The mineralization rate of black soil carbon in the deep layers of Japanese volcanic ash soil may be easily accelerated by labile carbon supply

ABSTRACT

The stability of black soil carbon in the deep layers of Japanese volcanic ash soil (i.e., buried A horizons) is often explained by its unique chemical (molecular structure) and physical (associated with short-range-order minerals) recalcitrance. However, the stability of black soil C in buried A horizons may be changed by labile C supply for soil microbes. Here, we hypothesized that the mineralization of black soil C in buried A horizons of Japanese volcanic ash soil could be easily accelerated by a supply of labile C (i.e., a priming effect; PE). To test our hypothesis, we investigated the direction and magnitude of the PE with a buried A horizon in Japan using ¹³C-labeled glucose (2.188 atom %) in a short-term (21 days) incubation study. We also investigated the effect of mineral nitrogen (N), which could contribute to microbial activity in this incubation study. We found that a positive PE occurred by glucose supply with (182%) or without (181%) mineral N input over the 21-day incubation, and its values were very similar to the PE ratios previously reported in other deep soils. The estimated mean residence time (MRT) of black soil C considering PE was clearly accelerated by glucose supply, regardless of mineral N input, compared with the initial soil MRT. These results strongly support our hypothesis that the mineralization rate of black soil C in buried A horizons is easily accelerated by a labile C supply, and it also demonstrates important implications for the effects of global warming on buried A horizons (e.g., increased root exudation, fine root biomass supply, and N deposition) in Japanese volcanic ash soils.

KEY WORDS:

• Deep black soil
• priming effect
• microbial biomass
• δ¹³C
• mean residence time

1. Introduction

It is well known that black soil carbon (C), particularly black humic acids in the deep layers (below 20–100 cm) of Japanese volcanic ash soils (i.e., buried A horizons), shows its unique chemical properties which are characterized by their high condensed aromatic C content with a large C-layer plane size and a high degree of darkness (i.e., A₆₀₀/C), such as black carbon (Shindo and Honma 2001; Watanabe and Takeda 2006; Watanabe and Fujitake 2008; Ikeya et al. 2011). These black soil C associated with short-range-order minerals (particularly allophane/imogolite) and/or monomeric Al and Fe ions (i.e., active Al and Fe) form unique aggregate hierarchy, which is physically very stable (Wada and Higashi 1976; Inoue 1990; Asano and Wagai 2014). These chemical and physical recalcitrant properties of black soil C result in the unavailability of C to soil microbes, which is often attributed to its strong stability in situ (Wada and Higashi 1976; Inoue 1990; Shindo and Honma 2001; Watanabe and Takeda 2006).

A new theory has recently been proposed that the stability of soil C is attributable to a lack of labile C for soil microbes. This phenomenon is known as the priming effect (PE), referring to a change in the mineralization of native recalcitrant soil C due to the addition of new labile C substrates, and has been observed in many laboratory and field studies (e.g., Kuzyakov, Friedel, and Stahr 2000; Fontaine et al. 2004, 2007, 2011; Chen et al. 2014). For example, Fontaine et al. (2007) found that cellulose input primed microbial activity, leading to faster decomposition of older recalcitrant soil C in the deep layers of grassland soil (60–80 cm). Wang et al. (2014) also reported similar results in forest soil by litter supply. These results are remarkable as new findings of soil C stabilization theory, and this theory probably applies to the stability of black soil C in buried A horizons of Japanese volcanic ash soil with rare black soil C characteristics.

Additionally, it is generally considered that land use change (e.g., reforestation) and/or global warming increases labile C input (e.g., root exudation and fine root biomass) to the soil, especially deep layers of soil, through increased photosynthesis and C allocation below ground (Jobbágy and Jackson, 2000; Lal, 2004; Wang et al. 2014; Karhu et al. 2016). Such a labile C input in the deep layers of soil through deepening of the root profile can break aggregates and decrease bulk density, increasing soil oxygen diffusion rates (Mueller, Jensen, and Megonigal 2015; Bernal et al. 2016). There is also a concern that an increase in N deposition (e.g., NH₄⁺ and NO₃⁻) to terrestrial ecosystems will accompany application of mineral N fertilizer and/or global warming (Galloway et al. 2004). Therefore, many researchers are focusing on the PE phenomenon and the effect of N deposition on the PE, especially in deep layers of soils.

We hypothesized that mineralization rate of black soil C in the deep layers of Japanese volcanic ash soil could be easily accelerated by a labile C supply with more aerobic condition despite its strong chemical recalcitrance and physical resistance. To test our hypothesis, we conducted a short-term (21 days) laboratory incubation study to investigate the direction and magnitude of PE in deep black soil C (below 40–50 cm) in Japanese volcanic ash soil with and without a supply of glucose (the most often released sugar in rhizodeposits; Derrien, Marol, and Balesdent 2004). We also investigated the relationships between PE and mineral nitrogen (N) application.

2. Materials and Methods

2.1. Soil sample

A soil sample was taken from the campus of the Takayama Experimental Forest (Takayama site), River Basin Research Center, Gifu University, Japan (36° 08′ N, 137° 25′ E, 1425 m above sea level). A detailed description of this site is provided elsewhere (Ohtsuka et al. 2005; Chen et al. 2017; Iimura et al. 2019). The soil at this study site was characterized as Andosols (Iimura et al. 2019). We collected our sample from the buried A horizon (40–50 cm). After collection, the fresh soil samples were passed through a 2-mm mesh sieve, which was followed by removal of almost all visible plant debris and roots using tweezers. Then, the fresh soil samples were immediately stored in a refrigerator at 5°C for approximately 10 days until use in the incubation experiment (see Section 2.3). Subsamples for soil characterization were air-dried and subjected to the same process of sieving and removal of roots as described above.

2.2. Soil analysis

Soil pH was measured using a solution of the air-dried soil (H₂O or 1 M KCl) at a ratio of 1:5 after shaking for 1 h. To measure the pH of NaF, 1 g of air-dried soil was mixed with 50 ml of 1 M NaF (pH 7.0), the suspension was stirred continuously for 2 min, and the pH of the suspension was measured. The C and N contents of the soil were measured with an NC analyzer (Sumigraph NC-22 F, Sumika Chemical Analysis Service, Japan). Microbial biomass C was determined by the ATP method (Aoyama 2005). The melanic index of soil humus (MI) was measured using the methods of Yamamoto, Honna, and Utsumi (2000). The δ¹³C value of the soil sample was measured by an elemental analysis/stable isotopic ratio mass spectrometer (EA/IRMS) continuous flow system (Thermo Scientific, USA). The δ¹³C measurements were repeated until the standard deviation reached <0.2‰. The ¹⁴C content of soil was determined by accelerator mass spectrometry (AMS) at the National Institute for Environmental Studies. Soil samples were first converted to CO₂ and purified cryogenically. The purified CO₂ was then reduced to graphite with H₂ over Fe and its ¹⁴C/¹²C ratio was measured. The ¹⁴C activities are determined with respect to the international standard of oxalic acid. Final ¹⁴C activities are reported in Δ¹⁴C. The Δ¹⁴C is defined as the deviation in parts per thousand from the modern standard (Stuiver and Polach 1977). Ages were calculated from the Δ¹⁴C values using the conventional ¹⁴C half-life.

2.3. Incubation experiment

The experimental units consisted of fresh sieved soil (50 g of oven-dried samples) placed in 500-ml jars and incubated at 25°C for 21 days until the CO₂ emission from the soil slowed down. After 10 days of preincubation, glucose was added to the soil (ca. 1 g C-glucose per kg dry soil was prepared according to Fontaine et al. 2007) and mixed. Mineral N ((NH₄)₂SO₄) was also applied to the soil at N rate of 50 mg NH₄⁺-N kg⁻¹, which was based on the predicted amount of mineral N (NH₄⁺, NO₃⁻) deposition to terrestrial ecosystems in the future (Galloway et al. 2004). Soil without glucose and mineral N was prepared as a control and was mixed to provide the same physical disturbance. The incubation experiment included four treatments (three replicates): soil only (S), soil with mineral N (SN), soil with glucose (SC), and soil with glucose plus mineral N (SCN). The moisture content of the soil was adjusted to 60% of the water-holding capacity with distilled water during incubation. The released CO₂ was trapped in a NaOH solution at 2, 4, 7, 14, and 21 days of incubation until CO₂ emission from the soil slowed down. A glass vial containing 20 ml of 0.2 M NaOH solution was placed in each jar to trap the CO₂ emitted from the soil, and the jars were sealed. The jars were flushed with CO₂-free air each time the NaOH solution was replaced. The C content of the NaOH solution was measured by titration (Wang et al. 2014). δ¹³C of CO₂ was analyzed by an EA/IRMS continuous flow system after precipitation of the carbonates with excess BaCl₂ and filtration (Fontaine et al. 2011) every 7 days of incubation. The δ¹³ C measurements were repeated until the standard deviation reached <0.2‰.

2.4. Calculation of priming effect (PE)

The substrates (glucose) added to the soil allowed the separation of the respiration rates (mg CO₂-C kg⁻¹ soil) of native soil C (R_soil) and of substrates (R_sub) using mass balance equations (Fontaine et al. 2011) as follows:

R_soil + R_sub=R_total

R_soil × δ¹³C _soil + R_sub × δ¹³ C _sub=R_total × δ¹³C _total

where δ¹³C _soil is the δ¹³C of soil C, δ¹³C _sub is the δ¹³C of glucose C, R_total is the total CO₂ emitted by soil with glucose, and δ¹³C _total is its δ¹³C.

The PE (mg C-CO₂ kg⁻¹ soil) induced by the addition of substrates was calculated as:

PE=(R_soil soil with substrates) – (R_soil control soil)

where (R_soil control soil) is the CO₂ emitted by the control soil.

2.5. Statistical analysis

For total CO₂ emission, soil-derived CO₂ emission, and PE among treatments in each incubation period, a one-way analysis of variance (ANOVA) was performed with StatPlus:mac Pro (version 5.9.92).

3. Results and discussion

In this study, we focused on the stability of black soil C in buried A horizon in a Japanese volcanic ash soil that is usually considered very stable in situ (Watanabe and Takeda 2006; Fujitake 2011), and this is the first study to quantify the effect of labile C addition on black soil C mineralization in the deep layers of Japanese volcanic ash soil. The soil is characterized as Andosols (FAO et al., 2014). It is weakly acidic, with a pH (NaF) of 9.5 or more, an MI of <1.7, and a high C content (Table 1). These results are consistent with the findings of a previous study (Chen et al. 2017; Iimura et al. 2019), which classified the soil at the Takayama site as an andic Japanese volcanic ash soil (Kuroboku). Additionally, the ¹⁴C age of this soil was calculated as 4764 to 5082 years before present (YrBP) (Table 1). From these results, we confirmed that the soil used in this study satisfied our study aims.

Table 1. Characteristics of the soil used in this study.

	pH
Depth (cm)	(H2O)	(KCl)	(NaF)	T-C (g kg^−1)	C:N Ratio	δ^13C (‰)	MI^a	^14C age^b (YrBP)
40–50	5.1	4.3	11.5	89.4	16.0	–19.8	1.64	4764–5082

^aThe melanic index of soil humus.

^bThe ¹⁴C age of soil samples from 40–45 cm and 45–50 cm, respectively.

The cumulative total CO₂ emission from each treatment in the 21-day incubation experiment is shown in Figure 1. Analysis by one-way ANOVA revealed that the total CO₂ emission in each incubation period was significantly different among treatments. Glucose application (SC and SCN), regardless of mineral N input, stimulated total CO₂ emission within a week and then gradually increased with incubation. In the soil with mineral N (SN), there was a gradual decrease in CO₂ emission compared with the control soil (S) during the 21-day incubation period; this result is in accordance with a meta-analysis where N deposition impedes organic C decomposition in temperate forest soils (Janssens et al. 2010). Chen et al. (2014), in studies of surface sandy loam soil, reported that total CO₂ emission significantly increased with sucrose (sugar) application, regardless of mineral N (NH₄⁺) input, but that the CO₂ emission from soil with mineral N showed a slightly lower trend than that of the control soil during 9 days of incubation, which was similar to our results. The reduction of CO₂ emission from soil with N deposition is often explained by N availability for microbes (Moorhead and Sinsabaugh 2006; Janssens et al. 2010) i.e., an increase in N availability with N deposition reduces microbial activity in mining for soil organic matter (SOM) and our result in SN strongly supported this theory.

Figure 1. Cumulative total CO₂ emissions from each treatment. S, soil only (control); SN, soil with mineral N; SC, soil with glucose; SCN, soil with glucose plus mineral N. Bars indicate the standard error of the mean (n = 3). Asterisks (***) indicate significant differences at P < 0.001 by a one-way analysis of variance (ANOVA).

We calculated the PE among each treatment of either glucose and/or mineral N during the incubation period (Figure 2a–c). One-way ANOVA revealed that the PE in each incubation period was significantly different among treatments. The PE for SC and SCN increased at a constant rate every week with the addition of glucose. The addition of glucose to soil, regardless of mineral N application, showed a significant positive effect on PE during the 21-day incubation period, resulting in an PE of +18.1 (183%) to +29.4 (181%) mg C kg⁻¹ for SC and +16.6 (176%) to 29.8 (182%) mg C kg⁻¹ for SCN (Figure 2a–c). On the other hand, the PE for SN was −11.5 (47%) to −13.5 (63%) mg C kg⁻¹ (Figure 2a–c). There was no significant difference in the amount of glucose degradation for SC (81.4–96.3 mg C kg⁻¹) and SCN (77.5–93.7 mg C kg⁻¹). Fontaine et al. (2007) reported that cellulose addition to the deep soil layer in French (Cambisols) soil caused a positive PE (ca. 172%) for 161 days of an incubation study. This corresponds well to our results, despite the different soil types, incubation periods, and substrates added. It also corresponds well to the results of Wang et al. (2014), who reported changes in CO₂ emission from Chinese forest deep soil (60–70 cm) with leaf litter (M. macclurei) application by 167–192% higher than that of the control for 129 days of an incubation study. Karhu et al. (2016) also reported that glucose addition to deep soil (32–47 cm) in a boreal forest caused a positive PE (ca. 130%) for 11 days of an incubation period. These results suggest that the reaction of black soil to the addition of glucose is not a rare case phenomenon, and it further suggests that glucose addition causes a positive PE on deep soils in a shorter period of time than higher molecule compounds, such as polysaccharide and plant litter.

Figure 2. Cumulative soil-derived CO₂ (gray), glucose-derived CO₂ (blank), and PEs (data above the bars) in each treatment. (a) days 7, (b) days 14, and (c) days 21. S, soil only (control); SN, soil with mineral N; SC, soil with glucose; SCN, soil with glucose plus mineral N. The number in parentheses (%) indicates the relative magnitude of PE. Bars indicate the standard error of the mean (n = 3). Values with different letters are significantly different (P < 0.05) by Tukey’s HSD post hoc test. Capital letters indicate glucose-derived CO₂, and small letters indicate soil-derived CO₂ and the PEs.

The PE can be separated into real effects (SOM decomposition) and apparent effects (changes in microbial biomass turnover excluding SOM decomposition) (Blagodatskaya and Kuzyakov 2008). Unfortunately, the experimental design does not distinguish between real and apparent PE according to differences in ¹³C signatures. The microbial turnover time usually ranged between 21 and 75 days (Blagodatskaya et al. 2009) and was approximately 30 days as determined by the natural abundance approach (Blagodatskaya et al. 2011). Considering the shorter monitoring period of the PE in this study (21 days) compared with microbial turnover, it is likely that only a small fraction of primed C comes from the microbial biomass turnover (i.e., apparent PE). In other words, most of the primed CO₂ is considered to be real PE.

PE mechanism is often discussed based on ‘microbial N, P mining’ theory that some microbes (especially slow-growing K-strategists) use labile C as energy source to mineralize recalcitrant SOM to acquire nutrient, such as N and phosphorus (P) (Fontaine et al. 2004; Chen et al. 2014; Wang et al. 2014). Based on this theory, microbial activity, particularly slow-growing K-strategist that benefit by utilizing recalcitrant SOM, is relatively lower in SN but is higher in SC and SCN than in control soil, resulting in the possible control in PE. Figure 3 shows the microbial biomass at the end of the incubation period for each treatment. Although one-way ANOVA did not reveal any significant differences among treatment and we did not focused only the domination of K-strategists in the microbial biomass community, the microbial biomass tended to increase (particularly SC and SCN) when glucose was added to black soil. Furthermore, microbial biomass C showed significant positive correlations (r = 0.742, P < 0.01, n = 12) with cumulative soil-derived CO₂ emission during the 21-day incubation period for each treatment (Figure 4). These results support the ‘microbial N, P mining’ theory and it suggests that K-strategists are more abundant in SC and SCN than in SN to acquire N and P as nutrient source.

Figure 3. Microbial biomass in each treatment. S, soil only (control); SN, soil with mineral N; SC, soil with glucose; SCN, soil with glucose plus mineral N. Bars indicate the standard error of the mean (n = 3).

Figure 4. Linear correlation coefficients (Pearson’s correlation coefficient) between microbial biomass C and cumulative soil-derived CO₂ emission during the 21-day incubation period for each treatment (n = 12). Asterisks(**) indicates significance level at P < 0.01.

Fontaine et al. (2007) calculated the mean residence time (MRT) of soil C in deep soil considering PE as 324 yr, which was much accelerated compared with the initial soil MRT (2560 yr). If the addition of a glucose supply and the resulting PE were repeated each year in the deep black soil C used in this study, we calculated that the MRT of SC and SCN according to Fontaine et al. (2007) would be 89400/29.4 or 29.8 = 3038 or 2999 yr, respectively, and it was clearly accelerated compared with the initial soil MRT (5960 yr) calculated from soil C stock (89400 mg C kg⁻¹) and annual dissolved organic C (DOC) input into buried A horizon at Takayama site (15.0 mg C kg⁻¹ yr⁻¹). The amount of glucose added to the soil in this study was approximately 1/10 of the annual DOC input into the surface mineral soil through the litter layer at the study site (Chen et al. 2017). If this level of glucose continues to be supplied to buried A horizon in the future at Takayama site, its MRT may further accelerate. Therefore, more detailed studies of the PE with labile C input (e.g., root exudation and fine root biomass), including the long-term PE reaction, on deep black soil C in Japanese volcanic ash soils are necessary for the future.

Acknowledgments

We thank the members of Laboratory of Soil Science, The University of Shiga Prefecture. The study was supported by JSPS KAKENHI Grant Number 15K16118 and 17K00524.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Japan Society for the Promotion of Science [15K16118, 17K00524].