The study site (54° 0′22.10″N, 12° 6′59.76″E) is a fen peatland and located in the low-lying grasslands of Pölchow, Mecklenburg-Western Pomerania, Germany. The dominant vegetation contributing to fen peat formation includes sedge (Cyperaceae), alder (Alnus), and undecomposed wood (Wang et al., 2021). Historically, this landscape underwent its first drainage intervention in the 13th century. For several centuries, the site was subject to continued drainage, which gradually transformed the peatland and led to topsoil degradation, while the deeper subsoil layers remain comparatively less affected (Wang et al., 2021). Since the second half of the 20th century, the site has been used as grassland for hay production (Gosch et al., 2018) and is not currently undergoing active ecological restoration.

Undisturbed soil cores (7.2 cm diameter, 6.12 cm height) were sampled from a 10 m × 10 m plot using a randomised sampling approach. A total of 29 sampling points were selected within the plot. At each sampling point, two cores were collected. The topsoil core was extracted first at 10–20 cm followed by the subsoil core from the same vertical profile at 60–70 cm in depth (Supplementary Figure S1). To extract the core, a 50 cm × 40 cm area was excavated at each sampling point. A knife was then used to carefully extract the core to preserve structural integrity. This sampling campaign yielded a total of 58 independent cores (29 cores collected per depth), representing two soil distinct degradation stages [graded according to von Post (1922)]: a degraded top- (H10) and less degraded subsoil (H4).

The pore water extraction procedure involved two main steps: peat soil saturation followed by pore water extraction using the suction plate system. To saturate the soil, deionised water was incrementally added every 24 h for 4 days to saturate the soil cores, ensuring uniform moisture distribution throughout the process. Although deionised water has been shown to cause pore structure changes and reduce Ks in bog peat due to pore constriction (Kettridge and Binley, 2010), Gosch et al. (2018) observed no such effects in degraded fen peat from Pölchow. Their results suggest that, for this peat type, low salinity conditions do not significantly alter pore structure or hydraulic conductivity.

After saturation, the soil cores were transferred to the suction plates and subjected to two pressure heads: - 60 hPa for 4 days followed by - 600 hPa for another 4 days. Filter paper was placed on the suction plates, and the cores were positioned upright with their bottom side resting on the plates. Porewater samples were collected after each suction period and the core weights were recorded accordingly to calculate bulk density and total porosity.

The suction plate system, illustrated in Supplementary Figure S2, consisted of eight circular borosilicate glass suction plates (EcoTech Bonn; diameter: 7.5 cm; 10-μm membrane filter; PTFE tubing) arranged within a container box. Tubes from the suction plates were connected to a vacuum system with dual functions: (1) directing pore water from the plates to collection bottles and (2) maintaining a constant pressure.

Previous (Okruszko, 1993; Walczak et al., 2002) studies defined the boundary value between macropores and mesopores as 30 μm and between mesopores and micropores as 0.2 μm. In this study, we used a pore size classification following Tassinari et al. (2022) which included macropores (equivalent pore diameter >30 µm), mesopores (equivalent pore diameter 3–30 µm) and micropores (equivalent pore diameter 0.2–3 µm). Thus, the applied tensions of - 60 hPa and - 600 hPa were estimated to correspond to pore radii of approximately 30 μm and 3 μm, respectively. A detailed explanation of this subsection can be found in the Supplementary Material (Supplementary Section S1).

The soil cores were dried at 105°C to constant weight for determining dry bulk density. The loss on ignition (LOI) method was employed to estimate soil organic matter (SOM) content. Dried samples were burned at 550°C for 4 hours, following EN 13039:2011 guidelines. Porewater was collected and stored at 4° prior analysis when not measured immediately. Porewater samples were filtered using pre-washed regenerated cellulose syringe filters (0.45 µm pore size; Rotilabo^® mini-tip, Carl Roth GmbH + Co. KG, Schoemperlenstraße 3–5, D-76185 Karlsruhe, Germany). DOC was measured with the AJ Multi 2100S (Analytik Jena, Germany). The instrument determines dissolved inorganic carbon (DIC, acidification with 10% phosphoric acid followed by CO₂ detection via infrared spectroscopy) and total dissolved carbon (TDC, catalytic high-temperature combustion at 850°C using a platinum catalyst). DOC was calculated as the difference between TDC and DIC. Thus, each soil core produced two pore water samples drained at - 60 and - 600 hPa, respectively. For each soil core, SOM content, BD, macroporosity, mesoporosity and total porosity were determined.

Descriptive statistics were calculated for the physical properties and DOC concentrations. Box plots of the DOC concentrations were generated to depict distribution differences across pressure heads and soil depths. Paired Wilcoxon tests were performed to assess the significant differences in DOC concentrations between - 60 hPa and - 600 hPa within topsoil and subsoil groups, while Wilcoxon rank-sum tests compared DOC concentrations between topsoil and subsoil at each pressure head. To further investigate structural differences between pore groups, paired Wilcoxon tests were used to test for significant differences between macroporosity and mesoporosity within topsoil and subsoil samples.

To examine the relationship between SOM content and DOC concentrations at - 600 hPa, a Pearson correlation analysis was performed. All statistical analyses and graphical illustrations were processed in R language. The statistical tests were implemented using the stats package, while the box plots and regression plots were generated using the ggplot2 package.

The physical properties, namely, SOM content, macroporosity, total porosity and BD observations align with findings from previous studies conducted on samples from the same study site (Liu and Lennartz, 2015; Gosch et al., 2018; Wang et al., 2021). Bulk density is widely recognised as an indicator of peat degradation and decomposition (Liu and Lennartz, 2019). In our study, a negative linear correlation was observed between SOM content and bulk density for both topsoil (r = - 0.75; p < 0.0001) and subsoil (r = - 0.88; p < 0.0001; Supplementary Figure S3; Supplementary Table S2). Previous studies reported that natural peatlands exhibit macropores formed by plant debris, which, in turn, facilitate water flow and solute transport (Quinton et al., 2008; Liu and Lennartz, 2019). However, as peatland degradation advances, the breakdown of parent material into fine particles decreases SOM content. The peat matrix becomes increasingly compacted, leading to an increase in BD and a corresponding decrease in total porosity. We observed a similar trend where bulk density was negatively correlated with total porosity (r = - 0.90; Supplementary Table S1). This trend was highly pronounced in the degraded topsoil (0.75 ± 0.04 cm³ cm^–3) relative to subsoil (0.81 ± 0.10 cm³ cm^–3). Additionally, Liu and Lennartz (2019) pointed out that in degraded peatlands, secondary macropores (e.g., root channels) can maintain relatively stable porosity. Wang et al. (2021) and Liu and Lennartz (2019) observed that macroporosity decreased with increasing bulk density from 0.01 g cm^-3 to approximately 0.2 g cm^-3, after which it remained nearly constant. This trend is consistent with the findings of the present study (Figure 2).

The transition to lower bulk density with increasing depth along with higher SOM content (22.48 wt% - 83.36 wt%), reflects degradation processes consistent with prior studies. Notably, fen peat degradation decreases with depth (Liu et al., 2016). Generally, peatlands exhibit a decline in pore size distribution and pore connectivity with increasing depth due to organic matter decomposition (McCarter et al., 2020). However, in Mecklenburg-Western Pomerania, this pattern is reversed, where historical drainage and intensive agricultural practices led to topsoil degradation, and less adverse degradation is observed in the subsoil domain leading to a site-specific “inverted profile” of fen peatlands (Kleimeier et al., 2014; Liu and Lennartz, 2015).

Our results showed that porewater DOC concentrations were significantly elevated in the topsoil relative to the subsoil depth irrespective of pressure (Figure 1A). From a depth perspective, this distinction was more evident in the topsoil, where DOC concentrations extracted from mesopores (-60 hPa to - 600 hPa) were significantly higher than those from macropores (0 hPa to - 60 hPa), despite mesopores accounting for only 5.37% of the total porosity compared to 15.3% for macropores (Supplementary Figure S5a and S5c). This suggests that there may be a greater DOC concentration per unit pore volume from mesopores. In contrast, no significant differences in DOC concentration were observed in the subsoil, suggesting that DOC accumulation in macro- and mesopore regions is less pronounced in deep, less degraded peat layers. Despite the suction experiment’s pressure limitation (up to −750 hPa) preventing drainage from micropores, their significance to DOC accumulation cannot be excluded. Based on the porosity measurements, microporosity accounts for approximately 70% of the pore volume in both top- and subsoil depth. Previous studies (Walczak et al., 2002; Tassinari et al., 2022) reported that micropores dominate pore volume in peat soils and can retain water for extended periods (>30 days; (Tassinari et al., 2022). These conditions are likely to promote microbial colonisation and by extension, DOC accumulation. Considering our results, it is clear that DOC concentration in micropores may likely be significantly greater relative to the meso- and macropores pore classes and warrants further research.

Bulk density had a pronounced to moderate correlation with DOC concentrations extracted from macropore (r = 0.44) and mesopore (r = 0.57; Supplementary Table S1) regions, respectively. This observation is supported by Liu et al.’s (2019) findings, who reported that greater bulk density in peat samples was related to higher DOC pore water concentrations. In addition, the degraded topsoil domain showed elevated and relatively more variable DOC concentrations at greater bulk density (SOM <40 wt%) under both pressure heads (Figure 1B; Supplementary Figure S4). In contrast, DOC concentrations were lower and less variable (ranging between 30 and 80  mg L^-1) at a lower bulk density (SOM >40 wt%) in the subsoil domain. Two plausible mechanisms may explain this observation. First, during drainage and peat mineralisation, organic carbon compounds accumulate in the topsoils through chemical binding with both amorphous and crystalline ferric compounds (Zak et al., 2018). Under saturated conditions, the reduction of ferric to ferrous iron leads to the release of organic carbon previously adsorbed onto iron hydroxides, thereby increasing DOC concentrations in the topsoils (Selle et al., 2019). Second, the topsoil contains a greater proportion of labile carbohydrates, likely due to continuous inputs of fresh plant litter. These labile compounds contribute to the elevated DOC concentrations observed in the topsoil (Liu et al., 2019).

In our study, micropores and mesopores together constitute approximately 78% of the total porosity in both topsoil and subsoil. This dominance of finer pores is a key factor in water retention within the soil matrix, as these pores hold water at high tensions and contribute to longer residence times. Longer residence times in finer pores have been shown to facilitate localised microhabitats favourable for microbial life. As a result, finer pores can likely act as hotspots for DOC accumulation. In line with these findings, it is plausible that the elevated DOC concentrations extracted from the mesopores in the topsoil region in our study are at least attributable to microbial transformation of organic matter microbial activity and breakdown of Fe³⁺-organic matter complexes within mesopores (Figure 3) (Zak and Gelbrecht, 2007).

Therefore, our study reveals that the pronounced DOC variability observed (50–200 mg L^-1) in samples of lower SOM but higher bulk density (Figure 1B) and specifically the high DOC concentrations extracted at mesoporosity (-600 hPa) in the topsoil compared to subsoil (Table 1; Figures 1A,B), illustrate that DOC accumulation is strongly associated with the physical pore structure particularly pore size class, with mesopores (-600 hPa) serving as hotspots for DOC concentration in degraded fen peat soils.

Liu et al. (2019) and Zak and Gelbrecht (2007) reported that even after restoration, highly degraded peat can still exhibit elevated DOC in porewater. In reference to our results, given the correlation observed between DOC accumulation and mesopores in degraded topsoil, restoration strategies such as rewetting should not only target reduction of carbon emissions but also consider their effects on water-bound carbon export. While topsoil removal has been suggested as a measure to reduce nutrient export risk of adjacent aquatic systems (Zak and Gelbrecht, 2007; Emsens et al., 2015; Zak et al., 2018), it seems an inappropriate method at the scale of current rewetting projects.