Chagan Lake is the largest natural lake in Jilin Province and the tenth-largest lake in China (Liu et al., 2021). It is located on the Songnen Plain in northeast China (124°03’ ~124°34’ E, 45°05’ ~45°30’ N) ( Figure 1 ), which serves as a crucial drainage area for commodity grain production. The surface area of the lake is 372 km², with an average depth of 2.5 m and a maximum depth of 6 m, storing approximately 5.89 ×10⁸ m³ of water. Chagan Lake is a typical meadow lake in Northeast China, characterized by a continental semi-arid monsoon climate. The regional climate exhibits dry and windy conditions in spring, precipitation concentrated in summer, rapid cooling in autumn, and long cold winters with a short frost-free period (~140 days) throughout the year.

Daily meteorological observations from the nearby Qianguo station (45°12′ N, 124°26′ E) for the period 1981–2010 report a mean annual temperature of 5.8°C, with monthly averages ranging from –16°C in January to 24°C in July. Annual precipitation ranges from 400 to 500 mm, with approximately 50% occurring between June and August. Annual evaporation is considerably higher, reaching 1140–1270 mm (China.Meteorological.Administration, 2010; Liu et al., 2020b, 2019).

The main rivers supplying Chagan Lake include the Huolin River, Songhua River, and Nenjiang River. Chagan Lake is connected to the Songhua River by irrigation canals, through which the lake receives a large amount of water from the Second Songhua River and from Qianguo Irrigation Area agricultural drainage. Since the 1950s, particularly from the 1960s onward, extensive farmland development surrounding Chagan Lake has resulted in significant amounts of irrigation drainage with high concentrations of total nitrogen (TN), total phosphorus (TP), and salt entering the lake, inevitably impacting its water quality and ecological security. In addition, natural precipitation and groundwater also play a role as ancillary water suppliers for the lake (Liu et al., 2020a, 2021; Zhang et al., 2017).

Three sediment cores of 78–81 cm depth were retrieved in parallel with a piston columnar sampler near the center of Chagan Lake (45°14’45.86’’N~124°16’48.33’’E) ( Figure 1 ) in December 2016. Considering the shallow water conditions at the lake’s edge, the sampling point was strategically positioned in the lake’s central region characterized by deeper water (~5 m) and minimal anthropogenic disturbances, thereby approximating its original ecological context.

During sampling, the coring device was maintained as vertically as possible relative to the water surface. The retrieved sediment cores retained structural integrity, and the overlying water column remained clear, collectively indicating an undisturbed water-sediment interface. From the surface to the bottom, the grain size of the sedimentary cores gradually changed from silty sand to clay, and the color changed from gray black to dark gray (Xie et al., 2021).

On-site, these cores were sliced at 1 cm intervals, sealed in plastic bags, and stored at –4°C. The samples were frozen in the laboratory before analysis.

To establish a core chronology, an 80-cm-long core was selected for dating analysis. The age of the samples was determined via the ²¹⁰Pb method, employing a high-purity germanium gamma spectrometer (AMETEK, United States). Standard samples were provided by the China Institute of Atomic Energy. The ORTEC high-voltage protection mode was utilized throughout this process to ensure a measurement range spanning from 10 KeV to 10 MeV. After a one-month storage of the sediments in sealed containers to allow radioactive equilibration, ²¹⁰Pb were determined at the energy peak efficiency scale of 46.84 KeV and measured lasting over 24 h. Sediments with ²¹⁰Pb activity were dated using the CRS (constant rate of supply) model (Xie et al., 2021), and the reliability of the age model was verified by the ¹³⁷Cs activity (Appleby, 2001; Zhang et al., 2018).

Approximately 1g of homogenized sediment sample was placed into a 50 mL beaker, after which 10 mL of 5% H₂O₂ was added. The mixture was stirred and allowed to stand for 12 h (or until effervescence ceased) to remove organic matter. The supernatant was decanted, and 10% HCl was added; the suspension was left overnight (until bubbling stopped) to dissolve carbonates. After discarding the acidified supernatant, a 10% (NaPO₃)₆ dispersant was introduced, and the sample was ultrasonicated for 10 min to ensure uniform dispersion. Grain size distribution was then measured using a Mastersizer 2000 laser diffraction analyzer (range: 0.02–2000 μm).

The diatom samples were subjected to standard procedures involving treatment with 10% HCl and 30% H₂O₂ (Birks et al., 2001). Subsequently, the cleaned diatom suspensions were carefully pipetted onto coverslips and mounted using Naphrax^®. A minimum of 300 valves for each sample (one each 1cm) were meticulously authenticated and counted at x1000 magnification using an Olympus microscope (BH) fitted with oil immersion objective. Diatom taxonomy was mainly based on the works of Krammer and Lange-Bertalot (1986–1991) (Krammer and Lange-Bertalot, 1986, 1988, 1991a, 1991b), and species names were updated to the currently accepted nomenclature using Algaebase (Guiry and Guiry, 2025).

This study applies the ecological classification system for diatoms developed by Van Dam et al. (1994) and updated by Mertens et al. (2025). Based on species-specific ecological traits, diatom taxa were assigned grading values for the following indicators: pH, salinity, oxygen requirement, saprobity, nitrogen uptake metabolism, and trophic state.

(1) pH index (R): Ranges from class 1 (strongly acidic) to class 5 (strongly alkaline), with an additional class X for taxa indifferent to pH levels. Lower classes (e.g., class 1: acidobiontic) indicate taxa adapted to pH < 5.5, while higher classes (e.g., class 5: alkalibiontic) represent taxa occurring exclusively at pH > 7 (Van Dam et al., 1994).

(2) Salinity index (H): Divided into seven classes (1–7), from freshwater (class 1) to marine conditions (class 7). Lower classes (1–2) reflect freshwater conditions, while higher classes (6–7) indicate marine/near-marine salinity.

(3) Oxygen requirement index (O): Includes five levels (1–5), reflecting tolerance to oxygen availability. Class 1 taxa require near-saturation oxygen (~100%), whereas class 5 taxa tolerate severely hypoxic conditions (~10% saturation).

(4) Saprobity index (S): Categorizes water quality based on organic pollution and oxygen demand. Class 1 (oligosaprobous) represents clean, oxygen-rich waters; classes 4–5 (polysaprobous) indicate heavily polluted, oxygen-depleted environments.

(5) Nitrogen uptake index (N): Four classes (1–4) reflecting organic nitrogen utilization capacity. Class 1 taxa are strictly autotrophic (inorganic nitrogen only), while class 4 taxa are obligate heterotrophs relying on organic nitrogen.

(6) Trophic status index (T): Ranges from class 1 (oligotrophic) to class 6 (hypertrophic), with an additional class X for taxa indifferent to nutrient levels. This index indicates nutrient enrichment, particularly nitrogen and phosphorus.

For each sample, the weighted average of these grading values was calculated for each ecological indicator, using the relative abundances of diatom species within the sample as weighting factors. This approach provides an integrated assessment of past environmental conditions. For Surirella hoefleri, species-specific ecological data were unavailable; therefore, genus-level values were used for all indicators except salinity, which was set to class 6 due to its known preference for high electrolyte concentrations (Krammer and Lange-Bertalot, 1988).

The relative abundances of diatom species were calculated as percentages of the total diatom count. A stratigraphic diagram illustrating diatom succession and diatom index was generated using Origin 2024b software, including only species with a relative abundance exceeding 2% in at least one sample (Cheng et al., 2023; Favot et al., 2024). Diatom assemblage zones were delineated through hierarchical cluster analysis, and principal component analysis (PCA) was subsequently performed to identify the major patterns of diatom distribution across the stratigraphic sequences; both analyses were conducted using Origin 2024b. Additionally, a chronological sequence diagram and a grain-size composition distribution diagram for the Chagan Lake sediment core were generated using the same software, offering insights into the temporal and hydrological dynamics within the sediment core.

The activity of excess ²¹⁰Pb (²¹⁰Pb_ex) in the sedimentary core ranged from 75.36 to 817.72 Bq·kg⁻¹ and generally exhibited a decreasing trend with depth ( Figure 2 ). The ¹³⁷Cs profile further revealed two distinct phases: (1) a pronounced peak at 57 cm depth, corresponding to the 1963–1964 global fallout maximum from atmospheric nuclear weapons testing (Pennington et al., 1973), and (2) persistently elevated activities in the upper 30 cm (post-1983 CE), indicating contributions beyond direct atmospheric deposition. Above ~30 cm, ¹³⁷Cs activities displayed multiple sub-peaks rather than a smooth decline toward the surface ( Figure 2 ). This pattern likely reflects intermittent riverine inputs of watershed-derived ¹³⁷Cs adsorbed to fine sediment particles, as observed in other lakes in Northeast China (Zhang et al., 2022). In addition, variations in sediment grain size modulate the adsorption and burial efficiency of ¹³⁷Cs; finer clay and silt components have a higher surface area and cation exchange capacity, which promotes greater ¹³⁷Cs retention (Xu et al., 2022). Thus, peaks in ¹³⁷Cs in the upper 1–30 cm interval likely record alternating phases of enhanced riverine supply during anthropogenic land-use disturbance and depositional shifts favoring fine-grained sediment accumulation.

The Chagan Lake core, collected in December 2016 with its surface assigned to 2016, was dated using the ²¹⁰Pb_ex-based CRS model calibrated with the ¹³⁷Cs marker. This chronology places the core base at 1859 ± 9 CE ( Figure 2 ), and the sedimentation history is divided into four distinct phases: (1) Very slow accumulation (ca.1860–1920 CE; 80–77 cm), with a mean rate of 0.07 ± 0.04 cm·yr⁻¹, indicating a stable depositional environment. (2) Moderate increase in sedimentation (ca.1920–1950 CE; 76–68 cm), with a mean rate of 0.34 ± 0.25 cm·yr⁻¹, suggesting enhanced sediment input. (3) Rapid but highly variable accumulation (ca.1950–1980 CE; 67–48 cm), with a mean rate of 0.68 ± 1.00 cm·yr⁻¹, reflecting intensified and unstable catchment disturbance. (4) High accumulation with reduced variability (ca.1980–2016 CE; 47–1 cm), with a mean rate of 1.31 ± 0.76 cm·yr⁻¹, indicating sustained but less fluctuating input. Over this 150-year period, sediment accumulation transitioned from relatively slower to rapid rates, a shift potentially linked to anthropogenic impacts and regional environmental changes (Xie et al., 2021).

The grain-size composition of the Chagan Lake sediment core (1–80 cm; 2016–1859 CE) revealed distinct variations, with differences in grain size distribution broadly corresponding to the staged changes in sedimentation rate ( Figure 3 ). These variations reflect both natural depositional environments and anthropogenic influence. Throughout the entire profile, the sediments are dominated by fine particles, particularly the 4–16 μm and <4 μm fractions.

Below 68 cm depth (pre-1950 CE), sediments are dominated by fine-grained particles, with the <4 μm and 4–16 μm fractions averaging 34% and 31%, respectively. Medium silt fractions (16–32 μm and 32–64 μm) account for approximately 9% and 10%, while coarser particles (>64 μm) comprise about 16%. The mean grain size (D [4, 3], volume-weighted) is relatively coarse, at ~31 ± 10 μm.

Between 67 and 48 cm depth (ca. 1950–1980 CE), the proportions of fine particles (<4 μm and 4–16 μm) slightly decrease to 31% and 27%, respectively. In contrast, the medium silt fractions (16–32 μm and 32–64 μm) increase to 10% and 15%. Coarse particles (>64 μm) show no significant change. The mean grain size increases slightly to ~32 ± 10μm, indicating a modest coarsening of the sediment.

Above 47 cm (post-1980 CE), the very fine fraction (< 4 μm) fell more markedly to 23%, while the 4–16 μm and 16–32 μm fractions rise to 34% and 21%, respectively. Notably, coarse particles (>64 μm) decrease to 7%. The mean grain size decreases significantly to ~23 ± 7μm, suggesting an enhanced input of fine-grained clay and silt material during this period.

The distribution of diatoms in the 80cm sediment core of Chagan Lake was found to be non-uniform. Diatom preservation was poor, and the frequency of diatom valve breakage was high in the layers above 59cm, whereas abundant and well-preserved diatoms with low breakage frequency were observed in the layers between 59–80 cm. Therefore, for this study, only samples in depth intervals 59–80 cm were analyzed for their diatom composition.

A total of 47 diatom species were identified in the sediment core spanning the 59–80 cm interval, among which 12 taxa had relative abundances ≥2% in at least one sediment layer ( Figure 4 ). Within these diatom assemblages, planktonic species exhibited limited taxonomic diversity but dominated in relative abundances. The planktonic community was primarily characterized by three taxa: Aulacoseira granulata (average ~ 44%), Aulacoseira ambigua (average ~ 11%), and Stephanocyclus meneghinianus (average ~ 6%). In contrast, benthic species constituted most taxa with relative abundances exceeding 2%, including Diploneis elliptica, Cocconeis placentula, Iconella tenera, Nitzschia palea, and Staurosirella pinnata. Notably, Staurosirella pinnata was present in multiple sediment core samples, with relative abundances ranging from 4% to 34%.

The hierarchical cluster analysis identified two major diatom zones in the 59–80cm sediment record of Chagan Lake: Zone I (80–68 cm, ca. 1859–1950 CE) and Zone II (67–59 cm, ca. 1951–1962 CE).

Zone I (ca. 1859–1950 CE) exhibited a pronounced dominance of Aulacoseira granulata, a common freshwater taxon, displaying relative abundances ranging from 43% to 72% across the interval. This zone was further characterized by four persistent subdominant species: Aulacoseira ambigua, Stephanocyclus meneghinianus, D. elliptica, and Staurosirella pinnata, all of which occurred consistently in all samples with mean abundances exceeding 5%. While N. palea and Surirella hoefleri were recurrent throughout multiple horizons, their mean abundances remained comparatively low. In contrast, five taxa (C. placentula, E. gibba, Nitzschia intermedia, Surirella brebissonii var. kuetzingii and I. tenera) had sporadic occurrences. Consistent with the dominance of planktonic taxa, this zone showed a relatively high planktonic-to-benthic diatom ratio (3.47 ± 1.78), indicative of deeper or well-mixed water conditions during this period.

Zone II (ca. 1951–1962 CE) marked a pronounced departure from Zone I in assemblage composition, most notably through the sharp decline of A. granulata (16–36% relative abundance vs. 43–72% in Zone I). This shift is further reflected in a substantially lower planktonic-to-benthic diatom ratio (1.26 ± 0.68), suggesting a transition toward shallower or more stratified water conditions. Concurrently, four taxa—Stephanocyclus meneghinianus (average ~ 18%), N. palea (12%), I. tenera (28%), and Staurosirella pinnata (13%)—emerged as dominant constituents. Notably, Iconella elegans first appeared within this interval, occurring exclusively in horizons 59, 60, and 63 (average ~ 6%). In contrast, six taxa (C. placentula, D. elliptica, Surirella brebissonii var. kuetzingii, Surirella hoefleri, E. gibba and Nitzschia intermedia) displayed sporadic occurrences.

The cluster analysis and PCA results revealed pronounced compositional shifts in diatom assemblages around 1950 ( Figure 5 ). The PCA of the diatom data revealed that axis 1 and axis 2 explained 57.7% and 21.8% of the variance respectively ( Figure 5b ). Axis 1 was dominated by strong negative loadings of the planktonic, eutrophication‐tolerant A. granulata (PC1 loading—0.67) and strong positive loadings of the benthic, pollution‐tolerant taxa I. tenera (+0.70) and N. palea (+0.16). Axis 2 was characterized by strong positive loadings of Stephanocyclus meneghinianus (+0.57) and Staurosirella pinnata (+0.34), and negative loadings of A. granulata (–0.47) and I. tenera (–0.55). The PCA1 scores ( Figure 4 ) provided a concise summary of the compositional turnover, facilitating the interpretation of the dynamic patterns observed in the diatom assemblages. Overall, both cluster analysis and PCA reveal the maximum reorganization of the diatom community around the 1950s, which was a period of initial land alteration, water pollution, and hydro management.

Based on the diatom ecological attribute classification method proposed by Mertens et al., historical environmental changes in Chagan Lake were inferred from sediment core samples. The ecological indicator classes—pH (R), salinity (H), oxygen requirement (O), saprobity (S), nitrogen uptake metabolism (N), and trophic state (T)—exhibited consistent values with minimal variation within the 80–68 cm depth interval. This stability suggests relatively stable environmental conditions during this period. However, above the 67 cm layer (i.e. 1951 CE), the values of O, S, N, and T began to show notable fluctuations. A gradual increase in these indicators was observed between 61 and 59 cm, corresponding to more recent sediment layers. This trend indicates that, since around 1950, Chagan Lake’s aquatic environment has experienced increasing external anthropogenic influences ( Figure 4 ).

pH (R): Values in the 59–80 cm interval ranged from 3.74 to 4.00, indicating diatom assemblages predominantly aligned with circumneutral to alkaliphilous preferences (classes 3–4). This suggests that, for most of the period from the mid-19th to the mid-20th century, the lake’s pH generally hovered near or above neutrality, with moderate alkaline tendencies.

Salinity (H): Values ranged from 1.94 to 2.62. These values indicated that the diatom assemblages predominantly fell within the boundary of fresh to brackish‐fresh conditions (classes 1–3), implying moderately low salinity. Throughout much of the mid-19th to mid-20th century, Chagan Lake likely maintained water chemistry characterized by limited chloride concentrations and salinity below the brackish threshold.

Oxygen requirement (O): Values ranged from 2.22 to 3.59, corresponding to classes 2–4 (Mertens et al., 2025). In Zone I (80–68 cm), O values predominantly lay in the class 2–3 intervals, indicating diatom assemblages adapted to roughly 50–75% oxygen saturation. By contrast, in Zone II (67–59 cm) the O values shifted downward into the class 2–4 range, reflecting a transition to diatom communities that tolerate low oxygen saturation (30–75%). Although several samples in Zone II (at 59, 60, 66, and 67 cm) exceeded 3.0, none reached the class 4 threshold.

Saprobity (S): Values ranged from 1.74 to 3.18. In Zone I, S values spanned classes 1–2, with most samples clustering around class 2 (β-mesosaprobous), indicating water quality class II, oxygen saturation of 70–85%, and BOD₅ ²⁰ of 2–4 mg·L⁻¹ (moderate organic pollution). In Zone II, however, S values shifted into classes 2–3; notably, samples at 59, 60, 64, 66, and 67 cm approached class 3 (α-mesosaprobous), indicating episodic organic enrichment (BOD₅ ²⁰: 4–13 mg·L⁻¹) and reduced oxygen saturation (25–70%).

Nitrogen uptake metabolism (N): Values ranged from 1.54 to 2.67. In Zone I, N values ranged from 1.64 to 2.04, with most samples falling in Class 2, indicative of nitrogen-autotrophic diatom assemblages relying primarily on inorganic nitrogen. In Zone II, however, N values extended up to 2.67, bringing samples into class 3 and reflecting a facultative nitrogen-heterotrophic capability. This transition from class 1–2 in Zone I to class 1–3 in Zone II suggests that, after ca. 1950s, Chagan Lake experienced more frequent organic nitrogen inputs alongside its baseline inorganic nitrogen regime.

Trophic status (T): Values were consistently high, ranging from 4.23 to 5.15. In Zone I, T values ranged from 4.23 to 4.93, corresponding to classes 4–5 (meso-eutrophic to eutrophic). In Zone II, the range extended from 4.74 up to 5.15, with samples at 59, 60, 62, and 67 cm exceeding 5.0, firmly within class 5 (eutrophic). This upward transition from a mixture of classes 4–5 in Zone I to predominantly class 5 in Zone II reflects an intensification of nutrient enrichment—and thus eutrophication—occurring after ca. 1950s.

The sediment core from Chagan Lake records clear shifts in sedimentation rate and grain size distribution from 1859 to 2016 CE, reflecting transitions in both natural depositional processes and anthropogenic disturbance.

Prior to the 1950s (80–68 cm), sediment accumulation was relatively slow, and the deposits were dominated by fine particles (<16 μm). However, the mean grain size (~31 μm) was relatively coarse due to intermittent inputs of coarser materials, possibly from seasonal runoff or wind-induced resuspension. The high standard deviation (± 10 μm) further implies a heterogeneous depositional regime under weak but variable hydrodynamic conditions typical of shallow semi-arid lakes.

Between the 1950s and 1980s (67–48 cm), the sedimentation rate increased, coinciding with regional agricultural expansion and the construction of dams ( Figure 6 ). Grain-size analysis indicates a slight coarsening trend, associated with intensified land use and reservoir development which reduced surface inflow, created a negative water balance (evaporation exceeding inflow), lowered lake levels, and contracted the lake surface—thereby promoting deposition of coarser material. The high standard deviation (± 10 μm) indicates sustained variability in sediment sources and more dynamic inflow conditions driven by catchment erosion, irrigation return flows, and watershed disturbances.

Interestingly, the sediment accumulation rate in Chagan Lake exhibits not only a significant long-term upward trend but also notable short-term variability. Particularly after the 1950s, the sediment record shows marked fluctuations. At a depth of 66 cm (early 1951 CE), the sedimentation rate peaked at 2.55 cm·yr⁻¹, with a mean grain size of 52.86 μm. These changes in sedimentary characteristics were likely caused by a catastrophic flood in the Liao River Basin, triggered by extreme rainfall over northeastern China in 1951. The event severely affected many parts of Jilin Province and likely caused extensive soil erosion and substantial sediment influx into the Chagan Lake. Shortly thereafter, at depths of 64–60 cm (ca. 1953–1962 CE), the sedimentation rate experienced a brief but significant decline. Historical records indicate that northeastern China suffered severe droughts between 1958 and 1962, during which reduced sediment input likely led to the observed decrease in accumulation. In the following decades, sedimentation rates remained highly variable, suggesting unstable sediment supply conditions. These high-amplitude variations likely reflect the combined influence of intense yet intermittent human activities—such as agricultural expansion, irrigation return flow, hydrological regulation, and increased pollution from population growth—and extreme climatic events ( Figure 6 ).

After the 1980s (47–1 cm), the sedimentation rate remained high, but with a shift toward finer particles and more uniform deposition. This transformation likely reflects more stable hydrological inputs and decreased variability in sediment sources, associated with improved water management and the implementation of ecological restoration measures. A key driver was the completion of an artificial canal in 1984 that diverted water from the Songhua River into Chagan Lake, introducing substantial new sediment inputs and raising the lake level. Concurrent climatic trends—namely, increasing regional precipitation and decreasing evaporation from 1960 to 2018—further contributed to the lake’s hydrological recovery (Liu et al., 2020b). Following the 1990s, growing public environmental awareness prompted a series of government-led conservation initiatives. In 2007, Chagan Lake was designated a national nature reserve with a 147 km² outer protection zone. Further efforts included the launch of the National Ecological Riparian Project in 2014, which targeted 23.6 km of degraded shoreline through measures designed to mitigate erosion, stabilize lake margins, and reduce soil loss due to freeze–thaw processes (Li and Zheng, 1997). Collectively, these anthropogenic and climatic influences led to a more regulated sedimentary regime and improved ecological conditions.

It is important to note that the sedimentation rates used in this study are based on volumetric (linear) accumulation estimates derived from ²¹⁰Pb dating. While these values provide valuable insights into temporal trends, they primarily reflect relative rather than absolute changes in erosional and depositional dynamics. Accordingly, our interpretation focuses on the comparative shifts in sediment input and transport regimes over time, rather than precise mass accumulation.

The PCA ordination highlights two distinct ecological gradients governing diatom assemblage shifts in Chagan Lake. Principal Axis 1 (57.7% variance) was primarily driven by A. granulata, I. tenera and N. palea. A. granulata is a planktonic species commonly found in deeper, well-mixed waters with moderate trophic conditions and dominated the assemblages before the 1950s (Zone I). Its prevalence suggests relatively stable hydrological conditions, weak alkaline pH (~7–8), moderate oxygen saturation, low organic pollution, and a meso-eutrophic state. In contrast, I. tenera and N. palea both epipelic species—increased after the 1950s (Zone II). These species are typically associated with fine-grained, unstable sediments and organically enriched, oxygen-poor conditions (Dalu et al., 2020; Klamt et al., 2020). Their increase reflects not only eutrophication but also a shift in dominant habitat types from planktonic to benthic, likely driven by shallowing of the lake, intensified siltation, and increased disturbance, in line with anthropogenic influences such as agricultural runoff and hydrological regulation. This shift is further supported by the decline in the planktonic-to-benthic diatom ratio, which dropped from 3.47 ± 1.78 in Zone I to 1.26 ± 0.68 in Zone II. This quantitative trend reinforces the interpretation of a functional community transition from planktonic dominance to benthic taxa, likely in response to decreased mixing depth and increased habitat instability. Thus, Axis 1 represents a complex ecological gradient: while it clearly captures the transition from meso-eutrophic to eutrophic conditions, it also reflects a shift in life-forms from planktonic (A. granulata) to benthic (Nitzschia, Iconella, and Surirella) diatoms.

Additionally, the notable rise of Cocconeis placentula, a typical epiphytic diatom, during the transition between Zones I and II may suggest an expansion of submerged macrophytes, potentially associated with a decline in lake level and improved light penetration. This period also coincides with a decline in sedimentation rate, further supporting the possibility of macrophyte proliferation during a more hydrologically stable or shallow phase. Notably, Surirella hoefleri only appeared in the layers with a lower sedimentation rate and reached its peak during this transition. This species is usually found in environments with high conductivity (Krammer and Lange-Bertalot, 1988; Stenger-Kovács and Lengyel, 2015). Its abundance likely reflects a temporary increase in salinity, providing additional evidence of changing hydrological conditions.

Axis 2 (21.8% variance) seems to represent the gradient changes influenced by salinity and oxygen dynamics. Positive loadings were associated with Stephanocyclus meneghinianus and Staurosirella pinnata. The tolerance of Stephanocyclus meneghinianus to brackish-fresh transitions and hypoxia indicates that it tends to adapt to salinity fluctuations or low-oxygen microhabitats, while the high-oxygen requirement of Staurosirella pinnata may suggest the presence of local oxygen-rich areas, which may be related to seasonal water level changes or sediment resuspension. Negative loadings for A. granulata and I. tenera (both with H=2) may reflect the dominance of freshwater. This means that Axis 2 captures the secondary gradients of salinity changes and redox heterogeneity, which may be influenced by intermittent hydrological inputs or human lake management.

Although A. granulata is often associated with eutrophic systems, its dominance in Zone I (negative Axis 1) was accompanied by low saprobity (S=2) and moderate oxygen levels (O=3), suggesting early-stage nutrient enrichment rather than severe organic pollution. In addition, A. granulata’s heavy silicified filamentous depend on wind-driven turbulence to maintain suspension—an increasingly unmet requirement as wind speeds decline. Previous studies have documented a significant decrease in near-surface wind velocities across mid-latitude regions from the mid-20th century to around 2010, a phenomenon known as “global terrestrial stilling” (Deng et al., 2018; McVicar et al., 2012). This “stilling” reduced mixing depth, disadvantaging A. granulata against benthic, heterotrophic taxa like N. palea (S=5, N=4), which thrived under complex stress conditions, even in eutrophic environments. Thus, the convergence of diminished turbulence, eutrophication, and hydrological modifications favored stress-tolerant assemblages, surpassing purely nutrient-driven succession patterns.

The diatom-derived ecological indicators provide a detailed understanding of the dynamic changes in the water environment of Chagan Lake from the mid-19th century to the mid-20th century. The interactions among pH, salinity, oxygen saturation, saprobity, nitrogen metabolism, and trophic state reveal a system that is constantly transforming under both natural and anthropogenic pressures, providing crucial insights into its ecological trajectory.

The weakly alkaline conditions (pH classification: 3.74–4.00) inferred from diatom assemblages are consistent with the dominance of alkaliphilic taxa such as Stephanocyclus meneghinianus and A. granulata. These taxa thrive in environments dominated by bicarbonate ions (HCO₃ ⁻) and carbonate ions (CO₃ ²⁻), which are characteristic of the geological origin of this lake in the saline-alkali Songnen Plain (Deng et al., 2018; Wang and Zheng, 1992). The persistence of alkaline conditions reflects ongoing inputs of Na⁺ and HCO₃ ⁻ through weathering of regional bedrock and interactions with Quaternary sediments. However, the absence of extreme alkalibiontic taxa (class 5) suggests pH remained below 9.0, avoiding conditions that could severely limit dissolved oxygen solubility (Passy et al., 2004).

Salinity classification values (1.94–2.62) reflect freshwater to brackish fresh environment (Cl⁻: ~100–500 mg·L⁻¹), which is consistent with modern measurement results (Sui et al., 2016). This moderate salinity may result from the balance between natural groundwater inputs (rich in Na⁺ and HCO₃ ⁻) and dilution by the Songhua River. However, following the 1950s, agricultural expansion and irrigation drainage likely contributed additional salts and nutrients to the lake, resulting in a slight increase in salinity and the ecological shifts observed in Zone II (Liu et al., 2021).

Oxygen saturation is a key water quality parameter that directly affects the functions of aquatic ecosystems. It exerts its influence by affecting the survival of organisms, biodiversity, and biogeochemical cycles. In Chagan Lake, sedimentary diatom proxies indicate moderate–good oxygenation (50–75% saturation) from the mid-19th to mid-20th century, but saturation fell to 30–50% by the 1960s under growing anthropogenic and climatic stress. Since the mid-20th century, Northeast China has experienced significant warming, with Chagan Lake’s regional annual temperature rising at a rate of 0.43 °C per decade (Liu et al., 2020a). This warming increases lake water temperature, reduces the solubility of oxygen, and enhances microbial metabolic rates—together contributing to lower oxygen saturation, particularly under the lake’ s alkaline conditions (pH>8). In addition, the long ice-cover period of Chagan Lake restricts atmospheric gas exchange and limits light availability for photosynthesis, leading to under-ice anoxia in shallow areas during winter months (Zdorovennova et al., 2021). Even during the ice-free season, reduced wind speeds may weaken vertical mixing, exacerbate stratification, and limit oxygen transport to deeper layers (Deng et al., 2018).

These climate-driven mechanisms (temperature rise, ice cover duration, and wind patterns) may synergize with anthropogenic pressures (e.g., nutrient input from agriculture), creating a compounding effect on oxygen saturation.

Historical records support this interpretation. As a former royal hunting reserve, Chagan Lake’s wetland ecosystem remained largely undisturbed until the mid-20th century, maintaining natural hydrological connectivity and vegetation coverage. This relatively pristine state is maintained through the interaction between undisturbed sediments and water bodies, as well as the maintenance of an effective oxygen regulation by aquatic plant communities. The decline in oxygen levels in the 1960s coincided with the beginning of regional development ( Figure 6 ), although the impact of pre-industrialization was still relatively limited compared to the changes in the late 20th century. This paleolimnological reconstruction demonstrates diatom assemblages’ sensitivity to oxygen regime changes, with their compositional variations effectively capturing early warning signals of ecosystem stress.

The distribution of saprobic indicator taxa in aquatic sediments is more responsive to increasing pollution than to improvements in water quality. Saprophilous taxa often have broad environmental tolerances and can persist even under low-saprobity conditions (Mertens et al., 2025). In the 59–80 cm interval of the sediment core, spanning from the mid-19th century (80 cm, ~1859 CE) to the mid-20th century (59 cm, ~1962 CE), shifts in diatom assemblages reflect temporal variations in organic loading (BOD₅ ²⁰) and associated oxygen demand. From the mid-19th to early 20th century (74–80 cm, S=1.74–1.99), oligosaprobous taxa, notably D. elliptica, dominated under low organic enrichment, which was consistent with stable oxygen conditions. Under these conditions, limited biodegradable organic matter reduced BOD₅ ²⁰ and maintained an aerobic habitat conducive to pollution-sensitive species. By the mid-20th century (59–64 cm, S=2.26–3.18), the increase in agricultural activities raised the input of organic matter, promoting the dominance of β−mesosaprobous A. granulata and polysaprobous taxa such as Stephanocyclus meneghinianus and N. palea. Elevated organic loads likely stimulated microbial decomposition, intensified BOD₅ ²⁰, and created anaerobic conditions that inhibited oligosaprobous diatoms. This can be seen from the peak saprobity index at 59 cm (S=3.18 at 59 cm). Between these two phases (65–73 cm, S=1.97–2.72), oligosaprobous taxa such as C. placentula showed intermittent recovery and the polysaprobous Nitzschia intermedia appeared sporadically, while Surirella hoefleri remained consistently abundant. Given its well-documented preference for high-conductivity, elevated-salinity waters (Krammer and Lange-Bertalot, 1988; Stenger-Kovács and Lengyel, 2015), the persistence of S. hoefleri in this interval likely reflects periodic salinity peaks superimposed on fluctuating organic loading. Together, these patterns of mixed saprobic and salinity-tolerant taxa indicate irregular oxygen demand and unstable redox conditions, creating a mosaic of ecological niches.

Because the saprobic index integrates both biodegradable organic load and dissolved oxygen, its gradient closely mirrors previously discussed oxygen saturation trends—ranging from 70–85% in β-mesosaprobous to 25–70% in α-mesosaprobous conditions. The increase in saprobity index—from S=1.79 at a depth of 80 cm (mid-19th century) to S=3.18 at a depth of 59 cm (mid-20th century)—reflects the gradual rise in the degree of organic pollution over time. The higher S values in the upper layers reflect elevated BOD₅ ²⁰, resulting from increased organic matter input that depletes dissolved oxygen and favors pollution-tolerant taxa. Conversely, the lower S values in the deeper layers correspond to reduced BOD₅ ²⁰ levels, preserving aerobic conditions for several oligosaprobous species.

The nitrogen uptake metabolism index increased from 1.70 at 80 cm to 2.67 at 59 cm, indicating a shift from dominance by nitrogen-autotrophic taxa to an increasing reliance on organically bound nitrogen after the 1950s. This transition is coherent with the simultaneous saprobity increase (S=1.79 → 3.18) and the rise in the oxygen‐requirement index (2.23 → 3.59), both of which testify to intensifying organic pollution and higher biochemical oxygen demand. Accordingly, the trophic state index climbed from 4.23 (mesotrophic) to 5.15 (eutrophic), which was consistent with the enrichment degree of nutrients and the increase in primary productivity. Taken together, the converging lines of evidence—from oxygen requirement trend, through saprobity, to nitrogen‐uptake metabolism and trophic state—paint a clear picture of progressive eutrophication and organic-N loading from the mid-19th to the mid-20th century. The diatom assemblages showed a rising proportion of facultative nitrogen-heterotrophic species, underscoring the value of nitrogen uptake metabolism as a sensitive indicator of nutritional regime shifts.

Furthermore, it should be emphasized that diatom remains were notably scarce in sediments above 59 cm (ca. 1962–2016 CE) in this study, including the top 10 cm (ca. 2008–2016 CE). However, field investigations of algal communities in Chagan Lake in 2008 (Sui et al., 2016) and another study in 2012 (Li et al., 2014) reported relatively abundant diatom populations, suggesting that the low diatom abundance in sediments may result from post-depositional dissolution processes rather than a decline in diatom productivity. Similar patterns have been documented in other saline-alkaline or alkaline lakes, where elevated pH, increased salinity, and intensified silica cycling can lead to significant dissolution of diatom frustules prior to their preservation in sediments (Lent and Lyons, 2001; Paull et al., 2017). These findings imply that the scarcity of diatom remains in Chagan Lake sediments may not reflect an actual reduction in diatom populations but rather a taphonomic bias caused by unfavorable preservation conditions. Future studies should integrate high-resolution environmental indicators (e.g., in situ salinity and dissolved oxygen monitoring) with modern diatom community surveys to further validate these mechanisms.