Is bog water chemistry affected by increasing N and S deposition from oil sands development in Northern Alberta, Canada?

Nitrogen and sulfur emissions from oil sands operations in northern Alberta, Canada have resulted in increasing deposition of N and S to the region’s ecosystems. To assess whether a changing N and S deposition regime affects bog porewater chemistry, we sampled bog porewater at sites at different distances from the oil sands industrial center from 2009 to 2012 (10-cm intervals to a depth of 1 m) and from 2009 to 2019 (top of the bog water table only). We hypothesized that: (1) as atmospheric N and S deposition increases with increasing proximity to the oil sands industrial center, surface porewater concentrations of NH4+, NO3−, DON, and SO42− would increase and (2) with increasing N and S deposition, elevated porewater concentrations of NH4+, NO3−, DON, and SO42− would be manifested increasingly deeper into the peat profile. We found weak evidence that oil sands N and S emissions affect bog porewater NH4+-N, NO3−-N, or DON concentrations. We found mixed evidence that increasing SO42− deposition results in increasing porewater SO42− concentrations. Current SO42− deposition, especially at bogs closest to the oil sands industrial center, likely exceeds the ability of the Sphagnum moss layer to retain S through net primary production, such that atmospherically deposited SO42− infiltrates downward into the peat column. Increasing porewater SO42− availability may stimulate dissimilatory sulfate reduction and/or inhibit CH4 production, potentially affecting carbon cycling and gaseous fluxes in these bogs. Supplementary information The online version contains supplementary material available at 10.1007/s10661-021-09555-4.


Introduction
Peatlands, ombrotrophic bogs and minerotrophic fens, cover 365,157 km 2 across northern Alberta, Saskatchewan, and Manitoba, Canada, with non-permafrost bogs covering 27,397 km 2 of the high boreal and midboreal regions . Most of these peatlands have persisted over 6000-7000 years  in areas of minimal human disturbance in a boreal continental climate with low atmospheric deposition of N and S (< 2 kg N or S ha −1 year −1 ). As is typical of bogs worldwide, Alberta bogs have acidic surface porewaters, with low concentrations of dissolved inorganic N, P, S, and base cations (Li & Vitt, 1997;Malmer et al., 1992;Vitt & Wieder, 2007;Vitt et al., 1995;Zoltai & Vitt, 1995).
Oil sands development in northern Alberta has grown markedly since 1967 when the first oil sands mine began operations. Annual emissions of gaseous nitrogen oxides (expressed as NO 2 ) from upgrader stacks and dieselfueled mine fleets have increased over time, reaching about 98,000 metric tonnes in 2019; gaseous SO 2 emissions peaked in 2009 at about 120,000 metric tonnes and decreased markedly to about 49,000 metric tonnes in 2014, flattening out at about 45,000 metric tonnes through 2019 (updated from Davidson & Spink, 2018;Wieder et al., 2021; Tables S1, S2). These emissions have led to wet and dry N and S deposition that is highest nearest to the oil sands industrial center and decreases with distance (Edgerton et al., 2020;Fenn et al., 2015;Hsu et al., 2016;Wieder et al., 2016a, b).
In 2009, we began monitoring bogs in northern Alberta for potential responses to changing atmospheric N and S deposition regimes, with sites situated at different distances from the oil sands industrial center. Bog porewater chemistry was one of the response variables included in the monitoring program. We hypothesized that: (1) as atmospheric N and S deposition increases with increasing proximity to the oil sands industrial center, surface porewater (at the top of the bog water table) concentrations of NH 4 + , NO 3 − , DON, and SO 4 2− would increase following a threshold response (cf. Lamers et al., 2000) and (2) with increasing N and S deposition, elevated porewater concentrations of NH 4 + , NO 3 , DON, and SO 4 2− would be manifested increasingly deeper in the peat profile. To test these hypotheses, we report withinyear, between-year, between-site, and depth (10-cm intervals to a depth of 1 m below the peat surface) patterns in porewater concentrations of NH 4 + , NO 3 − , DON, and SO 4 2− to a depth of 1 m at 5 peatlands from 2009 to 2012, and within-year, between-year, and between-site patterns in surface porewater chemistry at the top of the bog water table at 7 peatlands from 2009 to 2019.

Study sites
The peatlands selected for monitoring ranged from 11 to 77 km from the heart of the oil sands mining activity, designated as the midpoint between the Syncrude and Suncor upgrader stacks (Fig. 1). Six of the sites are ombrotrophic bogs, with acidic porewaters (Table 1), a nearly continuous cover of Sphagnum mosses (S. fuscum and S. capillifolium on hummocks, S. angustifolium in hollows), and an abundance of Rhododendron groenlandicum, Vaccinium oxycoccos, and Vaccinium vitis-idaea. Porewater at the top of the water table at the Mildred site has a circumneutral pH, although the site has ombrogenous hummocks dominated by S. fuscum and/or S. capillifolium, and a species composition similar to bogs Wieder et al., 2021). Recognizing the unique nature of the Mildred site, we will refer to all sites as bogs. Four sites were established for the original project in 2009 (Mildred, McMurray, McKay, and Anzac) with a fifth added in 2010 (JPH4). The Mildred site was slated for demolition and was decommissioned in 2013. The Horse Creek and Kearl Lake sites were added for sampling in the 2018 and 2019 seasons. Weather Mean daily temperature and daily precipitation were obtained for each site from 1989 to 2019 from Interpolated Weather Data Since 1961 for Alberta Townships (https:// acis. alber ta. ca/ towns hip-data-viewer. jsp). To assess meteorological drought conditions over the 30-year study period, we used the standardized precipitation index, using data from the Alberta government Mildred Lake Station (57.03°N, 111.45°W; https:// acis. alber ta. ca/ weath er-data-viewer. jsp) and the DrinC package (Tigkas et al., 2015, www. droug ht-softw are. com). Wind roses were generated using wind speed and wind direction data from the Mildred Lake Station using WRPLOT View, version 8.0.2 (Lakes Environmental Software, Waterloo, Ontario, Canada).

Sampling and analysis
At each site, we installed three, 1-m long porewater samplers, each consisting of 10 10-cm sections of thinly slotted 2.5 cm-diameter PVC pipe, with sections separated from each other by a polypropylene disk. Color-coded Tygon® tubing extended from each depth section to the peat surface. The samplers allowed us to collect peat porewater from discrete 10-cm depth intervals from the top of the water table to 1 m below the peat surface. Using a 60-mL  Wieder et al., 2016b -S (DIONEX ion chromatograph), and DON (dissolved organic nitrogen) (Shimadzu TOC-V CSH analyzer and TNM-1 total nitrogen detector, with prefiltration through 0.45-μm filters).
To simplify visual presentation of the 2009-2012 data, we prepared 2-dimensional plots showing changes in porewater chemical parameters as a function of depth below the water table and time using PROC G3GRID in SAS with spline interpolation, a smoothing factor of 0.05, and gridding at 5-cm depth intervals and 15-day time intervals. Interpolation was temporally bounded by the first and last sampling date of each year. Resulting grid data were used to generate contour plots in SigmaPlot 14.0. The position of the water table was inferred from the topmost segment of the samplers from which water could be collected on a given sampling date.
The standardized precipitation index (SPI; Guttman, 1999;McKee et al., 1993), an algorithm that calculates the number of standard deviations that observed precipitation deviates from the climatological average, also showed considerable variability over the 11-year monitoring period (Fig. 2). Of particular note is that the 2011 growing season had SPI values that were less than − 1 for every month, and this was preceded by a 6-month non-growing season in which monthly SPI values were negative in each month, indicating a year of relatively dry conditions.

N and S in bulk deposition
Across the seven sites, NO 3 − -N and SO 4 2− -S, but not NH 4 + -N, bulk deposition (kg ha year -1 ) and concentrations (mg L -1 ) in bulk deposition, show a general pattern of decreasing with increasing distance from the oil sands industrial center (Table 1) (Table S5). In general, NH 4 + -N concentrations increased with depth, although this increase was more pronounced at JPH4, McKay, and Anzac than at Mildred or McMurray  Guttman, 1999;McKee et al., 1993), calculated using data from the Alberta government Mildred Lake Station. Green shaded regions indi-cate growing season months (May-October). Numbers across the top are mean SPI values within a given 6-month period  Table S6). Although porewater NH 4 + -N concentrations exhibited some within-year seasonal variation, patterns were not consistent across all sites (Fig. 3).
Porewater NO 3 − -N concentrations varied between sites, years, and depth ( Fig. 4), exhibiting significant site by year (p < 0.0001) and year by depth (p = 0.0111) interactions (Tables S4, S7, S8). There was no clear pattern of NO 3 − -N concentrations decreasing with distance from the oil sands industrial center (Table S7). At all sites, porewater NO 3 − -N concentrations were substantially higher in 2011 than in the other 3 years (Table S7). No clear pattern of changing porewater NO 3 − -N concentrations with depth was observed at any of the sites; porewater NO 3 − -N concentrations in the uppermost peat did not decrease with increasing distance from the oil sands industrial center (Fig. 4, Tables S7, S8). Although there was within-year seasonal variation in NO 3 − -N concentrations, these patterns were not consistent across all sites (Fig. 4). Averaged across all sites, depths, and sampling dates, porewater NO 3 − -N concentrations were equivalent to 14% of porewater NH 4 + -N concentrations. Porewater DON concentrations varied between sites, years, and depth (Fig. 5), exhibiting significant site by year (p < 0.0001), site by depth (p < 0.0001), and year by depth (p = 0.0111) interactions (Tables S4,  S9, S10, S11). Although DON concentrations were highest at JPH4 from 2010 to 2012, there was no clear pattern of DON concentrations decreasing with distance from the oil sands industrial center (Fig. 5, Tables S9, S10). For all sites except Mildred, DON concentrations increased with depth, with the increase being most pronounced at JPH4 and least pronounced at McKay and Anzac (Fig. 4, Table S10). Increases in DON concentration with depth were slightly more evident in 2009 and 2011 than in 2010 and 2012 (Fig. 5, Table S11). Within-year seasonal variation in DON concentrations was not consistent across all of the sites (Fig. 5).
Averaged across all sites, depths, and sampling dates, porewater DON concentrations were equivalent   (Table 3).

Discussion
Porewater chemistry with depth over time, [2009][2010][2011][2012] Over the first 4 years of sampling, we found little evidence to support either of our hypotheses especially for N. Porewater concentrations of NH 4 + , NO 3 − , and DON at the top of the peatland water table did not clearly decrease with increasing distance from the oil sands industrial center. We also found no evidence to support that with increasing deposition, elevated concentrations of NH 4 + , NO 3 − , DON, and SO 4 2− would be manifested deeper in the peat profile. Porewater NH 4 + , DON, and SO 4 2− concentrations, however, did decrease with increasing distance from the oil sands industrial center. The pattern of increasing NH 4 + -N and DON, but not NO 3 − -N, concentrations with depth, observed across all sites, is consistent with findings from other boreal or north temperate bogs (e.g., Bleak Lake Bog, Alberta, Vitt et al., 1995; S1 bog in the Marcell Experimental Forest (MEF), Minnesota, Griffiths & Sebestyen, 2016), but this pattern has not been evident at all bogs (e.g., Mer Bleue, Ontario, Basiliko et al., 2005). Total soluble N concentrations increased with depth (up to 3 m) at the S1 bog in the MEF consistently over 3 years of measurement (Griffiths & Sebestyen, 2016).
Few studies have quantified porewater SO 4 2− -S concentrations as a function of depth in bogs. At Bleak Lake Bog, Alberta, porewater SO 4 2− -S concentrations were either uniform of decreased over the top 20 cm of peat; S-amended plots (25.6 kg S ha −1 year −1 ) had higher porewater SO 4 2− -S concentrations than control plots (Vile et al., 2003a). At the S2 bog in the MEF, porewater SO 4 2− -S concentrations were either uniform, or slightly higher near the top of the water table than deeper in 35-50 cm peat profiles (Urban et al., 1989). In 9 temperate peatlands in eastern North America and the Czech Republic, porewater SO 4 2− -S concentrations were either uniform across the top 45 cm of peat or higher in the top 5 cm of peat and uniform from 5 to 45 cm (Novák & Wieder, 1992). While the biogeochemical processes that affect N and S cycling in bogs are generally known (cf. Blodau et al., 2007;Limpens et al., 2006;Vile & Novák, 2006), how these processes generate the considerable variability in depth profiles of soluble inorganic N and S within sites, between sites, and between years is not yet fully understood.
One of the most striking findings from our study is the considerably higher concentration of porewater NH 4 + and NO 3 − , but not of DON or SO 4 2− , throughout the top 1 m of peat in 2011 relative to the other 3 years (Figs. 3-7), a pattern that we also observed in porewater at the top of the water table in control plots (receiving no added N) in a bog and poor fen near Mariana Lake, Alberta, 115 km south of oil sands industrial center (Wieder et al., 2019. The 2010-2011 water year was unusually warm and dry (Table S3). Moreover, the 2011 growing season was unusually dry and was preceded by an unusually dry winter (Fig. 2). It is possible that under these conditions, N mineralization in near-surface peat could have been stimulated by warmer than normal surface peat, leading to the anomalously high porewater NH 4 + -N and NO 3 − -N concentrations. However, it seems unlikely that N mineralization would have been stimulated in deeper peat as temperature variation is dampened with depth in peat, and where watersaturated conditions prevail below the bog water table. Yet in 2011, we observed elevated NH 4 + -N and NO 3 − -N concentrations throughout the watersaturated peat to a depth of 1 m (Fig. 3). Even if the unusually warm and dry conditions at our monitoring sites had stimulated N mineralization in unsaturated near-surface peat, the NH 4 + -N and NO 3 − -N produced by mineralization is likely to have been immobilized by cation exchange (for NH 4 + -N) and plant/microbial uptake rather than moving downward into the saturated peat (cf. Wieder et al., 2019Wieder et al., , 2020. Indeed, in the control plots at both the Mariana Lake bog and poor fen, leaf N concentrations in Andromeda polifolia and Chamaedaphne calyculata (but not in leaves of other major vascular plant species) and N concentrations in Sphagnum fuscum stems were higher in 2011 than in the four subsequent years (Wieder et al., 2019. However, we note that in the control plots of the Mariana Lake bog and poor fen, neither net N mineralization (net ammonification, net nitrification, and net DON production) nor KCl-extractable NH 4 + -N and NO 3 − -N concentrations in the top 10 cm of peat below the living Sphagnum layer were elevated in 2011 as compared to 2012-2015 (Wieder et al., 2019. Thus, it appears that stimulation of N mineralization by the relatively warm, dry conditions of the 2011 growing season is not sufficient to explain the anomalously high elevated NH 4 + -N and NO 3 − -N concentrations that we observed at all monitoring sites in bog porewater below the water table. At the bog monitoring sites, mean water table position, averaged across all sites and measurement dates, was 6 cm lower in 2011 than in 2012. During drought conditions, a decreased hydraulic head in bogs can lead to hydraulic reversals such that peatland porewater may not be completely isolated from deeper groundwater (DeVito et al., 1997;Romanowicz et al., 1993;Ulanowski & Branfireun, 2013). Thus, another possible explanation of the anomalously high porewater NH 4 + and NO 3 − concentrations in 2011 could be a hydraulic reversal such that bog porewater chemistry was influenced by deep bog porewater and/or groundwater with higher NH 4 + -N and NO 3 − -N concentrations than bog porewater throughout the upper 1 m of peat in normal years. We acknowledge, however, that at this point, we cannot offer a definitive mechanistic explanation for the high NH 4 + -N and NO 3 − -N concentrations in 2011 observed at all bog monitoring sites, throughout the top 1 m of peat, over the entire growing season.  (Table 4).
Three lines of evidence would support the hypothesis that porewater NH -N decreased exponentially with distance from the oil sands industrial center (midpoint between the Syncrude and Suncor stacks), albeit with quite low R 2 values for NH 4 + -N and DON (Fig. 8). Exponential decreases in concentrations with distance would be consistent with oil sands N and S emissions influencing porewater chemistry. For porewater NH 4 + -N, NO 3 − -N, and DON concentrations, these spatial patterns are strongly driven by high concentrations in porewater at the JPH4 site. For porewater SO 4 2− -S concentration, the decrease with distance was strongly driven by high concentrations at the Mildred site. Prevailing winds were from the south or south southeast 20% of the time over the 11-year 1 3 period, carrying N and S emissions toward JPH4; prevailing winds were from the north or north northeast 22% of the time, toward the Mildred site (Fig. 1). These patterns were consistent across all years (Fig.  S1). If wind-dispersion of N and S emissions was the primary cause of decreasing porewater concentrations with distance, we would expect high porewater concentrations of NH 4 + -N, NO 3 − -N, DON, and SO 4 2− -S at both JPH4 and Mildred, which was not the case. It is possible that the unusually high SO 4 2− -S concentrations at the Mildred site are, at least in part, the result of upward diffusion or advection of shallow groundwater from weathered tills with high SO 4 2− concentrations, known to occur in the oil sands region (Birks et al., 2019;Cowie et al., 2015). At the Mildred site, porewater reduced electrical conductivity (Sjörs, 1952) increased considerably with depth throughout the upper 1 m of peat, reaching values as high as 1000 µS cm −1 (Fig. S2), a value higher than found in extreme rich fens in Canada (cf. Gignac et al., 2004;Hartsock et al., 2019a;Slack et al., 1980;Vitt & Chee, 1990;Vitt et al., 1995), but lower than in saline fens in northern Alberta (cf. Hartsock et al., 2019b;Volik et al., 2017).
From 2009 through 2019, oil sands stack and fleet N emissions have steadily increased (from 33,437 to 97,873 metric tonnes year −1 ), while S emissions have dramatically decreased from 2009 through 2014 (from 120,052 to 48,923 metric tonnes year −1 ), and remaining relatively constant at about 45,000 metric tonnes year −1 through 2019. If oil sands emissions are influencing bog porewater chemistry, we would expect to -S concentrations were significantly correlated with growing season N and S deposition (quantified using ion exchange resin collectors), respectively, at each site and with calculated N and S concentrations in precipitation (Table 3). It is well established that wet and dry N and S deposition decrease with increasing distance from the oil sands industrial center (Edgerton et al., 2020;Fenn et al., 2015;Hsu et al., 2016;Wieder et al., 2016a, b). Annual net primary production (NPP) of Sphagnum fuscum across 6 bogs, 5 of which are the same as bogs used in the present study, was 259 ± 9 g dry mass m −2 year −1 (Wieder et al., 2016b). At these bogs, S concentrations in S. fuscum capitula ranged from 1.13 to 1.83 mg g −1 and decreased with increasing distance from the oil sands industrial center (Wieder et al., 2021). Therefore, the annual quantity of S taken up through NPP of S. fuscum would be 2.9-4.7 kg ha −1 year −1 . Especially at the bog sites closest to the oil sands industrial center, atmospheric SO 4 2− -S deposition (Table 1) is at least sufficient to support uptake by growing S. fuscum, and dry deposition of S would supply even more S than measured using ion exchange resin collectors (cf. Edgerton et al., 2020). Thus, the Sphagnum "filter" for S may have failed, leading to downward movement of atmospherically deposited SO 4 2− -S into the peat profile, affecting porewater SO 4 2− -S concentrations. Some of the SO 4 2− not used for Sphagnum growth is taken up by vascular plants. However, if S deposition is sufficient to cause the downward infiltration of sulfateenriched into the peat profile, dissimilatory sulfate reduction could be stimulated, changing the pathways of peat decomposition that ultimately produce and release CO 2 to the atmosphere (Vile et al., 2003b). At the same time, in bogs generally, increasing SO 4 2− -S deposition leads to a strong suppression of CH 4 fluxes to the atmosphere (Gauci et al., 2004). Therefore, increasing SO 4 2− -S deposition above background levels in the oil sands region may have broader consequences with respect to bog CO 2 and CH 4 emissions.

Comparisons with European studies
Several field or greenhouse studies in European bogs have shown strong evidence that NH 4 + -N, NO 3 − -N, and/or DON concentrations in surface pools or in porewater at the top of the bog water table increase with increasing N deposition, either along prevailing N and S deposition gradients or with experimental N addition (Bragazza & Limpens, 2004;Bragazza et al., 2003;Limpens et al., 2003;Tomassen et al., 2003;Williams et al., 1999;Yesmin et al., 1995). However, background N deposition at these bog sites in general was considerably higher (up to 35 kg N ha −1 year −1 ) than deposition in the oil sands region, and these high deposition levels may have persisted for decades. Further, for the European studies that included experimental N addition in the greenhouse or in the field, N treatments have often been at exceptionally high levels (up to 100 kg N ha −1 year −1 ). Thus, given the background and experimental N addition conditions in these European studies, their results may not have clear relevance to conditions in the oil sands region.
When NH 4 NO 3 was added to bog peat cores collected from a low N deposition site (Mariana Lake Bog, Alberta; 1.7 kg N ha −1 year −1 ) at rates up to 45 kg N ha −1 year −1 , consistently across all N treatment levels, 89% of the added NH 4 + -N and 50% of the added NO 3 − -N were retained in the peat, with no effect on DON leaching (Hartsock et al., 2018). When NH 4 NO 3 was added in the field to Mariana Lake Bog, 115 km south of the oil sands, region over a 4-year period at levels up to 25 kg N ha −1 year −1 , no increase in porewater NH 4 + -N, NO 3 − -N, or DON was observed 1 3 (Wieder et al., 2019). The absence of a porewater chemistry response in these instances may be related to the important role of biological N 2 -fixation in Alberta bogs (Vile et al., 2014). At Mariana Lake Bog, field NH 4 NO 3 addition caused a downregulation of N 2 -fixation that was similar in magnitude to the quantity of N added, such that the result was no change in the total amount of N added to the bog through both deposition and fixation (Wieder et al., 2019). None of the European studies mentioned above considered N 2 -fixation, but it is likely the case that with chronic high background N deposition, N 2 -fixation is of minimal importance at many of these bogs. Based on results from Mariana Lake Bog, and the bog monitoring sites in this study, N deposition rates today (Table 1) are not high enough to completely inhibit N 2 -fixation. However, if N emissions from oil sands operations continue to increase, N 2 -fixation should progressively decrease, which could eventually lead to increasing NH 4 + -N, NO 3 − -N, and/or DON concentrations in bog porewaters.
Of the European studies, two showed increasing porewater DON concentrations with increasing N deposition (Bragazza & Limpens, 2004;Williams et al., 1999). Bragazza and Limpens (2004) showed clear, linear, increases in both DIN and DON in bogs along a N deposition gradient of 2-20 kg N ha −1 year −1 . Along the gradient, DON concentrations were approximately 3 times higher than DIN concentrations at each site, with DON concentrations increasing to 65 µg L −1 at the highest N deposition site. The DON response was interpreted as resulting from the uptake of atmospherically deposited N by Sphagnum, its incorporation into amino acids, and the release of these organic N molecules into bog porewater (Bragazza & Limpens, 2004). It is unclear whether this sequence of events occurs in Alberta bogs, where Sphagnum capitulum N concentrations show only small increases in response to N deposition either at the bog monitoring sites (Table 1) (Wieder et al., 2021) or with experimental field N addition (Wieder et al., 2019), where porewater DON concentrations in general are substantially higher (Fig. 5) than those reported by Bragazza and Limpens (2004).

Conclusions
We found only weak evidence that bog porewater NH 4 + -N, NO 3 − -N, or DON concentrations at the top of the bog water table are strongly affected by N emissions from oil sand operations at this time. Although N concentrations in porewater at the top of the water table were correlated with NH 4 + -N and NO 3 − -N concentrations in deposition, and porewater NH 4 + -N and DON concentrations decreased exponentially with distance from the oil sands industrial center (with low R 2 values), the latter pattern was driven by high porewater concentrations at one site, JPH4. We did not observe an increased in bog porewater NH 4 + -N, NO 3 − -N, or DON concentrations over time, which would have been expected given the substantial increase in oil sands N emissions from 2009 to 2019.
In one of the 4 years where we examined porewater chemistry to a depth of 1 m (2011), we found high concentrations of NH 4 + -N, NO 3 − -N, and to a lesser extent DON that are atypical of Alberta bogs. Higher than usual N mineralization rates in peat in a warmer and drier growing season preceded by a drier than normal winter could have contributed to these anomalously high porewater N concentrations. It is also possible that the dry conditions led to a hydraulic reversal such that the bog porewaters became affected by groundwater more so than is typical of Alberta bogs. We cannot offer a definitive mechanistic biogeochemical explanation for the high porewater N concentrations across all bog sites in this one growing season.
Evidence that porewater SO 4 2− -S concentrations at the top of the bog water table may be affected by S emissions from oil sands operations was mixed. Bog porewater SO 4 2− -S concentrations generally decreased with distance from the oil sands industrial center (albeit largely driven by high concentrations at Mildred, which may be related to saline groundwater interactions with porewater), and were positively correlated with SO 4 2− -S concentrations in growing season rainfall and SO 4 2− -S deposition. However, given the considerable decrease in S emissions from 2009 to 2019, we would have expected a decrease in porewater SO 4 2− -S concentrations over time, which we did not observe. Current SO 4 2− -S deposition rates are at least adequate to meet the S demands for growing Sphagnum. Increasing S deposition may lead to downward movement of SO 4 2− -S into the peat, with potential consequences for CO 2 and CH 4 emission from bogs.