Introduction

Peatlands are important in the global carbon (C) cycle due to their function as a long-term carbon store. This store has built up through millennia due to very low decomposition rates that are surpassed by productivity rates (Clymo and Hayward 1982) making peatlands net carbon sinks. It’s been estimated that today’s peatlands store ca 30% (~ 550 Pg C) of the Earth’s terrestrial soil carbon while comprising only 3% of the Earths land area (Gorham 1991; Yu 2012).

Within peatlands in the northern hemisphere, Sphagnum mosses are often the dominant species, comprising up to 90% of peat (Clymo 1987). Sphagnum plays an important role in peatland formation and ecological resilience, through its low decomposition rate due to its recalcitrant and antibacterial nature (Kroken et al. 1996) and by creating acidic, anoxic and nutrient-poor conditions within peatlands (Amesbury 2013; Hajek and Vicherova 2014). Sphagnum’s extraordinary physiology allows it to hold 16–26 times its weight in water (Bold 1973; Amesbury 2013), in both live and dead non-decomposed Sphagnum, depending on species. This buffers the impact of dry weather conditions on peat moisture levels, preventing the peat from drying out and decomposing. Peatlands are classified as such once an organic soil layer of 30 cm or greater has formed; for northern peatlands, the average peat depth is 1.3–2.3 m, but peat depth can extend to 15–20 m in some peatlands (Clymo et al. 1998; Turunen et al. 2002). Accumulation rates of Sphagnum peat are approximately 0.61–3.9 mm year−1 (Royles et al. 2012; Amesbury 2013), due to Sphagnums slow growth rate within the cold, nutrient poor peatlands (Strack 2008). The slow growth rate is important in the context of peatland management as once the peat is lost, the peatland will only recover slowly over centuries/millennia. Degradation of peatlands through drainage, mining and removal of Sphagnum allows stored C to be released back into the atmosphere CO2, through decomposition of the exposed peat. It is estimated that UK peatlands are a net source of 3.72 Mtn CO2eq year−1 (Worrall et al. 2011), with only 1% of Englands deep peats considered to be in an undamaged state (Natural England 2010). This makes restoring peatlands a priority, as restored peatlands not only protect the already sequestered carbon, but can reinstate their carbon sink capacity (Strack 2008).

Atmospheric CO2 concentration has increased since pre-industrial times to ca. 408 ppm (ESRL NOAA 2015) and is predicted to increase to 800 ppm by the year 2100 (IPCC 2013). The effects of this elevation of atmospheric CO2 on Sphagnum moss are currently disputed, with studies indicating an increase in growth rate (Jauhiainen and Silvde 1999; Heijmans et al. 2001a; Saarnio et al. 2003), decreases in growth rate (Grosvernier et al. 2001; Fenner et al. 2007) and no response (Van der Hejiden et al. 2000; Hoosbeek et al. 2002; Toet et al. 2006). Contrasting responses between Sphagnum species are thought to be the cause of some of the variation. Toet et al. (2006), found increased abundance in S. recurve and no difference in S. palustre when they were subjected to an elevation of 180 ppm above ambient, while S. fuscum did not respond to elevated CO2 (350, 700, 1000 and 2000 ppm) at all (Jauhiainen et al. 1997). Furthermore, contrasting responses have been seen between studies conducted on the same species. For example, S. magellanicum, was found to increase in growth rate by Heijmans et al. (2001b) when subjected to 560 ppm CO2, while Heijmans et al. (2001a) reported a greatly reduced growth rate. This suggests factors other than species are affecting Sphagnum’s response to elevated CO2.

Hydrology has been identified as one of the most important factors controlling Sphagnum growth and peatland carbon dynamics (Strack and Waddington 2007; Turetsky et al. 2012; Lanta and Kantorova 2015 etc.). Water availability impacts both C sequestration and growth rate (Glime 2007), due to there being an optimum moisture content for Sphagnum photosynthesis (Rice and Giles 1996). Varying physiologies mean that different Sphagnum species vary in sensitivity to changes in water level, each often growing best within their own hydrological niche (Rydin and Mcdonald 1985; Grosvernier et al. 1997; Taylor et al. 2015). For example, mature S. fallax is observed to struggle at low water levels (Carroll et al. 2009), while having higher growth rates than any other species in wet conditions (Buttler et al. 1998). Recent studies have also highlighted the importance of precipitation on Sphagnum dominated peatlands carbon dynamics and species distribution (Robroek et al. 2009; Radu and Dual 2018). Nijp et al. (2014) found precipitation reduced the negative impacts of drought, while Robroek et al. (2007) found precipitation allowed species to survive outside of their hydrological niche. Precipitation also affects Sphagnum C uptake, with species responding differently to varying precipitation frequencies (Nijp et al. 2014). Water limitation has also been found to affect the impacts of other factors, such as temperature and vegetation’s impact on CO2 and CH4 fluxes (Silvola et al. 1996; Leppala et al. 2011).

Currently large tracts of damaged peatlands are being restored in the UK via application of Sphagnum containing beads and gel to encourage growth on degraded peatland sites (Hinde et al. 2010; Wittram et al. 2015). However, information is limited with regards to which species are most effective in quickly establishing and recreating peatland functions such as C sequestration, particularly in the context of global change, e.g., elevated atmospheric CO2 concentrations. Such information is urgently needed to ensure that restoration efforts are as effective as possible under future elevated atmospheric CO2 levels. For example, some studies indicate that elevated CO2 levels may result in increased CH4 fluxes peatlands dominated by Sphagnum species (Saarnio et al. 2000, 2003; Ellis et al. 2009) but effects differ among studies made across European peatland sites (Silvola et al. 2003). To date, a hand full of studies have considered a possible interaction between hydrology and raised atmospheric CO2 on the growth of Sphagnum moss (e.g., Jauhiainen et al. 1997; Toet et al. 2006) despite that interactions with hydrology may control how different Sphagnum species respond to elevated CO2. Indeed, interactions between hydrology and elevated atmospheric CO2 concentrations may cause the contrasting responses of Sphagnum species to elevated CO2 found in previous studies. Such interaction would have important implications for peatland restoration, suggesting that selecting the correct species for the environment is critical for restoration success.

The aim of this study was to determine whether there is an interaction between hydrology and elevated atmospheric CO2 on the growth and greenhouse gas (GHG) exchange of three Sphagnum moss species grown on peat-monoliths. To achieve this aim two specific hypotheses were tested: (i) elevated atmospheric CO2 will have a greater positive effect on S. fallax growth and GHG fluxes than other species; (ii) low water table will inhibit the effect of elevated CO2 on Sphagnum growth.

Materials and methods

Experimental set up

Peat cores were collected from Cadishead and Little Woolden Mosses (North West England, UK M44 5LR of Grid ref: SJ697953). This site is typical of drained and degraded cut over peatlands in the UK. The site is currently under restoration and some areas have been re-flooded. It is characterized by predominantly bare peat with cotton grass and birch growing around the edges with Sphagnum growing only in areas where reintroduction has been attempted.

To account for variation within the peat substrate, all 72 peat cores (approximately 18 cm × 16 cm) were collected from random locations within the non-flooded part of the peatbog. Collection and transportation of cores took place within 8 h, after collection cores were refrigerated at + 4 °C.

Three weeks after collection, peat cores were then cut into 17.4 cm × 11 cm cylinders and transplanted into 17.4 cm × 15.6 cm clear plastic, cylindrical polypropylene containers. The treatments were applied in a randomised block design (n = 4) with species (S. fallax, S. papillosum and S. capillifolium) and water level (1, 4 and 7 cm below the peat surface) as the treatments. Each combination of species and water level was repeated four times within each growth room. The randomized block design was used to account for any possible variations of conditions within each growth room. The two CO2 treatments used were 400 and 800 ppm CO2 (ambient and elevated), 800 ppm is the predicted 2100 concentration (IPCC 2013), due to the slow growth rates of Sphagnum this was chose as a good analogue for future restoration attempts. To achieve the two CO2 treatments cores were randomly allocated to one of two growth rooms in which the atmosphere was adjusted to the target CO2 concentration. Apart from the CO2 concentration the settings in the growth room were identical: Day length consisted of a 16 h photoperiod (300 µmol m−2s−1), including a 1 h dawn/dusk and 8 h night (0 µmol m−2s−1). Daytime conditions were set to a temperature of 21 °C and a relative humidity of 50%, while night-time conditions were set to 15 °C and 75% respectively.

Water levels

Once the peat cores were placed in the growth rooms, the three water level treatments were applied to the cores, which were then allowed to equilibrate for 1 week. All water applied to the cores initially and throughout the experiment was rainwater that was firstly collected as runoff from glasshouses located in East Leake, Loughborough, Leicestershire. The rainwater was then filtered through a sand gravity filter, leaving the rainwater free of particulate contaminants such as dirt washed from the glasshouse surface. Rainwater was used to simulate ombrotrophic peatland conditions. Water level treatments were monitored daily for the first month and then once every 2 days thereafter via visual inspection. If needed water was applied to the cores using a pressurised water sprayer to reduce disturbance to the peat surface. This kept variation in water level to within 1 cm throughout the course of the experiment.

Sphagnum application

After 1 week, each species of Sphagnum was applied to the relevant core using liquid gels, currently used for application to peatlands as part of restoration programmes, containing micro propagated material of each species. The Sphagnum suspended within the gel was micropropagated from source material originating from the Peak Districk, UK, by MicroPropagation Services (MicroPropagation Services (EM) Ltd 2015). The micro-propagated Sphagnum was suspended within a gel containing basic nutrients, at an approximate concentration of 150,000 individual plants per L, which ranged from 1 to 5 mm in length (for example gel see SI. 1). The gel was applied at a rate of 3L per m2 (0.072L per core) using a large pipette before being spread evenly over the core surface. The application rate is the same as trials conducted using the same gel in the restoration of the Cadishead peatland (Wright et al. 2012).

Measurements of Sphagnum performance and GHG fluxes

Measurements of Sphagnum performance and CO2 fluxes were made bi-weekly for each parameter apart from CH4 gas samples and dry weight, which were determined at the end of the experiment. Conditions within the growth rooms where consistent throughout the course of the experiment: Temperature was measured at 27.5 ± 0.1 °C in direct light with an air temperature of 21.2 ± 0.1 °C, relative humidity, 52 ± 1% and finally light was 234 ± 0.4 µmol m−2s−1.

Sphagnum height was measured using an adaption of the cranked wire method detailed by Clymo (1970). Initial height of the moss (peat level) was marked on the side of the clear plastic container; from this the height of the tallest Sphagnum plant in each individual container was measured to the nearest mm.

As Sphagnum was grown from micro propagated plants, the establishment rates of the different Sphagnum species could be quantified. This was inferred through the Normalized Difference Vegetation Index (NDVI), a numerical indicator that uses the visible and near-infrared bands of the electromagnetic spectrum to assess crop healthiness (Govaerts and Verhulst 2010). This was done using a hand held device (GreenSeeker®, Trimble Navigation Ltd., Sunnyvale California, USA), which gave a reading between 0 and 1 where 0 indicates bare soil and 1, established green vegetation (FSNAU 2016). The Green Seeker was held approximately 10 cm above the Sphagnum surface, with a reading for each core taken over 5 s while the device was continually moved over the whole surface of the core, giving an average reading for each core.

The total Sphagnum biomass production from each container was quantified at the end of the experiment. For this all Sphagnum growth in each container was removed, weighed (wet weight), dried at 70 °C and weighed again (dry weight).

CO2 flux and evaporation rate from each core was measured using an EGM-4 Infrared Gas Analyser (IRGA) connected to a CPY-3 using an open dynamic system (PP systems, Hitchin, UK). CO2 flux rates were recorded every minute over a 10 min period. IRGA time wishing was every 8 s, from which a linear regression is calculated to determine CO2 flux over each 1 min period. To ensure fluxes were not affected by disturbance following the installation of the CPY-3 chamber over the mesocosms, we used only the final three readings, these were averaged and used for subsequent data analysis. Positive fluxes represent a net release of CO2 from the mesocosms.

For CH4 flux determination, gas samples were collected from each container using a modified version of the closed static chamber method (Collier et al. 2014). Each container was carefully sealed using a lid with a septum in its centre giving a headspace of 0.11 dm3. Then, using a 30 ml syringe with a 25 gauge hypodermic needle, gas samples were then taken after 1, 30 and 60 min. Samples were then transferred into 12 ml pre-evacuated exetainers before each sample was analysed via gas chromatography (GC-2014; Shimadzu, UK) using a Flame Ionization Detector (FID), with hydrogen as the carrier gas. The oven temperature was 40 °C, the injector 80 °C and the FID 250 °C. The flow rate was 30 ml min−1; the column was a HayeSEp N packed column, 60/80 mesh. Measurements for each container were analysed using a linear regression model and the ideal gas law to calculate gas flux rate for each container, which was then used to calculate CH4 flux in mg m−2 h−1 where positive gas fluxes represent net CH4 emissions.

Statistical analysis

The effects of time, hydrology, species and CO2 treatment and their interactions on height, evaporation, CO2 flux and NDVI were tested using repeated measures analysis of variance (ANOVA) after box-cox transformations had been performed where necessary, to normalize the data. Time was treated as a categorical variable, with each measurement week treated as a separate category. Treatment effects on dry weight and CH4 flux were tested using ANOVA with hydrology, species, and CO2 treatment as fixed effects, where CO2 flux was normalised to the CO2 flux from the first set of measurements (week 3). This removed the impact of variation in the initial CO2 flux rates from the cores. All calculations were performed using the data analysis program GenStat 17th Edition for windows 10.1 (VSN International 2011).

Results

Growth

Different establishment rates were found between species (F(2,51) = 7.20, P = 0.002), with S. fallax establishing fastest with NDVI > 0.8 in week 7, which was 2 weeks earlier than S. papillosum and 3 weeks earlier than S. capillifolium (Fig. 1). Under eCO2 S. fallax and S. papillosum established faster, while S. capillifolium established slower than under the ambient treatment (F(2,51) = 2.96, P = 0.061).

Fig. 1
figure 1

The effect of elevated CO2 and species on NDVI (mean ± SE) over the 13 week experimental period. Black lines represent atmospheric CO2 of 400 ppm; light grey lines represent atmospheric CO2 of 800 ppm. Diamonds and solid lines represent S. fallax; squares and dashes lines represent S. papillosum; triangles and dotted lines represent S. capillifolium

Full size image

Sphagnum grew taller under eCO2 (F(1,51) = 16.34, P< 0.001) but this effect varied among species (F(2,270) = 5.31, P= 0.008) with S. fallax responding most significantly to eCO2 (F(2,51) = 48.78, P< 0.001) (Fig. 2a). S. papillosum height was also found to increase growth under eCO2, (2.9 mm taller by week 13) the effect was not as large as S. fallax (7.4 mm taller by week 13), while S. capillifolium indicated no difference in height between the treatments (Fig. 2a). The lowest water level (7 cm below peat surface) had shorter plants (near significant water level effect: F(2,51) = 2.49, P = 0.093) than both the medium (4 cm below peat surface) and high (1 cm below peat surface) water levels (Fig. 2b).

Fig. 2
figure 2

a The effect of elevated CO2 and species on plant height (mean ± SE) over the 13 week experimental period. Black lines represent atmospheric CO2 of 400 ppm; light grey lines represent atmospheric CO2 of 800 ppm. Diamonds and solid lines represent S. fallax; squares and dashes lines represent S. papillosum; triangles and dotted lines represent S. capillifolium. b The effect of water level on plant height (mean ± SE) over the 13 week experimental period. Diamonds and black line represents a water level 1 cm below peat surface; squares and dark grey line represent a water level 4 cm below peat surface; triangles and light grey line represent a water level 7 cm below peat surface

Full size image

The eCO2 treatment yielded 90.3% increase in biomass (294.2 ± 47.2 g m2) compared to the ambient treatment (154.6 ± 23.7 g m2) (F(1,51) = 30.94, P < 0.001). However, the effect size varied among species (F(2,51) = 4.17, P = 0.021), with S. fallax showing a 129.7% increase in biomass production under eCO2 (eCO2 142.4 ± 17.1 g m−2, ambient 62.0 ± 9.8 g m−2), followed by a 101.0% increase for S. capillifolium (eCO2 66.8 ± 16.1 g m−2, ambient 33.2 ± 6.4 g m−2) and finally a 43.2% increase for S. papillosum (eCO2 85.068 ± 14.0 g m−2, ambient 59.4 ± 7.6 g m−2) (Fig. 3a).

Fig. 3
figure 3

a The effect of species on dry weight (mean ± SE) recorded at the end of the 13 week experimental period. For each species, black bars represent growth under 400 ppm atmospheric CO2; light grey bars represent 800 ppm atmospheric CO2. b The effect of water level on dry weight (mean ± SE) recorded at the end of the 13 week experimental period. For each water level, black bars represent growth under 400 ppm atmospheric CO2; light grey bars represent 800 ppm atmospheric CO2

Full size image

Sphagnum grown at higher water level treatments produced more biomass (F(2,51) = 7.64, P = 0.001) (Fig. 3a). The eCO2 and water level treatments interacted (F(2,51) = 4.13, P = 0.022) with the high water level resulting in the largest increase of 142.0% (79.9 ± 9.2 g m−2) in biomass under eCO2 compared to a 66.0% increase (35.6 ± 8.1 g m−2) at the 4 cm and 19.2% (8.5 ± 5.4 g m−2) increase at the 7 cm water level treatments (Fig. 3b).

Gas flux

In parallel with the NDVI and height data, elevated and ambient CO2 treatments had almost identical CO2 flux rates over the first 7 weeks period. After 7 weeks there was a large increase in the CO2 uptake (increasingly negative CO2 flux) under eCO2 (F(1,51) = 14.59, P < 0.001). Specifically, CO2 uptake was more than 5-times greater in week 7 and consistently 3-times greater than the ambient treatment from week 9 to the end of the experiment (Fig. 4a; Time × eCO2 interaction: F(5,270) = 27.28, P < 0.001). Increasingly lower water levels were found to have a significantly larger CO2 uptake (F(2,51) = 5.00, P = 0.010) (Fig. 4b). CO2 fluxes were found to differ between species (F(2,51) = 4.29, P = 0.019), with S. capillifolium having the smallest uptake of CO2 (− 69.12 ± 14.74 mg m−2 h−1) compared to S. papillosum (− 101.30 ± 15.47 mg m−2 h−1) and S. fallax (− 91.56 ± 16.43 mg m−2 h−1) by week 13 (Fig. 4c). Over the course of the experiment, CO2 uptake increased more for S. fallax and S. papillosum over the course of the experiment than S. capillifolium (F(10,270) = 6.24, P < 0.001) (Fig. 4c).

Fig. 4
figure 4

a The effect of elevated CO2 on CO2 flux (mean ± SE) over the 13 week experimental period. Diamonds and black line represent 400 ppm atmospheric CO2; squares and grey line represent 800 ppm atmospheric CO2. b The effect of water level (cm below the peat surface) on CO2 flux (mean ± SE) over the 13 week experimental period. Diamonds and black line represent 1 cm below peat surface; squares and dark grey line represent 4 cm below the surface; triangles and light grey line represent 7 cm below the surface. c The effect of species on CO2 flux (mean ± SE) over the 13 week experimental period. Diamonds and solid line represent S. fallax; squares and dashed line represent S. papillosum; triangles and dotted line represent S. capillifolium. d The effect of species on CH4 flux (mean ± SE)

Full size image

Only the species treatment was found to have an effect on CH4 flux (F(2,123) = 3.66, P = 0.029), with S. fallax producing the most CH4 (5.28 ± 2.22 mg m−2 h−1), twice as much as S. papillosum (2.09 ± 1.09 mg m−2 h−1) and 10 times as much as S. capillifolium (0.48 ± 0.18 mg m−2 h−1) (Fig. 4d); the species and eCO2 treatments did not interact with the water level treatments.

Discussion

Growth responses

Our finding that eCO2 increased Sphagnum growth are in agreement with multiple studies (Jauhiainen and Silvola 1999; Heijmans et al. 2001a; Saarnio et al. 2003), however as this area of research is contested, the results are also in disagreement with several studies (Berendse et al. 2001; Grosvernier et al. 2001; Fenner et al. 2007).

The greater effect of elevated CO2 on S. fallax growth and establishment than either of the other species supports the first hypothesis which predicted that fast growing S. fallax would respond most positively to the eCO2 treatment. This is in agreement with previous studies (Berendse et al. 2001; Mitchell et al. 2002; Granath et al. 2012) which suggested that increased CO2 will increase the growth of certain species more than others. Variation in species responses may be caused by differences in the effectiveness of species to utilize the increased atmospheric CO2 concentration. Under elevated CO2 it’s possible that water became the limiting factor for growth, even under the highest water level treatments. Silvola (1991) showed that the differences in water needed for photosynthesis vary between species. Additionally, Nijp et al. (2014), found that Sphagnum species vary in their ability to exploit transient water resources. This could account for the differences seen in growth rates under increased CO2 between the species, with some species better able to exploit the elevated CO2 under the hydrological conditions.

Sphagnum also have higher water level limits due to blocking of CO2 diffusion to the leaf via the layer of water surrounding the leaf, with species differing on the extent of water level found limiting (Glime 2007). For example, optimum water content for photosynthesis for S. fuscum are lower than those required for S. angustifolium. This could have impacted Sphagnum growth within our experiment due to the high frequency of precipitation and surface water this provided to the Sphagnum.

We found that higher water levels coupled with eCO2 caused a greater increase in dry weight supporting our second hypothesis, which predicted that low water tables would inhibit the effect of elevated CO2 on Sphagnum growth. Similar results were found for S. fuscum by Jauhiainen et al. (1997), who suggested that this was due to water limitation of photosynthesis of Sphagnum at low water levels. We did not find an interaction effect of hydrology on species, in contrary to previous studies, which have identified that each species of Sphagnum grows best within its own hydrological niche (Grosvernier et al. 1997). It’s also been found that each species responds to variation in water tables differently (Jauhiainen et al. 1997; Carroll et al. 2009) due to differences in their physiologies. In particular, S. fallax has been found to be highly dependent on water level, holding 30% less water than S. capillifolium (Carroll et al. 2009). One reason for this could be the impact of precipitation rewetting on reducing the impact of drought by precipitation increasing the water contents of the top layer (Nijp et al. 2014), negating the impacts of a lower water table. However, differences between species utilization of precipitation was seen by Nijp et al. (2014) between S. majus, S. balticum and S. fuscum, which was not seen within our findings. This suggests that our experiment did not differentiate water level to a high enough degree to examine differences between the species.

CO2 and CH4 flux responses

Greater CO2 assimilation under eCO2 was found in S. fallax and S. papillosum than S. capillifolium, this lends additional support to our first hypothesis which predicted that elevated CO2 would increase C sequestration due to increased growth rates but that responses would differ among species. Our finding of increased CO2 uptake under eCO2 agrees in part with findings by Van der Heijden et al. (2000), who found that photosynthesis in mature plants of S. fallax was initially stimulated by elevated CO2 (700 ppm). However, in contrast to our findings, acclimation was observed with stimulated S. fallax returning to the levels of the controls after 3 days. This could have been due to differences between establishing and mature S. fallax.

Our findings somewhat contrasted with Harpenslager et al. (2015), who grew four Sphagnum species on peat monoliths incubated between 18 and 23 °C over a 12 week period. Harpenslager et al. (2015), found that S. fallax and S. palustre had positive net C effluxes similar to bare peat in spite of increased biomass. Harpenslager et al. (2015) suggested that these results were due to the release of CO2 derived from bicarbonate due to lowering of the monoliths pH. While this is dissimilar to our findings, these releases were species dependant, highlighting how different species can impact total CO2 flux beyond sequestration. These differences in growth and C fixation between the species were in congruence with our results.

Overall our findings suggest that in the context of future elevated CO2, all three species of Sphagnum will increase in CO2 assimilation due to increased growth, however, it is suggested that not all species will react the same, with certain species seeing a much greater increase in CO2 assimilation (in this case S. fallax and S. papillosum) compared to others. Whilst our study did not find acclimation (such as Van der Heijden et al. (2000), this could be due to our focus on establishing Sphagnum and not mature plants.

Greenhouse gas budgets for peatlands are often complex, with most peatlands being sources of CH4 (Moore and Roulet 1995; Saarnio et al. 2007; Leppala et al. 2011). As CH4 has a Global Warming Potential (GWP) of 28 over 100 years (IPCC 2013), the increased CO2 sink strength in response to eCO2 shown here could be offset if CH4 emissions increase. In our 13-week study CH4 emissions remained low, although they increased slightly at the end of the experiment.

The GWP of the CH44 emissions over 100 years (CO2 equivalents (CO2eq)) of the three species was 147.84 ± 62.12, 58.52 ± 23.98 and 13,44 ± 5.04 mg m−2 h−1 for S. fallax, S. papillosum and S. capillifolium respectively. If emissions of CO2 and CH4 from each species are taken into account, S. fallax would have a positive net emission (56.28 mg m−2 h−1 CO2eq), while S.papillosum and S.capillifolium would have a negative net emissions (− 42.78 and -55.68 mg m−2 h−1 CO2eq respectively). This suggests that while S. fallax was shown to have the greatest C assimilation over the course of the experiment, S. papillosum and S. capillifolium had lower net emissions. This highlights the importance of CH4 release within these systems which can offset carbon uptake to a substantial amount and even become a net source of emissions.

The increase in CH4 emissions towards the end of the experiment was likely due to recovery of methanogenic populations following the increased moisture content of the peat (Bellisario et al. 1999; Leppala et al. 2011) and increased C availability due to increased photosynthesis (Megonigal and Schlesinger 1997) and C transportation to the rhizosphere (Oremland 1988; Williams et al. 2000). The lack of response CH4 fluxes to eCO2 in our study contrasts with findings from FACE studies on UK and Finnish ombrotrophic peatlands which found modest increases in peatland CH4 fluxes in response to eCO2 (Saarnio and Silvola 1999; Ellis et al. 2009). An important difference between our study and those of Saarnio and Silvola (1999) and Ellis et al. (2009) are the fact that our mesocosms contained a sole Sphagnum species while the plots in the other two studies encompassed both Sphagnum mosses, sedges and dwarf shrubs making it difficult to isolate Sphagnum species responses in the two field studies.

The species-specific effects on CH4 fluxes of the three Sphagnum species used in this study is important as it suggests that the Sphagnum species composition per se, and not just the mix of functional groups (Strom et al. 2005; Leppala et al. 2011) is an important control of CH4. Although the mechanisms behind the species-specific effects on CH4 fluxes are unclear. They may relate to differences in production of labile carbon and/or phenolic compounds among Sphagnum species (Ellis et al. 2009). Transportation of C to the rhizosphere has been shown to vary with species (Reich et al. 2001). It’s possible that similar variation may account for the differences seen in emissions between the Sphagnum species.

Furthermore, relationships between species and specific methanogenic and methanotrophic communities could have also had an impact. This has been shown for plant functional types within peatlands by Robroek et al. (2015), which showed removal of plant functional types (gramminoids and ericoids) significantly alter microbial community structure. This fit’s with Opelt et al. (2007) who found that diversity of bacterial communities was higher in S. fallax than S. magellanicum.

Conclusions

Our study highlights the importance of Sphagnum species selection when undertaking restoration of a peatland. The clear differences in species establishment and growth under increased CO2 and CH4 flux between the species indicates that selecting the correct Sphagnum species or combination of species is critical to the success and GHG balance of recolonization attempts. While water level was not shown to interact with species, it’s interaction with CO2 suggests that environmental factors may also play an important role in establishment success.

Furthermore, differences between the species CO2 and CH4 fluxes within the initial establishment period, suggests that using different species will impact the overall GHG balance the peatland once restored. Within the establishment phase, the increase in CO2 uptake found in response to eCO2 suggests that the capacity to sequester C by Sphagnum dominated peatbogs may increase under future CO2 rich atmospheric conditions. Nevertheless, there may be trade-offs between plant performance and GHG emissions as our fastest growing species S. fallax also had the highest CH4 emissions, which offset assimilation and caused net GHG emission. While the estimation of the impacts to the final GHG balance of a restored peatland are beyond the scope of this paper, this highlights that species may be and important considered when determining which Sphagnum species should be used for peatland restoration.