Stream CO2 emissions contribute significantly to atmospheric climate forcing. While there are strong indications that groundwater inputs sustain these emissions, the specific biogeochemical pathways and timescales involved in this lateral CO2 export are still obscure. Here, via an extensive radiocarbon (14C) characterisation of CO2 and DOC in stream water and its groundwater sources in an old-growth boreal forest, we demonstrate that the 14C-CO2 is consistently in tune with the current atmospheric 14C-CO2 level and shows little association with the 14C-DOC in the same waters. Our findings thus indicate that stream CO2 emissions act as a shortcut that returns CO2 recently fixed by the forest vegetation to the atmosphere. Our results expose a positive feedback mechanism within the C budget of forested catchments, where stream CO2 emissions will be highly sensitive to changes in forest C allocation patterns associated with climate and land-use changes.
The flushing of terrestrially-derived C through runoff can represent up to 30% of the annual C balance of forested ecosystems, of which greenhouse gases such as CO2 hold a major contribution1,2,3. Once released from soils to surface waters, this CO2 is rapidly evaded to the atmosphere as a result of physical gas exchange4. This lateral CO2 flux is particularly relevant to headwater streams, which account for the bulk of surface water CO2 emissions5,6. Streams thus contribute actively to atmospheric climate forcing by returning terrestrially sequestered CO2 to the atmosphere. To date, the evidence supporting the terrestrial origin of stream CO2 has relied solely on mass balance exercises; demonstrating that the pool of groundwater CO2 is often sufficiently large to sustain stream CO2 fluxes7,8,9. Explicit demonstrations of this link are still absent. Most of all, the specific biogeochemical pathways giving rise to stream CO2, along with their associated timescales, have yet to be resolved. Without a clear assessment of the sources of lateral CO2 fluxes, the terrestrial and aquatic components of catchment C budgets cannot be reconciled.
Soil and groundwater CO2 can arise from several different biological sources, each confined across a spectrum between two main timescales10,11. Operating in the short timescales, is the current forest fixation of atmospheric CO2, which fuels autotrophic root respiration12 and heterotrophic mineralisation of recent photosynthates, transported to soils via throughfall, stemflow13 and root leachates14. Alternatively, over longer time-scales, saprotrophic decomposition embodies all forms of heterotrophic decomposition of older plant detritus and soil organic matter15. In groundwater, saprotrophic decomposition is supported mainly by dissolved organic C (DOC), which incorporates an assemblage of chemical properties and ages from the vegetation and soils traversed by the groundwater during its journey through the catchment16,17. Boreal forest catchments often comprise a peat-rich riparian zone, which serves as a repository of ancient soil organic matter that can support saprotrophic metabolism and may be remobilized through decomposition and runoff18,19,20. The source determination of lateral C fluxes may enable assessment of the vulnerability of these ancient C stocks19,21,22. Boreal forests also drive a considerable share of the global continental CO2 sink23, a process that is considered sensitive to a variety of anticipated disturbances24,25. The separation of timescales in the biological pathways governing lateral CO2 fluxes in boreal forested catchments is thus critical information, since different CO2 sources will likely follow distinct trajectories in response to environmental changes.
Here, we identify and apportion the sources of stream CO2 with particular emphasis on the separation of timescales and biogeochemical pathways involved in the lateral CO2 fluxes from an old-growth boreal forest catchment. We characterize the terrestrial and aquatic CO2 sources via repeated measurements of the groundwater and stream water radiocarbon (14C) content of CO2 and DOC (14C-CO2, 14C-DOC). Our sampling was designed following a three-level Upslope-Riparian-Stream transect, repeated over three different occasions during the growing season, thus allowing to associate spatio-temporal 14C patterns to different CO2 sources. The transect sampling was complemented with a year-round characterisation of the stream water 14C-CO2 and 14C-DOC, to further explore potential shifts in CO2 sources over time. Automated sensors recording hourly CO2 concentrations at each location along the transect and further downstream, allowed us to derive a complete annual C budget for this forested catchment using the age component of the lateral C fluxes to reveal links between the terrestrial and aquatic components. This study reveals that soil respiration, derived from the current forest C fixation, is the main source of stream CO2 fluxes.
Interpretation of 14C-contents
Radiocarbon analysis can be used to determine the average age of CO2 and DOC based on conventional 14C dating techniques and represents one of the most robust approaches for the separation of respiratory processes in soils10,11,26. However, the 14C content of gases or solute samples potentially originates from multiple combinations of sources, each with a different 14C-age, thus complicating the interpretation of the single average 14C-content. While the incorporation of post-bomb14C in the C cycle, resulting from the atmospheric testing in the 1950–60s, precludes a linear interpretation of 14C-content, it can provide clear evidence of carbon fixed from the atmosphere post ~AD1955 (i.e., when 14C concentration > 100%). Here, we specifically avoided referring to the measured 14C contents in terms of age from conventional 14C dating and instead focused our analysis on the relative differences in 14C contents between C species, in our case CO2 and DOC, as well as their changes over time and space to help define the stream C sources in connection to terrestrial processes.
Stream water 14C-CO2 and 14C-DOC
Stream water 14C-CO2 was surprisingly constant throughout the year, ranging from 102.5 to 105.3 %modern (n = 11) (Fig. 1a, Supplementary Table 1). In comparison, the stream water 14C-DOC was more variable and more 14C-enriched, ranging from 103.5 to 112.2 %modern (n = 8) (Fig. 1a, Supplementary Table 2). Stream water 14C-DOC was negatively related to the riparian water table position, indicating that more superficial water tables corresponded to more 14C-enriched DOC in the stream waters (Fig. 1b). There was no significant relationship between stream water 14C-CO2 and any of the measured variables included in the study, for example: temperature, discharge, water table position, C concentrations, net ecosystem exchange (NEE) or photosynthetic photon flux density (PPFD) (all p > 0.05, Fig. 1).
Groundwater 14C-CO2 and 14C-DOC
The patterns in 14C-CO2 and 14C-DOC in the connecting groundwater were generally similar to those of the stream waters. As such, groundwater 14C-CO2 was also remarkably homogenous across locations, depths and sampling dates. All samples were enriched in post-bomb 14C, together ranging from 101.1 to 106.6 %modern (n = 7), with the exception of one sample collected in the riparian deep soil water in August, where 14C-CO2 was 99.0 %modern (Fig. 2, Supplementary Table 2). Differences in 14C-CO2 between upslope and riparian groundwater were not significant (p = 0.6), despite a near doubling of the CO2 concentrations between the two locations. In contrast, the range in groundwater 14C-DOC was much larger than for 14C-CO2. Most of the groundwater 14C-DOC were enriched in post-bomb 14C, ranging from 100.8 to 116.8 %modern (n = 9), but there were two groundwater samples for which 14C-DOC was remarkably depleted (49.7 and 67.8 %modern, respectively) (Fig. 2, Supplementary Table 2). Both were collected in the upslope deep location, in August and October. Excluding these two groundwater samples, the 14C-DOC was significantly negatively correlated with the DOC concentration across the groundwater and stream waters (Fig. 3). There was no similar relationship between 14C-CO2 and CO2 concentrations in either the groundwater or stream water (Fig. 3).
The 14C-CO2 and 14C-DOC showed little correspondence in both groundwater and stream waters, with the DOC being on average 6 %modern more 14C-enriched than the CO2 across all discrete samples (n = 9), with the exception of the upslope deep groundwater in August and October (Figs. 1a, 2). In these groundwater, the difference in 14C content between CO2 and DOC was even larger, corresponding to 55 and 33 %modern for August and October respectively, where DOC was suddenly much older than CO2 (Fig. 2). In the riparian deep groundwater, the 14C-CO2 and 14C-DOC matched most closely and also showed the least temporal variability (only ~2 %modern across the three-sampling occasions despite a near doubling of the DOC concentrations between sampling occasions) (Fig. 2). In the stream waters, the differences in 14C content between CO2 and DOC, was largest during the spring freshet, but decreased during winter base flow conditions (Fig. 1a).
CO2 and DOC concentrations
The average CO2 concentrations doubled between the upslope mineral soils and the riparian organic-rich soils (d = 0.45) (Fig. 4e, f). In the riparian soils, the CO2 concentrations were similar between the shallow and deep layers (17.7 ± 3.5 and 14.9 ± 3.3 mg C L−1, respectively n = 8028 (d = 0.2)), while in the upslope soils, the CO2 concentrations were significantly lower in the shallow compared with the deep layers (4.6 ± 2.0 mg C L−1 n = 707, and 11.7 ± 3.3 mg C L−1 n = 8027 (d = 1.2), respectively) (Fig. 4e, f). Stream water CO2 concentrations during the open water season were significantly higher in the location adjacent to the transect (3.4 ± 1.1 n = 3496), compared with the stream gauging station, located 250 m downstream (2.8 ± 0.9 n = 3235), (d = 1.3) (Fig. 4d). Year-round hourly stream water CO2 concentrations were recorded at the downstream location, but showed no significant difference between the ice-covered and open-water period (d = 0.48, annual mean 2.8 ± 1.4 n = 8311) (Fig. 4d). The stream water CO2 concentrations never exceeded the groundwater CO2 concentrations in the riparian or upslope location (Fig. 4d–f).
Groundwater DOC concentrations were on average five times higher in the riparian compared with the upslope groundwater (39.2 ± 10.4 mg C L−1 n = 11, and 7.4 ± 6.7 mg C L−1 n = 8, respectively, p < 0.0001) (Fig. 4e, f). DOC concentrations were similar across the two groundwater depths, in both the riparian and upslope soils (p = 0.06, p = 0.5, respectively). The riparian DOC concentrations increased steadily between May and October, by 87 and 33% in the shallow and deep layers, respectively (Fig. 4e). This increase was not as clear in the upslope soils. The stream water DOC concentrations (19.8 ± 4.8 mg C L−1 n = 8) were significantly lower and never exceeded the riparian groundwater DOC concentrations (p ≤ 0.0001) (Fig. 4d, e). Together, the trends in DOC and CO2 concentrations along the upslope-riparian-stream transect resulted in a progressive shift, from a slight dominance of CO2 over DOC in the upslope groundwater (average CO2-C:DOC = 1.5), to a clear dominance of DOC over CO2 in the riparian groundwater (average CO2-C:DOC = 0.4), which was more pronounced in the stream waters (average CO2-C:DOC = 0.1).
Hydro-climatic conditions and catchment C budget
The NEE (determined by Eddy Covariance) of the forest ecosystem in the catchment for the study year was −205 g m−2 yr−1 (Fig. 5). The annual runoff for the catchment during the study year was 257 mm, representing about half of the annual precipitation (507 mm). The stream specific discharge (q) ranged from 0.036 to 10.8 mm d−1 (Fig. 1a). The modelled upslope and riparian water export within the upper one metre was 236 and 243 mm, respectively, which is similar to the catchment annual runoff (Supplementary Fig. 1c). Runoff through the shallow soil depths (0–0.5 m) contributed to 58 and 73% of the total runoff in the upslope and riparian location, respectively (Supplementary Fig. 1c).
The annual CO2 export from the riparian location was more than double that of the upslope location (Fig. 5). The contribution of the riparian soils was even more substantial for the annual DOC export, which was more than five times larger compared with the upslope site (Fig. 5). Downstream CO2 export represented only 13% of the initial riparian groundwater CO2 export, which can be attributed to rapid CO2 evasion to the atmosphere. The downstream DOC export was more comparable to the riparian groundwater DOC export, representing about 70% of the initial flux. The cumulative age of the lateral C fluxes was 104 and 110 %modern for CO2 and DOC, respectively. The difference in 14C content between the annual CO2 and DOC exported from the catchment could be explained by a 75% contribution from currently fixed CO2 from the atmosphere (i.e., during the last growing season (2015–2016), with the remaining fraction originating from the bulk DOC mineralization.
This study provides, to our knowledge, the first explicit evidence that stream CO2 fluxes are sustained by currently fixed CO2 from atmosphere via the forest vegetation’s photosynthetic activity (i.e., during the last growing season (2015–2016). The first piece of evidence supporting our interpretation was the persistent gaps between 14C-CO2 and 14C-DOC in groundwater and stream water, highlighting a major disconnect in the cycling of the two C species. Secondly, the homogeneity of the 14C-CO2 in groundwater and stream water, which remained systematically close to the current atmospheric 14C level, indicated that CO2 was sustained by a large and steady source, likely associated with current photosynthesis. Previous studies have provided indications that groundwater inflow of soil-derived CO2 is sufficient to support stream CO2 sources7,8,9. However, explicit demonstrations of the link between soil and stream CO2, along with determination of the biogeochemical pathways involved, were still lacking. Our results imply that stream CO2 fluxes are cycled rapidly, and likely provide a fast pathway for returning CO2 fixed from the atmosphere by the forest vegetation during this year’s growing season. The main implication of this work is that anticipated alterations in boreal forest growth and ecosystem level C allocation patterns, driven by climate change and other disturbances, will produce a rapid response in the stream CO2 fluxes, since both processes are tightly linked by the current forest activity.
The systematic offset between the 14C-DOC and 14C-CO2 in groundwater and stream water is key evidence of the limited overlap in their respective sources and controls (Figs. 1a, 2). The substantial variability in groundwater and stream water 14C-DOC, which contained both severely 14C-depleted and 14C-enriched post-bomb values (Figs. 1, 2), indicated that DOC arises from more diverse sources than CO2 and is cycled more slowly, up to millennia. There was a clear connection between the 14C-DOC and hydrological retention and flowpaths (Figs. 1b, 3), as suggested by previous studies in this catchment using independent methods19,27,28. Activation of fast flowing superficial flowpaths29 was associated with the transport of modern 14C-DOC, with rising water tables leading to an increase proportion of post-bomb 14C and dilution of the DOC, for example during spring freshet (Figs. 1b, 3). These dynamic superficial flowpaths are supplemented by intermittent activations of deeper flowpaths associated with longer water retention times30 and the transport of aged-DOC, for example, the upslope groundwater later in the growing season (Figs. 2, 3). Despite these profound changes in hydrological flowpaths across locations and seasons, the groundwater and stream water 14C-CO2 remained relatively unchanged (Figs. 1, 2). In fact, the groundwater 14C-CO2 was similar between the upslope mineral soils and the riparian organic soils, despite major contrasts in soil properties and contributing flowpaths (e.g., a near doubling of the CO2 concentrations and a five times increase in DOC concentrations (Fig. 4e, f), and a shift from severely 14C-depleted to post-bomb enriched DOC between the two locations later in the growing seasons (Fig. 3). This suggest that the sources governing groundwater CO2 can override these dynamics in hydrological flowpaths and soil chemistry.
Inconsistencies between the 14C content of DOC and CO2 have been reported in other catchments including the Amazon river network31 and various peatland dominated catchments32,33,34 (Supplementary Fig. 2). Other studies comparing the 14C content of DIC and DOC in surface waters also concur with these observations35. Although the form of these isotopic inconsistencies may vary across catchments, the 14C-DOC often reveals greater levels of post-bomb C than the CO2, suggesting more association with moderately old organic C reservoirs31,35. Severely 14C-depleted CO2 and DIC in surface waters are also more frequently reported than DOC, but these are typically connected to weathering of carbonate-containing minerals36,37. Such geological sources of CO2 are absent in this catchment, as indicated by the δ13C-CO2 values that were consistent with the C3 plant metabolic pathway (Supplementary Fig. 238). A considerable number of observations in the literature demonstrate a close agreement between the surface water 14C-CO2 and the current atmospheric 14C level, together with a clear photosynthetic δ13C value31,35. These observations, comply with ours, and suggests that surface water CO2 sources may often arise from rapid C cycling processes within catchment soils.
The uniformity of 14C-CO2 and its similarity with the current atmospheric 14C level, underlined that groundwater and stream water CO2 was sustained by a large and steady source, with rapid turnover times and omnipresent across this catchment (Figs. 1a, 2). Forest C fixation can fuel groundwater CO2 via autotrophic root respiration12 or the transport of recent photosynthates to soils by root exudates14, throughfall and stemflow13, thereafter mineralized by the soil microbial communities. Invasion of atmospheric CO2 in groundwater and stream water could not explain this close agreement between the 14C-CO2 content of the current atmospheric 14C level, as indicated by high CO2 concentrations in soil and stream waters, consistently above atmospheric saturations (Fig. 4d–f) and the low δ13C-CO2 values, conformed with C3 derived organic matter (Supplementary Fig. 238). The stability of the 14C-CO2 content was remarkable, considering the dramatic seasonal changes in groundwater and stream water C concentrations, forest productivity, precipitation, runoff and temperature over the year. In fact, the 14C-CO2 was independent of all measured environmental variables monitored in this study and showed no connection to seasonal patterns, further supporting a steady source (Figs. 1–3). Seasonal changes in hydro-climatic conditions are major controls on forest C fixation, soil respiratory processes39,40 and hydrological connectivity between soils and streams41, but the large amount of CO2 fixed by the forest vegetation during the growing season is likely sufficient to support the lateral CO2 export throughout the year (Fig. 5).
Based on the cumulative 14C content of stream CO2, an estimated 75% of the lateral CO2 export from the catchment possibly originated from the forest C fixation during the last growing season, with the remaining fraction arising from saprotrophic respiration, subsidized by the cumulative bulk 14C-DOC export (Fig. 5). This rough estimate agrees well with studies partitioning the vertical soil CO2 fluxes, where autotrophic root respiration contributes 50–64% of the total soil respiration in various forested catchments12,15, as well as in a nearby boreal forest catchment26. This flux is likely supplemented by heterotrophic mineralization of recent photosynthates, but the relative importance of these processes cannot be separated here. While there is general consensus that autotrophic root activity makes up a major fraction of vertical soil CO2 efflux42, there is also awareness that mineralization of root exudates and DOC transported via throughfall and stemflow may not be adequately quantifiable due to rapid turnover rates14. The specific contribution of these different biological pathways to lateral CO2 export remains to be further investigated, but our results support a steady and widespread connection to rapidly cycling CO2 sources fuelling lateral CO2 export. Only a small fraction (ca. 2%) of the net CO2 fixed from the atmosphere by the vegetation that year (NEE: −205 g C m−2 yr−1) was lost through the lateral CO2 export (Fig. 5). This is consistent with mass balance estimates from the same catchment reported by Öquist, Bishop1 during years with similar precipitation. Considering that precipitation and runoff were low during our study year, the fraction of currently fixed CO2 lost through lateral CO2 export in this forested catchment is likely higher in other years (up to 9%1).
Since the bulk DOC pool contains a mixture of C ages, the 14C-DOC in this catchment could reflect refractory DOC compounds, not mineralized during transit, masking an underlying connection between 14C-CO2 and 14C-DOC. Further studies are needed to assess the age-composition of DOC, but this does not invalidate that the current atmopsheric 14C-CO2 was reflected in the stream and groundwater of this boreal forest catchement. Saprotrophic respiration may be limited in theses CO2-rich groundwater, owing to lack of oxygen38 and short water transit time in the transiently saturated zone43,44. The best agreement between 14C-CO2 and 14C-DOC was observed in the deep riparian groundwater, which is found below the dominant source layer. The latter is responsible for the majority of DOC19, DIC41 and water export to the stream43. Longer residence time in the deeper groundwater may promote DOC mineralization, likely through fermentative processes38, hence a closer overlap between 14C-CO2 and 14C-DOC.
Closer agreements between the 14C-DIC and 14C-DOC than those observed in this study have sometimes been reported in larger river catchments and lakes45,46,47. This highlights that the terrestrial DOC export may remerge as an important source sustaining CO2 emissions further downstream; when connectivity with the catchment soils decreases8,48 and longer water residence time allows for in-situ mineralisation to occur49. Both aged and modern DOC were exported from this catchment during the study year, but our results do not indicate any significant incorporation of severely 14C-depleted or highly 14C-enriched post-bomb DOC in the lateral CO2 export. Emerging new research is now demonstrating that aged-DOC may be bioavailable for freshwater microbial communities leading to the production of aged-CO222,50,51.The remobilization of ancient DOC in the upslope soils of this catchment is a concern, and more studies are needed to assess its sources and fate. While previous studies have shown that aquatic DOC mineralization remains low across the Krycklan catchment52,53,54, there is also mounting evidence of deeper groundwater contribution increasing further downstream and potentially transporting aged-DOC19,48,55. Further studies are needed to address the generality of our findings, and the potential mineralization of aged-DOC within higher order rivers and streams.
This study reveals that stream CO2 fluxes are fuelled by the current forest C fixation and its associated soil respiration processes in a boreal forest catchment. This close connection between the forest C sink and lateral CO2 fluxes had already been put forward based on inter-annual coupling in flux measurements1. Our study provides further description of the mechanistic underpinning of this connection between terrestrial and aquatic C fluxes. The lateral CO2 export rapidly mobilises a significant fraction of the C currently fixed from the atmosphere by the forest vegetation. Groundwater CO2 inputs support surface waters CO2 emissions across multiple types of aquatic ecosystems7,8,9, thus suggesting that rapidly cycling CO2 sources may have a widespread contribution to aquatic CO2 emissions. Owing to the prevalence of root respiration and mineralization of recent photosynthates in forested soils, our results may also be generalized across a large number of forested ecosystems and biomes. Forest C balance and ecosystem level C allocation patterns in the northern high latitudes are vulnerable to a large number of disturbances including global warming, increased forest fire frequency, insect outbreaks and industrial and commercial exploitation56,57,58. Stream CO2 emissions will quickly feedback on these disturbances, because of the speed of stream CO2 cycling and its close connection to the current forest activity.
The study was conducted in a 0.13 km2 catchment located in northern Sweden within the Krycklan Catchment Study (64°14′N, 19°46′E)59. The catchment has been heavily studied for more than two decades and is occasionally referred to as “Västrabäcken” or “C2” across the literature. The catchment is almost completely forested (99%), with Scots pine (Pinus sylvestris) (64%) and Norway spruce (Picea abies) (36%). The active root depth is mostly distributed above the average groundwater table position60. The average tree stand age is 103 years old59. Man-made ditching of the stream to improve forest productivity occurred about a century ago. The stream is adjacent a peat-rich riparian zone, with the soil profile consisting of ~70 cm thick peat transitioning to the underlying till at ~90 cm depth. The age of the accumulated solid peat ranges from modern near the surface to 2810 years BP at 70 cm depth (Bishop, unpublished data). The organic soil content is > 80% in the riparian zone, which is considerably higher than the upslope podzols (< 5%)61. The latter is composed of well-developed iron podzols on sandy till, comprising a 5 cm humus layer at the surface, overlying a 12 cm thick sandy bleached E-horizon and a 60 cm thick B-horizon61. The underlying bedrock is composed predominantly of base-poor Sveco-fennian metasediments-metagraywacke and holds no known carbonate containing minerals. Carbonate alkalinity is rather produced by weathering of silicate minerals38.
The climate is cold temperate humid and bears a persistent snow cover from November to April. The 30 year mean annual precipitation is 640 mm (1981–2010), of which 35% falls as snow59. The annual precipitation during our study year (507 mm) corresponds to the bottom 6% of the previous 30 years observations (1981–2012). The 30 years mean annual, July and January temperatures are +1.8, +14.7 and −9.5 °C, respectively59. The annual peak stream discharge in the region typically occurs during spring in connection to snow melt, but storm events during summer and autumn can also generate peak flows in some years. The winter is typically dominated by low flow conditions.
Soil and stream instrumentation
Groundwater and stream water sampling was carried out following an upslope-riparian-stream transect, as described in ref. 38,41. Groundwater wells were installed along the assumed hydrological flowpaths, with a first set located at 1–2 m (riparian) and the second at 10–12 m (upslope) distance from the stream. The 10-year mean water travel time for the entire catchment is estimated at 690, ranging between 470 and 2064 days62. The water turnover time, from the water divide to the stream and with depths ranging between 0.5 and 3 m, is estimated to 4.6 years43. The area represented by the groundwater transect was estimated at 2540 m2 with an average width of 17.7 m, occupying 2% of the total catchment area43. The estimated time for water to exit this transect is in the scale of a month (near the upslope mineral soils) to hours (near the stream)43.
The groundwater characterisation at both locations was focused to the upper one metre of the soil profile, where most of the runoff generation is confined43. The groundwater installation comprised a set of wells with screening of the upper 0–0.5 m (shallow) and lower 0.5–1 m (deep), respectively. For each of the two locations and depths, two identical sets of groundwater wells were installed, with one allowing manual sampling of the groundwater and the other containing sensors for continuous reading of groundwater temperature and CO2 concentration. Alongside, a fifth well was installed at each of the upslope and riparian locations where water table position was recorded.
Continuous reading of dissolved CO2 concentration at each location and depth was enabled using Vaisala CARBOCAP GMP221 nondispersive infrared (NDIR) CO2 sensors (range 0–3%, 0–5%, in the stream and groundwater, respectively, except for the upslope shallow groundwater where the range was 0–1%). The sensors were enclosed inside a water-tight, gas-permeable Teflon membrane (PTFE) and sealed with Plasti Dip (Plasti Dip international, Baine, MN, USA) to ensure that the sensor was protected from water, but remained exposed to dissolved gas. The groundwater CO2 concentration in the riparian and upslope location at the deep and shallow depth was recorded during the full year. The upslope shallow groundwater well was completely dry for 245 days out of the studied year, but all other sensors remained below the groundwater table at all times. The measurements in the stream adjacent to the groundwater wells were restricted to the open-water season. However, year-round measurements of stream water CO2 concentration was undertaken in a heated dam house (C2), located 250 m downstream from the soil transect location. The continuous reading of groundwater and stream water CO2 concentrations were validated against manual snapshot measurements, using an acidified headspace method63, which showed an average 11 and 9% difference in the groundwater and stream water, respectively. All three stations along the transect (upslope, riparian and stream) were instrumented with pressure transducers (MJK 1400, 0–1 m, MJK Automation AB) recording water table height, and temperature sensors (TO3R, TOJO Skogsteknik). Groundwater temperature was measured at 0.8 m from the ground surface, and stream water temperature was recorded in two locations, adjacent to the transect and 250 m downstream. All continuously measured data were collected hourly and stored on external data loggers (CR1000, Campbell Scientific, USA).
Discharge was determined at a V-notch weir in the downstream heated dam house at C259. Stream discharge was determined by applying stage height-discharge rating curves to hourly water level measurements. The discharge for each of the four groundwater sampling points (riparian/upslope, deep/shallow) was modelled according to Amvrosiadi, Seibert43. To calculate total discharge through the riparian and upslope profiles, Darcy’s law was applied in combination with the transmissivity profiles, and the local water table gradients43. The fraction of discharge flowing through the shallow (0–0.5 m) and deep (0.5–1 m) soil layers was estimated based on the same transmissivity profiles.
Soil and stream water chemical analysis
Groundwater and stream water samples were collected monthly for DOC at each location and depth during the ice-free season (May–November 2015). The DOC concentration was analysed from 10 ml of ground and stream water, filtered through glass-fiber Whatman GF/F filters (0.7 μm) in the field and stored in high-density polyethylene bottles. Prior to analysis, samples were acidified and sparged to remove inorganic carbon. The samples were analysed using a Shimadzu Total Organic Carbon Analyzer TOC-VCPH, following storage at 4 °C for 2–3 days64.
The radiocarbon sampling was carried out in two phases, first a repeated catchment scale characterisation of groundwater and stream water in May (spring), August (summer) and October (autumn), and consecutively, a complete year characterisation in the stream waters. In total, 21 samples were collected and analysed for 14C-CO2 and 16 for 14C-DOC in groundwater and stream water between May 2015 and June 2016. The stream water samples were collected directly adjacent to the groundwater transect during the open water periods and downstream in the heated dam house (C2) during the ice-covered periods for accessibility. Simultaneous measurement of 14C-CO2 at both stations in July 2015 showed close correspondence, with only a <0.1 %modern difference, which was within the range of measurement precision. The hillslope transect sampling included all four groundwater locations and depths; riparian/upslope and deep/shallow. However, the available water volume in the shallow riparian and upslope locations was insufficient for radiocarbon analysis in August and October.
Sample collection for 14C-CO2 included two different methods, one using manual spot measurements applied to groundwater and stream water (n = 11) and the other allowing for time-integrated sampling in the stream water (n = 10). Manual spot measurements of 14C-CO2 were carried out with the super headspace method whereby manually equilibrated CO2 samples were trapped onto molecular sieve cartridges (MSCs) (see ref. 65 for further details). The integrated measurements of 14C-CO2 were performed using passive samplers comprising MSCs installed below the water surface, which slowly collects stream water CO2 over extended time periods (see ref. 66 for further details). The samplers were based on the MSC described above, but attached to a gas permeable hydrophobic filter (Accurel PP V8/2 HF tubing; Membrana GmbH, Germany67). These passive samplers were deployed for periods ranging from 26 to 72 days; collectively they cover more than a full year (May 2015–June 2016). The trapping capacity of the MSC was never exceeded (< 100 ml CO2). Unfortunately, two time integrated 14C-CO2 samples were discarded due to contamination that resulted from cracks in the MSC glass casing during deployment or transportation that caused stream water or air contamination of the sample. At the NERC Radiocarbon Facility (East Kilbride, UK), CO2 samples were recovered from the MSCs by heating and cryogenically purified.
The 14C-DOC analysis was performed on 1 L samples of groundwater and stream water collected in acid-washed glass bottles. The samples were filtered in the laboratory through 0.7 μm glass fibre filters, rotary evaporated and freeze-dried. Acid-fumigation of samples was undertaken to guard against carbonate contamination, and the dried DOC was combusted to CO2 in an elemental analyser (Costech ECS 4010, Italy) and cryogenically recovered. Manual spot measurements of stream water 14C-DOC were taken at each change of the passive 14C-CO2 samplers in order to characterize cumulative 14C-DOC under a large range of hydrological conditions.
All radiocarbon samples were converted to graphite using Fe-Zn reduction and measured by accelerator mass spectrometry at the Scottish Universities Environmental Research Centre (East Kilbride, UK). Stable carbon isotope measurement (δ13C) was performed on an aliquot of the recovered CO2 using isotope ratio mass spectrometry (IRMS; Thermo-Fisher Delta V, Germany) and reported relative to the Vienna PDB standard. All radiocarbon results were normalised to a δ13C of −25 ‰ using the measured δ13C values, and expressed as %modern and conventional radiocarbon age (years before present (BP), where 0 BP = AD 1950), with ± 1σ analytical precision. The passively collected 14C-CO2 samples were additionally corrected for the +4.2 ‰ isotopic fractionation effect caused by the gas trapping into the molecular sieves66. The northern hemisphere atmospheric 14C-CO2 content during the study period ranged from 101.4 to 100.8 according to ref. 68.
Catchment C budget and statistical analysis
The contribution of currently fixed CO2 from the atmosphere by the forest vegetation during the growing season of the studied year (2015–2016; Atm14C (%)) was estimated using a simple two-endmember mixing model, solving for the mass of current atmospheric 14C-CO2 (Atm14C) required to explain the observed gap between the cumulative lateral 14C-CO2 and 14C-DOC export as follows:
The forest NEE and PPFD were obtained from the ICOS data from the Svartberget site (http://www.icos-sweden.se/). The groundwater C export for each soil depth and location was estimated by combining the modelled specific discharge according to Amvrosiadi, Seibert 43, along with the measured hourly CO2 concentrations and the interpolated monthly point measurements of DOC concentrations. The total 14C content export at the upslope, riparian and stream locations was estimated by calculating the weighted average 14C content of both CO2 and DOC export over the full year. Significant differences in soil or stream water chemistry were tested using the non-parametric Wilcoxon test, with p-values reported in brackets in the result sections. For large sample sizes, such as hourly measurements of CO2 concentrations, statistical differences between means were tested using Cohen’s d test for effect size. Differences were considered statistically significant when p-value < 0.01. Mean values followed with the standard deviation and the number of observations are presented in brackets in the text. All analyses were performed using R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.
Öquist, M. G. et al. The full annual carbon balance of boreal forests is highly sensitive to precipitation. Environ. Sci. Technol. Lett. 1, 315–319 (2014).
Wallin, M. B. et al. Evasion of CO2 from streams—the dominant component of the carbon export through the aquatic conduit in a boreal landscape. Glob. Change Biol. 19, 785–797 (2013).
Richey, J. E., Melack, J. M., Aufdenkampe, A. K., Ballester, V. M. & Hess, L. L. Outgassing from Amazonian rivers and wetlands as a large tropical source of atmospheric CO2. Nature 416, 617–620 (2002).
Oquist, M. G., Wallin, M., Seibert, J., Bishop, K. & Laudon, H. Dissolved inorganic carbon export across the soil/stream interface and its fate in a boreal headwater stream. Environ. Sci. Technol. Lett. 43, 7364–7369 (2009).
Raymond, P. A. et al. Global carbon dioxide emissions from inland waters. Nature 503, 355–359 (2013).
Wallin, M. B. et al. Carbon dioxide and methane emissions of Swedish low-order streams-a national estimate and lessons learnt from more than a decade of observations. Limnol. Oceanogr. Lett. 3, 156–167 (2018).
Duvert, C., Butman, D. E., Marx, A., Ribolzi, O. & Hutley, L. B. CO2 evasion along streams driven by groundwater inputs and geomorphic controls. Nat. Geosci. 11, 813–818 (2018).
Hotchkiss, E. R. et al. Sources of and processes controlling CO2 emissions change with the size of streams and rivers. Nat. Geosci. 8, 696–699 (2015).
Weyhenmeyer, G. A. et al. Significant fraction of CO2 emissions from boreal lakes derived from hydrologic inorganic carbon inputs. Nat. Geosci. 8, 933–936 (2015).
Kuzyakov, Y. Sources of CO2 efflux from soil and review of partitioning methods. Soil Biol. Biochem. 38, 425–448 (2006).
Hanson, P. J., Edwards, N. T., Garten, C. T. & Andrews, J. A. Separating root and soil microbial contributions to soil respiration: A review of methods and observations. Biogeochemistry 48, 115–146 (2000).
Hogberg, P. et al. Large-scale forest girdling shows that current photosynthesis drives soil respiration. Nature 411, 789–792 (2001).
Van Stan, J. T. & Stubbins, A. Tree-DOM: Dissolved organic matter in throughfall and stemflow. Limnol. Oceanogr. Lett. 3, 199–214 (2018).
Van Hees, P. A. W., Jones, D. L., Finlay, R., Godbold, D. L. & Lundström, U.S.The carbon we do not see—the impact of low molecular weight compounds on carbon dynamics and respiration in forest soils: a review. Soil Biol. Biochem. 37, 1–13 (2005).
Trumbore, S. Age of soil organic matter and soil respiration: radiocarbon constraints on belowground C dynamics. Ecol. Appl. 10, 399–411 (2000).
Baker, M. A., Valett, H. M. & Dahm, C. N. Organic carbon supply and metabolism in a shallow groundwater ecosystem. Ecology 81, 3133–3148 (2000).
Downing, J. A. & Striegl, R. G. Size, age, renewal, and discharge of groundwater carbon. Inland Waters 8, 122–127 (2018).
Gurwick, N. P. et al. Mineralization of ancient carbon in the subsurface of riparian forests. J. Geophys. Res.: Biogeosciences 113, n/a–n/a (2008).
Ledesma, J. L., Grabs, T., Bishop, K. H., Schiff, S. L. & Kohler, S. J. Potential for long-term transfer of dissolved organic carbon from riparian zones to streams in boreal catchments. Glob. Change Biol. 21, 2963–2979 (2015).
Lambert, T. et al. DOC sources and DOC transport pathways in a small headwater catchment as revealed by carbon isotope fluctuation during storm events. Biogeosciences 11, 3043–3056 (2014).
Grosse G. et al. Vulnerability of high-latitude soil organic carbon in North America to disturbance. J. Geophys.Res. 116, G00K06 (2011).
Mann, P. J. et al. Utilization of ancient permafrost carbon in headwaters of Arctic fluvial networks. Nature. Communications 6, 7856 (2015).
Winkler, A. J., Myneni, R. B., Alexandrov, G. A. & Brovkin, V. Earth system models underestimate carbon fixation by plants in the high latitudes. Nat. Commun. 10, 885 (2019).
Bonan, G. B. Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science 320, 1444–1449 (2008).
Chapin, F. S. et al. Arctic and boreal ecosystems of western North America as components of the climate system. Glob. Change Biol. 6, 211–223 (2000).
Hahn, V., Hogberg, P. & Buchmann, N. 14C - a tool for separation of autotrophic and heterotrophic soil respiration. Glob. Change Biol. 12, 972–982 (2006).
Köhler, S. J., Buffam, I., Seibert, J., Bishop, K. H. & Laudon, H. Dynamics of stream water TOC concentrations in a boreal headwater catchment: Controlling factors and implications for climate scenarios. J. Hydrol. 373, 44–56 (2009).
Pacific, V. J., Jencso, K. G. & McGlynn, B. L. Variable flushing mechanisms and landscape structure control stream DOC export during snowmelt in a set of nested catchments. Biogeochemistry 99, 193–211 (2010).
Ameli, A. A., McDonnell, J. J. & Bishop, K. The exponential decline in saturated hydraulic conductivity with depth: a novel method for exploring its effect on water flow paths and transit time distribution. Hydrol. Process. 30, 2438–2450 (2016).
Peralta-Tapia, A., Sponseller, R. A., Tetzlaff, D., Soulsby, C. & Laudon, H. Connecting precipitation inputs and soil flow pathways to stream water in contrasting boreal catchments. Hydrol. Process. 29, 3546–3555 (2015).
Mayorga, E. et al. Young organic matter as a source of carbon dioxide outgassing from Amazonian rivers. Nature 436, 538–541 (2005).
Leith, F. I., Garnett, M. H., Dinsmore, K. J., Billett, M. F. & Heal, K. V. Source and age of dissolved and gaseous carbon in a peatland–riparian–stream continuum: a dual isotope (14C and δ13C) analysis. Biogeochemistry 119, 415–433 (2014).
Billett, M. F., Garnett, M. H. & Dinsmore, K. J. Should aquatic CO2 evasion be included in contemporary carbon budgets for peatland ecosystems? Ecosystems 18, 471–480 (2015).
Campeau, A. et al. Aquatic export of young dissolved and gaseous carbon from a pristine boreal fen: Implications for peat carbon stock stability. Glob. Change Biol. 23, 5523–5536 (2017).
Marwick, T. R. et al. The age of river-transported carbon: A global perspective. Glob. Biogeochem. Cycles 29, 122–137 (2015).
Vihermaa, L. E., Waldron, S., Garnett, M. H. & Newton, J. Old carbon contributes to aquatic emissions of carbon dioxide in the Amazon. Biogeosciences 11, 1773–1800 (2014).
Billett M. F., Garnett M. H. & Harvey F. UK peatland streams release old carbon dioxide to the atmosphere and young dissolved organic carbon to rivers. Geophys. Res. Lett. 34, L23401 (2007).
Campeau, A. et al. Stable carbon isotopes reveal soil-stream DIC linkages in contrasting headwater catchments. J. Geophys. Res.: Biogeosciences 123, 149–167 (2018).
Boone, R. D., Nadelhoffer, K. J., Canary, J. D. & Kaye, J. P. Roots exert a strong influence on the temperature sensitivityof soil respiration. Nature 396, 570–572 (1998).
Bond-Lamberty, B. & Thomson, A. Temperature-associated increases in the global soil respiration record. Nature 464, 579–582 (2010).
Leith, F. I. et al. Carbon dioxide transport across the hillslope–riparian–stream continuum in a boreal headwater catchment. Biogeosciences 12, 1881–1892 (2015).
Bond-Lamberty, B., Wang, C. & Gower, S. T. A global relationship between the heterotrophic and autotrophic components of soil respiration? Glob. Change Biol. 10, 1756–1766 (2004).
Amvrosiadi, N., Seibert, J., Grabs, T. & Bishop, K. Water storage dynamics in a till hillslope: the foundation for modeling flows and turnover times. Hydrol. Process. 31, 4–14 (2017).
Ameli, A. A. et al. Hillslope permeability architecture controls on subsurface transit time distribution and flow paths. J. Hydrol. 543, 17–30 (2016).
Raymond, P. A. et al. Controls on the variability of organic matter and dissolved inorganic carbon ages in northeast US rivers. Mar. Chem. 92, 353–366 (2004).
Hossler K. & Bauer J. E. Estimation of riverine carbon and organic matter source contributions using time-based isotope mixing models. J. Geophys. Res. Biogeosci. 117, G03035 (2012).
Caraco, N., Bauer, J. E., Cole, J. J., Petsch, S. & Raymond, P. Millennial-aged organic carbon subsidies to a modern river food web. Ecology 91, 2385–2393 (2010).
Peralta-Tapia, A. et al. Scale-dependent groundwater contributions influence patterns of winter baseflow stream chemistry in boreal catchments. J. Geophys. Res.: Biogeosciences 120, 847–858 (2015).
Catalán, N., Marcé, R., Kothawala, D. N. & Tranvik, L. J. Organic carbon decomposition rates controlled by water retention time across inland waters. Nat. Geosci. 9, 501–504 (2016).
Dean J. F., Garnett M. H., Spyrakos E. & Billett M. F. The potential hidden age of dissolved organic carbon exported by peatland streams. Geophys. Res. Biogeosci. 124, 328–341 (2019).
McCallister, S. L. & del Giorgio, P. A. Evidence for the respiration of ancient terrestrial organic C in northern temperate lakes and streams. Proc. Natl Acad. Sci. USA 109, 16963–16968 (2012).
Winterdahl, M. et al. Decoupling of carbon dioxide and dissolved organic carbon in boreal headwater streams. J. Geophys. Res.: Biogeosciences 121, 2630–2651 (2016).
Kothawala, D. N. et al. The relative influence of land cover, hydrology, and in-stream processing on the composition of dissolved organic matter in boreal streams. J. Geophys. Res.: Biogeosciences 120, 1491–1505 (2015).
Köhler, S., Buffam, I., Jonsson, A. & Bishop, K. Photochemical and microbial processing of stream and soil water dissolved organic matter in a boreal forested catchment in northern Sweden. Aquat. Sci. 64, 269–281 (2002).
Tiwari, T., Laudon, H., Beven, K. & Ågren, A. M. Downstream changes in DOC: Inferring contributions in the face of model uncertainties. Water Resour. Res. 50, 514–525 (2014).
Gauthier, S., Bernier, P., Kuuluvainen, T., Shvidenko, A. Z. & Schepaschenko, D. G. Boreal forest health and global change. Science 349, 819–822 (2015).
Lenton, T. M. et al. Tipping elements in the Earth’s climate system. Proc. Natl Acad. Sci. USA 105, 1786–1793 (2008).
Kurz, W. A. et al. Mountain pine beetle and forest carbon feedback to climate change. Nature 452, 987–990 (2008).
Laudon, H. et al. The Krycklan Catchment Study-A flagship infrastructure for hydrology, biogeochemistry, and climate research in the boreal landscape. Water Resour. Res. 49, 7154–7158 (2013).
Bishop, K. & Dambrine, E. Localization of tree water uptake in Scots pine and Norway spruce with hydroiogical tracers. Can. J. For. Res. 25, 286–297 (1995).
Lidman F., Boily Å., Laudon H. & Köhler S. J. From soil water to surface water—how the riparian zone controls the transport of major and trace elements from a boreal forest to a stream. Biogeosciences, 14, 3001–3014 (2017).
Peralta-Tapia, A. et al. Hydroclimatic influences on non-stationary transit time distributions in a boreal headwater catchment. J. Hydrol. 543, 7–16 (2016).
Åberg, J. & Wallin, M. B. Evaluating a fast headspace method for measuring DIC and subsequent calculation of pCO2 in freshwater systems. Inland Waters 4, 157–166 (2014).
Agren, A., Buffam, I., Jansson, M. & Laudon, H. Importance of seasonality and small streams for the landscape regulation of dissolved organic carbon export. J. Geophys. Res.-Biogeosciences 112, 2156–2202 (2007).
Garnett, M. H., Gulliver, P. & Billett, M. F. A rapid method to collect methane from peatland streams for radiocarbon analysis. Ecohydrology 9, 113–121 (2016).
Garnett, M. H., Dinsmore, K. J. & Billett, M. F. Annual variability in the radiocarbon age and source of dissolved CO2 in a peatland stream. Sci. Total Environ. 427-428, 277–285 (2012).
Gut, A. et al. A new membrane tube technique (METT) for continuous gas measurements in soils. Plant Soil 198, 79–88 (1998).
Graven, H. et al. Compiled records of carbon isotopes in atmospheric CO2 for historical simulations in CMIP6. Geoscientific Model. Development 10, 4405–4417 (2017).
This study was supported by the Swedish Research Council (contract: 2012-3919 to K. Bishop), the Natural Environment Research Council (NERC) Radiocarbon Facility NRCF010001 (allocation number 1947.1015) and the Department of Earth Sciences at Uppsala University. Financial support from the Swedish Research Council and contributing research institutes to the Swedish Integrated Carbon Observation System (ICOS-Sweden) research infrastructure and the Swedish Infrastructure for Ecosystem Science (SITES) are also acknowledged. The study further benefitted from support by the Knut and Alice Wallenberg foundation. We also thank Joachim Audet and the Krycklan Catchment Study crew for field support.
The authors declare no competing interests.
Journal peer review information: Nature Communications thanks Aron Stubbins and the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Campeau, A., Bishop, K., Amvrosiadi, N. et al. Current forest carbon fixation fuels stream CO2 emissions. Nat Commun 10, 1876 (2019). https://doi.org/10.1038/s41467-019-09922-3
This article is cited by
Integrating terrestrial and aquatic ecosystems to constrain estimates of land-atmosphere carbon exchange
Nature Communications (2023)
Anthropogenic Transformation Disconnects a Lowland River From Contemporary Carbon Stores in Its Catchment
The undetected loss of aged carbon from boreal mineral soils
Scientific Reports (2021)
The relative importance of seasonality versus regional and network-specific properties in determining the variability of fluvial CO$$_2$$, CH$$_4$$ and dissolved organic carbon across boreal Québec
Aquatic Sciences (2021)
Divergent Controls on Stream Greenhouse Gas Concentrations Across a Land-Use Gradient
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.