2.1 Sediment Coring and Sampling

The Upo Wetland, located across Yuo-myeon, Ibang-myeon, Daehap-myeon, and Daiji-myeon in Changnyeong-gun, Gyeongsangnam-do, is the oldest inland wetland in Korea and the largest natural wetland in the country, covering a total surface area of approximately 2.3 km² (230 ha) ^{(Kang et al., 2007; Lee et al., 2020; Lee et al., 2011)}. The wetland complex consists of four interconnected sub-basins―Upo, Mokpo, Sajipo, and Jjokjibeol―of which Upo, the largest basin, represents the main body of the wetland ^{(Lee et al., 2020; Lee et al., 2011)}.

The Upo Wetland was designated as a Natural Reserve in 2011 and subsequently designated as a Wetland Improvement Area in 2012, when the officially protected wetland area was expanded from 8.54 km² to 8.547 km^{2 (Lee et al., 2020; Lee et al., 2011)}. In March 1998, it became the second site in Korea to be registered under the Ramsar Convention as a habitat for migratory waterbirds ^{(Lee et al., 2020; Lee et al., 2011)}.

To compare the characteristics of organic carbon storage and diatom abundance according to depositional environments at different depths and locations, four sediment cores were collected from the Upo Wetland at sites UPW02–UPW05 (Table 1). The UPW02 site was in a shallow marginal zone with relatively low water depth, whereas UPW03, UPW04, and UPW05 were positioned in the central basin of the wetland, representing deeper depositional environments (Fig. 1).

Coring was conducted in May 2023 using a double-tube rotary corer equipped with an NX-size barrel (outer diameter = 75.4 mm; inner core diameter = 54.7 mm), which was selected to maximize core recovery and minimize disturbance in the soft, water-saturated sediments of the Upo Wetland. The apparatus consisted of a steel outer barrel and a polyvinyl chloride (PVC) inner liner to ensure stable recovery of unconsolidated sediments. The collected sediment cores reached depths of 9.4 m (UPW02), 7.2 m (UPW03), 8.4 m (UPW04), and 7.2 m (UPW05), respectively. Immediately after coring, each core was vacuum sealed in waterproof plastic bags to prevent desiccation and oxidation and subsequently transported to the laboratory for further analysis. In the laboratory, each core was split lengthwise into two halves: one preserved for archival storage and the other used for analysis. Sedimentological descriptions and subsampling were performed on the working half of each core following standard procedures ^{(Cohen, 2003; Last and Smol, 2002)}.

2.2 Diatom Analysis

Subsamples for diatom analysis (∼1 g each) were taken at 10 cm intervals from the split sediment cores. In the laboratory, approximately 20 g of wet sediment was weighed using a microbalance (OHAUS Co., USA) and transferred to a 50 mL conical tube. Each tube was then filled to 50 mL with Ludox-HS 40 solution (Sigma-Aldrich, USA) and stirred thoroughly to disperse sediment particles and expose diatom cells to the solution. The mixture was centrifuged at 2,000 rpm for 15 min to separate the diatom frustules from heavier mineral particles. The supernatant containing diatom cells was poured through a 20 µm sieve to collect the diatom fraction, which was subsequently rinsed several times with distilled water to remove any remaining Ludox solution. The cleaned diatom suspension was concentrated to approximately 10 mL with distilled water and stored in 20 mL amber glass vials prior to mounting.

Fig. 1. Location map of the Upo Wetland and sediment core sampling sites (UPW02–UPW05).

For quantitative microscopic observation, 1 mL of the concentrated diatom suspension was pipetted into a Sedgwick–Rafter (SR) counting chamber, and all 1,000 grids were examined. To remove organic matter, additional subsamples were oxidized with 30 mL of 30% hydrogen peroxide (H₂O₂) and rinsed five times with distilled water. The cleaned material was mounted in Pleurax (Mountmedia, Wako, Japan) for observation using a light microscope (LM) (Eclipse Ni, Nikon, Japan) equipped with Nomarski differential interference contrast (DIC) optics. Observations were conducted at magnifications of ×400 and ×1,000, and digital photomicrographs were taken with a DS-Ri2 camera (Ni2, Nikon, Japan). For ultrastructural examination of fine morphological features not visible under the LM, hydrogen-peroxide-cleaned samples were filtered onto 2.0 µm polycarbonate membrane filters (Nuclepore, Whatman, UK), mounted on aluminum stubs, and sputter-coated with gold–palladium. These were analyzed using a field emission scanning electron microscope (FE-SEM) (MIRA 3, TESCAN, Czech Republic).

2.3 Total Organic Carbon (TOC) Analysis

The collected sediment cores stored frozen until analysis. The frozen cores were split longitudinally, and sub-samples were obtained from one half of each core, Sub-samples were transported in a frozen state to the Isotope Ecology and Environmental Science Laboratory at Hanyang University. Samples were freeze-dried and homogenized using a mortar and pestle. Approximately 150 mg of the homogenized material was used for pretreatment prior to TOC analysis. The samples were decalcified with 1N HCl (12 h) to remove inorganic carbon, rinsed with triple-distilled water, and freeze-dried again. The dried samples were packed into tin capsules, with 20 - 30 mg of each weighed for analysis. TOC contents were determined using an elemental analyzer (Vario Select, Elementar, Germany) combined with a continuous-flow isotope ratio mass spectrometer (VisION, Elementar, Germany). The IAEA CH-3 standards were analyzed every 12 samples for calibration and quality control.

2.4 Statistical Analyses of TOC–Diatom Relationship

Statistical analyses were conducted to examine the spatial and vertical variations in the relationship between TOC content and diatom cell abundance in surface sediments. The data were first tested for normality and homogeneity of variances using Shapiro–Wilk and Levene’s tests, respectively.

Spearman’s rank correlation analysis was then used to assess the association between TOC and diatom cell counts for the entire dataset, as well as within each site and depth stratum. A two-way ANOVA was performed to evaluate the effects of site and depth, as well as their interaction, on TOC and diatom cell count. This was followed by a Tukey’s HSD post hoc comparison to identify specific group differences. When the assumptions of parametric tests were not satisfied, the nonparametric Scheirer–Ray–Hare test was applied as an alternative two-factor analysis. To further assess how the TOC– ‘diatom cell counts’ relationship varied among sites and depths, an analysis of covariance (ANCOVA) was performed, with TOC as the dependent variable and ‘diatom cell counts’ as a covariate, while site and depth were treated as categorical factors. Additionally, a linear mixed-effects model was fitted, treating depth as a random factor (TOC∼ ‘diatom cell counts’ × Site + (1 | Depth)), to account for nested variability among depth layers. All analyses and graphical visualizations were conducted in R (version 4.5.2) using the dplyr, ggplot2, car, lme4, broom and companion packages. Pairwise comparisons of diatom cell count among sampling sites were conducted based on estimated marginal means using post hoc t-tests.

3.1 Vertical and Spatial Variations in Sedimentary Characteristics

Four sediment cores (UPW02–UPW05) collected from the Upo Wetland exhibited distinct vertical and spatial variations in color, grain size, and organic carbon content (Figs. 2, 3). The sediments were mainly composed of fine silt and clay, reflecting the low-energy depositional environment characteristic of lacustrine basins.

The upper layers of all cores showed higher organic content and finer textures, suggesting enhanced primary productivity and reduced oxidation during more recent depositional periods. Such vertical transitions in facies correspond well with previous studies that interpreted the Upo Wetland as a Holocene lacustrine–palustrine system formed through interactions among sediment supply, hydrology, and water-level fluctuations ^{(Lee et al., 2020; Nahm et al., 2005)}.

Fig. 2. Stratigraphic sections of sediment cores (UPW02–UPW05) from the Upo Wetland. Each core profile shows vertical variations in lithology and sediment color. Gravel deposits are predominant at the shallow marginal site (UPW02), whereas relatively fine-grained sediments are more extensively developed in the central basin cores (UPW03–UPW05).

3.2 Vertical and Spatial Variations and Correlations between TOC and Diatom Abundance

Depth profiles of TOC (%) and diatom cell abundance (cells g^-1) were obtained from the four sediment cores (Fig. 3). One core (UPW02) represented the shallow marsh, and three cores (UPW03–UPW05) represented the deep marsh of the wetland. Both TOC and diatom abundance exhibited pronounced vertical variations across all sites.

At the shallow marsh (UPW02), the upper sediments exhibited the highest TOC values (up to 5%) and diatom densities (up to 3 × 10⁷ cells g^-1) gradually decreased with depth (Fig. 3A). In contrast, the deep marsh cores (UPW03–UPW05) exhibited generally lower TOC (< 2.5%) and reduced diatom abundance (Fig. 3B–D). The vertical patterns were more uniform or patchy in these deeper cores, reflecting differences in hydrodynamic stability and productivity. Horizons with elevated diatom cell abundance often coincide with peaks in TOC, indicating a close association between diatom productivity and sedimentary organic carbon accumulation.

Fig. 3. Vertical profiles of total organic carbon (TOC, black line) and diatom cell count (green area) at four sampling sites (A: UPW02, B: UPW03, C: UPW04, D: UPW05). All variables are plotted against sediment depth (m).

Scatter plots revealed spatial variability in TOC – ‘diatom cell counts’ relationship (Fig. 4). A strong positive correlation was identified in the shallow marsh (UPW02; R² = 0.54), while weaker relationships were observed in the central basins―UPW03 (R² = 0.59), UPW04 (R² = 0.43), and UPW05 (R² = 0.05). These patterns suggest that the strength of coupling between TOC and diatom abundance depends on local depositional and ecological conditions. The highest values and strongest covariation occurred at the shallow site, implying more active biogenic carbon input associated with enhanced diatom growth.

Fig. 4. Relationship between TOC and diatom cell count across the four sampling sites (A: UPW02, B: UPW03, C: UPW04, D: UPW05). Solid blue lines indicate the linear regression fits with the 95% confidence intervals (gray dashed lines). The regression equations and coefficients of determination (R²) are shown for each site.

The correlation between TOC and diatom abundance at each site suggests that diatom productivity plays a significant role in organic carbon accumulation in the Upo wetland sediments. This is particularly evident in the marked covariation between TOC and diatom cell density in the shallow marsh (UPW02; R² = 0.54) (Fig. 4.). This close correlation suggests that microalgal blooms may periodically stimulate the inflow of biogenic carbon into the sediment reservoir. This interpretation is consistent with previous studies highlighting the importance of microalgae in carbon fixation and sequestration through photosynthesis ^{(Bhola et al., 2014; Ighalo et al., 2022)}.

The differences in TOC content, diatom abundance, and their correlations between shallow and deep marshes reflect water column depth effects (Table 2, 3., Table S1, S2). The longer sediment remains suspended in the water column, the more it undergoes biodegradation, uptake, and various chemical and biological tansformations ^{(Arndt et al., 2013; Ramondenc et al., 2024)}. Consequently, sediments that sink more quickly are more likely to be deposited rapidly. Therefore, the difference in TOC content between shallow marshes and deep marshes is attributed to shallower water depths. These results suggest the need to consider water depth when utilizing wetlands for teal carbon.

TOC content and diatom cell counts generally decline with sediment depth, indicative of temporal and environmental changes in organic matter accumulation and decomposition (Fig. 3). The surface layer (0-5 cm) is supplied by phytoplankton deposited from the water column. Consequently, TOC is high and diatom frustules are well preserved. However, as the sediment depth increases, the sediment ages and organic matter loss due to microbial decomposition accumulates, resulting in a decrease in organic matter concentration and diatom counts ^{(Bao et al., 2021; Bhattacharya et al., 2021)}. This decreased is maintained by microbial residues of heterotrophic bacteria and archaea or refractory organic matter from terrestrial or humus sources ^{(Cai et al., 2019; Oni et al., 2015)}. Ultimately, the simultaneous decrease in TOC and diatom cell counts with depth reflects a continuous sedimentation-decomposition process: high surface productivity and fresh organic matter supply → loss of organic carbon through burial and decomposition → preservation of refractory organic matter at stable depths ^{(Cai et al., 2019; Mackay et al., 2012; Oni et al., 2015; Stief et al., 2013; Veuger and van Oevele, 2011)}.

3.3 Statistical Evaluation and Implications for Carbon Sequestration

Nonparametric analyses confirmed significant relationships between TOC and diatom abundance (Spearman ρ = 0.46, p = 0.001). Site-specific correlations ranged from ρ = 0.38–0.52 (p < 0.05), and depth-stratified analyses showed similar trends (ρ = 0.40–0.49). The Shapiro–Wilk tests indicated non-normal data distributions, whereas Levene’s test showed no significant heterogeneity of variance among sites and depths (p > 0.3). Two-way ANOVA results demonstrated that both site (F = 3.21, p = 0.028) and depth (F = 4.85, p = 0.011) significantly affected TOC, while Scheirer–Ray–Hare tests revealed significant site (H = 8.13, p = 0.043) and depth (H = 12.55, p = 0.002) effects for diatom abundance. ANCOVA further confirmed an interaction between site and depth on the TOC–diatom relationship (F = 7.33, p = 0.008), consistent with the linear mixed-effects model (LMM) results (t = 3.44, p = 0.001).

Pairwise comparisons (Table 3) showed that diatom cell abundance at the shallow marsh (UPW02) was significantly higher than that at the deep marsh sites UPW03 (t = 3.62, p < 0.01) and UPW05 (t = 3.77, p < 0.01), with a marginal difference from UPW04 (t = 1.77, p ≈ 0.05). No significant differences were detected among the three deep marsh sites (p > 0.5), indicating generally homogeneous depositional conditions within the central basin.

The spatial contrast between the shallow and deep zones highlights the influence of geomorphology and hydrology on carbon accumulation within the wetland. Higher TOC and diatom densities in UPW02 likely reflect stronger nutrient influx, greater light availability, and periodic resuspension, all of which enhance microalgal productivity and promote the burial of biogenic organic matter. Conversely, lower values in the central basins correspond to reduced primary productivity and more stable, mineral-dominated sedimentation.

These findings support the role of microalgae―particularly diatoms―as key biological agents of Teal Carbon sequestration in inland wetlands. Similar biogenic processes have been reported from coastal Blue Carbon ecosystems. In Korean tidal flats and Chinese intertidal zones, microphytobenthos―including benthic diatoms―enhance organic carbon accumulation through high primary productivity and sediment-binding capacity ^{(Chen et al., 2020; Lee, Kim et al., 2021)}. In the present study, the significant co-variation between sedimentary TOC and diatom abundance across different depths and sites in the Upo Wetland suggests that microalgal productivity may likewise contribute to carbon accumulation in freshwater inland wetlands. These findings provide ecological parallels between coastal Blue Carbon systems and inland Teal Carbon environments, highlighting a common microalgae-associated pathway for sedimentary carbon storage. The tight coupling between diatom abundance and TOC supports previous evidence that diatom-derived organic matter constitutes a significant portion of sedimentary carbon pools in low-energy freshwater environments ^{(Cohen, 2003; Round et al., 1990)}. Moreover, the spatial heterogeneity observed within the Upo Wetland underscores the importance of integrating geomorphological and hydrological variability into carbon budget assessments. This study provides a foundational dataset for improving national-scale greenhouse gas inventory frameworks and developing Teal Carbon assessment methodologies applicable to Korean freshwater ecosystems ^{(IPCC, 2013)}.

Furthermore, Teal Carbon―representing carbon stored in freshwater inland wetlands―constitutes a globally significant carbon sink, storing approximately 30% of global soil organic carbon despite wetlands comprising only about 6% of the land area ^{(Kumar et al., 2025; Nahlik et al., 2016;} Kumar). This highlights the critical importance of conserving and restoring such wetlands as part of national and international climate mitigation strategies ^{(Zhang et al., 2025)}. The current dataset from Upo Wetland contributes valuable baseline information toward establishing comprehensive greenhouse gas inventories and carbon offset frameworks specific to Korean inland wetland ecosystems.

As primary producers, diatom plays a vital role in carbon fixation, being responsible for approximately 20% of global atmospheric CO₂ assimilation. Their unique biological mechanisms—including specialized proteinaceous shells (PyShell) and the ability to concentrate CO₂ to high levels—help Rubisco, the carbon-fixing enzyme in diatoms, function more effectively ^{(Catherall et al., 2025; Shimakawa et al., 2024)}. As a result, diatoms synthesize more organic matter (carbon), and the increase in TOC in sediments as a byproduct of photosynthesis is closely related to the abundance of diatoms.

When normalized to the wetland area, the estimated carbon accumulation in Upo sediments aligns with the lower end of global Teal Carbon sequestration rates (126–217 g C m^⁻2 yr^⁻1; ^{Kumar et al., 2025)}. Although small in areal extent, the Upo Wetland thus represents a micro‐scale analogue of inland freshwater systems that collectively store about 452–524 Pg C worldwide. This highlights the significant but often overlooked contribution of small inland wetlands to regional carbon budgets and to Korea’s emerging Teal Carbon framework.

Lastly, the observed spatial variability, driven by geomorphological gradients and hydrological dynamics, emphasizes that wetland carbon storage is not uniform but is influenced by site-specific sedimentation regimes. Accounting for these factors is essential to model carbon budgets accurately and to develop effective management plans aimed at maximizing carbon sequestration in freshwater wetlands.