Ecological distribution of modern diatom in peatlands in the northern Greater Khingan Mountains and its environmental implications

Relationships between diatom assemblages and environmental variables in peatlands of the northern Greater Khingan Mountains are helpful for understanding the indicative significance of diatoms to environment changes, and potentially provide a reference for environmental monitoring and paleoenvironment reconstruction in the edge of monsoon region. In this study, we analyzed modern diatom assemblages and their relationships with environment factors in 30 surface samples from shrubby-herbaceous and herbaceous peatlands based on ordination analysis. Benthic and epiphytic Pennatae diatoms are mainly ecological types, whereas planktonic Centricae diatoms are relatively fewer. The most diverse genera are Pinnularia and Eunotia. Eunotia paludosa and Achnanthidium minutissimum dominated in shrubby-herbaceous peatlands, while Navicula minima and Fragilaria capucina dominated in herbaceous peatlands. The diatom community structures are different in different vegetation types and the diatom species diversity in herbaceous peatlands is higher than that in shrubby-herbaceous peatlands. CODMn and pH are the most important environmental factors affecting diatom species composition and diversity. Eunotia bilunaris, Eunotia mucophila, and Eunotia paludosa can be used as indicators of acidic water environments. Caloneis silicula, Fragilaria capucina, Hantzschia amphioxys, and Navicula radiosa can be applied to indicate the weak alkaline conditions. Eunotia bilunaris and Eunotia paludosa can indicate low conductivity, while Sellaphora pupula indicates the medium–high conductivity. Fragilaria capucina and Navicula radiosa can indicate oligotrophic habitats.


Introduction
Peatland is a complex natural ecosystem which generally developed at the junction of land and water; it not only has some properties of aquatic ecosystem, but also grows plants similar to those in terrestrial ecosystem (Rydin and Jeglum 2013). It has hydrological functions such as water storage, drought prevention, runoff regulation, and flood peak reduction; it also plays an important role in water purification, material circulation, climate regulation, and biodiversity conservation (Ma et al. 2018), while hydrological variability caused by human disturbance and climate change has led to the degradation of peatland ecosystem services (Chen et al. 2020).
Diatom is a kind of eukaryotic algae that breed quickly; it is widely found in peatlands (Hargan et al. 2014). The frustule of diatoms is highly silicified and it can be well stored in sediments; the valves of different genera diatoms are easy to be identified (Kienel et al. 1999;Chen et al. 2021). Except for the wide distribution range of some species, diatom ecological amplitude is narrow and is highly sensitive to environmental changes; even slight changes in environmental characteristics may cause changes in diatom assemblages and dominant species (Rühland et al. 2000). Different species of diatoms can adapt to different habitats; when habitats change, diatom species that are sensitive to new environment may die rapidly and would be replaced by other diatom species that adapt to the new environment (Mangadze et al. 2017). Diatom species in different regions are generally varied, and diatom species living in diverse ecological environments are also discrepant (Verleyen et al. 2009). The distribution of diatom communities is controlled by many environmental variables such as conductivity, water depth, salinity, nutrients, pH, and temperature (Van Dam et al. 1994;Battarbee et al. 2001;Shurtliff et al. 2017). Due to these characteristics, diatoms have been widely used in water quality monitoring, geological research, and paleoclimatic and paleoenvironmental reconstruction (Rühland et al. 2000;Cubizolle et al. 2005;Hargan et al. 2015;Proske et al. 2017;Chen et al. 2021).
Researchers are concerned about lake acidification and eutrophication (Yang et al. 2005;Wang et al. 2012;Dong et al. 2016). In the middle of the twentieth century, European scholars began quantitative diatom-pH studies to reveal the effects of acid deposition on lake ecosystems (Charles and Whitehead 1986). The species of alkalipilous or acidophilous diatoms are effective indicators for monitoring the pH level of water. Additionally, diatoms are more likely to grow in eutrophic water, and diatom biomass is usually higher in eutrophic water than in oligotrophic water (Kingston et al. 1983). The diatom community composition is also an effective indicator to evaluate the pollution degree of river water. Diatom community composition changes from oligo-mesotrophic diatom population to eutrophic diatom population with the increase of the pollution degree of river (Mangadze et al. 2017). The aforementioned researches show that diatom has been widely utilized as bioindicator of changes in hydrology and water quality in lakes and rivers, while it is seldom studied in peatlands (Carballeira and Pontevedra-Pombal 2020; Chen et al. 2020), even if it has been proven that water level and pH are important factors affecting diatom community structure in peatlands (Poulíèková et al. 2004;Hargan et al. 2014;Chen et al. 2014). Poulíèková et al. (2004) found moisture and pH were primary factors influencing the distribution of diatom species in acidic, mineralpoor spring fens. Hargan et al. (2014) investigated the spatial distribution of diatom in response to water-table depth, pH, and vegetation type from peatlands in Canada. Chen et al. (2020) and Ma et al. (2018) explored relationships between diatom composition and hydrological changes and quantitatively reconstructed depth to the water table in peatlands in northeastern China.
The Greater Khingan Mountains, located in the high latitude region in the northern China, are one of the main peatlands distribution areas, and are one of the most sensitive regions to global climate change; the climate in study area is unique and complex due to the combined action of East Asian summer monsoon and Central Asian westerly winds (Han et al. 2019). Previous studies (Li et al. 2022a, b) have suggested that there existed chemical characteristic differences in peatlands with different vegetation types. Accordingly, our hypothesis is that the variation in diatom community from shrubby and herbaceous peatlands can be expected. To test the hypothesis, we selected 30 samples from surface sediments of shrubby-herbaceous and herbaceous peatlands in the northern Greater Khingan Mountains, combined with 11 environment factors to investigate the relationships between diatom community distribution and environmental variables by using ordination analysis. We aim to (1) explore the modern diatom species diversity and community composition in peatlands with different vegetation types, (2) identify the main environmental factors controlling diatom assemblages, and (3) verify whether diatoms can be used as indicators of aquatic environment changes within peatlands in the northern Greater Khingan Mountains. The study of diatom community structure, diversity, and ecological preference could enhance their capacity as bioindicators and would provide important references for environmental monitoring and paleo-environment reconstruction in local and worldwide peatlands.

Study area and sampling
The research area is located in permafrost region of the northern Greater Khingan Mountains, which belong to the southern part of permafrost region in Eurasia continent. In recent years, due to climate warming, permafrost has a trend of moving northward (Jin et al. 2007). Under the control of the Mongolian cold high pressure, the climate of the Greater Khingan Mountains in winter is cold and dry owing to the influence of polar continental air mass. Precipitation is plentiful and concentrated in summer under the control of the Pacific High (Han et al. 2019). The average annual temperature in this area is about − 4.3 ℃, the average temperature of the coldest month is − 28.5 ℃, and the average temperature of the hottest month is 18.4 ℃ (Han et al. 2019). The main soil type in Greater Khingan Mountains is peat soil. The plant species in study area included deciduous shrubs such as Betula fruticose Pall. and Vaccinium uliginosum Linn.; evergreen shrubs, e.g., Ledum palustre L. and Chamaedaphne calyculata Linn.; and herbaceous taxa, i.e., Carex spp., Phragmites australis (Cav.) Trin. ex Steud. and Glyceria spiculosa (Schmidt) Roshev. Main vegetation and location information of sampling sites are shown in Table S1.
A total of 30 surface sediment samples and corresponding water samples were taken from peatlands in the northern Greater Khingan Mountains in late July 2017; the locations are shown in Fig. 1. Latitude, longitude, and altitude were recorded by using GPS and plant sample survey was conducted at each sampling site (Table S1). The samples were collected along the water-table gradient in peatlands; the differences of microtopography and vegetation type were also taken into account. Fifteen surface sediments were taken from 15 different peatlands which are dominated by herbs, another 15 surface sediments were all collected in Tuqiang Peatland which is dominated by shrubs and herbs. Three transects were constructed from center to edge within Tuqiang shrubby-herbaceous peatland, and five samples were taken along each transect. At each subsite, surface sediment samples with a thickness of 1 cm were sealed in plastic bags for diatom analysis. The corresponding surface water samples were collected with 500-ml sampling bottles and brought back to the laboratory immediately after the field work for chemical analysis (Ma et al. 2018).

Physicochemical characteristics
Depth to water table (DWT) was measured with a meter stick relative to the surface of substrate. Electrical conductivity (EC) and pH were measured in situ by using an EXO multiparameter water quality analyzer. Silica (SiO 2 ) concentration was determined applying the molybdenum blue colorimetric method. Total phosphorus (TP) was determined by using the ammonium molybdate spectrophotometric method. Iron (Fe) was measured using the flame atomic absorption spectroscopy method. Total nitrogen (TN) was digested with alkaline potassium persulfate and determined by using ultraviolet spectrophotometry. Concentrations of ammonium nitrogen (NH 4 -N) and nitrate nitrogen (NO 3 -N) were measured using a continuous flow auto-analyzer. Phosphate (PO 4 3− ) concentration was analyzed using a Bray-2 extract method. Chemical oxygen demand based on the potassium permanganate index (COD Mn ) was determined using the acid method. Above water chemical parameters measurements are based on standard methods (State Environmental Protection Administration of China 2006).

Diatom analysis
The sample processing followed standard methods by using 10% HCl to remove carbonate and 30% H 2 O 2 to remove organic matter (Battarbee et al. 2001). Diatom identification and counting were carried out using an Olympus BX-53 light microscope with oil immersion under 1000 times magnification. At least 300 valves (301-394 valves, 322 valves in average) were counted for each sample and diatom taxa abundances were represented as percentages. The identification of diatoms was guided by reference materials and published studies (Krammer and Lange-Bertalot, 1986, 1988, 1991a, 1991b. Diatom assemblage diagrams were drawn using Tilia version 1.7.16. The diatom species diversity was determined by Shannon-Wiener, Simpson, and Pielou diversity indices (Kingston et al. 1983;Poulíèková et al. 2004;Stanek-Tarkowska et al. 2015) using R version 3.3 (R Core Team 2016) and package vegan version 2.4-4 (Oksanen et al. 2017).

Numerical analyses
To reduce the bias due to the presence of rare species in statistical analysis, diatom taxa which occurred in at least two samples with a percentage of > 5% were selected. A total of 20 diatom taxa were selected and the diatom percentages data were "square-root" transformed before analysis. Detrended correspondence analysis (DCA) was used to determine whether linear-or unimodal-based techniques should be employed in the subsequent ordination analysis. The gradient length of the first axis in DCA was 2.95 (less than 3), which showing a linear structure of data set and suggesting the use of linear-based redundancy analysis (RDA). RDA between diatom species and water physicochemical characteristics was performed to identify significant explanatory variables using automatic forward selection and Monte Carlo permutation tests (999 random permutations). DCA and RDA were carried out using Canoco 4.5 (TerBraak and Smilauer 2003). Spearman rank correlations between environmental variables and diatom species diversity were analyzed using IBM SPSS 19.0 software. Indicator species analysis (ISA) was applied to distinguish indicator diatom taxa in shrubby-herbaceous and herbaceous peatlands using R software. Optima and tolerance ranges of environmental Mountains. b Fifteen shrubbyherbaceous samples were all collected in Tuqiang Peatland, three transects (in the direction of arrows) were constructed from center to edge and five samples were taken along each transect variables for main diatom species based on weighted averaging inverse deshrinking (WA.inv) and tolerance downweighting WA.inv (WA.inv.tol) models were performed by C2 software 1.7.7 version (Juggins 2007).
Spearman rank correlation coefficients among environmental variables are shown in Table 2. pH is significantly negatively correlated with COD Mn and is positively correlated with EC. TN and NH 4 -N are significantly positively correlated; TN is also positively correlated with TP, PO 4 3− , COD Mn , and Fe. EC is significantly negatively correlated with COD Mn . TP is significantly correlated with PO 4 3− , COD Mn , NH 4 -N, TN, and Fe. Besides, Shannon-Weiner, Simpson, and Pielou diversity indices are positively correlated with pH, EC, and SiO 2 , negatively correlated with COD Mn , TP, and PO 4 3− . RDA based on 11 environmental variables and 20 main diatom taxa in 30 surface sedimentary samples is analyzed to better understand the relationships between diatom assemblages and environment conditions in shrubby-herbaceous and herbaceous peatlands (Fig. 4). All environmental variables account for 63.8% of the total eigenvalues of diatom species. COD Mn , pH, and TN are screened out as significant explanatory variables (explained 48.5%). COD Mn , pH, and TN are all statistically significant at the 95% level (p < 0.05), and have significant edge effects in RDA analysis when only limited one environmental variable at a time. The variance inflation factor of each variable was lower than 5, indicating that the degree of collinearity among COD Mn , pH, and TN was low. COD Mn , pH, and TN capture 38.5%, 33.8%, and 3.2% of the total variance in diatoms distribution, respectively. There existed significant differences in diatom assemblages and environments between shrubby-herbaceous and herbaceous peatlands (Fig. 4)

Distribution and diversity of diatom communities and their relationships with water environments
Diatom communities are characterized by epiphytic and benthic species, with a few planktonic species owing to shallow water environments in peatlands (Carballeira and Pontevedra-Pombal 2020). The most diverse genera in both shrubby-herbaceous and herbaceous peatlands are Pinnularia and Eunotia, followed by Nitzschia, Navicula, and Gomphonema (Fig. 2). Kingston (1982) found similar diatom assemblage (Eunotia, Navicula, and Pinnularia) from peatlands in northern Minnesota. Hargan et al. (2014; and Chen et al. (2014; . palea, N. minima, F. capucina, H. amphioxys, G. parvulum, N. radiosa, N. terrestris, P. pulchra, C. silicula, and A. minutissimum which are suitable for neutral-alkaline habitats (Hargan et al. 2014;Proske et al. 2017;Mangadze et al. 2017). Besides, the species diversity of diatoms in herbaceous peatlands is higher than that in shrubby-herbaceous peatlands. What are the major reasons for the variation in diatom species community and diversity between herbaceous and shrubby-herbaceous peatlands in Greater Khingan Mountains? There exist great diversity of microhabitats and spatial heterogeneity in peatlands which caused small-scale chemical gradients, discontinuous water availability, and vegetation distribution (Carballeira and Pontevedra-Pombal 2020; Li et al. 2022a, b). The COD Mn refers to the amount of oxygen consumed when treating the sample with potassium manganate as oxidant under certain conditions, which is often used as a comprehensive indicator of the degree of organic pollution of waterbody (Ndiritu et al. 2006). The maximum COD Mn is 58.7 mg/L at sample 15 which belonging to shrubby-herbaceous peatland (Table S2), E. paludosa (64.18%) and A. minutissimum (33.8%) are the dominant species at sample 15 (Fig. 2). E. paludosa and A. minutissimum have the highest optima and tolerance ranges of COD Mn among the main diatom taxa (Fig. 5). As we can see in Fig. 4, there are significantly positive correlations between COD Mn and E. paludosa and A. minutissimum. Ndiritu et al. (2006) showed heavily polluted (elevated COD level) downstream of Nairobi River in Central Kenya was also characterized by A. minutissimum. The minimum value of COD Mn is 1.03 mg/L, and appears at sample 16 which is herbaceous peatland (Table S2). The

Fig. 5
Optima and tolerance ranges of COD Mn , pH, TN, and EC for main diatom species based on WA.inv and WA.inv.tol models dominated diatoms at sample 16 are H. amphioxys (17.2%) and F. capucina (12.9%) which have the lower optima and tolerance ranges of COD Mn (Fig. 5). COD Mn suggests negative correlations with H. amphioxys and F. capucina (Fig. 4). Sample 15 has lower diatom species diversity, with Shannon-Weiner index of 0.8, Simpson index of 0.5, and Pielou index of 0.4. The species diversity at sample 16 is higher, and the Shannon-Weiner, Simpson, and Pielou indices are 2.7, 0.9, and 0.8, respectively (Table S2). Ndiritu et al. (2006) concluded negative correlation between COD and diatom species richness in urban tropical stream in Kenya. The negative correlation between diatom species diversity and COD Mn in Table 2 may indicate that freshwater with high cleanness in the study area is more suitable for diatom species flourishing.
The maximum pH value of 7.7 appears at sample 16 (herbaceous peatland) (Table S2), which is dominated by H. amphioxys (17.2%), F. capucina (12.9%), N. terrestris (11.2%), N. palea (7.9%), L. mutica (7.3%), M. circulare (5.6%), and C. silicula (4.3%) (Fig. 2). Most of these aforementioned diatom species are alkaliphilous or circum-neutral types (Hargan et al. 2014;Proske et al. 2017;Mangadze et al. 2017), and they are positively correlated with water pH (Fig. 4) and have higher optima and tolerance ranges of water pH (Fig. 5). The minimum pH of 4.8 occurs at sample 11 (shrubby-herbaceous peatland) (Table S2); E. paludosa accounts for 68.4%, which is a typical species in acidic peatlands (Poulíèková et al. 2004;Hargan et al. 2014). Hargan et al. (2014) implied the acidobiontic diatoms E. paludosa and E. mucophila dominated when water pH is lower than 5.5 in peatlands of the Boreal Shield and Hudson Plains, Canada. As we could see in Fig. 5, the pH optima and tolerance ranges of E. mucophila and E. paludosa are 5-6 in our study area. Besides, the diatoms species diversity in shrubby-herbaceous peatlands is lower than that in herbaceous peatlands (Fig. 3g,h,i). Diatom species diversity suggests significant positive correlation with water pH (Table 2). Shurtliff et al. (2017) noted that alkaline and peatrich and nutrient-rich wetland waters were consistent with high diatom species diversity. Yang et al. (2005) suggested that diatom species diversity increased from mesotrophic to meso-eutrophic waters, but decreased in eutrophic lakes; the number of diatom species in hypereutrophic lakes was lowest. Kingston (1982) and Hargan et al. (2014) also suggested that diatoms diversity and richness on hummocks were lower than in hollows, in acidic bogs were lower than in rich fens due to the increase of minerotrophy. Water pH has a significant impact on the growth, reproduction, and metabolism of diatoms. Many diatom species have specific optima and tolerance ranges on the pH of water habitats (Chen et al. 2014). If the pH of the habitat exceeds this specific range, these species will quickly change or disappear. Peatlands are very complex ecosystems, with a great diversity of spatial heterogeneity and microhabitats (Carballeira and Pontevedra-Pombal 2020), so there might exist differences in optima and tolerance ranges of each diatom species from different regions. The maximum pH in our study area is 7.7; therefore, we infer that the increase of pH in this range is conducive to the increase of diatom species diversity.
TN maximum value is 3.04 mg/L, with higher contents of N. palea, N. minima, H. amphioxys, and N. terrestris at sample 28 (herbaceous peatland); the minimum TN content is 0.55 mg/L at sample 17 which also belongs to herbaceous peatland and consists of N. minima, N. palea, C. silicula, and N. radiosa (Table S2 and Fig. 2). As we could see in Fig. 4, there exist positive correlations between TN and N. palea, N. terrestris, N. minima, and H. amphioxys, and negative correlations between TN and C. silicula and N. radiosa. Tan et al. (2013) found positive correlation between TN and N. palea in a subtropical river, China. Zhang et al. (2011) also implied TN was a driving factor for the variation of diatom composition in Honghe Wetland, China. TN is positively correlated with H. amphioxys but negatively correlated with N. radiosa. Besides, N. terrestris, H. amphioxys, N. minima, and N. palea have higher optima and tolerance ranges of TN, while C. silicula and N. radiosa have lower optima and tolerance ranges of TN (Fig. 5).

Diatoms as biomonitors of hydrological environments
Caloneis silicula content in the samples with pH > 7 is much higher than that in the samples with pH < 7; the highest content (over 10%) of this species is at sample 17 in herbaceous peatland (Fig. 2). C. silicula is positively correlated with pH ( Fig. 4), which has the highest pH optima and tolerance range in our study area (Fig. 5), indicating that C. silicula is not suitable for acidic water and can be used as an indicator of high pH water environment in the northern Greater Khingan Mountains.
Eunotia bilunaris is an epiphytic species and distributed worldwide, and prefers to live in acidic, oligo-mesotrophic, low conductivity swamps, lakes, and shallow waters (Ma et al. 2018;Chen et al. 2020). The relative content of E. bilunaris is the highest (> 40%) in samples 5 and 12 (Fig. 2). The EC is 43 and 39 μS/cm, and pH is 5.5 and 5.2 in samples 5 and 12, respectively; both EC and pH in samples 5 and 12 are less than the average EC and pH in shrubby-herbaceous peatlands (Table S2). The lower optima and tolerance ranges along the pH and EC gradients (Fig. 5) indicate that E. bilunaris can be used as an indicator of low conductivity and acidic environments.
Eunotia mucophila is epiphytic and generally found in oligotrophic and acidic lakes and bogs (Chen et al. 2014;Hargan et al. 2014Hargan et al. , 2015Ma et al. 2018). The relative content of E. mucophila are the highest at samples 10 and 13, exceeding 15% (Fig. 2). The pH values of these two samples are less than 5.5 (Table S2); the lowest pH optima and tolerance range of E. mucophila (Fig. 5) among the main diatom taxa in our study area indicated that this species is dominant in acidic environments, little or no presence in alkaline conditions. E. mucophila prefers acidic water conditions and is negatively correlated with water pH (Fig. 4); it could be used as indicator species in acidic environments.
Eunotia paludosa prefers to live in low pH and oligotrophic pools and streams, and can be found in water with moderate EC (Poulíèková et al. 2004;Hargan et al. 2014Hargan et al. , 2015. E. paludosa has the highest relative content of more than 70% in samples 1, 4, and 6 ( Fig. 2). The pH values of these three samples are all less than 6 and the average EC is 53 μS/cm (Table S2). Little or no content of E. paludosa is found in samples when EC over 110 μS/cm. The relative content of E. paludosa in the samples with pH > 6 is much lower than that in the samples with pH < 6. The lower optima and tolerance ranges of pH and EC in Fig. 5 indicated that E. paludosa prefers acidic water and can be used as an indicator of low conductivity and acidic environments.
Fragilaria capucina which epiphytic or planktonic is a universal species in oligotrophic freshwater (Chen et al. 2021). The species cannot exist in the samples with pH less than 6.5, and it is positively correlated with pH ( Fig. 4) and has relatively high pH optima and tolerance range (Fig. 5). In samples 23 and 29, the concentrations of NH 4 -N, NO 3 -N, and PO 4 3− are the lowest, TN and TP are also close to the lowest value in the study area (Table S2), while the relative content of F. capucina is the highest (over 25%) (Fig. 2). The relative content of F. capucina in the samples with high TN and TP in the study area is much lower than that in the samples with low TN and TP. F. capucina has low TN optima and tolerance range (Fig. 5) and prefers nutrient-poor habitats, and hence it can be used as an indicator of nutrientpoor environments.
Hantzschia amphioxys is widespread in circum-neutral pH conditions (Hargan et al. 2014;Proske et al. 2017). The relative content of H. amphioxys is the highest (> 10%) at samples 16, 18, 19, and 30 (Fig. 2). This species tends to grow in herbaceous peatlands with higher pH than in shrubby-herbaceous peatlands in the northern Greater Khingan Mountains and it exhibits positive correlation with pH ( Fig. 4) and has relatively high pH optima and tolerance range (Fig. 5). It can be used as an indicator of high pH water environments in the study area.
Sellaphora pupula prefers medium-high conductivity, eutrophic, and weakly alkaline water (Dong et al. 2006). At samples 20 and 29, the relative contents of S. pupula are the highest (more than 3%), and the EC of sample 20 is up to 208 μS/cm (Table S2). The relative contents of S. pupula in the samples with low water conductivity in the study area are lower than those in the samples with high conductivity, indicating that S. pupula prefers water conditions with medium to high conductivity and can be used as an indicator of medium-high conductivity.
Navicula radiosa is a benthic, alkaliphilous species and is widely distributed (Hargan et al. 2014). In sample 23, the relative content of N. radiosa is the highest, over 17% (Fig. 2). The concentrations of TP and PO 4 3− at this point are the lowest; TN and COD Mn are also relatively low (Table S2), and lower optima and tolerance ranges of TN and COD Mn (Fig. 5) indicated that N. radiosa may be suitable for living in clean and nutrient-poor water. The content of N. radiosa is little or absent in the samples with pH less than 6; higher pH optima and tolerance range (Fig. 5) indicate that N. radiosa is alkaliphilous and prefers water with higher pH, which can be used as an indicator of clean water environment with high pH in the northern Greater Khingan Mountains.

Conclusions
Through extraction and identification of diatoms in surface peat sediments, combined with physicochemical parameters of water environments, we explored how environmental factors affecting diatom community and diversity in peatlands from northern Greater Khingan Mountains.