Fluxes of dissolved organic carbon and nitrogen throughout Andisol, Spodosol and Inceptisol profiles under forest in Japan

Abstract

Leaching of dissolved organic matter (DOM) is an important process in the translocation and stabilization of organic carbon (C) and in influencing nitrogen (N) availability in forest soils. The roles of DOM in soil carbon and nitrogen cycles were evaluated by quantifying the fluxes of dissolved organic carbon (DOC) and nitrogen (DON) entering and leaving the organic (O), A and B horizons. In Spodosol and Inceptisol soils, DOC fluxes were highest in the O horizon (149 to 344 kg C ha⁻¹yr⁻¹), decreasing in the A and B horizons. In Andisol soils, DOC fluxes were low throughout the profile because of low DOC production in the O horizon (53 kg C ha⁻¹yr⁻¹) and the high adsorption capacity of amorphous aluminum (Al) and iron (Fe) (hydr)oxides in the mineral horizons. In Spodosol soil, DOC from the O horizon represented a large proportion of C input into the mineral soil, whereas this contribution appeared to be small in Andisol soil. The DOM was enriched in nitrogen during decomposition and humification of the foliar litter, but DON was a small proportion (5–31%) of total dissolved nitrogen (TDN) in surface soil solutions. The narrow DOC/DON and DON/TDN ratios were attributable to the low C/N ratios of the foliar litter (33–40). It was quantitatively shown that the importance of DOM in C and N cycles in forest soils varied depending on soil types and litter C/N ratio.

Keywords:

• adsorption
• dissolved organic carbon
• dissolved organic nitrogen
• soil organic matter
• volcanic soil.

Introduction

Most of the organic matter supplied to the organic (O) horizon of forest soils is mineralized, but a proportion is leached into the mineral horizons as dissolved organic matter (DOM) (McDowell and Likens 1988, Zech and Guggenberger 1996). The DOM transported into the mineral horizons may be mineralized, leached or adsorbed onto mineral surfaces (Kalbitz et al. 2000). Leaching of DOM, including dissolved organic carbon (DOC) and nitrogen (DON), is an important process in the translocation and stabilization of soil organic matter (SOM) (McDowell and Likens 1988) and the transport or loss of plant available nitrogen (N) (Qualls et al. 1991) and may influence soil formation (esp., podzolization and acidification) (Guggenberger and Kaiser 1998, Fujii et al. 2009a, b).

Easily biodegradable, low molecular weight organic substances such as organic acids, sugars and amino acids constitute a small fraction of DOM (Van Hees et al. 2005, Fujii et al. 2010), while the remaining fraction is recalcitrant, high molecular weight humic substances (Qualls and Bridgham 2005). This recalcitrant DOM is produced mainly through the solubilization of macromolecules (e.g., lignin and cellulose) in litter and is translocated down through the soil profile (Currie and Aber 1997). The differences in concentrations and fluxes of DOC between soil profiles under boreal and temperate forests demonstrate that the roles of DOM in the carbon (C) and N cycles vary depending on climate, vegetation types (broad-leaved vs. coniferous forests), and soil types (Zech and Guggenberger 1996, Michalzik et al. 2001, Fujii et al. 2009b). Fluxes of DOC and DON from the O horizons have been shown to vary from 6 to 30% and 1 to 53% of litterfall-C and -N, respectively (Michalzik et al. 2001). In temperate coniferous forests, 70 to 95% of total dissolved nitrogen (TDN) in solution was DON (Currie et al. 1996), whereas inorganic N was reported to be a dominant form of N in the broad-leaved forests of temperate and tropical regions (Solinger et al. 2001, Schwendenmann and Veldkamp 2005). In mineral soil horizons, DOM leached from the O horizon may be partly stabilized by adsorption onto mineral surfaces (Sollins et al. 1996), accounting for 38–89% of SOM formation (Michalzik et al. 2003, Kleja et al. 2008). This DOM adsorption is increased by a clayey texture or an abundance of amorphous aluminum (Al)/iron (Fe) (hydr)oxide sorbents (Kaiser and Zech 2000). Japanese soils are strongly influenced by volcanic activities; therefore, the roles of DOM in soil C and N cycles may vary with the amorphous Al/Fe (hydr)oxides derived from volcanic materials.

To date, only a few studies have investigated the importance of DOM fluxes in forest ecosystems in Japan (Kawasaki et al. 2005, Shibata et al. 2005). The objectives of our study are to quantify the fluxes of DOC and DON in throughfall and soil solutions, and to evaluate the role of DOM in the C and N cycles of Andisol, Spodosol and Inceptisol forest soils in Japan.

Materials and Methods

Experimental sites

Site and soil descriptions are presented in Tables 1 and 2. Three sites were selected for ando soil under cold temperate forest, podzolic soil under cold temperate forest and brown forest soil under warm temperate forest. These represent 91.7% of the total forest area in Japan (11.5% for ando soils, 3.6% for podzolic soils and 76.5% for brown forest soils) (Forest Soil Division 1976).

Table 1. Site description

Site	NG (Nagano)	TG (Tango)	KT (Kyoto)
Coordinates	N35°57′, E138°28′	N35°37′, E1135°10′	N35°1′, E135°47′
Elevation (m)	1440	675	90
Mean annual air temperature (°C)	6.9‡	10.7	15.9‡
Mean annual precipitation (mm)	1422‡	1782‡	1490‡
Vegetation	Quercus crispula	Fagus crenata	Quercus serrata Castanopsis cuspidate
Soil type†	Acrudoxic Melanudands	Andic Haplohumods	Typic Dystrudepts
Parent material	Volcanic ash	Sedimentary rocks, granite	Sedimentary rocks
†Soils were classified according to Soil Taxonomy (Soil Survey Staff 2006).
‡Average of 1984–2004 taken from nearest meteological stations, i.e. Nobeyama, Miyazu and Kyoto.

Table 2. Chemical properties of soils

Site	Horizon	Depth (cm)	pH (H2O)	TC	TN	C/N	Sio†	Feo†	Alo†	Sip‡	Fep‡	Alp‡
				(g kg^−1)			(g kg^−1)			(g kg^−1)
NG	O	+2–0	6.3	448	18.4	24	–	–	–	–	–	–
Andisol	A1	0–5	4.6	212	14.2	15	3.5	16.3	24.7	1.7	12.3	17.7
	A2	5–15	4.6	167	10.0	17	6.1	14.4	25.1	1.1	8.6	15.2
	A3	15–35	4.8	134	8.1	16	6.8	20.2	32.8	0.9	9.0	16.3
	BA	35–50	4.9	98	6.0	16	13.2	20.1	36.7	0.4	4.5	8.0
	Bw	50–76+	5.0	23	1.9	12	16.5	18.7	36.3	0.8	1.2	3.9
TG	Oi	+10–+8	5.4	501	14.7	34	–	–	–	–	–	–
Spodosol	Oe	+8–+3	4.3	468	18.4	25	–	–	–	–	–	–
	Oa	+3–0	4.0	432	20.8	21	–	–	–	–	–	–
	AE	0–9	3.8	125	6.8	17	0.1	11.6	4.9	1.2	14.7	4.8
	Bhs	9–35	4.4	66	3.4	17	2.2	14.9	17.5	0.8	10.9	11.9
	Bs	35–72	4.5	18	1.1	17	2.2	7.0	11.7	0.8	4.7	6.9
	BC	72–86+	4.6	5	0.3	15	0.3	2.8	3.1	0.4	2.7	2.8
KT	O	+2–0	5.2	483	18.7	26	–	–	–	–	–	–
Inceptisol	A	0–7	4.2	45	2.4	19	0.1	2.6	4.4	0.4	2.0	4.1
	Bw1	7–15	4.2	13	0.8	16	0.1	1.9	2.1	0.1	1.5	2.0
	Bw2	15–45	4.2	8	0.6	12	0.2	1.6	2.2	0.2	1.4	1.5
	Bw3	45–65+	4.2	5	0.5	11	0.2	1.7	1.8	0.4	1.4	1.5
Oven-dried basis.
†The oxalate-extractable Si, Fe and Al compounds were determined by extraction in the dark with acid ( P H 3) 0.2 M ammonium oxalate (McKeague & Day 1966).
‡The oxalate-extractable Si, Fe and Al compounds were determined by extraction with 0.1 M pyrophosphate at pH 10 (Schuppli et al. 1983). C, carbon; N, nitrogen; H2O, water; TC, total carbon; TN, total nitrogen; NG, Nagano; TG, Tango; KT, Kyoto.

The Nagano site (NG) is located on the eastern slope of Mt. Yatsugakate (Yatsugatake Experimental Forest of Tsukuba University, Nagano Prefecture), and has a mean annual air temperature of 6.9°C and annual precipitation of 1422 mm yr⁻¹. The dominant overstory vegetation is Quercus crispula with a herbaceous understory. Andisols are derived from secondary sediments of volcanic products. The C stocks in the mineral soil horizons are higher at the NG site than the other two sites, because of the higher amorphous Al and Fe (hydr)oxides (Imaya et al. 2010). The distribution pattern of Alo and Alp with depth suggests that an organo-mineral complex is dominant in surface soil horizons (Table 2). In subsoil, the occurrence of allophane is suggested by the ratio of (Alo−Alp)/(Sio−Sip) (Table 2), which is similar to the ideal Al/Si ratio of 2 for allophane (Ugolini et al. 1988).

The Tango site (TG) is located on the Tango Peninsula of Kyoto Prefecture, and has a mean annual air temperature of 10.7°C and annual precipitation of 1782 mm yr⁻¹. The soil surface is covered with 1–2 m of snow from mid December to late March. The dominant overstory vegetation is Fagus crenata, while the understory vegetation is Sasa kurilensis. Spodosols are derived from sedimentary rocks and granite. The site is covered with a 10-cm morder-type O horizon. The A horizon is strongly acidic (pH 3.8) but pH increases with depth (Table 2). The distribution pattern of Alo and Alp with depth suggests that there is some eluviation of Al and OM from the AE horizon and accumulation in the Bhs and Bs horizons (Table 2), as would occur with podzolization (De Coninck 1980).

The Kyoto site (KT) is located on Mt. Yoshida, Kyoto Prefecture, and has a mean annual air temperature of 15.9°C and annual precipitation of 1490 mm yr⁻¹. The vegetation is dominated by Quercus serrata and Castanopsis cuspidate, with little understory vegetation. The Inceptisols are affected by brunification and weak podzolization (Fujii et al. 2008). The higher contents of Alo and Feo in surface soil horizons suggest that Al and Fe oxides were immobilized in an organo-mineral complex (Table 2), as would occur during brunification (Ugolini et al. 1990).

Plant materials

Five litterfall replicates were collected in 60-cm diameter circular litter traps. Fine root biomass in the O horizon was measured by collecting the roots from five replicate 30-cm quadrats. Fine root biomass in the mineral soil was measured by collecting roots at 5-cm depth intervals in five replicate cores of 0.1-L volume and rinsing in distilled water to remove soil materials. Plant samples were oven-dried at 70°C for 48 h, weighed and milled. The C and N contents were determined using a CN analyzer (Vario Max CN, Elementar Analysensystem GmbH). Klason lignin concentrations of foliar litter were determined by digestion with sulfuric acid (Allen et al. 1974).

Monitoring temperature and volumetric water content in soils

Soil temperature at 5 cm was measured with a thermistor probe (107 Temperature Probe, Campbell Scientific, Inc.) in two replicates. Volumetric water content of soils at 5, 15 and 45 cm was measured with time domain reflectometer probes (CS615 Water Content Reflectometer, Campbell Scientific, Inc.) in three replicates. Data were recorded at 30-min intervals using dataloggers (Campbell Scientific, Inc., CR-10X).

Concentrations of dissolved organic carbon and nitrogen in throughfall and soil solution

Five soil solution replicates were collected by tension-free lysimeters beneath the O, A and B (or A2 in NG) horizons (0, 5 and 45 cm depths, respectively). The lysimeters were installed horizontally in small soil pits by inserting 200 cm² plates into precut openings and with collecting bottles connected by tubes. Throughfall was collected using a precipitation collector which consisted of a 20-cm diameter funnel attached to the collection bottle. A fiberglass mesh covered the top of the five replicate funnels and a plug of glass wool was placed in each funnel neck to exclude particulate matter from the collection bottles. Sample collection was carried out 15 times from June 2003–November 2004 at NG, 17 times from June 2003–December 2004 at TG and 19 times from June 2003–February 2005 at KT. Sample solutions were filtered through a 0.45 -µm filter, and stored at 1°C in the dark prior to analyses. Solution pH was determined with a glass electrode. The concentrations of DOC and TDN in solution were determined using a total organic carbon analyzer (TOC-V_CSH, Shimadzu, Japan). The concentrations of dissolved inorganic nitrogen (DIN) ( and ), in solution were determined by high performance liquid chromatography (HPLC; Ion chromatograph HIC-6A, Shimadzu; shim-pack IC-C3 for , shim-pack IC-A1 for , conductivity detector CDD-6A). The concentrations of DON were calculated by subtracting DIN from TDN concentrations (DON = TDN −  − ).

Fluxes of dissolved organic carbon and nitrogen

Fluxes of DOC, DON and DIN from each horizon were calculated each month by multiplying the water fluxes by the concentrations of DOC, DON and DIN in the throughfall and soil solutions. Throughfall was measured using precipitation collectors, and the half-hourly fluxes of soil water percolating at depths of 5, 15 and 30 cm were estimated by applying Darcy's law to the unsaturated hydraulic conductivity and the gradient of the hydraulic heads at each depth. The unsaturated hydraulic conductivity and soil water pressure heads at depths of 5, 15 and 45 cm were estimated using the saturated hydraulic conductivity, and water retention curves of soil and volumetric water content monitored at each depth at 30-min intervals (Mualem and Dagan 1978, Van Genuchten 1980). Details of the calculations of soil water fluxes are presented in our recent study (Fujii et al. 2008). The soil water fluxes for each month were calculated by summing the half-hourly water fluxes.

Calculations and statistics

The fluxes of DOC and DON are calculated and expressed per area on an annual basis. All results are the means ± standard error (SE) of five determinations. The statistical differences between mean values between groups (horizons or sites) were tested using analysis of variance (ANOVA) at a P < 0.05 significance level for the mean concentrations and fluxes of C and N in soil solution. The linear regression analyses were performed to assess the relations between DOC concentration and soil temperature and between DOC/DON or DON/TDN ratio in soil solution and C/N ratio of the foliar litter. The statistical analyses were performed with SYSTAT 11.0 (SPSS Inc., 2008).

Results

Aboveground biomass, fine root biomass and litterfall

Aboveground biomass and total fine root biomass were lowest at NG and highest at KT. Fine root biomass was concentrated in the O horizon of TG, while it was also distributed in the mineral soil horizons of NG and KT. The C/N ratios of the foliar litter of TG and KT were higher than NG and the lignin concentration was higher at TG than at KT and NG.

Soil solution composition

At TG, the O horizon solution was strongly acidic (pH 3.9), but pH increased with depth (Table 4). Solution pH at NG was relatively high (6.0–6.1) throughout the soil profile, but was moderately low at KT. The concentrations of DOC and DON varied widely between sites, soil depths, and seasons (Table 4, Fig. 1). At TG and KT, the concentrations of DOC and DON were highest in the O horizons, and decreased with depth (Table 4, Fig. 1). At NG, the DOC concentrations were consistently low throughout the profile (Fig. 1). The concentrations of DOC and DON varied seasonally at TG (Fig. 1), and the DOC concentrations in the O horizons increased with soil temperature, being highest during summer (Fig. 2). At NG and KT, the seasonal variation of DOC and DON was relatively small (Fig. 1) and the annual mean DOC and DON concentrations in the O horizons were lower than at TG (Table 3). The DOC concentrations at TG are comparable to previous reports from temperate forests [20–90 mg C L⁻¹; 41.6 ± 19.5 mg C L⁻¹ (mean ± SD)] and DOC concentrations at NG were the lowest of our study and low compared with previous data (Michalzik et al. 2001).

Figure 1. Seasonal fluctuation of the concentrations of dissolved organic carbon (DOC), dissolved organic nitrogen (DON) and dissolved inorganic nitrogen (DIN) in throughfall and soil solution. Bars indicate standard errors (n = 5). NG, Nagano; TG, Tango; KT, Kyoto; N, nitrogen; O, organic.

Figure 2. Relationship between soil temperature and dissolved organic carbon (DOC) concentrations in the organic (O) horizon solutions in Tango. The soil temperature used in analysis was the mean soil temperature monitored during sampling period. Bars indicate standard errors (n = 5).

Table 3. Aboveground biomass, fine root biomass and litterfall

	Above ground biomass†	Root biomass (Mg C ha^−1)		Litterfall
Site				C	N	C/N	Lignin
	(Mg C ha^−1)	O horizon	Mineral soil	(Mg C ha^−1yr^−1)	(kg N ha^−1 yr^−1)		(%)
NG	78	0.4	1.6	1.6	49.8	32.8	32.2
TG	83	2.1	0.8	2.1	51.0	40.4	38.2
KT	115	1.4	2.8	3.2	81.6	39.1	33.8
†The data were cited from Shinjo et al. (2006). C, carbon; Mg, magnesium; N, nitrogen; O, organic; NG, Nagano; TG, Tango; KT, Kyoto.

Table 4. Annual volume-weighted mean concentrations and fluxes of dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) in throughfall and soil solution

Site	Horizon†	pH	Concentration‡					DOC/ DON	DON/ TDN	Flux‡
			DOC		DON		DIN			Water	DOC		DON		DIN
			(mg C L^−1)		(mg N L^−1)				(%)	(mm yr^−1)	(kg C ha^−1yr^−1)		(kg C ha^−1yr^−1)
NG	TF	6.0 ± 0.0	5.3 ± 0.3	aA	0.23 ± 0.01	aB	0.59 ± 0.09	23	28	1155 ± 17	61.3 ± 2.9	aB	2.6 ± 0.1	aC	6.8 ± 1. 0
	O	6.0 ± 0.1	6.8 ± 1.7	aC	0.30 ± 0.08	aC	1.90 ± 0.27	22	14	779 ± 87	52.6 ± 12.1	aC	2.4 ± 0.6	aC	14.8 ± 1.3
	A1	6.1 ± 0.1	1.5 ± 0.5	bB	0.07 ± 0.03	bB	1.38 ± 0.10	21	5	718 ± 43	10.5 ± 3.7	bB	0.5 ± 0.2	bC	9.9 ± 0.4
	A2	6.0 ± 0.1	1.4 ± 0.1	bC	0.09 ± 0.02	bC	1.38 ± 0.14	15	6	616 ± 54	8.6 ± 0.4	bC	0.6 ± 0.1	bC	8.5 ± 0.4
TG	TF	5.4 ± 0.0	4.7 ± 0.2	cA	0.23 ± 0.01	cB	0.53 ± 0.07	20	30	1676 ± 33	78.5 ± 2.4	cA	3.9 ± 0.1	cA	8.9 ± 1.1
	O	3.9 ± 0.1	28.9 ± 3.5	aA	1.16 ± 0.14	aA	6.48 ± 0.84	25	15	1189 ± 116	343.6 ± 24.9	aA	13.8 ± 1.0	aA	77.0 ± 6.6
	A	4.1 ± 0.1	13.0 ± 1.2	bA	0.73 ± 0.07	bA	4.23 ± 0.51	18	15	945 ± 62	122.7 ± 8.6	bA	6.9 ± 0.5	bA	40.0 ± 4.0
	B	4.6 ± 0.0	2.3 ± 0.2	cB	0.16 ± 0.02	cB	1.89 ± 0.26	15	8	873 ± 78	20.5 ± 1.4	dB	1.4 ± 0.1	dA	16.5 ± 1.7
KT	TF	5.2 ± 0.0	5.6 ± 0.4	bA	0.34 ± 0.03	bA	1.07 ± 0.14	16	24	1657 ± 33	92.7 ± 7.1	bA	5.6 ± 0.5	aB	17.8 ± 2.3
	O	4.3 ± 0.1	15.4 ± 1.6	aB	0.66 ± 0.07	aB	1.87 ± 0.41	23	26	973 ± 76	149.4 ± 10.3	aB	6.4 ± 0.4	aB	18.2 ± 3.7
	A	4.4 ± 0.1	14.0 ± 1.6	aA	0.65 ± 0.08	aA	1.81 ± 0.28	22	26	805 ± 82	112.8 ± 6.3	aA	5.2 ± 0.3	aB	14.6 ± 1.7
	B	4.5 ± 0.1	4.3 ± 0.8	bA	0.34 ± 0.06	bA	0.17 ± 0.04	12	67	825 ± 86	28.2 ± 4.0	cA	2.3 ± 0.3	bB	1.1 ± 0.2
†TF represents throughfall and O, A and B represent present soil horizons.
‡All results are the mean ± standard error of five replicates. The different lowercase letters (a, b, c, d) mean that mean values are significantly different (P < 0.05) between horizons at each site, whereas different uppercase letters (A, B, C) indicate that mean values of each horizon are significantly different between sites. C, carbon; N, nitrogen; DIN, dissolved inorganic nitrogen; TDN, total dissolved nitrogen; O, organic; NG, Nagano; TG, Tango; KT, Kyoto.

Table 5. Comparison of dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) fluxes in the organic (O) horizons between our study and the published data

Location	Vegetation type†	Soil type	Annual precipitation	Litterfall C/N	Flux (O horizon)			DOC/ DON	DON/ TDN	Reference
					DOC	DON	DIN
			(mm yr^−1)		(kg C ha^−1)	(kg C ha^−1)			(%)
NG	B	Andisols	1422	33	52.6	2.4	14.8	22.2	14	The present study
TG	B	Spodosols	1782	40	343.6	13.8	77.0	24.8	15
KT	B	Inceptisols	1490	39	149.4	6.4	18.2	23.2	26
USA	C	Andisols	2370	67	482.0	8.9	n.d.	54.0	n.d.	Yano et al. (2004)
USA	C	Spodosols	1400	57	216.5	3.9	2.6	55.5	60	Dittman et al. (2007)
USA	C	Inceptisols	1100	71	398.0	9.5	2.0	41.9	83	Currie et al. (1996)
Norway	C	Spodosols	1400	n.d.	363.0	14.0	n.d.	25.9	n.d.	Mulder et al. (2000)
Finland	C	Spodosols	503	80	125.0	2.3	0.4	54.0	85	Piirainen et al. (2002)
Germany	C	Spodosols	1100	33	114.7	4.7	12.3	24.4	28	Michalzik & Matzner (1999)
USA	B	Spodosols	1400	42	157.9	4.7	6.8	33.6	41	Dittman etal. (2007)
USA	B	Inceptisols	1100	58	225.0	6.1	3.0	36.9	67	Currie et al. (1996)
USA	B	Inceptisols	1770	61	405.0	10.1	1.2	40.1	89	Qualls et al. (1991)
Germany	B	Inceptisols	752	42	353.1	8.9	45.9	39.9	16	Solinger et al. (2001)
cf.Tropical forest
Taiwan	C	Entisols	4815	69	962.0	15.7	3.6	61.3	81	Schmidt et al. (2010)
Costa Rica	B	Oxisols	4200	25	277.0	13.0	200.0	21.3	6	Schwendenmann & Veldcamp (2005)
†C and B indicate coniferous and broad-leaved forests, respectively. C, carbon; N, nitrogen; DIN, dissolved inorganic nitrogen; TDN, total dissolved nitrogen.

The DON concentrations in the O horizon solutions are comparable to other reported values (0.40– 2.45 mg N L⁻¹ from Michalzik et al. 2001). The concentrations of DON in soil solution fluctuated similarly to DOC at each site (Fig. 1). The DOC/DON ratio decreased from 22.2–24.8 in the O horizons to 12.4–15.4 in the B horizons (Table 4). The DOC/DON ratios in the O horizons in our study are comparable to the lower reported values from published data (Table 5).

The concentrations of both DON and DIN in soil solution decreased with depth. The contribution of DON to TDN (DON, and ) decreased from the O horizons to the mineral soil horizons, except in KT. The DON/TDN ratios of the O horizon solutions in our study are in the lower range of published data (Table 5).

Fluxes of DOC and DON in the soil profiles

The fluxes of DOC and DON increased markedly in the O horizon of TG (Table 4; Fig. 3), indicating that the O horizon was the main source of DOC and DON. The higher fluxes of DOC and DON in the throughfall of KT compared with TG, and the slight increase in DOC and DON fluxes in the O horizon, indicated that both the canopy and the O horizon were the main sources of DOC and DON at this site. In contrast, no net increase in DOC and DON fluxes was observed in the O horizon of NG (Fig. 3). The DOC and DON fluxes from the O horizons at TG are comparable to the highest values recorded in temperate forests [100–482 kg C ha⁻¹yr⁻¹ and 0.2–18.0 kg N ha⁻¹yr⁻¹ (Michalzik et al. 2001)].

Figure 3. The stock and the annual fluxes of carbon (C) (a) and nitrogen (N) (b) via litterfall, organic matter (OM) decomposition, throughfall and soil solution. ^†The data of OM decomposition rates were cited from Shinjo et al. (2006). ^‡The stocks of C and N in soil at the depths of 0 to 45 cm were counted. DOC, dissolved organic carbon; O, organic; NG, Nagano; TG, Tango; KT, Kyoto.

The TDN fluxes were largest in the O horizons at all sites (Fig. 3). The O horizon was the main source of DON and DIN at TG, whereas the canopy was the main source of DON at NG and KT. Compared with the C and N fluxes via litterfall (Fig. 3), the DOC fluxes from the O horizon corresponded to 3.2, 16.7 and 4.7% of litterfall-C at NG, TG and KT, respectively. The corresponding values for N were higher at each site, being 4.8, 27.6 and 8.1% (Fig. 3). The annual net inputs of DOC into the upper 45 cm of mineral soil (DOC fluxes from O horizon minus those from B horizon) were calculated to be 44, 323 and 121 kg C ha⁻¹yr⁻¹ at NG, TG and KT, respectively (Fig. 3, Table 4). These DOC inputs corresponded to 0.01, 0.20 and 0.13% of the C stocks in the mineral soil (Fig. 3). The values at TG and KT are lower than but close to the 0.30% boreal forest soil (McDowell and Likens 1988), whereas DOC input at NG is small compared with the C stock in the mineral soil (Fig. 3).

Discussion

Seasonal fluctuation in concentrations and fluxes of DOC in soil solution

In temperate forests, DOC concentrations in the O horizon solution fluctuate seasonally depending on soil moisture (Guggenberger and Zech 1993), temperature (Guggenberger and Zech 1994, Fujii et al. 2008) and litterfall pattern (Qualls et al. 1991). In our study, the seasonal variation of DOC concentrations in the O horizon solution in TG (Fig. 1) indicates the dependence of DOC concentrations on soil temperature, with higher temperatures increasing DOC production (Fig. 2). Production of DOC (litter solubilization) depends on microbial activities, which are enhanced by temperature (Gödde et al. 1996, Shinjo et al. 2006). On the other hand, mineralization of DOC is unaffected by temperature because of the limited ability of microorganisms to mineralize DOC (Gödde et al. 1996). These could account for the higher concentrations of DOC in the O horizon solutions of TG in summer. It should be noted that temperature might also affect the rate of DOM diffusion into soil solution (Stutter et al. 2007). At NG and KT, however, DOC concentrations did appear to be unaffected by temperature (Fig. 1). Production of recalcitrant DOC during litter decomposition is lower throughout the year at NG and KT, likely because of the lower lignin concentrations in the foliar litter.

Controls of concentration and fluxes of DOC and DON in soil solution

The concentrations and fluxes of DOC are largest in the O horizons (Table 4), consistent with previous reports (Michalzik et al. 2001, Fujii et al. 2009b). The annual fluxes of DOC and DON from the O horizons depend on annual precipitation, vegetation types (lignin concentration, C/N ratio), chemical properties of soils, the amounts of litterfall, and the type of O horizon (Gödde et al. 1996, Currie and Aber 1997, Michalzik et al. 2001, Fujii et al. 2009b). In our study, the high DOC fluxes from the O horizon of TG may be due to the presence of humified Oe and Oa horizons. The humus accumulated in Oe and Oa horizons produces and leaches a greater amount of recalcitrant DOC, compared with fresh litter (Park et al. 2002, Fröberg et al. 2007). The DOC leached from the O horizons is generally recalcitrant because a large fraction of DOC is composed of water-soluble lignin fragments that are covalently bonded to carbohydrates (Guggenberger and Zech 1994, Guggenberger et al. 1994). The higher lignin concentration in foliar litter in TG, compared with NG and KT, might contribute to the accumulation of humified materials and the consequent release of recalcitrant DOC in the O horizons (Tables 3 and 4). In NG, the lowest concentrations and fluxes of DOC are attributable to low production of recalcitrant DOC in the O horizon. The low production of recalcitrant DOC in the O horizon is caused by lower lignin concentrations in the foliar litter (Table 3) and the smaller stocks of the humified materials. As observed in NG (3.6 Mg C ha⁻¹; Fig. 3), the C stocks of the O horizons are smaller in Andisols than the other soil types in Japan (Kaneko 2011). The low production of recalcitrant DOC from the thin O horizon and microbial mineralization of DOC are considered to result in the smaller DOC fluxes from the O horizon of Andisol in NG (Table 4; Fig. 3).

At three sites, a decrease in the concentrations and fluxes of DOC with depth is caused by adsorption, rather than mineralization, as reported by Qualls and Haines (1992). Our recent study of adsorption experiment showed that adsorption capacities of organic acids increased with the amorphous Al and Fe (hydr)oxides in the order of NG (Andisol) >TG (Spodosol) >KT(Inceptisol) (Table 2) (Fujii et al. 2010). The lowest DOC concentrations in the A2 horizon of NG are considered to be caused by the high adsorption of DOC onto amorphous Al and Fe (hydr)oxides.

Among the three sites in Japan, the DOC/DON ratios in soil solution are similar, irrespective of differences in climate, vegetation, and soil types (Table 4). When compared to the global data from published sources (Table 5), the DOC/DON ratios in the O horizon solutions are variable between 15 sites, and the values of the three sites are close to the minimum values of previous reports (Table 5). Typically, the broad-leaved forests had lower C/N ratios in foliar litter and DOM than coniferous forests (Dittman et al. 2007). The regression analysis of the global dataset including the published data of 12 sites and those from our study shows that the DOC/DON ratio depends on C/N ratio of foliar litter, rather than on vegetation type (Fig. 4). The lower DOC/DON ratios of the O horizon solutions in our study are attributable to the lower C/N ratio of foliar litter (Fig. 4, Table 5). The DOM in the O horizon solutions had a lower C/N ratio (22–25) than the fallen litter (33–40) (Tables 3 and 4). This suggests that the organic matter being translocated downwards is enriched in N, as also reported by Schoenau and Bettany (1987) and Qualls et al. (1991). The incorporation of the extra N in DOM occurs on all three sites and is characteristic of decomposition and humification (Qualls and Haines 1992).

Figure 4. Relationship between carbon/nitrogen (C/N) of the foliar litter and the ratio of dissolved organic carbon (DOC) to dissolved organic nitrogen (DON) in the organic (O) horizon solution.

Roles of DOC and DON in soil C and N cycles

Leaching of DOC from the O horizon of forest soils and root litter input are important carbon sources into the mineral soil. Annual net inputs of DOC make up 11–75% of total C input (sum of net DOC input from the O horizon and root litter input) in boreal and temperate forests (Zech and Guggenberger 1996, Kleja et al. 2008). In our study, assuming a similar rate of root litter input (20% of fine root biomass) to that reported by Nakane (1980), this will result in a carbon input of 321, 167 and 558 kg C ha⁻¹yr⁻¹ in NG, TG and KT, respectively, meaning the corresponding annual net inputs of DOC to the mineral soil constitute 12.9, 65.8 and 17.9% of the total C inputs (Tables 3 and 5). Our data indicate that DOC from the O horizon is the dominant source of C into the mineral soil at TG, whereas root litter is the dominant source at NG and KT. The quantitative importance of DOM to SOM accumulation varied from soil to soil. The large contribution of DOM to SOM in Spodosol of TG is consistent with the results in Spodosols under boreal coniferous forests (Michalzik et al. 2003). As suggested by the accumulation of organo-mineral complex in the Bhs horizon of TG (Table 2), leaching of DOC from the O horizon and its adsorption onto amorphous Al and/or Fe (hydr)oxides might be major SOM-accumulating process in Spodosols.

On the other hand, the low fluxes of DOC into the mineral soil in NG suggest that root litter, rather than DOC flux from the O horizon, plays an important role in SOM accumulation of Andisols (Fig. 3). The importance of root litter to SOM formation was a common process in grassland ecosystems (Don and Schulze 2008, Sanderman and Amundson 2008). Although the large stock of SOM in Andisols has been accumulated under grassland (esp. Miscanthus sinensis) in the past through input of root litter or charred materials (Ugolini and Dahlgren 2002), the input of root litter into the mineral soil is still a dominant ongoing process in Andisol of NG. However, the large fluxes of DOC input into the mineral soil (Table 5; 482 kg C ha⁻¹yr⁻¹) play an important role in SOM formation of Andisol under coniferous forests in Oregon, USA (Yano et al. 2004). The wide variation within vegetation or soil types should be considered before generalization of SOM-accumulating process in Japanese forest soils.

In O horizon solutions, DON is generally the dominant proportion of TDN in temperate coniferous forests (Currie et al. 1996, Yu et al. 2002), whereas inorganic N is the dominant form of N in broad-leaved forests (Solinger et al. 2001, Schwendenmann and Veldkamp 2005). In the three sites in Japan, DON constitutes a minor proportion (14–26%) of TDN in the O horizon solution (Table 4; Fig. 3). The proportions of DON relative to TDN in soil solution are small due to: (1) the predominance of mineralization over plant uptake in theO horizon of TG and (2) high adsorption of DOM in the mineral soil horizons of NG and TG (Fujii et al. 2010). The higher proportion of DON relative to TDN at KT may partly be due to the lower adsorption capacities of DOM (Fujii et al. 2010). The regression analysis of the global dataset including the published data of 10 sites and those from our study shows that the DON/TDN ratio of the O horizon solution depends on C/N ratio of foliar litter, rather than vegetation types (Fig. 5). The DON/TDN ratios in the O horizon solutions are variable between 13 sites and the values of our three sites are close to the lower values of previous reports (Table 5). In forests which have a high C/N ratio in foliar litter, the substantial C substrates enhance N immobilization by microorganisms and limit N mineralization (Jones et al. 2004). Because amino acids generally account for a small proportion (< 10%) of DON (Yu et al. 2002), plant available N is limited by the slow decomposition rates of high molecular weight DON to amino acids (Jones et al. 2004). In our three sites, DIN is available to plants in broad-leaved forests (Fig. 3), which have a low C/N ratio in foliar litter or receive N deposition as in the case of KT. It was quantitatively shown that the role of DOM in C and N cycling varies depending on soil types or litter C/N ratio.

Figure 5. Relationship between carbon/nitrogen (C/N) of the foliar litter and the ratio of dissolved organic carbon (DOC) to dissolved organic nitrogen (DON) in the organic (O) horizon solution.

Conclusions

In Spodosol and Inceptisol soils, DOC fluxes were highest in the O horizon, and decreased in the A and B horizons. In Andisol soil, the DOC fluxes were low throughout the profile because of low DOC production in the O horizon and the high adsorption capacity of amorphous Al and Fe (hydr)oxides in the mineral horizons. In Spodosol soil, DOC leached from the O horizon represented a large proportion of C input into the mineral soil, whereas it had only a small contribution in Andisol soil. DOM was enriched in nitrogen during decomposition and humification of the foliar litter; however, DON is a small proportion (5–31%) of TDN in surface soil solutions. The narrow ratios of DOC/DON and DON/TDN are attributable to the low C/N ratios of the foliar litter (33–40). It was quantitatively shown that the importance of DOM in C and N cycles varies depending on soil types or litter C/N ratio.

Acknowledgments

The authors would like to thank Yoshida Shrine, the Kinki and Chugoku Regional Forest Office of the Ministry of Agriculture, Forestry and Fisheries of Japan, and the Yatsugatake University Forest of Tsukuba University for allowing them to conduct studies at the experimental sites. We are also deeply indebted to the editors and two anonymous reviewers for improving our manuscript.