa were identified. Chironomid communities of the investigated core were very diverse, with N2 varying from 24.4 at the bottom of the core (ca. 2200 to 1900 BP) to 13.0 at ca. 1983 CE. The median N2 is 18.2. The chironomid record was subdivided into four chironomid assemblage zones (Fig. 8).
Relative proportions of the most abundant chironomid taxa in the sediments of Lake Zapovednoye, PCA axes 1 scores for chironomid data, N2 diversity, chironomid-inferred T July, and goodness-of-fit results (squared residual distance) tests for reconstructed T July, with 95th percentile of the residual distances of all the modern samples that are identified as samples with a “poor fit” with the reconstructed T July (dashed line). Horizontal red line shows the sediment layer of the 1908 Tunguska impact event.
In СН I (2200–1600 BP), the chironomid communities consisted of taxa characteristic of moderate temperature conditions (Tanytarsus pallidicornis-type, Microtendipes pedellus-type) and cold-stenotherm taxa (Tanytarsus chinyensis-, Micropsectra insignilobus-, and Corynoneura arctica-types). The taxa characteristic of flowing water (Tvetenia bavarica-, Eukiefferiella fittkaui-, and Nanocladius rectinervis-types and Synorthocladius), indicative of water-level fluctuations, as well as the semiterrestrial taxa Limnophyes–Paralimnophyes were well represented. Many taxa from the genera Limnophyes and Paralimnophyes live in rapidly changing environments and are often found in semiterrestrial conditions. These detritivorous taxa live mainly near the water surface and among thickets of vegetation (Brooks et al., Reference Brooks, Langdon and Heiri2007; Moller Pillot, Reference Moller Pillot2013).
In CH II (1600–800 BP) after 1400 BP, the abundance of M. pedellus-type increased to 12.5% and remained high throughout this zone. As a rule, this taxon is widely distributed in stagnant or slow-flowing water bodies with a high oxygen content and low nutrient load in temperate climatic conditions. Some species from this group are acid tolerant (Moller Pillot, Reference Moller Pillot2013). Many taxa that prefer to live among macrophytes, including in flooded areas, were present in this zone and include Polypedilum nubeculosum-, Cricotopus intersectus-, Cladotanytarsus mancus-, E. fittkaui-, and Co. arctica-types, Paratanytarsus, etc. After 1000 BP, there was a decline in M. pedellus-type and an increase in the taxa characteristic of colder climates and oligotrophic conditions. These included Mi. insignilobus-, Paratanytarsus penicillatus-, T. chinyensis-, and Nanocladius branchicolus-types. The number of species characteristic of flowing water decreased, except for Tv. bavarica-type, which was numerous at about 1500 BP and fell thereafter.
In СН III (800–300 BP), the abundance of Limnophyes–Paralimnophyes and Co. arctica- and Cricotopus cylindraceus-types increased compared to the previous period. However, the abundance of some other species indicative of flowing water decreased (Synorthocladius, Tv. bavarica-type, Nanocladius taxa, Eukiefferiella).
In СН IV (300 to −64 BP), two intervals could be distinguished in this zone: from 300 to ca. 50 BP, and from ca. 50 BP to −64 BP. At the beginning of the first interval, at about 250 BP, there was a sharp increase in such cold-condition taxa such as Sergentia coracina-type and Corynocera ambigua that prefer mainly cold oligotrophic lakes, which may have indicated climate cooling as well. This interval probably corresponds to the Little Ice Age.
Around the depth of 16 cm that, according to our age model, corresponded to 1904 ± 5 CE, we observed a substantial shift in the taxonomic composition and strong short-term rise in the diversity of chironomids, which was also indicated by a sharp decrease in the values of the PCA 1 axis. At this time, the number of taxa indicative of flowing water (Synorthocladius, Nanocladius, Eukiefferiella) and of unstable water level or shoreline erosion (terrestrial and semiterrestrial taxa; Limnophyes–Paralimnophyes and Smittia–Parasmittia) increased significantly. After about 40 BP, the number of oligotrophic M. pedellus-type and other taxa, also characteristic of moderate temperatures (Eukiefferiella claripennis-, T. pallidicornis-, and Cr. intersectus-types) rose.
Cladocera
A total of 30 cladoceran taxa were found in the subfossil Сladocera community of Lake Zapovednoye, of which 23 were identified to the species level. Such species as Camptocercus fennicus, Camptocercus lilljeborgi, and Phreatalona protzi were new for the region. Throughout the studied period of the history of the lake, the community was dominated by planktonic species. Benthic or lithophilic taxa were much less represented, reflecting the predominance of pelagic biocenoses over the littoral zones and a considerable lake depth. Also, in the CZ I–CZ III cladoceran zones, the species P. protzi indicative of flowing water was found. The community was dominated by the taxa preferring neutral pH conditions and indicative of oligo- and mesotrophic conditions. Four Cladocera zones were distinguished (Fig. 9).
Cladocera taxa in the sediments of Lake Zapovednoye, ratio of littoral and pelagic species, PCA axes 1 scores for cladoceran data, and N2 diversity. Horizontal purple line shows the sediment layer of the 1908 Tunguska impact event.
CZ I (2200–1100 BP) is characterized by a clear dominance of the planktonic taxa Bosmina longispina (59–78%) preferring cold and temperate conditions. The taxon is very sensitive to changes in the trophic status of the environment and can serve as an indicator of oligotrophic and oligosaprobic conditions (Flößner, Reference Flößner2000). In the lower part of the core, a high proportion of the relatively warm-water taxa Leptodora kindtii and Camptocercus rectirostris (up to 7%) was detected, the abundances of which decreased towards the top part of the zone (Fig. 9).
Around 1350–1200 BP, a rare oligosaprobic cold-stenotherm taxon Ophryoxus gracilis was found. The species has a northern distribution and is most common in the littoral zone of lakes; it is indicative of oligotrophic conditions and relatively electrolyte-poor waters (calcium [Ca2+] concentrations <26 mg/L) (Flößner, Reference Flößner2000). The presence of the rare and highly specialized taxon P. protzi, typical of flowing or interstitial waters, indicated that the lake was influenced by flowing water (Frolova et al., Reference Frolova, Nigmatullin, Frolova and Nazarova2019).
CZ II (1100–500 BP) is characterized by changes in the dominant taxon: Daphnia longispina gr./Ceriodaphnia sp. increased up to 60% and became dominant while the true planktonic taxa of B. longispina-type decreased down to 20%. The abundance of littoral, cold-stenothermic, or the so-called “northern” taxa (Harmsworth, Reference Harmsworth1968) Acroperus harpae and Biapertura affinis attained maximal occurrences in the core (approx. 7% and 2%, respectively) at the top of CZ II after 600 yr BP.
In CZ III (500–100 BP), the cool-water planktonic taxa B. longispina increased to 40%, while the proportion of more warm-water Da. longispina gr./Ceriodaphnia sp. decreased to 20–40%. In the beginning of the zone, the number and abundance of littoral taxa including cold-water A. harpae, Alona guttata, and Coronatella rectangula increased. These changes in the composition of cladoceran communities after 600 BP at the end of CZ II and CZ III probably indicate a deterioration in the climatic conditions and reflect the harsher conditions of the Little Ice Age.
In CZ IV (100 to −64 BP), substantial changes led to yet another change of the dominant taxa in the cladoceran community when the oligosaprobic and cold-water B. longispina was replaced again by Da. longispina gr., preferring more moderate conditions. Daphnia longispina gr. composed from 66.0 to 82.2% of the total abundance of cladoceran remains in this zone. Around the depth of 16.5 cm, during the period corresponding to the Tunguska event, we observed a prominent shift in the composition of cladoceran communities, namely a decrease in the PCA 1 values, reflecting a sharp decrease in species richness. The cladoceran taxa in this layer drop to six, although generally, for CZ IV, the number varied from 8 to 13. The changes occurred primarily due to the loss of the littoral taxa (Fig. 9), mostly closely associated with littoral vegetation (Eurycercus sp., Bi. affinis, Ca. rectirostris) that reappeared in the lake in subsequent layers. Water flow indicators were not found in this zone, which may be only due to the relative rarity of such indicators among cladocerans.
In the uppermost sample there was an increase in the abundance of littoral taxa, such as Eurycercus sp. and Ceriodaphnia sp., closely associated with well-developed littoral vegetation.
Diatoms
Cluster analysis divided the Zapovednoye Lake core into four diatom zones (Fig. 10).
Most abundant diatom taxa in the sediments of Lake Zapovednoye, PCA axes 1 scores for diatom data, N2 diversity and N for the number of taxa. С/Р is Centric/Pennate diatoms ratio. The horizontal red line shows the sediment layer of the 1908 Tunguska impact event.
Zone D I (2200–1100 BP) contains a complex of planktonic–benthic species dominated by Tabellaria fenestrata. Although this zone was characterized by the lowest proportion of benthic species in the core, the benthic Staurosirella pinnata was among the dominant ones. Within D I, we observed a clear increase in the proportion of planktonic species (Ta. fenestrata, Aulacoseira granulata var. angustissima) to reach the highest abundance in the core.
Lindavia lemanensis was subdominant in the lower part of the zone, but increased significantly and became dominant in the upper part of the zone. The proportion of the small pennate diatoms S. pinnata, Staurosira venter, and Pseudostaurosira elliptica decreased in this zone.
The lower part of the zone is dominated by halophobic and salinity-indifferent species preferring alkaline conditions, while the proportion of acidophilic species increased towards the top of the zone. The dominant species were cosmopolitan, but the arctic–alpine species Aulacoseira valida and cold-stenotherm Diploneis parma were present in small numbers.
The diatom composition of D I is indicative of flowing water with low mineralization under cool temperature conditions. There was also a tendency towards increasing water levels and the formation of a zone of acidic shallow waters.
Zone D II (1100–350 BP) is characterized by an increase in the proportion of halophilic and alkaline species while the abundance of the halophobic planktonic–benthic Ta. fenestrata decreased, to be replaced by another halophobic planktonic–benthic species, Meridion constrictum Ralfs, that prefers alkaline environments. Abundancies of the alkaliphilic species Aulacoseira ambigua, Au. granulata var. angustissima, and Aulacoseira subarctica gradually increased alongside the increasing proportion of benthic species dominated by S. pinnata, Ulnaria ulna, and the planktonic–epiphytic Fragilaria vaucheriae. In the upper part of the zone, the highest increase in the proportion of Discostella stelligera for the entire core was detected. The findings indicate the lake had a high and constant level of warm water of increased productivity.
Zone D III (350–100 BP). Within this time interval, the trend was towards an increase in the abundance of planktonic–benthic species dominated by Ta. fenestrata and Cocconeis placentula var. euglypta and benthic species dominated by S. pinnata, St. venter, and U. ulna. The planktonic species L. lemanensis and Au. subarctica were found to dominate in this zone. However, at the top part of the zone, the proportion of C. placentula var. euglypta reduced to be replaced by L. lemanensis. Moreover, such taxa as Encyonema minutum, Epithemia gibba, and Epithemia spp. appeared for the first time. The zone was characterized by an increase in the proportion of halophilic species and the preservation of weakly alkaline conditions. The cold-stenotherm Di. parma was constantly recorded within the zone. The diatom composition reflects a cooling environment, stable or lowered water level, and increased mineralization.
Zone D IV (100 to −67 BP). The proportion of benthic and planktonic–benthic species, predominantly halophilic and alkaliphilic, continued to increase with such species as Ta. fenestrata, S. pinnata, St. venter, and U. ulna becoming dominant, which may have reflected the lowering of the water level, a warmer environment, more alkaline water, and shallow-water littorals overgrown by macrophytes. Mineralization remained similar to that in D III.
There was only a slight taxonomic shift in the diatom assemblages corresponding to ca. 18 cm, or ca. 1908 CE, reflected as well by PCA 1 (Fig. 10). The abundance of Ta. fenestrata dropped from 24 to 2%, S. pinnata from 10 to 0–4%, and U. ulna from 9 to 4%. At the same time, the abundances of otherwise less-abundant species grew for a short time: F. vaucheriae (2 to 6%), Fragilaria capucina var. distans (2 to 9%), Ps. elliptica (8 to 18%), Staurosira incerta (0 to 6%), and St. venter (6 to 15%). However, the diversity of diatom assemblages remained similar to those found in adjacent sediment layers.
Mean July air temperature reconstruction
The application of the chironomid-based NR transfer function resulted in reconstructed T July fluctuations of approximately 2.4°C over the last ca. 2.2 ka (Fig. 8) with the lowest temperatures (14–15°C) reconstructed for the period between ca. 200 BCE and 650 CE, and gradual warming towards 950 CE with the warmest T July calculated between ca. 1000 and 1350 CE (max T July 16.5°C, median T July 15.6°C). There were coolings between ca. 1690 and 1920 CE (median T July 14.1°C), and T July rose to 16.8°C in the more recent period, which is close to the modern T July, taking into account the RMSEP of the applied model (1.43oC).
All 77 identified chironomid taxa were represented in the modern training sets. Three of 77 taxa had a Hill's N2 < 5 and therefore were defined as not well represented in the training sets: Tv. bavarica-type (N2 = 2 in the NR data set and 8.8 in the core), Synorthocladius (N2 = 1 in the NR data set and 12.4 in the core), and E. fittkaui-type (N2 = 1 in the NR data set and 12.1 in the core). These three taxa were frequently found in the core, and their low representation in the NR data set hampered the reconstruction quality, which is confirmed by GoF statistic results. GoF for T July reconstruction revealed that five samples had a “poor” fit with the temperature (Fig. 8), while the rest of the samples showed either a “good” or “moderate” fit. The poorest fit was demonstrated by the sample at ca. 1700 CE containing Tv. bavarica- and E. fittkaui-types, the taxa of high abundance that were poorly represented in the NR data set. The high representation of the taxa in the training set and primarily good GoF results indicated that the temperature reconstruction from the Lake Zapovednoye record was reliable. However, the samples with poor GoF tests should be interpreted with caution.
DISCUSSION
Lake morphology
The rounded-funnel-shape of the lake, its significant depth, and large depth/diameter ratio indicate its origin is not associated with thermokarst processes (Rogozin et al., Reference Rogozin, Krylov, Dautov, Darin, Kalugin, Meydus and Degermendzhy2023). Two other lakes of unknown origin and similar shape, Cheko (Gasperini et al., Reference Gasperini, Alvisi, Biasini, Bonatti, Longo, Pipan, Ravaioli and Serra2007; Rogozin et al., Reference Rogozin, Darin, Kalugin, Melgunov, Meydus and Degermendzhi2017) and Peyungda (Rogozin et al., Reference Rogozin, Krylov, Dautov, Darin, Kalugin, Meydus and Degermendzhy2023), have been described in the vicinity of Zapovednoye. According to one version, Lake Cheko is a crater from the fall of a cometary body at the time of the Tunguska event (Gasperini et al., Reference Gasperini, Alvisi, Biasini, Bonatti, Longo, Pipan, Ravaioli and Serra2007), but others assert that the lake is older than the 1908 Tunguska impact event (Rogozin et al., Reference Rogozin, Darin, Kalugin, Melgunov, Meydus and Degermendzhi2017). The similar funnel shape, large depth-to-diameter ratio, and location of river channels indicate all three lakes have a similar origin that is geologic rather than impact in nature. However, the question about the origin of these lakes remains open. Seismoacoustic studies have shown that the bottom sediment thickness of Lake Zapovednoye is about 4 m, and that of Lake Peyungda is over 6 m, meaning the age of both lakes spans several thousand years (Rogozin et al., Reference Rogozin, Krylov, Dautov, Darin, Kalugin, Meydus and Degermendzhy2023).
Seismoacoustic profiling of the littoral slopes of Lake Zapovednoye has revealed several horizontal “shelves” covered in sediments (Rogozin et al., Reference Rogozin, Krylov, Dautov, Darin, Kalugin, Meydus and Degermendzhy2023), which means its basin could have subsided gradually increasing the lake volume and depth. Therefore, the bottom sediment composition could be affected directly by climatic factors but also indirectly by basin subsidence that may have occurred as a result of geologic factors.
Reconstruction of the climate and limnological conditions
The biological and geochemical data obtained during our study make it possible to discuss several environmental and ecological aspects of Lake Zapovednoye such as temperature, lake level fluctuations, water flow, terrigenous material input, redox conditions, and trophic status. Figure 11 summarizes the results for all the proxies and shows several event boundaries.
Summary of environmental and ecosystem changes reconstructed from the sediment proxy of Lake Zapovednoye. LIA, Little Ice Age.
The time intervals of statistically significant zones for the different proxies do not coincide in the studied core, but the most prominent periods have been confirmed by several paleo-indicators. The climatic changes are most clearly seen in the chironomid-based T July profile (Figs. 8 and 11). The identified temperature changes are consistent with those from neighboring Siberian regions and reflect the known trend of alternating cold and warm periods in northern Eurasia (Fig. 11).
During approximately the last 1600 years, neither the climate nor the environment have changed much, which is generally consistent with published paleoenvironmental reconstructions for Central Siberia that cover a similar period (i.e., Glückler et al., Reference Glückler, Geng, Grimm, Baisheva, Herzschuh, Stoof-Leichsenring, Kruse, Andreev, Pestryakova and Dietze2022). However, the lowest layers of the core, separated by a clear boundary at a depth of about 94 cm (ca. 1600 BP), differ markedly from its other parts. Below we provide a chronological description of the major changes in the external conditions of the lake and its ecosystem identified from a combination of examined proxies.
Lower zone: 2200–1600 BP
As inferred from the chironomids, T July before 1600 BP was lower than it is today. The composition of cladoceran communities during this time interval confirms cool climatic conditions. The presence of lotic species of chironomids and Cladocera can indicate that the influent river had stronger runoff. The absence of a developed littoral zone and a small amount of littoral vegetation is strengthened by the apparent dominance of planktonic species among the Cladocera. The rare oligosaprobic cold-stenotherm taxon O. gracilis, indicative of electrolyte-poor water (Smirnov, Reference Smirnov1971; Flößner, Reference Flößner2000), and another rare and highly specialized taxon P. protzi that inhabits flowing or interstitial waters (Frolova et al., Reference Frolova, Nigmatullin, Frolova and Nazarova2019), support the presence of rivers flowing into and out of the lake.
The dominance of the oligotrophic diatom species Ta. fenestrata may also indicate water flow (Trifonova, Reference Trifonova1990). The appearance and dominance of the benthic species S. pinnata reflects water mixing, water-level fluctuations, and the influx of nutrients from the catchment (Watchorn et al., Reference Watchorn, Hamilton, Anderson, Roe and Patterson2008). The decreased proportion of this species reflects the stabilization of environmental conditions, decreasing water inflow, and mixing.
However, it could have been that the greater flow was due to a smaller lake volume and not to intensive rainfall. At the same time, the signs of water-level fluctuations in the chironomid and diatom compositions testify in favor of more intense precipitation. The abundance of the hydrophilic Abies may also be indicative of a more humid climate in that period. However, it is unknown which of the factors, temperature or humidity, controlled its dynamics to a greater extent. Since Abies does not grow on dry or waterlogged soils (Burlakov et al., Reference Burlakov, Khmara and Belyayev2009), its disappearance during the latter periods could be a consequence of cooling and/or drying up or waterlogging of the soil. Given that only three pollen samples have been detected in the lowermost pollen zone (2200–1600 BP), one should be cautious about making conclusions based on these scarce data.
Before 1600 BP, the high iron and manganese content (presented as Fe/Ti and Mn/Fe) was a clear sign of an oxidative environment, indicating an ecosystem with intensive water flow (Wirth et al., Reference Wirth, Gilli, Niemann, Dahl, Ravasi, Sax and Hamann2013; Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018). Fe and Mn are generally assumed to behave similarly in lake water environments, except when Mn precipitates at a higher redox potential than that of iron (Klemm et al., Reference Klemm, Herzschuh and Pestryakova2015). Fe and Mn become increasingly soluble in reducing lake conditions, with Mn more readily reduced than Fe under anaerobic conditions (Klemm et al., Reference Klemm, Herzschuh and Pestryakova2015; Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018). Hence, an increase in Mn/Fe is related to lake water mixing, and may serve as an indication of, e.g., water depth or wind-speed variations. The Mn/Fe ratio is assumed to reflect the lake water mixing of, i.e., the supply of oxygen to the water–sediment interface. Based on these same properties, increases in Fe/Ti ratios have been interpreted as enhanced oxic conditions. Reduced mixing may have occurred during times of more stable lake conditions with relatively higher water levels. Br is considered a tracer of organic content in sediments (Klemm et al., Reference Klemm, Herzschuh and Pestryakova2015; Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018; Baisheva et al., Reference Baisheva, Biskaborn, Stoof-Leichsenring, Andreev, Heim, Meucci and Ushnitskaya2024).
Mo is readily scavenged under sulfidic conditions, and therefore is a well-established proxy for reduced conditions and the presence of sulfide in the water column (Wirth et al., Reference Wirth, Gilli, Niemann, Dahl, Ravasi, Sax and Hamann2013), whereas enhanced Mn content signifies oxic conditions (Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018). For that reason, the Mo/Mn ratio best reflects changes in the redox potential of the lake. A low Mo content is also a sign of oxic conditions (Wirth et al., Reference Wirth, Gilli, Niemann, Dahl, Ravasi, Sax and Hamann2013; Zhen et al., Reference Zhen, Li, Xu, Wang and Zhao2020). It is possible that between 2200 and 1600 BP, the lake basin had not yet taken its modern form. The lake could have been shallower and smaller in size, which probably ensured a greater water flow, a more oxic environment in the near-bottom water layers, a small area of the littoral zone, and a lower density of littoral vegetation that is confirmed by the composition of the chironomid fauna including more taxa preferring lotic conditions (Fig. 11). Some discrepancy between the pollen and chironomid reconstructions could have resulted from the fact that chironomids respond most to the mean July air temperatures (Nazarova et al., Reference Nazarova, Syrykh, Grekov, Sapelko, Krasheninnikov and Solovieva2023), while vegetation reacts not only to temperature fluctuations but, in inland Siberia, also to changes in complex climate–permafrost–fire–vegetation interactions (Cao et al., Reference Cao, Tian, Li, Gaillard, Rudaya, Xu and Herzschuh2019). Consequently, the decrease in average annual temperature detected by pollen may be accompanied by an increase in the July temperatures detected by chironomids, if climate continentality had increased. In general, the results of our multiproxy analysis suggest that this period can be characterized as relatively cool and humid.
Middle zone: 1600–350 BP
An evident change in the ecological conditions can be traced for a period after 1600 BP. During this time, the amount of precipitation might have fluctuated significantly, which led to an uneven supply of allochthonous matter and surface runoff. As a rule, a slight increase in Br is associated with increased organic matter (Fedotov et al., Reference Fedotov, Phedorin, Enushchenko, Vershinin, Krapivina, Vologina, Petrovskii, Melgunov and Sklyarova2013) and may also indicate an increase in lake productivity. On the other hand, a decrease in Fe and Mn is indicative of a reduced flow and a possible increase in the trophic status of the lake. An increase in the frequency and amplitude of fluctuations in Ti, Rb, Sr, Zr, and Y concentrations indicates unstable conditions. These elements are usually associated with clastic indicators of weathering in surrounding rocks (Vegas-Vilarrúbia et al., Reference Vegas-Vilarrúbia, Corella, Pérez-Zanón, Buchaca, Trapote, López, Sigró and Valentí Rull2018; Sorrel et al., Reference Sorrel, Jacq, Van Exem, Escarguel, Dietred, Debret, McGowan, Ducept, Gauthier and Oberhänsli2021).
An increase in Mo/Mn, indicative of reducing conditions often associated with organic matter (Wirth et al., Reference Wirth, Gilli, Niemann, Dahl, Ravasi, Sax and Hamann2013), was observed after ca. 1100 BP. Most likely, it also indicates a further decrease in the oxygen content of the water column, which may have been a consequence of warming and/or an increased trophic status of the lake (Wirth et al., Reference Wirth, Gilli, Niemann, Dahl, Ravasi, Sax and Hamann2013). It may also have been due to the development of an anaerobic hypolimnion related to the lake's increased depth. The decrease in Abies pollen and Larix appearance are indicative of a cool climate and low humidity.
Between 1000 and 350 BP, a shift towards more mesotrophic chironomid taxa was observed, and the chironomid-inferred temperature reached a maximum of ca. 1.5°C above the modern level. The thermophilic species that appeared in the zooplankton were also indicative of warming. At the same time, the decrease in the planktonic taxa of the cladoceran community suggests a change in the ratio of deep pelagic and littoral shallow areas in favor of the latter. If compared to the earlier sediment layers, the species richness of cladoceran remains in the bottom sediments increased, indicating the community had become more complex and balanced under the relatively favorable conditions in the presence of diverse biotopes.
Analysis of the diatoms also showed signs of warmer conditions, as indicated by Au. ambigua, Au. granulata, and their varieties (Trifonova, Reference Trifonova1990; Stenina, Reference Stenina2009). There was also a succession of diatom species from a poor planktonic assemblage dominated by the planktonic, centric, acidophilic, and halophobic species of diatoms preferring high mineralization, biogenic elements, and pH. The high abundance of D. stelligera indicates an increase in trophic status (Trifonova, Reference Trifonova1990) and/or pH (Camburn and Charles, Reference Camburn and Charles2000). A significant increase in D. stelligera could also have been associated with a decrease in the water level in the lake and an increase in mineralization following the warming. This period presumably corresponds to the Medieval Climate Optimum.
Top zone: 350 BP–current time
The observed decrease in Mo, indicative of a more oxic environment, may have reflected the cooling of the lake associated with a decrease in its trophic status. The cold-water chironomid and cladoceran taxa that appeared at ca. 350 BP and a total disappearance of Abies between 300 and 100 BP could have been related to the cooling corresponding to the Little Ice Age. Among the diatoms, the return of the freshwater epiphyte Ta. fenestrata to its dominant position reflected the lowering of the water level and the formation of shallow-water zones overgrown by macrophytes (Patrick and Reimer, Reference Patrick and Reimer1966).
In the more modern time, after 100–50 BP, an increase in Betula, Pinus, and herbaceous pollen proportions, an almost complete disappearance of Larix pollen, and a decrease in Picea may indicate climate warming in the region.
Thermophilic cladoceran species, such as Le. kindtii, also appeared. In the chironomid communities, a shift occurred from cold-stenotherm to more temperate species. The decline in Se. coracina-type and an increase in the chironomids indicative of moderate conditions after 0 BP indicate climate amelioration. Thus, the composition of sediments and shifts in biological communities during the twentieth century reflects the transition from the Little Ice Age to modern warming.
Traces of the 1908 Tunguska impact event
The investigated core had a geochemical anomaly observed at the depth corresponding to ca. 1908 CE. The layer was visually distinctive and enriched with the elements indicative of terrigenous inflow (Fig. 6). We assume the layer originated due to the 1908 Tunguska event. This powerful atmospheric explosion that happened on June 30, 1908, caused a forest fire of about 700 km2 and felled the forest trees around the epicenter in an area of about 2200 km2 (Vasiliev et al., Reference Vasiliev, Lvov, Plekhanov, Logunova, Muldiyarov, Bibikova, Volkov, Vasiliev, Plekhanov, Logunova, Boyarko, Ivanova and Muldiyarov2003; Gladysheva, Reference Gladysheva2020). According to the chronicles, observations following the event, and further investigations, the density of the fallen trees at the border of the fire zone was about 1400 trees per hectare, i.e., on average, one tree fell in a 3 × 3 m plot, and the holes from uprooted trees could reach 3 m in diameter (Abramov et al., Reference Abramov, Arkayev, Russkikh, Vasiliev, Plekhanov and Logunova2003) (Fig. 12), which means that the soil was significantly disturbed over a large area leading to increased terrigenous material supply from the catchment (Darin et al., Reference Darin, Rogozin, Meydus, Babich, Kalugin, Markovich and Rakshun2020).
Trees knocked down and burned around the site of the 1908 Tunguska impact event. Copied from the Around The World magazine, 1931. The original photo was taken in May 1929.
Lake Zapovednoye is located about 20 km from the border of the forest fall zone. However, the upper reaches of the Verkhnyaya Lakura River flowing into Lake Zapovednoye are located precisely at the border of this zone, and the terrigenous material could have easily been transported into the lake by the river. It means the layer high in K, Ti, Rb, Y, Nb, and Zr that formed at a depth of 16 cm (Fig. 6) appeared through the increased supply of terrigenous material from the catchment area after the Tunguska event.
No geochemical anomalies, namely elevated concentrations of Ni and Ir, that could be associated with extraterrestrial matter were identified in the lake sediments. The Ni profile was uniform, and that of Ir was not measured in this study.
There were substantial changes in the taxonomic composition of chironomids and Cladocera around 1908 CE. Right after this time, the chironomid taxa that prefer flowing conditions and are indicative of an unstable water level or soil/coastline erosion (terrestrial and semiterrestrial taxa) appeared. As for the Сladocera, their numbers dropped by almost two times in the period following the Event, and these changes occurred primarily due to the disappearance of the taxa associated with littoral vegetation. Though their abundance was restored in the overlying sediment layers, no flow indicators were detected in this interval, which may have been due to a relative rarity of such indicators among the Cladocera.
The observed short-term decline in pollen concentration is presumably related to the Event. The increased proportions of the valves of several alkaliphile diatom species could have been due to increased reservoir productivity in the specified time period. On the other hand, it may have also been associated with increased turbidity due to intensive surface runoff following the Event. Based on the time interval between adjacent 1 cm samples, it is possible to estimate the rate of recovery of chironomid, cladoceran, and diatom communities after the Tunguska event as 6–10 years. However, communities did not recover to be similar to those pre-Event, i.e., the Event had a more lasting paleoecological effect.
Comparison to the neighboring regions
The earlier reconstructions based on peatland plant macrofossil analysis and soil sections from the Nizhnyaya Tunguska basin (Kutafyeva, Reference Kutafyeva1974; Koshkarova and Koshkarov, Reference Koshkarova and Koshkarov2005; Koshkarov and Koshkarova, Reference Koshkarov and Koshkarova2018) demonstrated that between ca. 2400 and1600 BP, the studied area was covered in Picea, Larix–Pinus, and Larix–Betula forests. Similar conclusions were drawn from pollen spectra (Kol'tsova, Reference Kol'tsova1981; Koshkarova and Koshkarov, Reference Koshkarova and Koshkarov2005). The modern period in this region is characterized by the predominance of Pinus–Larix forests, while until ca. 500 BCE, there were dark coniferous forests, including Picea and Abies (Koshkarov and Koshkarova, Reference Koshkarov and Koshkarova2018).
In the pollen spectra of CH-12 (Khatanga-12), a small lake located 1000 km to the north of Lake Zapovednoye at the tree line border of the Taimyr Peninsula (72°N, 102°E), Larix was replaced by tundra vegetation at ca. 2000 BP, which indicates a transition from a warmer to a cooler climate (Klemm et al., Reference Klemm, Herzschuh and Pestryakova2015). In our study, a decrease in the amount of Abies pollen and the appearance of Larix pollen at about 1600 BP have been observed (Fig. 7). Thus, it is possible that in the pollen spectra of both lakes, there are signs of Larix shifting to the south that is indicative of climate cooling. This phenomenon is generally consistent with the Late Holocene cooling observed southeast of the Tunguska basin and in the region of Lake Baikal (Prokopenko et al., Reference Prokopenko, Khursevich, Bezrukova, Kuzmin, Boes, Williams, Fedenya, Kulagina, Letunova and Abzaeva2007; Fedotov et al., Reference Fedotov, Vorobyeva, Vershinin, Nurgaliev, Enushchenko, Krapivina, Tarakanova, Ziborova, Yassonov and Borissov2012).
However, results of chironomid analysis from the same Taymyr core (CH-12) did not show any significant climatic change at ca. 2000 BP, and the reconstructed T July remained similar to the modern one (Syrykh et al., Reference Syrykh, Nazarova, Herzschuh, Subetto and Grekov2017). At about ca. 2500 BP, pollen spectra have shown some warming that started earlier in central Evenkia (Koshkarova and Koshkarov, Reference Koshkarova and Koshkarov2005). A transition from a cold and dry to a warm and humid climate at 2050 BP was reconstructed for Western Siberia (55°N, 82°E) (Krivonogov et al., Reference Krivonogov, Takahara, Yamamuro, Preis, Khazina, Khazin, Kuzmin, Safonova and Ignatova2012).
Inferred from the chironomids of Lake Zapovednoye, T July revealed cool conditions until ca. 1400 BP that developed later into a warming trend (Fig. 8). The reconstruction based on chironomids from the small shallow lakes of the Putorana Plateau (68°N, 92°E) demonstrated a cool temperature and an increase in continentality shortly before 2000 BP (Self et al., Reference Self, Jones and Brooks2015).
Therefore, we have found that between 2200 and 1600 BP, the investigated area had a humid and cool climate corresponding to the Late Holocene cooling. This period matches well with the cool period observed in Siberia about 2000 years ago that manifested itself by glacier advances in the Altai Mountains (Solomina, Reference Solomina1999; Rudoy et al., Reference Rudoy, Lysenkova, Rudski and Shishin2000; Agatova et al., Reference Agatova, Nazarov, Nepop and Rodnight2012). The end of this period corresponds to the Late Antique cooling, or the Late Antique Little Ice Age, observed in Europe and the Altai during this period (Büntgen et al., Reference Büntgen, Myglan, Ljungqvist, McCormick, Di Cosmo, Sigl and Jungclaus2016).
Evidence shows that in the more northern regions of Taimyr and Putorana, the warm period between 1100 and 600 BP (Naurzbaev and Vaganov, Reference Naurzbaev and Vaganov2000; Naurzbaev et al., Reference Naurzbaev, Vaganov, Sidorova and Schweingruber2002; Syrykh et al., Reference Syrykh, Nazarova, Herzschuh, Subetto and Grekov2017) was followed by cooling (Andreev et al., Reference Andreev, Siegert, Klimanov, Derevyagin, Shilova and Melles2002; Hantemirov and Shiyatov, Reference Hantemirov and Shiyatov2002; Syrykh et al., Reference Syrykh, Nazarova, Herzschuh, Subetto and Grekov2017)
The Medieval Climate Optimum (between ca. 1000 and 700 BP) characterized by strong climatic fluctuations in the Northern Hemisphere was generally relatively warm (Stine, Reference Stine, Issar and Brown1998; Mackay et al., Reference Mackay, Bezrukova, Leng, Meaney, Nunes, Piotrowska and Self2012). The warm period preceding 700 BP and followed by a sharp cooling in the Eastern Sayan mountains is evidenced by the taxonomic composition of and the mean T July inferred from chironomids (Mackay et al., Reference Mackay, Bezrukova, Leng, Meaney, Nunes, Piotrowska and Self2012). This reconstruction is generally consistent with the T July reconstructed from chironomids and signals from other bio- and geochemical proxies of Lake Zapovednoye. Tree rings (Naurzbaev and Vaganov, Reference Naurzbaev and Vaganov2000; Naurzbaev et al., Reference Naurzbaev, Vaganov, Sidorova and Schweingruber2002) and glaciomarine sediment records (Zeeberg et al., Reference Zeeberg, Forman and Polyak2003) indicate the climate became cooler in Northern Siberia during this period.
The T July inferred from chironomids from the lakes of the Putorana Plateau has also shown climatic warming during the last 50 years (Self et al., Reference Self, Jones and Brooks2015). Signs of modern warming have been found in the sediments of many lakes in Northern Siberia (Nazarova et al., Reference Nazarova, Lüpfert, Subetto, Pestryakova and Diekmann2013; Biskaborn et al., Reference Biskaborn, Subetto, Savelieva, Vakhrameeva, Hansche, Hezschuh and Klemm2015 and references therein). In the southern part of Eastern Siberia, a sharp transition from the Little Ice Age to modern warming was observed (Fedotov et al., Reference Fedotov, Phedorin, Enushchenko, Vershinin, Krapivina, Vologina, Petrovskii, Melgunov and Sklyarova2013).
CONCLUSIONS
This multiproxy investigation of the lake sediments aiming at paleoclimate and paleoenvironment reconstruction has been conducted for the first time for an unexplored taiga region located in the Central Siberian Plateau. Special attention has been paid to possible traces of the 1908 Tunguska event. The studied sediment core covers a period of about 2200 years. Our investigation has shown that, in general, the inferred climatic and limnological changes are consistent with those that occurred in the neighboring regions of Siberia and reflect the known trend of interchanging cold and warm periods in northern Eurasia. A cool and humid period occurred between 2200 and 1000 BP, though a tendency towards warming and drying had already appeared at ca. 1400 BP. A warm period corresponding to the Medieval Climatic Optimum and accompanied by a rising lake trophic status is reconstructed to have occurred between 1000 and 350 BP. An evident climate cooling corresponding to the Little Ice Age has been found to have happened after 350 BP. A trend towards climate amelioration appeared after ca. 50 BP.
Geochemical traces of the 1908 Tunguska event have been identified in the investigated core and used as a stratigraphic marker for the developed age model. No geochemical anomalies that could be associated with the extraterrestrial origin of the Tunguska event were identified. Substantial taxonomic shifts occurred in the communities of benthic chironomids and zooplankton (Cladocera) after 1908. The abundance of lotic and semiterrestrial chironomids that survive unstable water levels or soil erosion increased. Cladoceran abundance declined almost two times, and the taxa associated with littoral vegetation disappeared. Pollen and diatom assemblages demonstrate weaker compositional shifts. Pollen concentration declined and recovered quickly. An increase in the abundance of several alkaliphile diatom species can reflect an increase in the lake's productivity. We assume this can be related to the substantial influx of terrestrial material and the rise of water turbidity from the catchment due to the tree fall and fire caused by the Tunguska event. The reaction of the biological communities of the lake and surrounding vegetation requires further investigation.
Acknowledgments
The authors thank the staff of Tungussky State Nature Reserve for their great help during the fieldwork. We would like to express our deep gratitude to Prof. Kulikovsky MS (Institute of Inland Waters RAS, Borok) for his help in diatom identification. This study was supported by the Russian Science Foundation, Grant no. 22-17-00185 https://rscf.ru/en/project/22-17-00185/. Pollen analysis was carried out by N. Rudaya and Cladocera analysis was carried out by L. Frolova under the R&D Project No. FWZG-2022-0010 of the Institute of Archaeology and Ethnography SB RAS.
