This study was conducted for the period 2017–2021 in Mycklemossen mire and Erssjön lake (58°21′ N, 12°10′ E; 80 m a.s.l.), which are part of the Skogaryd research catchment in the hemiboreal forest zone in southwestern Sweden. The Skogaryd research catchment is part of the Swedish Infrastructure for Ecosystem Science (SITES). Air temperatures range between -0.8°C in February and 16.8 °C in July and precipitation is highest in July to December with approximately 75 mm per month and driest January to March with approximately 54 mm per month (Swedish Meterological and Hydrological Institute normal means for the period 1991–2020, Fig. S1 in the Supplement). During the survey period, the highest annual precipitation was in 2019 and 2020 with 1150 and 1100 mm annually, respectively. The lowest annual precipitation was 580 mm in 2017 (https://meta.fieldsites.se/resources/stations/Skogaryd, last access: 22 August 2025). Due to climate change, summers (June, July and August) at the west coast of Sweden are predicted to receive 7 % less rain in June–August, 13 %–18 % more rain in December to February and annually be 2.2–3.1 °C warmer by 2041–2060 in reference to preindustrial levels (IPCC, 2021).

Mycklemossen mire is a nutrient-poor fen with a surface topography composed of a mosaic of hummocks and hollows, referring to parts of the vegetation found above and below the water table respectively, and transition zones called “intermediate” in our study, with hummocks as the dominating topography (Rinne et al., 2022). The hummocks are mainly dominated by Eriophorum vaginatum, Calluna vulgaris and Erica tetralix and the hollows consist of different Sphagnum species, mainly S. rubellum, S. fallax and S. austinii as well as Rhyncospora alba, and the peat samples were sampled approximately within a square-meter plots (Kelly et al., 2021). Two streams lead away from Mycklemossen. One runs from the north into the lake Erssjön (this stream is hereafter referred to as S1) and the other runs into a neighbouring catchment. Erssjön is 0.061 km2, 4.5 m deep in the deepest part and has characteristics of a mesotrophic lake. Erssjön is sourced from Mycklemossen and from a neighbouring mostly forested catchment and drains through a stream (hereafter referred to as S6). The total catchment area of Mycklemossen (S1) is 0.595 km2 and the catchment of Erssjön (S6) is 1.337 km2 (Table S1 in the Supplement). The sizes of the catchments have been generated based on detailed topographic maps in relation to the location of the measuring stations (out-flow). The Skogaryd Research Catchment is located in one of the densest populated areas in Sweden and east of a highway and a former industrial ship wharf. In Europe, gasoline contained Pb from 1922 until it was banned in 1998. Mycklemossen is downwind from a potential source of industrial pollution, Uddevalla ship wharf. The ship wharf actively produced ships in the years between 1946 and 1985 and used Pb-containing antifouling paint. The area can therefore be expected to have been subjected to considerable Pb pollution during a period of 70 years.

On 3 March 2021, one peat core of 4.5 m depth was sampled at five locations in three topography types: hummock, intermediate and hollow with 1–2 m apart. This resulted in 15 peat cores in total. The five locations were chosen to represent Mycklemossen (Fig. 1). The top 50 cm of the peat was sampled with a hand saw. The deeper peat was collected with a stainless-steel, cylinder-shaped Russian peat corer (inner dimensions: 5 cm in diameter). Just after sampling, the peat cores were sliced into 5 cm segments and frozen.

Peat age was determined by radiocarbon dating in samples from 20–25, 120–125, and 420–425 cm depth from hummock, intermediate and hollow from sampling location 3 (Fig. 1c). Samples were freeze-dried and homogenized prior to 14C analysis at the Tandem laboratory (Uppsala, Sweden), using a MICADAS accurate accelerator mass spectrometer, following routine sample preparation by the tandem laboratory (Salehpour et al., 2013).

For all 15 peat cores (locations 1–5, Fig. 1c) the top 5 cm of each 100 cm increment from 0–5 cm below surface to 400 cm (0–5, 100–105, 200–205, 300–305, and 400–405 cm) were analysed for physical and chemical properties. pH was analysed in filtered peat pore water (Metrodam 691 pH Meter; Switzerland). Subsamples of peat were oven-dried at 40 °C until constant mass. Dry matter (DM) content was used to determine peat bulk density. Soil organic matter content (SOM) was determined from loss on ignition at 550 °C for 8 h in a muffle furnace (Heraeus Instruments). For total C, total N, δ13C and δ15N, dried samples were ground and weighed (5 mg) into tin capsules and analysed with a GSL elemental analyzer coupled to an isotope ratio mass spectrometer (IRMS, Sercon 20–22, Sercon Ltd., UK).

To map the three-dimensional layout of the mire, bathymetry, which is based on ground-penetrating radar was performed. The radar measurements were carried out with Malå Geoscience Ramac ground-penetrating radar with a shielded 250 MHz antenna. Measurements were collected at a sampling frequency of 2600 MHz and radar depth scans were made every 0.1 m in the different transects. For the radar measurements, the mire was divided with 4 lines from south to the north, and 10 lines east to west, forming a grid of 27 intersections, which were surveyed. The radar measurements were calibrated against 5 soil cores taken in grid intersections using a “Russian peat corer”. Bathymetry data was illustrated with QGIS 3.18 and the data is available on the SITES data portal (https://hdl.handle.net/11676.1/lmj-WqszXOi5T_XafhHU75Ps, last access: 27 February 2025). The bathymetry data were combined with peat bulk density and C and N content to calculate total C and N in increments of 50 cm depth. Specifically, total soil volume layer (m3) at each 50 cm increment was multiplied by depth-specific peat bulk density (gm-3) and the average fraction of C or N.

The concentration of Pb that is considered one of the highest ranked heavy metals to consider for public health (Tchounwou et al., 2012) and Fe were quantified in technical triplicates of the peat samples at 10–15, 15–20, 20–25, 120–125, and 420–425 cm depth with inductively coupled plasma-mass spectrometry (ICP-MS) (iCAP, Thermoscientific) (Table S2). Prior to analysis, peat samples were freeze-dried and homogenized with a mortar and pestle. A subsample of 0.3 g was weighed into a Teflon bomb and 10 mL of concentrated nitric acid (70 %) was added. The samples were heated in a microwave with a ramping phase of 20 min and a hold phase of 25 min at 180 °C. After cooling, the solutions were first diluted to 50 mL with MilliQ water and further diluted 10-fold before analysis.

Manual grab samples from the streams were taken fortnightly during the ice-free period for chemical analysis. Water samples were filtered through 0.7 µm pore size filters prior to analysis. DOC was measured via spectrometry Shimadzu TOC-VCPH. Lead and Fe concentrations were analysed using an Agilent 8800 Triple Quadrupole ICP-MS at the Microgeochemistry Laboratory, University of Gothenburg. Samples, reference waters, and blanks were prepared in 2 % HNO₃ with internal standards (Re, Rh, Ge) added for drift and matrix correction. The ICP-MS operated in MS/MS mode with N₂O as the reaction gas. Each analyte was analyzed in four replicates of 25 sweeps at three points per peak (0.02 a.m.u. spacing). Pb was measured on mass (206Pb) and Fe as oxide (56→72Fe), with dwell times of 0.5 and 0.3 s respectively. Quantification was done in the MassHunter Workstation Software v4.4. using a six-point calibration curve (0–1000 µgL-1) from the Periodic Table Mix 1 (TraceCERT®, Sigma Aldrich).

Discharge calculations were based on continuous stream level measurements using an ISCO 2110 ultrasonic flow module (Mire north, S1) and Mjk 1400 0–1 m (Erssjön outlet, S6), respectively. Location-specific stream level discharge relation curves have been established. The data were gap-filled when necessary either via linear interpolation for shorter gaps, or by cross-correlations with nearby measurement locations within the Skogaryd catchment for longer gaps. Longer gaps usually occurred during low-flow conditions. The export of DOC and metals were calculated for each manual water sampling as the product of discharge and concentration. Linear interpolations between sampling dates were done to estimate annual export rates.

To determine the source of Pb in the mire catchment the composition of Pb isotopes were determined in stream water from both the mire and lake from the manual grab samples collected between February 2022 and May 2024 (N=25) for the mire, and between May 2022 and May 2024 (N=21) for the lake, respectively. The Pb isotopic composition was analysed in the Microgeochemical Laboratory at the University of Gothenburg using an Agilent 8800 ICP-MS/MS coupled to an ASX-500 autosampler, a peristaltic pump, a concentric glass nebulizer (MicroMist) and a Scott double-pass quartz glass spray chamber. The autosampler was placed in an ISO 5 clean room and was connected to the ICP-MS in the adjacent laboratory with an approximately 1 m long tubing. All sample preparations were performed in the ISO 5 clean room using Milli-Q water and ultrapure nitric acid (NORMATOM^®, VWR chemicals). The ICP-MS was run in MS/MS mode with N2O as reaction gas. A comprehensive description of Pb isotopic analysis is given in the supplementary information together with a summary of ICP-MS/MS settings, acquisition parameters and reference materials used (Table S3–S5). Uncertainties of the Pb isotopic composition are reported at 2 s and include quadratic addition of the standard error for the 10 replicates, the excess scatter of the primary reference solution NIST SRM 981 for the respective ratio from the measurement session and the uncertainty in the 204Hg correction. The uncertainties from the published ratios for the primary reference material (approximately 0.04 %; NIST SRM 981; Cantanzaro et al., 1968) would need to be propagated to add systematic uncertainties. These uncertainties are considerably smaller than the random uncertainties (approximately 0.30 %) and thus do not significantly contribute to the total uncertainties when added in quadrature.

To determine the fate of anthropogenic Pb in the mire-lake complex, we applied an isotope mixing model (Eq. 1). Assuming Erssjön receives Pb from two different sources, water from Mycklemossen and other sources from the surrounding forest with distinct isotopic value the mixing model used is: 1RErsjön=f1×RMycklemossen+f2Rforest With R being the isotope ratio of Pb and f being the relative contribution of the two sources, with f1+f2 equal to 1. The model was applied both with the 206Pb/207Pb and 208Pb/206Pb ratios, using the IsoError tool (Phillips and Gregg, 2001) that provides mean contributions in addition to an uncertainty estimate. 11 samples from a forest stream within the Skogaryd catchment were analysed in 2023.

Data were analysed and visualized in R version 4.3.3 (R Core Team, 2024). For parameters measured in peat, we tested the effects of depth and topography type using mixed-effects models with the nlme package (Pinheiro et al., 2023). Depth, topography type and their interaction were included as fixed, independent predictors. For all dependent variables, except pH, depth was included as an orthogonal quadratic polynomial term. Sampling location was included as a random factor to take the nested sampling design into account. The relationships between annual export of DOC and annual export of Fe and Pb were tested with linear regression including DOC and stream identity as independent variables. For the regression model with Pb as dependent variable, there was a significant interaction, and the model was run separately for the two streams.

The maximum depth of Mycklemossen was about 600 cm (Fig. 2a) and the age of the mire centre was determined to be at least 3900 years by 14C dating at 420–425 cm depth (Table S6). The total surface area of the mire was 0.23 km2 in 2021. For the deepest part, the 550–600 cm depth interval, the area was only 0.0069 km2 which is 0.03 % of the total area. The total C pool of Mycklemossen was 34806±1633.23 t of which 18 % (6340±4.76 t) was in the top 50 cm of the mire (Fig. 2b). The C and N pool declined to 1570t±2.28 and 35t±48 at 550–600 cm depth, respectively.

Physicochemical parameters were measured to a depth of 400 cm. Peat bulk density (gcm-3) significantly decreased with depth from 0.085±0.023 at the surface to 0.04±0.009 at 200 cm depth and then increased to 0.064±0.011 to 400 cm (p=0.001, Fig. 3 and Table S7). The SOM content was above 90 % at all depths and locations and was highest in hummock. The SOM content for intermediate and hollow was lowest at the mire surface and at 400 cm depth, and highest at 200 cm depth (p=0.04, Table S7). Peat pH was lowest at the mire surface and increased with depth from the surface to 400 cm depth, spanning from a pH of 4.5±0.18 to 5.5±0.11 (Fig. 3). The increase in pH with depth was most pronounced for the hollow topography (p=0.04, Fig. 3 and Table S7). Peat C content increased slightly with depth from around 47±0.8% at the top to 51±1.9% at 400 cm, while N content decreased from 1.27±0.23% at the surface of the mire to 0.78±0.01% in 200 cm depth and increased to 1.59±0.21% at 400 cm depth (p=0.001, Table S7). The change in N with depth was less extreme for hummock compared to intermediate and hollow (p=0.04, Fig. 3 and Table S7). Peat δ13C increased from -27.5±1.5 ‰ at the mire surface to -26.03±0.74 ‰ at 100 cm depth followed by a decrease to -27.71±1.22 ‰ at 400 cm depth (p=0.001, Fig. 3 and Table S7). Peat δ15N decreased from -1.9±1.8 ‰ at the surface to -2.6±1.06 ‰ at 200 cm depth and then increased to 0.06±1.18 ‰ at 400 cm depth (p=0.001, Fig. 3 and Table S7).

Peat Pb content was highest in the top of the mire and was essentially absent in peat samples at 420–425 cm depth (Fig. 4). The highest Pb content was generally measured in intermediate, which was more than twice as high as in hollows, while the single highest measurement of Pb content was measured in hummock at 25–50 cm depth. Specifically in hummocks, the Pb content increased from 28.26 mgkg-1 in 15–20 cm depth to 91.8 mgkg-1 at 25–50 cm depth (Fig. 4 and Table S2). In intermediate and hollow topographies Pb content was highest at 15–20 cm: 64.25 and 32.21 mgkg-1 and decreased to 58.5 and 14.89 mgkg-1 at 25–50 cm depth, respectively (p=0.02, Fig. 4 and Table S8). Peat Fe concentrations at the top of the mire were between 606 and 1237 mgkg-1 and were measured to between 1434 and 1474 mgkg-1 at 120 cm depth (Fig. 4, Table S2). At 420–425 cm depth, Fe content was between 7147 and 8481 mgkg-1 and contents were similar between topography types.

The discharge from Mycklemossen (S1) was lowest during summer and highest in winter (Fig. 5). The water table concurrently decreased during summer and caused periods without discharge from S1. Stream water concentration of Pb, Fe and DOC in water from S1 were highest during summer at low discharge and low water table level (Fig. 5). The highest concentrations were in general measured during summer after a drought period when discharge became measurable again. Stream water pH increased at low water table depth during summer and decreased again in winter (Fig. S2). The seasonal tendencies in Fe, Pb and DOC concentrations seemed alike, but did not follow the same pattern all the way through the study period. Lead concentrations in S1 were highest in 2018 at the end of summer and after a long drought period and the Pb concentration measured was almost twice as high as in 2017 and 2019. During the study period, Pb concentrations varied between 0.4 and 14.3 µgL-1 (Fig. 5). Iron also peaked in 2018 at 4.6 mgL-1, which was clearly higher than the other years, whereas DOC peaked in the summer of 2019.

During the period 2017–2020 the average annual DOC export was 5834 kg from the north part of Mycklemossen (S1) and 12 616 kg leaving Erssjön (S6) (Table S1). The DOC export varied considerably between years and especially from Erssjön (Fig. 6). The highest DOC export from both Mycklemossen and Erssjön was in 2019, while the lowest export was in 2018. On average, annual DOC export from Erssjön in 2017–2020 was almost double of what it received from Mycklemossen, although this pattern was not consistent for individual years. In 2017 and 2018 DOC export from Mycklemossen and Erssjön was very similar, whereas in 2019 and 2020 export from S6 was around 3-fold higher than from S1. In 2017–2020, the average annual Fe export was 150 kg from S1 and 325 kg from Erssjön (S6) (Table S1). The annual Fe export from Erssjön was therefore 215 % higher than the input from Mycklemossen. The average annual export of Pb was 0.705 kg from Mycklemossen and 0.681 kg from Erssjön (Table S1).

Within streams, the inter-annual variation in Pb export mirrored the patterns observed for Fe and DOC. However, opposite to DOC and Fe, the export of Pb was only slightly higher from Mycklemossen than Erssjön. The annual export of DOC linearly correlated with annual Fe export in a similar way across the two streams (p<0.001, R2=0.96). DOC was also linearly correlated with Pb, but the correlation slopes differed between streams (p<0.001). Specifically, the annual DOC export in S6 was more than twice as high as the export from S1.

The Pb isotopic composition from the 57 water samples was analysed to determine whether the Pb exported from Mycklemossen and Erssjön was of anthropogenic origin (Table S9). The 208Pb/206Pb and 206Pb/207Pb composition from Mycklemossen (S1, n=25), Erssjön (S6, n=21) and forested catchment (s12, n=11) is plotted in Fig. 7, and for water leaving Mycklemossen, the 208Pb/206Pb ratio ranged between 2.106±0.004 (2se) and 2.117±0.006 (2se) and has a weighted mean ratio of 2.11175±0.00095 (n=25, MSWD=1.4). The 206Pb/207Pb ratio ranged between 1.143±0.004 (2Se) and 1.150±0.003 (2Se) and a weighted mean ratio of 1.1460±0.0008 (n=25, MSWD=1.5). For the water leaving Erssjön, the 208Pb/206Pb ratio ranged between 2.101±0.005 (2Se) and 2.109±0.010 (2Se) with a weighted mean ratio of 2.10690±0.0014 (n=21, MSWD=0.6). The 206Pb/207Pb ratio was between 1.149±0.004 (2Se) and 1.154±0.004 (2Se) with a weighted mean ratio of 1.1527±0.0009 (n=21, MSWD=0.4). For the forested neighbouring catchment, the 208Pb/206Pb ratio was between 2.062±0.009 (2Se) and 2.088±0.008 (2Se) with a weighted men ratio of 2.08010±0.0046 (n=11, MSWD=3.8). For the 206Pb/207Pb ratio, it ranged between 1.170±0.005 (2Se) and 1.191±0.004 (2Se) with a weighted mean of 1.778±0.0052 (n=11, MSWD=9). The isotope mixing model revealed that Mycklemossen contributed by 84.2±2.6% of the total Pb load to Erssjön (S6), despite the much lower total annual discharge (Table S1).

The peat in Mycklemossen was dated to 3900 years in 4 m depth and is thus likely closer to 6000 years in the mire centre at almost 6 m deep. This is within the age range of mires in Sweden (Ehnvall et al., 2024). However, Mycklemossen is in the younger end of the range and seems to have accumulated peat at a faster rate compared to other Swedish mires (Andersson and Schoning, 2010; Hansson et al., 2024). The difference in dominant vegetation of hollow and hummock likely controls the difference in C to N ratio of the top layer, as hummock was mostly dominated by vascular plants (Eriophorum vaginatum, Calluna vulgaris and Erica tetralix) and the hollows mainly consist of different Sphagnum species as S. rubellum, S. fallax and S. austinii as well as Rhyncospora alba (Kelly et al., 2021). The higher C to N ratio in hummocks is likely due to higher lignin content from the vascular plants, as Sphagnum does not contain lignin (Biester et al., 2014; Weng and Chapple, 2010; Zeh et al., 2020). This effect was also reflected in the δ13C in the top layer, with slightly more 13C depletion at the hummock with Calluna vulgaris than hollow with Sphagnum dominance (Biester et al., 2014). Furthermore, plant-sourced C near the surface is more 13C depleted nowadays than before the beginning of the industrial age (“the Suess effect”) (Drollinger et al., 2019; Esmeijer-Liu et al., 2012). Below the top layer and the water table (8–25 cm depth; Rinne et al., 2022) in the permanently anaerobic zone, the slightly higher 13C enrichment may also reflect methane production activity as methanogens discriminate against 13C, which consequently enriches the peat in 13C (Keane et al., 2021; Krull and Retallack, 2000). Further down, the decreasing C to N ratio likely reflects increasing degree of degradation with depth as C is removed (Biester et al., 2014; Smeds et al., 2025). While the 15N enrichment increased with depth, reflecting discrimination during peat decomposition (Biester et al., 2014), the 13C on the other hand decreased with depth. This latter can be explained by climatic changes/variation during the Holocene, in particular changes in soil moisture which in turn affects photosynthetic 13C discrimination and consequently peat 13C (Kumar Das et al., 2024; Nykänen et al., 2020).

The Fe content in the peat increased about 10-fold from the top of the mire to 400 cm depth, supporting that Fe in the Mycklemossen is primarily sourced from bedrock and groundwater and to a small extent from deposition (Steinnes et al., 2005). 18 % of Mycklemossen C pool was in the top 50 cm, partly because of the large mire area at this depth interval and a higher peat bulk density in the top one meter of the mire. This part of the mire was exposed to oxic conditions during summers, which makes a large part of the peat susceptible to decay (Fenner and Freeman, 2011). However, static oxic conditions during summer could also stabilise peat and DOC through adsorption and complexation interactions with Fe (Riedel et al., 2013; Song et al., 2022; Wang et al., 2017), though most Fe in Mycklemossen is placed in deep anoxic peat layers. Nevertheless, Fe-OM complexes are formed in both oxic and anoxic peat (Bhattacharyya et al., 2018). The stabilising effect of Fe might therefore not be limited to the oxic layer. However, in the presence of oxygen, oxidative reactions catalysed by Fe can lead to production of hydroxyl radicals that can promote degradation of peat (Qin et al., 2022; Trusiak et al., 2018). Such reactions might be driven by small concentrations of Fe (between 280–2.300 mgkg-1 peat; Curtinrich et al., 2024), which is within the range of Fe measured in the top part of Mycklemossen. In the top part of Mycklemossen mire there might therefore be both stabilising and destabilising reactions with Fe taking place, but it is unclear how important the destabilising reactions compared to the stabilising reactions is for peat stability and mire C dynamics.

Iron exhibited strong and linear correlations with DOC in the streams from Mycklemossen (S1) and Erssjön (S6). The average ratio between DOC and Fe of 0.03 in both S1 and S6 is the same as the median ratio value reported for Fe and DOC ratios in Swedish lakes from a national investigation (Weyhenmeyer et al., 2014). The Fe-DOC export relationship in the mire-lake complex of Mycklemossen and Erssjön therefore likely represents similar ecosystems in Sweden. In general, Fe in boreal rivers is primarily transported in colloidal form (between 1 kDa and 0.22 µm), for instance as Fe-DOC complexes with humic substances, and Fe as Fe2+ or Fe(oxy)hydroxides and to a much lesser extent found as “truly dissolved” (<1kDa) (Aleshina et al., 2024; Heikkinen et al., 2022).

In addition, water that leaves Sphagnum-dominated bogs is typically rich in polyphenolic DOM that is seemingly derived from Sphagnum and binds strongly to metals (Kaal et al., 2017). The mobility of the different forms of Fe is primarily affected by redox conditions (oxic/anoxic) and pH, whereas oxic conditions favor co-precipitation of DOC and Fe (Riedel et al., 2013) and acidic pH (4–5) favors soluble Fe-DOC complexes, while a pH of 6 and higher favors precipitation of Fe-DOC complexes (Neubauer et al., 2013). The Fe-DOC relationship reported in this study is probably best explained by colloidal transport of Fe-DOC complexes, with essentially all Fe bound to an organic ligand and a small proportion as Fe(oxy)hydroxides. Björnerås et al. (2021) investigated Fe inflow to a lake from a Mire in the south of Sweden. Here, Fe was found bound mainly in colloidal form, but also a noteworthy proportion as Fe(oxy)hydroxides, but the inflow water had a pH of 6.7 and above, which favors precipitation of Fe-DOC complexes (Björnerås et al., 2021). At the pH we measured (4–5), a higher fraction of soluble Fe-DOC complexes would be expected.

DOC production is highest under oxic conditions when the water table is low (Fenner and Freeman, 2011; Strack et al., 2008) and at high temperatures (Rosset et al., 2022). Thus, the notably warm and dry year of 2018, where the mire was exposed to oxic conditions for the longest period (Fig. S2) could therefore be expected to yield the highest DOC production also leading to the highest export of DOC together with Fe and Pb considering the strong correlations presented (Fig. 6), but that was not the case. The asynchronous timing of Fe (2018) and DOC (2019) peaks was probably due to a higher soil respiration of DOC in 2018. Both Fe and Pb concentrations were highest in 2018, so the peat degradation was likely highest that year. The year of 2018 was an extreme warm and dry summer in Scandinavia, and the soil respiration in the study site was 15 % higher compared to “normal” years (Keane et al., 2021) making it likely that the higher soil respiration was due to mineralization of DOC, leading the lower DOC peaks in 2018 compared to 2019. A higher peat degradation in 2018, which is likely the main mobilisation process of Fe and Pb in the top part of bogs (Broder and Biester, 2017), would also explain the higher release of Pb and Fe in 2018. The large DOC peak in 2019 is harder to explain. One possible explanation is that the stream was supplemented by DOC rich water from the catchment, as supported by Fig. S2, that shows the pH in the stream water in 2019 (pH 5–6.5) was notably higher than the stream water pH in 2018 in the summer period. The high pH is not what we expect from water coming from the mire. Furthermore, there was a much higher precipitation in 2019 compared to 2018 (Fig. 6f), that could lead to a higher runoff, rich in DOC, from the forested catchment. One point this study could not address is whether longer periods of drought, despite not increasing DOC export from Mycklemossen, increased the overall C loss through increased mineralization under oxic conditions (Fenner and Freeman, 2011).

In agreement with other studies, we observed highest Pb content in the top 50 cm which then dropped and reached a non-detectable content below 120 cm depth in all topographies (Klaminder et al., 2006; Nieminen et al., 2002). This indicates that Pb was deposited onto Mycklemossen via wet and dry deposition. Hummock forming species have higher biomass production rates (Malmer and Wallén, 1999), but typically degrade slower compared to hollow species, which is generally due to a higher content of recalcitrant components as lignin-like compounds and secondary metabolites in hummock (Limpens et al., 2017; Mäkilä et al., 2018; Turetsky et al., 2008). This is supported by the higher SOM and C content in the top of the mire compared to hollow (Fig. 3). Under the assumption that the hummock and hollow topographies have remained stable at least since the start of anthropogenic pollution (Nungesser, 2003), a slower decomposition of hummock material could be what led to a higher accumulation of Pb in the top part of Mycklemossen in hummock compared to hollow. Peat in hummock was too young to date with 14C at 20–25 cm and was likely younger than the ban of lead. Thus, it is likely that new hummock material formed on top of the polluted layer during the last few decades since Pb was banned from gasoline and antifouling paint, explaining why the highest concentration of Pb was measured in 25–50 cm depth in hummocks and not in 15–20 cm depth. The observed local differences in Pb content might lead to differences in peat decomposition depending on topography.

Lead is toxic for mosses (Choudhury and Panda, 2005) and hampers the growth of microbial communities associated with Sphagnum fallax (Nguyen-Viet et al., 2007). Lead contents of 80–300 mgkg-1 soil can affect microbial decomposition of C and nutrient turnover (Bååth, 1989; Zheng et al., 2017). The Pb content of 92 mgkg-1 in hummocks at 25–50 cm depth is thus likely to affect the microbial community and the biomass turnover rate. The studies by Nguyen-Viet et al. (2007); Bååth (1989), and Zheng et al. (2017) report effects of Pb based on contents (mgkg-1) and not bioavailable Pb. Therefore, the high Pb content measured in hummock in this study, could have a potentially harmful effect on microbes. The bioavailability of Pb can be assessed in different ways, but requires some form of bioassay (Fleming et al., 2013; Hoang et al., 2025). In peatlands, Pb is mostly found strongly bound to organic matter and minerals with a low available fraction and is generally very immobile due to high binding affinity for OM (Hou et al., 2019; Lu et al., 2025; Novak et al., 2011; Vile et al., 1999). The bioavailable Pb in hummock is therefore certainly lower than the 92 mgkg-1, but Pb can be made available by microbial decomposition and accumulate in organisms over time. Most studies that report toxic effects of Pb on microbial decomposition perform experiments in periods of weeks or months, but toxicity of Pb increases with exposure time (Connell et al., 2016). Therefore, the toxic effect of Pb could be more severe in long term in situ decomposition studies. In that regard, Sphagnum species associated with hummock commonly decompose slower than hollow species under the same environmental conditions, but to which extent contents of Pb, or other heavy metals, could affect decomposability has not previously been considered (Bengtsson et al., 2016; Johnson et al., 1990; Johnson and Damman, 1991). In Mycklemossen, we only detected contents of Pb exceeding the toxicity threshold in hummocks. Decomposition experiments with peat were not tested in this study, but different pollution loads between topographies and local degradation dynamics might be necessary to understand Pb cycling in mires exposed to Pb pollution.

The concentration of Pb in stream water from Mycklemossen was up to 10-fold higher compared to streams from pristine Siberian bogs located 200 km from anthropogenic sources of heavy metals (Kharanzhevskaya et al., 2023), and slightly lower compared to a bog in Germany exposed to pollution from a nearby shooting range (Broder and Biester, 2015). The average Pb concentrations in stream water from Mycklemossen in the period of 2017 to 2020 were 4-fold higher than the long-term accepted values of 1.2 µgL-1 and suggests that this stream, according to European standards, is of poor environmental quality (EU directive, 2013). In stream water from Erssjön, the concentration of Pb was drastically lower (1.1±0.1µgL-1), and can be considered good quality for long term exposure (EU directive, 2013). The highest concentration of Pb in stream water from Mycklemossen was measured in late 2018 and reached 14 µgL-1, which is the short-term maximum limit value for fresh water (EU directive, 2013). As already pointed out, the high Pb concentration in S1 (14 µgL-1) in 2018 was likely due to a higher degree of Pb mobilisation (Broder and Biester, 2017). Beside the European standard for Pb thresholds, a comprehensive analysis taking bioavailability into account and including 7 European surface water scenarios and 25 freshwater species, found Pb toxicity thresholds within 6.1 and 31 µgL-1 (Van Sprang et al., 2016). Therefore, Pb in S1 reached concentrations on short-term basis that can be considered potentially harmful for freshwater fauna. Longer periods of drought in a future climate could lead to higher mobilisation of Pb and thus stream water concentrations of Pb exceeding toxicity thresholds (Szkokan-Emilson et al., 2013).

While twice as much Fe and DOC was exported from Erssjön through S6 compared to S1 leaving Mycklemossen, which roughly follows differences in catchment size, the export of Pb was similar between the two streams in the period 2017 to 2020. At the same time, about 2-fold more Pb per amount of DOC left S1 compared to S6. The main explanation is probably found in how Pb in the two streams was sourced. The expected 206Pb/207Pb ratio in the area around Skogaryd is in the range of 1.245–1.326 (Reimann et al., 2012), which is higher than the data from Mycklemossen of 1.146. The value at Mycklemossen is at the lowest end for Northern Europe (Reimann et al., 2012) and is indicative for anthropogenic pollution (Fig. 7). For the 208Pb/206Pb ratio the average value of 2.112 for Mycklemossen is close to the median for Northern Europe (Reimann et al., 2012). Generally, the Pb isotope ratios revealed that Pb in S1 and S6 primarily was sourced from gasoline (Fig. 7), but the Pb isotopic signature in S6 was slightly shifted towards the signature of forest soil which overlaps with European coal, marine antifouling paint and natural geogenic sources. As isotope signature of marine antifouling paint falls on the mixing line of geological and gasoline Pb, its contribution is unclear. Erssjön also receives water from other land-uses (forest) within its catchment, which could be dominated by natural Pb isotopes, which supplies Pb to S6 at lower concentrations than S1. Except for one sampling, the 208Pb/206Pb ratio is always lower in Erssjön than Mycklemossen, indicating that the forest land almost always contributes to the lead in Erssjön. No previous estimates of the water residence time have been made but given the size of the Erssjön (6.2 ha) and the mean depth of 1.7 m (Milberg et al., 2017) the volume would be ∼100000m3, which gives a mean residence time of about 3 months. Conducting mixing model on the individual sampling dates (N=10), Mycklemossen contributed by 68–100 % to the lead in Erssjön outflow. Furthermore, the mixing model showed that around 19 % of the Pb coming from Mycklemossen was retained in Erssjön, indicating that Erssjön acts as a sink for the Pb from Mycklemossen. Thus, the similar Pb export from S1 and S6 is likely explained by an additional Pb input from the forested catchment to S6 of a lower concentration than what Mycklemossen supplies, and sedimentation/retention of Pb in Erssjön, which we estimated to be around 0.132 kgyr-1.