Long-term carbon and nitrogen dynamics in boreal peatlands are affected by both vegetation production and decomposition processes. Here, we examined the carbon accumulation rate (CAR), nitrogen accumulation rate (NAR) and δ¹³C, δ¹⁵N of plant residuals in a peat core dated back to ~8500 cal yr BP in a temperate peatland in Northeast China. Impacted by the tephra during 1160 and 789 cal yr BP and climate change, the peatland changed from a fen dominated by vascular plants to a bog dominated by Sphagnum mosses. We used the Clymo model to quantify peat addition rate and decay constant for acrotelm and catotelm layers during both bog and fen phases. Our studied peatland was dominated by Sphagnum fuscum during the bog phase (789 ~ -59 cal yr BP) and lower accumulation rates for the upper sections in the acrotelm layer during this phase, suggesting the dominant role of volcanic eruption in the CAR of the peat core. Both mean CAR and NAR were higher during the bog phase than during the fen phase in our study, consistent with the results of the only one similar study in the literature. Because the input rate of organic matter was considered to be lower during the bog phase, the decomposition process must have been much lower during the bog phase than during the fen phase and potentially controlled CAR and NAR. During the fen phase, CAR was also lower under higher temperature and summer insolation, conditions beneficial for decomposition. δ¹⁵N of Sphagnum hinted that nitrogen fixation had positive effect on nitrogen accumulation, particular in recent decades. Our study suggested that decomposition is more important for carbon and nitrogen sequestration than production in peatlands in most conditions and if future climate changes or human disturbance increase decomposition rate, carbon sequestration in peatlands will be jeopardized.

Study site description and sampling

The peat core (42.185° N, 128.312° E) was collected in Dongfanghong peatland in the basalt plateau platform of Changbai Mountains region, with a present mean elevation of 1730 m, a total area of 16 km2, MAP of 780 mm, and MAT of 2.2 ℃. The present dominant plants include Vaccinium uliginosum, Cyperaceae (Carex and Eriophorum angustifolium), S. fallax, S. magellanicum and S. fuscum. The length of the peat core was 2 m. The core was divided into 1 cm slices for plant macrofossil and elemental analysis. The average DWT near the studied peat core was ~18 cm based on the monitor for three years (Yang et al., 2022).

AMS ¹⁴C dating of the peat core

The chronology of the peat core was based on the ¹⁴C accelerator mass spectrometer (AMS) analyses of plant residual stems and leaves subjecting to an acid–alkali–acid treatment (Table 1). The graphite samples were finished in the AMS¹⁴C dating preparation lab in Northeast Normal University and then were sent to the NTUAMS laboratory at the National Taiwan University for Accelerator Mass Spectrometer (AMS) measurement with a HVE 1.0 MV Tandetron Model 4110 BO-AMS. The measured 14C ages were converted to calibrated calendar ages (BP = years before 1950 AD) with the IntCal20 calibration curve52 using the CALIB 8.10 program and expressed as cal yr BP with 2σ ranges. The calendar age was calibrated using the Bacon age-depth model (Figure 1)

Plant macrofossil analysis

Sub-samples (1 cm³) for plant macrofossils were analyzed using the standard method (Mauquoy et al., 2010). Plant residues were washed with 5% NaOH and then examined with a stereomicroscope at 10 magnification and macrofossils were identified. Moss leaves and small seeds were examined at high magnifications (×100 ~ 400). Three plant groups: Ericaceae, Carex and S. fuscum were picked out for further isotope analyses because of their high abundance (> 10 mg for each sample of each group) along the whole core and their residues were stored in a 5 ml centrifuge tube in a 4 ℃ refrigerator. Tilia 2.0.4154 was used to calculate plant macrofossil percentages and to plot the figure.

Carbon and nitrogen isotope analysis of plant residuals 

Each residue sample of the three selected groups was dried at 50 ℃ for 48 h, homogenized in a ball mill, and analyzed for C and N concentrations and isotopic abundance in an elemental analyzer coupled with an isotope ratio mass spectrometer (DELTA V Advantage; ThermoFisher, Germany).

C and N isotopic abundance was expressed as δ¹³C and δ¹⁵N and was calculated as follows: δ (‰) = [(R_sample/R_standard)-1)], where R is the ¹³C/¹²C and ¹⁵N/¹⁴N ratio of the sample and the correspondent standard, Peedee belemnite for δ¹³C and air N₂ for δ¹⁵N. To check the system stability, standard material with known δ¹³C and δ¹⁵N values were analyzed between every ten samples. Standard deviations of laboratory standards for δ¹³C and δ¹⁵N were less than 0.15‰.

Δ¹³C of S. fuscum was calculated according to Δ¹³C = 1000* (δ¹³C_a - δ¹³C_moss)/(1000 + δ¹³C_moss) (Farquhar et al., 1982). δ¹³C_moss: δ¹³C of Sphagnum moss; δ¹³Ca : δ¹³C of atmospheric CO_2. The value of δ¹³C of atmospheric CO₂ was reconstructed from ice cores (Schmitt et al., 2012); Δ¹³C: the net photosynthetic discrimination, which is usually dominated by fractionation against ¹³CO₂ by Rubisco (O’Leary, 1988). Higher Δ¹³C value of S. fuscum means higher photosynthesis rate (Raghoebarsing et al., 2005).

Analysis of bulk peat and calculation of measured CAR, NAR

Each peat sample was weighed before and after drying at 105 °C until constant weight was reached. Dry bulk density (DBD) was calculated from the dry weight and sample volume. The dried samples were milled and homogenized for C and N content analysis. C and N accumulation rates were calculated by the following equation based on raw data including peat accumulation rates (PAR), DBD and C, N contents. CAR = PAR * DBD * C (%) * 100；NAR = PAR * DBD * N (%)* 100. Peat accumulation rates (PAR) (cm yr-1) for each profile were calculated for peat layers between the dated depths of the peat profile (Tolonen and Turunen, 1996).

Literature data

We compared our results with literature data, which was from literature search on Web of Science. The criteria of literature search were: There was a fen to bog transition after disturbance such as volcanic activity or fire, and both CAR and NAR of the peat core were provided. Although several peat cores were reported in the literature, which developed from a fen phase dominant by vascular plants to a bog phase dominant by Sphagnum mosses due to disturbance, only the MIL peat core in van Bellen et al., 2020 had both CAR and NAR values. This was why only one similar study in the literature was finally used for comparison with our study. Based on the plant macrofossil of MIL peat core (Magnan et al., 2018), the core was divided into fen (960 ~ 1800 AD) and bog (1800 ~ 2013 AD) phases. The average, maximum and minimum values of CAR and NAR were recalculated for both phases.

Peat decay and modeling of peat addition and decomposition rates

Peat surface is always incompletely decomposed and this needs to be carefully taken into account when comparing recent CARs with those of the past. Based on the bacon age and measured DWT of our studied peat core, the switch between acrotelm and catotelm of occurred at ~16 cm (after 1850 AD) during the bog phase. For the catotelm, peat addition rate (PAR) was quite different between fen and bog phases. The PAR of fen phase should follow the concave model and PAR of bog phase should follow the convex model (Yu eal., 2003). For the fen phase, peat decomposition can be modeled as catotelm following Clymo (1984). The exponential decay model developed by Clymo was used to derive PAR (p) and peat decay constant (α) from the observed cumulative peat organic matter pool (M). The exponential decay model assumes that p and α remain constant over time and follow the equation of M = (p/α)* (1-e^-α*t). For the bog phase of catotelm, CAR varies with peat age and the exponential decay model should be modified using the concave model following Yu et al., 2003. The extended model we used was: M = (p/α-b)* (e^-b*t-e^-α*t), where b is the coefficient of PAR modifier. The convex model was also used to model PAR of the acrotelm during the bog phase. CAR was then calculated following Frolking et al. (2001): M(t) = p/(1+α*t), where M was the remaining amount of peat that was eventually buried into deeper layers and was multiplied by the assumed 50% C content in peat organic matter to calculate modeled CAR (Loisel and Yu, 2013). Modeled NARs of the studied peat core were recalculated based on bulk peat C/N.

Other data sources

Based on the plant macrofossil of the peat core, quantitative DWT was reconstructed by mire plant assemblages-based transfer function of WATOL-inv model (Yang et al., 2022). Higher DWT means drier hydroclimate and lower mire surface wetness. Atmospheric CO₂ concentrations were reconstructed from the ice core data previously reported (Köhler et al., 2017). Mean temperature of the warmest month (Mtwa) were reconstructed by the pollen record from Sihailongwan Lake, which is very close to our study site (Stebich et al., 2015). Summer insolation data were from Berger and Loutre (1991).

Calculations and statistics

The difference in peat properties between the bog and fen phases was tested using one-way analysis of variance (ANOVA) and a Tukey HSD post-hoc test. The level of significance (α) was set at 0.05. Pearson’s correlation analysis between CAR, NAR and environmental variables were conducted. All statistical analyses were conducted using SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Power spectral analyses of Carex coverage, Sphagnum coverage, reconstructed DWT and Carex δ¹³C were calculated using the REDFIT 38 spectral analyses program to detect periodicity (Schulz and Mudelsee, 2002). Peaks above the false-alarm level (90%, 95% and 99%) were labeled as their cyclicity.