Comparison of organic carbon properties in extracted soil solutions obtained underneath Cryptomeria japonica and Quercus acutissima and its implication on stream dissolved organic carbon

Abstract

Dissolved organic carbon (DOC) in soils is released into streams, working as a main component of the carbon cycle. DOC in terrestrial and aquatic ecosystems participates in many biogeochemical reactions and processes such as heterotrophic respiration, sorption of metals, and transport of pollutants. In order to understand the connectivity of organic carbon among soil, soil water, and forest streams, we investigated the concentrations and dual carbon isotope ratios (δ¹³C and Δ¹⁴C) of soil organic carbon (SOC) and water-extractable organic carbon (WEOC) obtained from soils beneath two tree species stands, and compared these with stream Δ¹⁴C-DOC of the forest watershed. Soil samples were collected at different depths (0–10, 10–30, and 30–50 cm) beneath Japanese cedar (Cryptomeria japonica) and sawtooth oak (Quercus acutissima). Although the SOC concentration was not significantly different between the two tree species, the WEOC concentration ([WEOC]) of soil at 0–10 cm depth under Cryptomeria japonica was higher than that of Quercus acutissima, in general. The δ¹³C-SOC and δ¹³C-WEOC increased, while the Δ¹⁴C-SOC and Δ¹⁴C-WEOC decreased with increasing soil depth. The Δ¹⁴C-WEOC was higher than the Δ¹⁴C-SOC, indicating that WEOC could be primarily derived from the young, hydrophilic, and exchangeable fraction of SOC, rather than from SOC strongly bonded to mineral soils. However, the stream Δ¹⁴C-DOC was lower than Δ¹⁴C-WEOC in general, except during summer storms. The ¹⁴C-depleted DOC released from deep soils or groundwater might lower Δ¹⁴C-DOC of a stream, suggesting that relatively old DOC could be released into streams during baseflow. This is contrary to the results of previous studies that have reported positive stream Δ¹⁴C-DOC from temperate forests. The discrepancy warrants future research on forest stream Δ¹⁴C-DOC across entire seasons, particularly under Asian monsoon climates.

Keywords:

• Soil organic carbon (SOC)
• water-extractable organic carbon (WEOC)
• dissolved organic carbon (DOC)
• ¹⁴C
• ¹³C
• stream

1. Introduction

Dissolved organic carbon (DOC) is an important component of many reactions and processes, such as, the transport of pollutants from soils to surface waters, heterotrophic respiration, and soil formation (Magee et al. 1991; Kalbitz et al. 2000; Lundstrom et al. 2000; Delle Site 2001). DOC can be derived from throughfall, litter leaching, root exudates, or decomposition of soil organic carbon (SOC) (Bolan et al. 2011). The soil DOC can be leached to groundwater or transported to streams (Kalbitz et al. 2000; Bolan et al. 2011), thereby potentially altering the regional or global carbon cycle (Raymond et al. 2013; Cory et al. 2015).

Many studies have investigated factors which can influence the leaching of DOC from soils. Tree species may influence soil DOC, as such higher concentration of DOC ([DOC]) was observed in soils beneath coniferous stands compared to broadleaf stands (Currie et al. 1996; Borken et al. 2011; Fröberg et al. 2011). The content of organic carbon (OC) in the forest floor was apparently higher in pine stands as compared to broadleaf stands (Cha et al. 2019), supporting that higher [DOC] could be observed in soils beneath coniferous forests. Precipitation and hydrologic regime also play important roles in the release of soil DOC because more DOC can be released by surface runoff during high-flow periods, such as, rainfall and snowmelt (Lambert et al. 2013; Raymond et al. 2016).

The studies on soil DOC have been conducted by directly sampling soil water using a lysimeter or by extracting soils with a solution in the laboratory (Table 1). Although the lysimeter can collect soil water in situ, it is difficult to obtain a sufficient amount of soil water for isotope analysis, and thus, extracted soil solution has been widely used to analyze carbon isotope ratios. Both the concentrations and characteristics of extracted DOC (i.e. water-extractable organic carbon: WEOC), are dependent on extracting conditions, including soil-to-water ratio, extracting time, and water temperature (Hagedorn et al. 2004; Jones and Willett 2006; Chantigny et al. 2014; Guigue et al. 2014). In general, a higher soil-to-water ratio, longer extraction time, and higher water temperature increased the concentration of WEOC ([WEOC]) (Jones and Willett 2006; Chantigny et al. 2014; Guigue et al. 2014).

Table 1. Carbon isotope analysis on SOC and soil DOC in forest ecosystems.

References	Collected/ Extracted^a	Location	Extracting method (soil:water mass ratio)	Controlling factors	Depth (cm)	Δ^14C-SOC (‰)	δ^13C-SOC (‰)	Δ^14C-DOC (‰)	δ^13C-DOC (‰)
van der Voort et al. (2019)	Extracted	46–47^oN, 6–10^oE	1:4	Climate, Soil type	soil: 0–40 soil DOC: 0–100	−416 to 211		−164 to 205
Nakanishi et al. (2012)	Extracted	40^o00′N, 140^o56′E	1:5	Depth, OC fraction	0–60	−19.8 to 78.6	−28.3 to −25.8	6–102	−27.5 to −24.7
Nakanishi et al. (2014)	Extracted	40^o00′N, 140^o56′E	1:5	Depth, Climate	0–15	−19.8 to 78.6	−28.3 to −25.8	0–120	−28 to −24.5
Sanderman et al. (2008)	Extracted		1:50	Depth, LULC	0–∼70	−442 to 107	−28 to −23	−488 to 111	−28 to −23
Tegen & Dörr (1996)	Collected	51.6°N, 7.4°E 49.2°N, 8.4°E	lysimeter	Tree species, Climate	soil: 0–13 soil DOC: 0–50	40–275	−30.5 to −24.2	29–105	−29.9 to −27.6
Dutta et al. (2006)	Extracted	69°31′N, 161°52′E 69°39′N, 162°31′E	1:10	Temperature, Depth	surface: 0–10 mineral soil: 200–1300	−993 to −421	−26 to −22	−938 to −20	−26.5 to −24.6
Trumbore et al. (1992)	Collected/ Extracted	45°23′N, 79°08′W	from 1:10 to ∼ 1:2 lysimeter (only high flow period)	Depth, Acidity	detrital layers, 0–40	175 (litter)	−30 to −28 (litter), < −150 (10–40 cm)	175^b (litter), 10^b (10–25 cm)	−27^b
Fröberg et al. (2007)	Collected	35°58′N, 84°16′W	lysimeter	Soil type, Litter addition, Depth	15, 70			42–399
Fröberg et al. (2009)	Collected	35°58′N, 84°16′W	drain (the day after a major rainfall)	Litter addition, Depth	10 (O and A horizon)	123–167 (A horizon)		112–453 (A horizon)
Michalzik et al. (2003)	Collected	58°22′N, 08°13′E 50°08′N, 11°52′E	lysimeter	Depth	9–45	99–126		101–120
Karltun et al. (2005)^c	Collected		lysimeter	LULC, Stand age	soil: 5–15 soil DOC: 15–35	−50 to 100		−10 to 110
Hensgens et al. (2021)	Extracted	64°14′N, 19°46′E	from 1:40 to ∼1:4	Depth	up to ∼75	−130 to 126		493–195

^aCollected: DOC collected in situ from soils using lysimeter, sampling probe, or glass bottle. Extracted: DOC extracted from soil samples with deionized water in laboratory. Only van der Voort et al. (2019) used 0.5% NaCl solution to extract DOC from soils.

^bArithmetic mean.

^cΔ¹⁴C-SOC or Δ¹⁴C-DOC was calculated from ¹⁴C enrichment of % absolute modern provided by the reference.

Dual carbon isotope analysis, especially radiocarbon (¹⁴C) analysis, is a powerful tool used to track the fates of organic carbon (Trumbore 2000; Raymond and Bauer 2001b; Trumbore 2009). The stable carbon isotope ratio (δ¹³C) provides information on whether the organic carbon has been derived from C₃ plants or C₄ plants, with about −28‰ and −12‰, respectively (Cerling et al. 1997; Kohn 2010; Marwick et al. 2015). Forests are mainly covered by trees which are mostly C₃ plants, and thus δ¹³C alone cannot clearly distinguish the sources of organic carbon. In contrast, the radiocarbon isotope ratio (Δ¹⁴C) has a wide range (−1000‰ to ∼200‰) and provides information on when the organic carbon was photosynthesized (Trumbore et al. 1992; Marwick et al. 2015). The Δ¹⁴C-SOC and Δ¹⁴C-DOC in soil water or extracted solution decreases as soil depth increases (Table 1 and the references therein; Shi et al. 2020). Thus, varying sources of soil DOC and stram DOC depending on hydrological pathway can be investigated using dual carbon isotope analysis (Barnes et al. 2018). However, most of the previous studies using dual carbon isotope analysis have been conducted only for soils or streams, and there are few studies linking the SOC, soil DOC, and stream DOC.

Despite the merits of the dual carbon isotope analysis, to our knowledge, only approximately 10 studies have reported both Δ¹⁴C-SOC and Δ¹⁴C-DOC in forest ecosystems (Table 1). These studies reported that as soil depth increased both Δ¹⁴C-SOC and Δ¹⁴C-DOC decreased, and that Δ¹⁴C-SOC and Δ¹⁴C-DOC varied significantly depending on soil type, tree species, temperature, and precipiation (Table 1 and the references therein). However, compared to the other factors, the effect of tree species on Δ¹⁴C-SOC and Δ¹⁴C-DOC have been rarely studied, although tree species can affect concentrations and biogeochemical properties of OC in soil, soil water, and streams.

The objectives of this study are (1) to compare and contrast the amount and properties of WEOC of soils collected underneath two common tree species in the southern part of South Korea, Cryptomeria japonica and Quercus acutissima, under three different extraction conditions and (2) to compare dual carbon isotope ratios among SOC, soil DOC, and stream DOC in small forested watersheds. Despite of the limited number of samples, to our knowledge, this is the first attempt in the Republic of Korea to investigate the relationship among Δ¹⁴C-SOC, Δ¹⁴C-WEOC from soils, and stream Δ¹⁴C-DOC within a forested watershed, enabling a deeper understanding on the connectivity of organic carbon among soil, soil water, and streams in a temperate forest ecosystem.

2. Materials and methods

2.1. Site description

The study sites are located in the Bukmoongol (BM) and Baramgol (BR) forested watersheds, which are part of the Seoul National University Forest in the Baekwoon Mountain, Gwangyang, Jeollanam-do, South Korea (35.0319° N, 127.6050° E) (Figure 1). The size of the BM watershed is 0.33 km² and the mean depths of the O and A horizons are 5.9 and 24.1 cm, respectively (Im et al. 2007). The watershed area of BR is 0.15 km² and the mean depths of the O and A horizons are 3.4 and 23.3 cm, respectively (Choi 2001). The soils of both watersheds are mainly composed of sandy loam, with relatively higher clay content in the BR watershed than BM watershed (Choi 2001, Korea Soil Information System: http://soil.rda.go.kr). The soils of both watersheds are categorized as moderately well-drained soils (Korea Soil Information System: http://soil.rda.go.kr). The ratios of stream discharge to precipitation were 50% and 36% for the BM and BR watersheds, respectively (Choi 2001). The bedrock consists of mostly granite and partially gneiss within the two watersheds (Park et al. 2000; Choi 2001). The slopes were 29.1% and 31.0% for the BM and BR watersheds, respectively (Choi 2001).

Figure 1. The location (red circle) and the boundary of the forested watersheds (red line) of this study. Soil samples were collected at orange triangle (Cryptomeria japonica stand) and circle (Quercus acutissima stand) points. The blue lines are streams. Stream samples were collected at green circles in the Bukmoongol (BM) and Baramgol (BR) watersheds.

The BM and BR watersheds are fully covered by forest, and there was no wetland. Both sites were reforested since the 1920s, and the BM watershed consists of 70% and 30% of coniferous and deciduous trees, respectively, whereas that of the BR watershed consists of 95% and 5% of coniferous and deciduous trees, respectively (Choi 2001; Im et al. 2007). Stands of Cryptomeria japonica (Japanese cedar) and Quercus acutissima (sawtooth oak) in the watershed were planted in 1959 and 1926, respectively. Cryptomeria japonica is a main afforestation tree species in the southern part of South Korea and Quercus acutissima is widely distributed throughout South Korea (Lee et al. 2007; Kim et al. 2020). They were selected as a conifer and a broadleaf tree species in this study to investigate interactions between soil organic carbon and DOC in soil solution.

The region is under a temperate monsoon climate, which is characterized by large rainfall during summer. From 2009 to 2018, the annual precipitation at the watersheds ranged from 874 to 2162 mm, with a mean of 1509 mm yr⁻¹, as measured at the weather station of the study sites. The mean precipitation during summer (from June to August) was 789 mm for the years from 2009 to 2018, accounting for approximately 50% of the annual precipitation. The mean daily temperature was around 23 °C in summer, while 4 °C in winter.

2.2. Laboratory experiments

2.2.1. Soil sampling

Soil samples were collected from mineral soil horizons beneath Cryptomeria japonica and Quercus acutissima stands at three different depths (0–10, 10–30, and 30–50 cm) from three pits for each species at the BM watershed on March 15, 2019 (Figure 1). The O horizon was not sampled because organic carbon content varies depending on the season, and we focused on the mineral horizon where the OC content remained relatively stable. Soil sampling plots between two tree species are about 400 m apart (Figure 1). There was no significant difference in mostly sandy soil texture beneath the two tree species. Collected soil samples were well mixed for each depth and sieved using a 5.6 mm screen to remove stones (Kalbitz et al. 2007). All samples were stored at 4 °C before extraction.

2.2.2. Soil solution extraction

The extraction methods were adapted from Jones and Willett (2006) and Kalbitz et al. (2007). Soil solution was extracted with deionized water (DI) to measure the concentration of DOC and major ions. In order to assess the effect of varying extraction conditions on the properties of WEOC, three different extraction conditions (EC) were applied as described below:

• EC1: A soil-to-water ratio of 1:10 (20 g of soil mixed with 200 ml of DI) was used. The mixture was stirred using a glass rod for 1 minute. The method presents relatively high water content and short extraction time conditions.

• EC2: A soil-to-water ratio of 1:10 (20 g of soil mixed with 200 ml of DI) was used. The mixture was stirred using a shaker for 18 hours. The method presents relatively high water content and long extraction time conditions.

• EC3: A soil-to-water ratio of 1:3 (50g of soil mixed with 150 ml of DI) was used. The mixture was stirred using a shaker for 18 hours. The method presents relatively low water content and long extraction time conditions.

The soil extractants were filtered using a Macherey-Nagel glass fiber filter (pore size: 0.4 μm) which was pre-combusted at 400 °C for 4 hours. Three replicates were prepared for each sample. A total of 54 samples were utilized to analyze the concentrations of WEOC and major ions. Extractants of EC1 and EC2 were stored in polyethylene bottles, while those from EC3 were stored in glass bottles for dual carbon isotope analysis. Every sample was stored at 4 °C.

2.2.3. Stream water sampling

Stream water was collected at the pool next to the U-shaped weir in the BM watershed from May 2012 to April 2015 and in the BR watershed from May 2012 to April 2014 on a weekly basis, in order to compare the [DOC] between the two watersheds. The water sample was filtered using pre-combusted Whatman glass fiber filter (pore size: 0.7 μm). About 300 ml of filtered sample was stored in polyethylene bottle to measure [DOC] and the concentration of major cations (Na⁺, K⁺, Mg²⁺, and Ca²⁺), and the remaining was frozen in a polycarbonate bottle for dual carbon isotope analysis. The same volume of weekly stream samples at the BM watershed were composited into one sample for each season (spring: March–May, summer: June–August, fall: September–November, and winter: December–February) during January 2014–April 2015 to increase the volume of the samples, which were analyzed for Δ¹⁴C-DOC.

Stream water samples were also collected at 2–24 hours intervals during two storms with high amount and intensity of precipitation to understand the effect of soil DOC on stream DOC during summer storms. The first storm was a typical summer storm in 03–08 July 2013 (123 mm of precipitation) which was non-separable, two consecutive rain events, and the second was an unusual storm event during a typhoon hit in 02–09 August 2014 (240 mm of precipitation). The stream water samples that were collected when discharge reached the peaks in each storm event were used for Δ¹⁴C-DOC analysis (i.e. a total of three samples including two samples for the first storm and one sample for the second storm). In sum, a total of 9 stream water samples were analyzed for Δ¹⁴C-DOC: 2 in spring, 1 in summer, 3 in summer storms, 1 in fall, and 2 in winter.

2.2.4. Chemical properties

Concentration of WEOC ([WEOC]) or stream DOC ([DOC]) was measured using Shimadzu TOC-V_CPH (Shimadzu Corporation, Japan). The concentration of major cations (Na⁺, K⁺, Mg²⁺, and Ca²⁺) was measured by ICS-1600 using Dionex CS12A column with 20 mM methanesulfonic acid eluent (Dionex, Sunnyvale, CA, USA). Soil organic carbon concentration ([SOC]) was determined using SSM-5000A (Shimadzu Corporation, Japan). The soils were air dried and ground using a Pulverisette 23 ball mill (Fritsch, Germany). Each soil sample was fumigated with concentrated HCl in a desiccator for 6 hours to remove inorganic carbon (Komada et al. 2008). The samples were then dried at 50 °C overnight and used to measure [SOC]. The following equation was used to calculate the ratio between WEOC and SOC (WEOC/SOC): (1) $$\text{WEOC}/\text{SOC }({\text{mg g}}^{-1})=\left[\text{WEOC}\right]/\text{extraction ratio}/\left[\text{SOC}\right]$$(1) where, the [WEOC] is concentration of WEOC in mg-C L⁻¹, the extraction ratio is the mass of soil divided by the volume of extractant, in g-soil L⁻¹, and [SOC] is in g-C g-soil⁻¹. Thus, the WEOC/SOC is the mass of WEOC which can be released from a unit mass of SOC.

2.2.5. Dual carbon isotope analysis

The dual carbon isotope analysis (δ¹³C and Δ¹⁴C) was conducted for the SOC, WEOC, and stream DOC. For SOC, each ground and acid-fumigated soil was sealed in a pre-baked quartz tube with a few strands of silver wire and CuO, and was oxidized at 850 °C in a furnace (Druffel et al. 1992).

The WEOC from EC3 for each soil depth was composited, which was the highest [WEOC] among ECs, and stream DOC for each season were composited. Only EC3 provided enough organic carbon to be analyzed for dual carbon isotope analysis. About 0.125 ml of 40% of H₃PO₄ solution was added to 100 ml of water sample and helium gas was purged to remove dissolved inorganic carbon. Then, the samples were oxidized with ultrapure O₂ gas by a high-energy (2400 W) UV lamp for 4 hours to convert DOC to CO₂ (Raymond and Bauer 2001a). The generated CO₂ gas was cryogenically purified using liquid N₂ in a vacuum line and was sealed in a pyrex tube.

The sealed quartz tubes and pyrex tubes were sent to the National Ocean Sciences Accelerator Mass Spectrometry (NOSAMS) facility at Woods Hole Oceanographic Institution in the United States for dual carbon isotope analysis (Lee et al. 2021). Carbonates (IAEA-603, IAEA-610, and USGS-44) and oxalic acid (IAEA-C8) were used as reference materials to calculate δ¹³C and Δ¹⁴C of the samples, respectively.

2.2.6. Statistical analysis

The concentrations of WEOC and major cations under different conditions (soil depth, ECs, and tree species) were compared using t-test, one-way analysis of variance, and Tukey’s test. The Mann-Whitney U test and Kruskal–Wallis test were used as non-parametric methods when the normality assumption was not satisfied. The statistical significance level was determined by p-value. A p-value less than 0.05 was considered as significant, 0.05 ≤ p-value < 0.10 as marginal, and 0.10 ≤ p-value as insignificant. The relationships between the stream [DOC] in the BM and BR watersheds and the Δ¹⁴C of SOC, WEOC, and stream DOC were analyzed using linear regression. We identified four outliers among the weekly collected samples in which stream [DOC] exceeded the threshold defined as “mean + 3 times the standard deviation (S.D.)” during the study period (Oh et al. 2013). These outliers were excluded from the linear regression analysis. Every statistical analysis was conducted using R 3.6.0. (R Core Team, 2019).

3. Results

3.1. Concentration of SOC and WEOC

The [SOC] decreased with increasing soil depth (p < 0.01; Table S1) and did not significantly vary between tree species at each soil depth (Table S1). The [SOC] were 2.6–4.4, 1.4–2.7, and 0.5–1.2% beneath Cryptomeria japonica, and 2.2–4.2, 1.4–2.7, and 0.5–1.0% beneath Quercus acutissima at 0–10, 10–30, and 30–50 cm, respectively.

A significantly higher [WEOC] was obtained from EC3 than EC1 (p < 0.01) (Figure 2). The mean [WEOC] was about twice higher under Cryptomeria japonica (3.0 and 7.9 mg L⁻¹ for EC2 and EC3, respectively) compared to Quercus acutissima (1.8 and 3.8 mg L⁻¹ for EC2 and EC3, respectively) (Table 2, Figure 2). For the [WEOC], significant differences among soil depth were observed under both tree species (p < 0.05; Table 2). The [WEOC] at 0–10 cm was significantly higher compared to 10–30 and 30–50 cm (p < 0.05).

Figure 2. The relationships between SOC concentration and [WEOC] by (a) EC1, (b) EC2, and (c) EC3. Black triangles and dashed line are for Cryptomeria japonica and red circles and dashed line are for Quercus acutissima. Dashed lines are linear regression lines.

Table 2. Arithmetic mean and SE (standard error of the mean, n = 3) of the [WEOC] and cations.

Tree species	EC	Soil depth (cm)	[WEOC] (mg L^−1)	WEOC/SOC^a (mg g^−1)	[Na^+] (µmolc L^−1)	[K^+] (µmolc L^−1)	[Mg^2+] (µmolc L^−1)	[Ca^2+] (µmolc L^−1)	[Sum of cations] (µmolc L^−1)	[WEOC]: [Sum of cations] (mmol mmolc^−1)
Cryptomeria japonica	1	0–10	1.1 ± 0.1	0.3 ± 0.1	33.6 ± 5.5	9.1 ± 1.5	10.0 ± 1.0	48.7 ± 5.2	101.3 ± 2.7	0.9 ± 0.1
		10–30	0.4 ± 0.0	0.2 ± 0.0	17.9 ± 4.9	7.0 ± 0.9	7.4 ± 0.9	38.2 ± 7.1	70.5 ± 7.2	0.5 ± 0.0
		30–50	0.3 ± 0.1	0.4 ± 0.2	19.2 ± 5.4	8.0 ± 0.5	9.6 ± 1.7	40.7 ± 8.3	77.5 ± 6.3	0.4 ± 0.1
	2	0–10	6.4 ± 0.9	1.9 ± 0.1	98.4 ± 15.9	29.0 ± 3.1	21.4 ± 2.1	114.8 ± 8.1	263.6 ± 26.8	2.0 ± 0.1
		10–30	1.5 ± 0.3	0.7 ± 0.0	38.4 ± 6.3	25.2 ± 4.6	17.8 ± 2.3	74.7 ± 16.1	156.1 ± 8.1	0.8 ± 0.2
		30–50	0.9 ± 0.4	1.4 ± 1.0	33.3 ± 11.9	15.5 ± 3.7	16.2 ± 2.8	73.8 ± 15.6	138.8 ± 15.0	0.5 ± 0.2
	3	0–10	15.7 ± 3.0	1.4 ± 0.3	157.0 ± 16.0	59.7 ± 2.3	52.9 ± 4.2	221.5 ± 27.8	491.1 ± 39.8	2.7 ± 0.7
		10–30	4.9 ± 1.8	0.7 ± 0.1	105.1 ± 18.5	65.6 ± 10.9	48.1 ± 7.5	123.7 ± 30.5	342.5 ± 8.0	1.2 ± 0.5
		30–50	2.9 ± 1.5	1.4 ± 1.1	104.3 ± 35.5	49.6 ± 16.9	43.8 ± 16.2	107.8 ± 34.2	305.6 ± 17.4	0.8 ± 0.4
Quercus acutissima	1	0–10	1.0 ± 0.2	0.3 ± 0.1	29.2 ± 5.8	16.7 ± 2.4	8.4 ± 3.3	12.1 ± 0.5	66.4 ± 5.5	1.3 ± 0.4
		10–30	0.4 ± 0.0	0.2 ± 0.1	22.0 ± 4.1	11.1 ± 0.3	10.7 ± 2.2	10.4 ± 1.1	54.2 ± 3.1	0.6 ± 0.0
		30–50	0.2 ± 0.0	0.3 ± 0.1	16.2 ± 0.5	7.5 ± 0.5	12.9 ± 1.5	11.6 ± 3.3	48.2 ± 3.8	0.4 ± 0.0
	2	0–10	4.1 ± 1.0	1.4 ± 0.2	53.9 ± 5.1	37.8 ± 2.4	18.5 ± 6.4	18.8 ± 3.3	129.0 ± 10.7	2.6 ± 0.5
		10–30	1.1 ± 0.4	0.5 ± 0.1	30.0 ± 5.4	25.3 ± 1.9	20.7 ± 2.6	13.9 ± 3.5	89.9 ± 8.1	1.0 ± 0.2
		30–50	0.3 ± 0.0	0.4 ± 0.1	25.2 ± 1.6	14.6 ± 2.1	22.2 ± 2.2	9.7 ± 2.7	71.7 ± 3.8	0.3 ± 0.0
	3	0–10	7.8 ± 2.8	0.7 ± 0.1	103.1 ± 7.8	85.8 ± 3.7	46.9 ± 6.6	52.7 ± 10.9	288.5 ± 21.1	2.2 ± 0.6
		10–30	2.0 ± 0.5	0.3 ± 0.0	69.2 ± 6.7	48.7 ± 6.5	43.6 ± 3.3	28.2 ± 7.6	189.8 ± 17.3	0.8 ± 0.1
		30–50	1.6 ± 0.7	0.6 ± 0.2	76.9 ± 8.6	25.8 ± 3.9	33.5 ± 3.3	17.3 ± 1.2	153.5 ± 14.2	0.8 ± 0.3

^a The WEOC/SOC ratio is calculated using Equation (1)(1) $$\text{WEOC}/\text{SOC }({\text{mg g}}^{-1})=\left[\text{WEOC}\right]/\text{extraction ratio}/\left[\text{SOC}\right]$$(1) .

The WEOC to SOC ratio (WEOC/SOC in Table 2) shows how much organic carbon in soils is dissolved into water on a unit mass basis. The WEOC/SOC ratio of soils collected beneath the two tree species from EC1 was lower than values obtained from EC2 and EC3 (p < 0.01 for Cryptomeria japonica, p = 0.02 for Quercus acutissima), probably due to insufficient water retention time to extract OC from soils using EC1 (Table 2). There was no significant difference in the WEOC/SOC ratio between EC2 and EC3. The mean WEOC/SOC ratio of soils at 0–50 cm, following EC1 was 0.3 mg⁻¹ g⁻¹ for both Cryptomeria japonica and Quercus acutissima, while 1.2 and 0.5 mg⁻¹ g⁻¹ were obtained for soils collected beneath Cryptomeria japonica and Quercus acutissima, respectively, following EC3. The WEOC/SOC ratio was not dependent on soil depth nor tree species in general (Table 2).

3.2. Concentrations of major cations

The concentration of the sum of cations ([sum of cations]) ranged between 59–107 µmol_c L⁻¹, 111–313 µmol_c L⁻¹, and 273–543 µmol_c L⁻¹, following EC1, EC2, and EC3, respectively, under Cryptomeria japonica and 43–77 µmol_c L⁻¹, 64–146 µmol_c L⁻¹, and 126–328 µmol_c L⁻¹, following EC1, EC2, and EC3, respectively, under Quercus acutissima (Table 2). A significantly larger amount of cations was extracted under Cryptomeria japonica as compared to Quercus acutissima (p < 0.01). The concentration of each cation increased with long water retention time and high soil to water ratio, such that a significantly higher concentration of cations was obtained for samples treated following EC3 than EC1 and EC2, regardless of the tree species (p < 0.01).

The [WEOC]/[sum of cations] ratios under Cryptomeria japonica at 0–50 cm were 0.2–1.0 mmol mmol_c⁻¹, 0.3–2.2 mmol mmol_c⁻¹, and 0.4–3.9 mmol mmol_c⁻¹, following EC1, EC2, and EC3, respectively (Table 2). For soils collected underneath Quercus acutissima, the [WEOC]/[sum of cations] ratio was not significantly different from Cryptomeria japonica. The [WEOC]/[sum of cations] decreased with increasing soil depth such that the ratio at 0–10 cm was about 2–4 times higher than at 10–30 and 30–50 cm, under both tree species (p < 0.01).

3.3. Dual carbon isotope ratios of WEOC and SOC

The δ¹³C-WEOC ranged from −27.1‰ to −23.3‰ and the δ¹³C-SOC ranged from −27.1‰ to −23.3‰ (Table S2). Both δ¹³C-WEOC and δ¹³C-SOC increased with increasing soil depth (Figure 3(a); p < 0.01). The δ¹³C-WEOC and δ¹³C-SOC of soil sample collected beneath Cryptomeria japonica tend to be higher than that of Quercus acutissima at all soil depths. However, the difference in δ¹³C-WEOC or δ¹³C-SOC between the two tree species was marginal (p = 0.08 for δ¹³C-SOC; Figure 3(a)).

Figure 3. The relationships between (a) δ¹³C-SOC and δ¹³C-WEOC and (b) Δ¹⁴C-SOC and Δ¹⁴C-WEOC. Black triangles are Cryptomeria japonica and red circles are Quercus acutissima. Dashed line is 1:1 line.

The Δ¹⁴C-WEOC was mostly modern under the two tree species, ranging from −53.3‰ to 70.5‰ (Table S2). Depleted Δ¹⁴C-WEOC was only observed at 30−50 cm soil under Quercus acutissima. The Δ¹⁴C-WEOC appeared to decrease with increasing soil depth under both tree species, but, the tendency was marginal for Cryptomeria japonica (p = 0.08) and not significant for Quercus acutissima (p = 0.26). There was no significant difference of Δ¹⁴C-WEOC between the two tree species.

The Δ¹⁴C-SOC of the soil sample collected beneath Cryptomeria japonica decreased with increasing soil depth (p = 0.05) whereas that beneath Quercus acutissima showed a similar pattern (p = 0.10; Figure 3(b)). The Δ¹⁴C-SOC was up to 78.6‰ at 0−10 cm of soil depth, whereas the Δ¹⁴C-SOC decreased to −87.0‰ under Cryptomeria japonica and −73.7‰ under Quercus acutissima at 30−50 cm of soil depth (Table S2). The ¹⁴C-ages of SOC ranged from 665 yBP (years before present) to modern. No significant difference of Δ¹⁴C-SOC was observed between the two tree species although this could be due to the low number of analysis.

3.4. Stream water chemistry

Weekly collected stream [DOC] ranged from 0.1 to 2.3 mg L⁻¹ (mean ± S.E.: 0.5 ± 0.0 mg L⁻¹) for BM watershed, and from 0.2 to 3.8 mg L⁻¹ (mean ± S.E.: 0.6 ± 0.0 mg L⁻¹) for the BR watershed during May 2012–April 2014. Stream [DOC] increased to 2.6 mg L⁻¹ in the BM watershed and 4.1 mg L⁻¹ for the BR watershed during storm events in 2012. The [DOC] of the BR watershed was significantly higher than that of the BM watershed over the study period, including the storm events (Figure 4; p < 0.01). The difference of [DOC] between the two watersheds was higher in storm samples than weekly collected samples (Figure 4; slopes: 1.27 (storms) and 1.13 (weekly)). However, there were no significant differences in [sum of cations] between the two watersheds. Stream Δ¹⁴C-DOC in the BM watershed ranged from −68.1 to −18.3‰ for composited samples for each season during 2014–2015, which corresponds to 85–505 yBP with a mean of 290 yBP. A relatively high Δ¹⁴C-DOC, ranging from −6.2 to 45.5‰, was observed during summer storms (Figure 5).

Figure 4. The relationships between stream [DOC] of the BM watershed ([DOC]_BM) and stream [DOC] of the BR watershed ([DOC]_BR). Black circles and red squares are weekly collected stream DOC and stream DOC during storm events, respectively. The four white circles are outliers of weekly collected samples (> mean + 3 S.D.) which were excluded in the linear regression. Black and red dashed lines are regression lines of weekly collected samples and storm samples, respectively. Gray solid line is 1:1 line.

Figure 5. The relationships between the ratio of [DOC] to [sum of cations] and Δ¹⁴C-DOC of extracted soil solution (i.e. WEOC) under the two tree species and stream water. Black triangles are Cryptomeria japonica, and red circles are Quercus acutissima. Light to dark green symbols are stream water samples collected weekly and composited for each season, and during summer storms. Blue dashed line is the linear regression line for stream and extracted soil solution.

4. Discussion

4.1. [WEOC] of soils collected underneath Cryptomeria japonica and Quercus acutissima

Many studies have reported larger [DOC] in soil solution under coniferous forests than under deciduous forests, at least for top soils (Currie et al. 1996; Borken et al. 2011; Fröberg et al. 2011; Camino-Serrano et al. 2014; Thieme et al. 2019). For example, soil [DOC] under coniferous trees was 23% higher than those under broadleaved trees at 0–80 cm soil depth in 365 sites which are mainly located in Europe (Camino-Serrano et al. 2014). However, the observed difference in the [DOC] between conifers and deciduous stands cannot be solely due to the tree species, but also to variations in precipitation, temperature, and SOC concentration. In this study the Cryptomeria japonica and Quercus acutissima stands are approximately only 400 m apart (Figure 1), thus the precipitation and temperature of the two sites are almost identical.

Soil organic carbon concentration can be dependent on tree species (Cha et al. 2019). Although a study including 595 forest sites in South Korea reported that at a mineral soil depth of 0–30 cm the [SOC] under deciduous trees was ∼30% higher than that under conifers (Cha et al. 2019), we have not observed a significant difference in [SOC] between the two tree species at all soil depths in this study (Table S1). Even though high [SOC] can increase [DOC] in soil solution, our results suggest that the quality, rather than the quantity, of SOC can be a controlling factor on the extracted [DOC].

Despite of similar [SOC], the [WEOC] of soils under Cryptomeria japonica was larger by approximately 50% and 100%, compared to Quercus acutissima under EC2 and EC3, respectively, and the difference was mainly observed for 0–10 cm depth soils (Figure 2 and Table 2). The WEOC/SOC of Cryptomeria japonica was about two times larger than that of Quercus acutissima under EC3 (Table 2). These suggest that the OC of soils under Cryptomeria japonica can be easily dissolved than soils under Quercus acutissima. Although the soils can contain OC formed even before the two species were planted and we do not know the distribution of tree species in the watersheds before 1920, at least SOC in surface soils is likely to be influenced by present tree species rather than tree species of the past, considering that SOC can accumulate quickly in young growing stands (Cha et al. 2019). Thus, if stream DOC is mainly derived from DOC in soil water with other conditions (e.g. distance between tree stands to the stream) being equal, soils under Cryptomeria japonica stands would have a higher potential to increase stream [DOC] than soils under Quercus acutissima.

The mean of weekly collected stream [DOC] in the BR watershed was 30% higher than that of the BM watershed from May 2012 to April 2014 (p = 0.01, Figure 4) and the difference in stream [DOC] between the two watersheds was amplified during the storm events (p = 0.02, Figure 4). Given that the proportion of conifers in the BR watershed is 95% compared to 70% in the BM watershed, this suggests that watersheds dominated by conifers might release higher [DOC] than watersheds with a higher proportion of broadleaf trees. Soil [DOC] in coniferous trees tended to be higher than those of deciduous trees (Camino-Serrano et al. 2014), and thus stream [DOC] within watersheds dominated by conifers could be higher than those by deciduous trees.

Forest stream [DOC] can be influenced by many factors such as wetland cover, slopes of the watersheds, hydrologic flow path, and proximity of tree species to streams. The BM and BR watersheds are entirely covered by forest and there is no wetland within the watersheds. The slopes are similar in the BM and BR watersheds. Although we did not investigate all the tree species within the watersheds and the proximity of them to the streams, the effects of watershed hydrology could be inferred from the results of WEOC and stream DOC. Results of the extraction experiments demonstrated that [WEOC] from Cryptomeria japonica was up to about twice of that from Quercus acutissima, for the shallowest 0–10 cm depth soils (Table 2). Stream [DOC] during storms are preferentially from surface flow (Hood et al. 2006; Jeong et al. 2012), and thus, the higher stream [DOC] in the BR watershed than the BM watershed during summer storms could be due to higher [WEOC] of the conifer-dominant BR than the BM (Figure 4). The ratios of stream discharge to precipitation in the BM and BR were 50% and 36%, respectively. A longer residence time of water in the BR, compared to the BM, could contribute to higher stream [DOC] in the BR watershed. The observed difference of stream [DOC] between two watersheds may be attributed not only to variations in tree species but also to other factors, such as discharge and soil texture throughout the watershed (Choi 2001).

4.2. Dual carbon isotope ratios of WEOC and SOC

There were significant negative relationships between δ¹³C and Δ¹⁴C of both SOC and WEOC (p < 0.05; Figure S1). As soil depth increased, the δ¹³C-SOC and δ¹³C-WEOC increased while the Δ¹⁴C-SOC and Δ¹⁴C-WEOC decreased (Table S2). This could be explained by (1) microbial degradation, (2) contribution of roots to SOC, and/or (3) selective sorption and leaching of OC with increasing soil depth (Kaiser et al. 2001; Nakanishi et al. 2012; Ahrens et al. 2015; Paul et al. 2020).

The Increase of δ¹³C-SOC and δ¹³C-WEOC with soil depth (Table S2) could be explained if soil microbes preferentially use lighter carbon (¹²C) in SOC, resulting in depleted δ¹³C in respired CO₂ and enriched δ¹³C in SOC (Wynn et al. 2006; Diochon and Kellman 2009). The δ¹³C of roots of C3 plants was ∼1.2‰ higher than δ¹³C of shoots on average in studies using a variety of C3 plants (Werth and Kuzyakov 2010), and the δ¹³C of roots of four tree species, including fir, pine, beech, and oak, was higher than δ¹³C of surface litters by 4.6‰ (Lorenz et al. 2020). This suggests that deep roots could also contribute to high δ¹³C-SOC as soil depth increases because of post-photosynthesis fractionation (Paul et al. 2020).

Preferential dissolution and sorption of OC in soils can also be a reason of δ¹³C-enrichment of SOC and WEOC with soil depth. The δ¹³C of the hydrophilic portion of soil DOC (HPIC), including carbohydrates and amino acids, can be higher than that of the hydrophobic portion of soil DOC (HPOC), including lignin, lipid, and cutin, by approximately 3–4‰ (Guggenberger et al. 1994; Kaiser et al. 2001; Nakanishi et al. 2012). The proportion of HPIC to WEOC can increase as soil depth increases in forest soils (Kaiser et al. 2001; Nakanishi et al. 2012). Thus, the higher δ¹³C-WEOC and δ¹³C-SOC in deeper soils could be explained by a higher proportion of HPIC which is leached and then deposited to the deeper soil horizons, in contrast to the preferentially adsorbed HPOC by soil minerals.

The decline of Δ¹⁴C-SOC and Δ¹⁴C-WEOC with increasing depth can be explained by the repeated decomposition of SOC as microbes turn OC into more preserved status after young OC enters from litters or roots to soils (Rumpel et al. 2002; Rumpel et al. 2004; Nakanishi et al. 2012; Ahrens et al. 2015). Soil minerals in subsurface horizons can bind SOC, preventing microbial access to SOC, which can result in low Δ¹⁴C-SOC and Δ¹⁴C-WEOC (Schmidt et al. 2011; Lehmann and Kleber 2015). Similar results have been reported regardless of the tree species, indicating that soil depth is a controlling factor on the Δ¹⁴C of SOC and soil DOC (Table 1 and Figure 6; Shi et al. 2020).

Figure 6. The relationships between Δ¹⁴C-SOC and Δ¹⁴C of soil DOC or soil WEOC. Data from nine previous studies are also included (Table 1). Blue triangles, purple circles, green diamonds, and gray stars are for conifers, deciduous trees, mixed forests, and other plants such as shrubs and grasses. Black triangles are for Cryptomeria japonica whereas red circles are for Quercus acutissima in this study. Gray and black dashed lines are linear regression line and 1:1 line, respectively.

δ¹³C-WEOC can be higher than δ¹³C-SOC because of preferential dissolution of lighter carbon in SOC (Nakanishi et al. 2012). However, the δ¹³C-WEOC was similar with δ¹³C-SOC in this study (Table S2), suggesting that dominant source of both SOC and WEOC are C₃ plants. In contrast, the Δ¹⁴C-WEOC was higher than Δ¹⁴C-SOC in general (Figure 3(b)), which agrees with the other studies (Figure 6), demonstrating that the relatively young SOC was preferentially extracted by water. Or this could be because hydrophilic young OC in subsurface horizons, which had been transferred from the surface by water and deposited, are re-dissolved during extraction (Kaiser and Kalbitz 2012). The difference between Δ¹⁴C-WEOC and Δ¹⁴C-SOC was larger for soils collected under Cryptomeria japonica compared to Quercus acutissima (Figure 3 and Table S2), suggesting that more young OC could be released as DOC from soils under Cryptomeria japonica.

Although the effects of tree species on either Δ¹⁴C-SOC or Δ¹⁴C-DOC were not significant in this study (black triangles and red circles in Figure 6), when the results of other studies were combined (Figure 6, Table 1, and the references therein), both the Δ¹⁴C-SOC and Δ¹⁴C-DOC of soils under deciduous trees were slightly higher compared to soils beneath conifers (Figure 6; p < 0.07 for SOC and p < 0.05 for DOC). This suggests that tree species has a weak potential to influence the Δ¹⁴C-SOC and Δ¹⁴C-DOC whereas soil depth can be a strong factor on the Δ¹⁴C-SOC and Δ¹⁴C-DOC. However, site history, soils properties, climates, or hydrological characteristics can be different each other (Figure 6, Table 1; Camino-Serrano et al. 2014; Nguyen and Marschner 2016; van der Voort et al. 2019; Shi et al. 2020), and thus caution is needed to draw a conclusion with a limited number of samples.

4.3. Linkage between soil DOC and stream DOC

The stream Δ¹⁴C-DOC in this study ranging from −68.1 to 45.5‰ (Figure 5), was significantly lower than that of other forest streams, where contributions of the OC inputs from surface litters or shallow soils were relatively larger (Schiff et al. 1997; Longworth et al. 2007; Lu et al. 2014). A positive log-linear relationship between [DOC]/[sum of cations] and Δ¹⁴C-WEOC was observed (Figure 5). A similar relationship was also observed between [DOC]/[Na⁺] and Δ¹⁴C-WEOC, where [Na⁺] was used as a proxy of weathering products (Figure S2). This agrees with the results reported by Barnes et al. (2018) which suggest that aged DOC could be released from deep soils where the concentrations of weathering products (i.e. major cations) were higher than those of surface soils.

Despite Δ¹⁴C of WEOC was mostly positive (Table S2), the stream Δ¹⁴C-DOC of the BM watershed (Figure 5) corresponded to ¹⁴C-age dating up to 505 ± 15 yBP during baseflow, suggesting that contribution from mineral soil horizons or possibly groundwater to the stream [DOC] could be relatively large. The ¹⁴C-age of SOC at a soil depth of 30–50 cm was more than 500 yBP (Table S2), which also suggested leaching of old SOC during baseflow with long water residence time. In contrast, positive stream Δ¹⁴C-DOC was observed during summer storms (Figure 5) which indicated that contribution of fresh OC from litters or surface soils to stream DOC could increase with a rise in precipitation (Yang et al. 2015; Denis et al. 2017; Hernes et al. 2017).

In temperate forests, the inferred mean Δ¹⁴C-SOC from global radiocarbon profile data was −9‰ at a soil depth of 0–30 cm and −229‰ at soil depth of 30–100 cm (Shi et al. 2020). Groundwater is reported to have low Δ¹⁴C-DOC, down to −400‰ (Trumbore et al. 1992), because DOC can be leached from aged OM in deep soils or because ¹⁴C-enriched OC can be preferentially decomposed or sorbed while water percolates down to deep soils. This suggests that more ¹⁴C-depleted DOC can be released from temperate forests when baseflow is the dominant hydrological pathway, although many studies have reported positive Δ¹⁴C-DOC from temperate forest streams (Longworth et al. 2007; Lu et al. 2014).

5. Conclusions

We investigated the concentrations and properties of SOC and WEOC from soils collected under two different tree species stands, Cryptomeria japonica and Quercus acutissima, and compared these with the properties of stream DOC. There was no difference between [SOC] in soils collected under the two tree species, but [WEOC] tends to be higher under Cryptomeria japonica compared to Quercus acutissima, suggesting that more stream DOC could be released from the forested watershed with Cryptomeria japonica if other conditions are equal.

Neither Δ¹⁴C-WEOC nor Δ¹⁴C-SOC were significantly different under the two tree species in this study. However, the insignificant difference in Δ¹⁴C-WEOC or Δ¹⁴C-SOC in the two tree species can be due to a limited number of radiocarbon analysis. When the results of other studies were combined, the Δ¹⁴C-DOC and Δ¹⁴C-SOC under deciduous trees were slightly higher than under conifers, suggesting that tree species may not be a strong factor controlling the bulk ¹⁴C-ages of soil DOC and SOC. Instead, soil depth was a controlling factor on the Δ¹⁴C-SOC and Δ¹⁴C-WEOC whose corresponding ¹⁴C ages were up to 665 and 375 yBP, respectively, at 30–50 cm depth, in contrast to modern carbon was observed in both SOC and WEOC at 0–10 cm depth. The ¹⁴C-age of the forest stream was up to 505 yBP, highlighting the contribution of old OC from deep soils or possibly groundwater to streams during baseflow, although many studies have reported release of mostly modern DOC from temperate forest streams.

Acknowledgements

We thank the staff at Seoul National University Forest for weekly stream water sampling. We also thank Minjung Go for help with sampling, laboratory analyses, and discussion. This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2016R1D1A1B03930149), and by NRF grant funded by the Korea government (MSIT) (No. 2021R1A2C1006783 and NRF-2021R1A4A1025553).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Research Foundation of Korea [NRF-2016R1D1A1B03930149; 2021R1A2C1006783; 2021R1A4A1025553].