Release of Carbon in Different Molecule Size Fractions from Decomposing Boreal Mor and Peat as Affected by Enchytraeid Worms

Abstract

Terrestrial export of dissolved organic carbon (DOC) to watercourses has increased in boreal zone. Effect of decomposing material and soil food webs on the release rate and quality of DOC are poorly known. We quantified carbon (C) release in CO₂, and DOC in different molecular weights from the most common organic soils in boreal zone; and explored the effect of soil type and enchytraeid worms on the release rates. Two types of mor and four types of peat were incubated in laboratory with and without enchytraeid worms for 154 days at + 15 °C. Carbon was mostly released as CO₂; DOC contributed to 2–9% of C release. The share of DOC was higher in peat than in mor. The release rate of CO₂ was three times higher in mor than in highly decomposed peat. Enchytraeids enhanced the release of CO₂ by 31–43% and of DOC by 46–77% in mor. High molecular weight fraction dominated the DOC release. Upscaling the laboratory results into catchment level allowed us to conclude that peatlands are the main source of DOC, low molecular weight DOC originates close to watercourse, and that enchytraeids substantially influence DOC leaching to watercourse and ultimately to aquatic CO₂ emissions.

Appendix 1. ‘Numerical Discussion’

We present a numerical discussion on how our results, combined with existing literature, could reflect to DOC quality and quantity in water courses. The calculation procedure includes the following steps: (1) scaling from soil sample level to site level; (2) scaling from constant to changing temperature; (3) scaling from site to catchment level; (4) computing transport time, biodegradation and DOC export load; 5) results and comparison to literature; and (6) evaluation of the computation.

1.1 Scaling from Soil Sample to Site Level

A catchment consists of upland sites (subscript u) with a mor layer on the top of mineral soil, and peatland sites (subscript p). For upland sites, we set the thickness of mor to 0.04 m according to a typical mor depth in this study. For peatland sites, we assumed that DOC release takes place above the mean water table level, which was set to 0.31 m according to Ojanen et al. (2010). This peat column sets the frame into which the peat results from this study were embedded: the top part of the peat column (depth 0–0.2 m) was parameterized with the results of slightly decomposed peat and the rest (depth 0.2 to 0.31 m) with highly decomposed peat. The parameters for the organic soil layers were derived from this study by averaging (separately with and without enchytraeids) soil types 1 and to 2 to ‘mor’, soil types 3 and 5 to ‘slightly decomposed peat’, and soil types 4 and 6 to ‘highly decomposed peat’ (Table 3). The hectare-based dry mass (M_i, kg ha⁻¹) was obtained as a product of the layer thickness and bulk density, and was 35,400 kg ha⁻¹ for mor, 135,600 kg ha⁻¹ for slightly decomposed peat and 149,600 kg ha⁻¹for highly decomposed peat.

1.2 Scaling from Constant to Changing Temperature

The DOC release rates from this study (γ_ik, Table 3) were adjusted for temperature using Q₁₀ approach (Laurén et al. 2012) and mean monthly air temperature in Finland. The annual release of DOC was obtained by multiplying the adjusted release rate with M_i and summing over the months.

$$ {\mathrm{DOC}}_{\mathrm{annual}\_ ik}={\sum}_{m=1}^{12}{M}_i{\gamma}_{ik}{Q}_{10}^{\left(\left({T}_m-{T}_{ref}\right)/10\right)}{dt}_m\ast {10}^{-6}, $$

(A1)

where DOC_{annual_ik} is the annual DOC release (kg ha⁻¹ year⁻¹) for soil type i (mor, slightly decomposed peat, highly decomposed peat) and DOC fraction k (LMW, HMW), M_i is the hectare-based dry mass of soil type i (kg ha⁻¹), γ_ik is the release rate (μg g⁻¹ dry mass day⁻¹) of DOC fraction k for soil type i, T_m is monthly mean air temperature (°C, − 9.3, − 9.3, − 4.8, 1.0, 7.4, 12.6, 15.6, 13.4, 8.3, 2.8, − 3.2, − 7.3, http://ilmatieteenlaitos.fi/kuukausitilastot), dt_m is the length of month m in days, T_ref is the reference temperature (15 °C) and Q₁₀ = 3.0 is a parameter. Now, the annual DOC release for upland sites (DOC_{annual_uk}) is directly obtained from Eq. A1 solved with i = ‘mor’:

$$ {\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{uk}}={\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{mor}\_\mathrm{k}} $$

(A2)

and for peatland sites the release (DOC_{annual_pk}) is obtained as a sum of Eq. A1 solved for i = ‘slightly decomposed peat’ and ‘highly decomposed peat’.

$$ {\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{pk}}={\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{slightly}\_\mathrm{decomposed}\_\mathrm{peat}\_\mathrm{k}}+{\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{highly}\_\mathrm{decomposed}\_\mathrm{peat}\_\mathrm{k}} $$

(A3)

1.3 Scaling from Site to Catchment Level

In this calculation, we used average characteristics of water flow paths in head water catchments in Central Finland as analysed by Korkalainen et al. (2007). The authors used 782 head water catchments to determine water flow path length, elevation above the receiving water body, site type (upland, peatland) and relative catchment area as a function of distance to receiving water body (Fig. 4). Peatlands, comprising on average 17% of the area, were located close to water bodies; and uplands had steeper slope than peatlands did. The total length of the characteristic hillslope was 925 m, and for the computation it was discretized into 25-m intervals (dx = 25 m). The centre points of the intervals are called nodes (number of nodes N = 37). To upscale the DOC release from site to catchment, we located DOC_{annual_uk} to upland nodes and DOC_{annual_pk} to peatland nodes using the soil type information in Fig. 4 (node DOC release in node n and fraction k is referred as DOC_{annual_kn}), and the relative area (A_n) for node n. Now the total catchment scale release of DOC was obtained as area weighted average of DOC_{annual_kn} along the flowpath

$$ {\mathrm{DOC}}_{tot\_k}=\frac{\sum_{n=1}^N{\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{kn}}{\mathrm{A}}_{\mathrm{n}}}{\sum_{n=1}^N{A}_n}, $$

(A4)

where DOC_{tot_k} is the total catchment scale release of DOC (kg ha⁻¹ year⁻¹) in fraction k (LMW, HMW), n is the computation node, N is the number of nodes, DOC_{annual_kn} is DOC release in fraction k (kg ha⁻¹ year⁻¹) in node n and A_n is the relative area of the dx interval where node n is situated.

Fig. 4

Measured DOC export load after Kortelainen et al. (2006) (black circles), fitting Eq. A7 to black circles (grey line), and the export load estimated here (with worms: red circles, without worms: green circles). The red area represents sensitivity of the obtained export to hydraulic conductivity of soil k_sat (for the upper limit is applied k_sat*10 and for the lower limit k_sat*0.1)

1.4 Transport Time, Biodegradation and DOC Export

Next, we assumed that water flow in soil follows the surface gradient described in Fig. 4. We computed the time needed for DOC transport (with water, omitting sorption reactions) from node n to receiving node (n = 0 in watercourse) as:

$$ {t}_n={\sum}_{m=1}^n\frac{dx}{k_{s\mathrm{a}t\_m}\ {g}_m\varphi\ 86400} $$

(A5)

where t_n is the transport time from a node n through all nodes m between the watercourse and node n (days), dx is discretization interval (25 m), k_{sat_m} is the horizontal saturated hydraulic conductivity and φ is porosity in node m (if peat k_sat = 3.4*10⁻⁴ ms⁻¹ and φ = 0.9 Koivusalo et al. 2008; if mineral soil k_sat = 1.5*10⁻⁴ ms⁻¹ and φ = 0.5 Laine-Kaulio 2011) and g_m is the slope gradient around node m (Fig. 4, m m⁻¹). The transport time is shown in Fig. 4.

Now biodegradation of LMW- and HMW-DOC can be computed using the decay function presented by Kalbitz et al. (2003a):

$$ {\mathrm{DOC}}_{\mathrm{rem}\_\mathrm{kn}}={\mathrm{DOC}}_{\mathrm{annual}\_\mathrm{kn}}{e}^{-{d}_k{t}_n} $$

(A6)

where DOC_{rem_kn} is the remaining DOC in fraction k from node n after biodegradation during the transport time t_n (days), DOC_{annual_kn} is the release of DOC in fraction k in node n, d_k is the biodegradation rate constant for fraction k (d_LMW = 0.15 day⁻¹, d_HMW = 0.0004 day⁻¹, Kalbitz et al. 2003a). Now it is possible to compute the share of the produced DOC that remains nondegradated after the transport (Fig. 4). By scaling DOC_{rem_kn} with the relative area A_n, we obtain an estimate of DOC export to water course and its origins from the catchment in LMW and HMW fractions (Fig. 4).

1.5 Results and comparison to literature

The total DOC release from peatland was 183.5 kg ha⁻¹ year⁻¹ and for upland 42.3 kg ha⁻¹ year⁻¹, thus the area weighted average was 66.1 kg ha⁻¹ year⁻¹ for the whole catchment. From this amount, 12.6 kg was degraded during the transport and 53.5 kg reached the water course. Peatland contributed to 45.5% of the DOC export even though it covered only 17% of the catchment area. HMW-DOC dominated the export, and negligible amount of LMW-DOC reached the watercourse, even though in average 11.9 kg ha⁻¹ year⁻¹ LMW-DOC was released at the catchment scale. Therefore, if any LMW DOC is present in watercourse, it has to originate from the close proximity of the water body (< 12.5 m).

According to Kortelainen et al. (2006), the range of DOC export in undisturbed catchment in Finland is 10–140 kg ha⁻¹ year⁻¹, and the export increases with increasing proportion of peatland in area. For the same data, Palviainen et al. (2016) presented the following dependency between the DOC export load (DOC_export, kg ha⁻¹ year⁻¹) and peatland proportion (A_p %):

$$ {\mathrm{DOC}}_{\mathrm{export}}=13.97{\left({\mathrm{A}}_{\mathrm{p}}+1\right)}^{0.45} $$

(A7)

Plugging in the average peatland proportion of 17% in this example gives 51.3 kg ha⁻¹ year⁻¹ which is remarkably close to the export load of 53.5 kg ha⁻¹ year⁻¹ gained in our simple computation. When the same analysis was computed with the DOC release rates obtained from the incubation without enchytraeid worms (Table 3), the DOC export was 14 kg ha⁻¹ year⁻¹ lower.

To test whether our estimate holds with different shares of catchment peatland area, and different k_sat values, we repeated the above computation by extending gradually the peatland coverage from node 1 to node 25 (ref. Fig 4) giving peatland coverages of 9.4 to 88.3%. Plotting the computed export loads with the data presented by Kortelainen et al. (2006), and Eq. A7, reveals a remarkably good correspondence (Fig. 4). Repeating the computation with the no worms–parameter set (Table 3), the export load was from 14 to 20 kg ha⁻¹ year⁻¹ lower.

Our numerical discussion allows concluding that the decomposing materials, i.e. mor and peat, and division of the released C into CO₂, LMW-DOC and HMW-DOC, can play an important role in DOC export to water courses, and ultimately in ecosystem C balance. Enchytraeid worms can substantially enhance the DOC leaching from terrestrial ecosystem to watercourse.

1.6 Evaluation of the computation

The set-up of our computation represents a steady-state situation of DOC fluxes, and the described processes are simplified and many other processes, such as DOC retention in soil, have been omitted. DOC sorption in mineral soil follows a saturating curve as demonstrated by, e.g. Kothawala et al. (2008), indicating that DOC is retained efficiently into pristine soil and thereafter gradually the net DOC sorption decreases. It is likely that soil is after 10,000 years of DOC input in a slower phase of sorption. When interpreted in this context, the outcome is interesting: the magnitudes of DOC release and biodegradation seems plausible even if the role of DOC retention was neglected, suggesting a small net retention of DOC.

The transport mechanism was treated following an equally simplistic way. Implicitly, water is moving in a deep saturated layer with velocity determined by the soil surface gradient, hydraulic conductivity and soil porosity. This transport takes place below the soil frost layer which typically extends to less than 50 cm depth in Finland (Venäläinen et al. 2001), and therefore the transport is mainly unaffected by winter conditions.

Table 3 Parameters derived from the results of the incubation experiment and used in the numeric discussion