Soil organic matter dynamics in lands continuously cultivated with maize and soybean in Heilongjiang Province, Northeast China: Estimation from natural ¹³C abundance

Abstract

Northeast China is the main production area of maize and soybean in China. In the present study, the rates of decomposition and replacement of soil organic carbon (SOC) were estimated using the soil inventory collected since 1991 from long-term maize and soybean cultivation plots in Heilongjiang Province, Northeast China, to evaluate the sustainability of the present cultivation system. The total carbon (C) content in soil was stable without any significant changes in the plots (approximately 28.5 g C kg⁻¹). The δ¹³C value of soil organic matter under continuous maize cultivation increased linearly with an annual increment of 0.07 from −23.9 in 1991, which indicated that approximately 13% of the initial SOC was decomposed during the 13-year period of maize cultivation, with a half-life of 65 years. Slow decomposition of SOC was considered to result from the low annual mean temperature (1.5°C) and long freezing period (170–180 days year⁻¹) in the study area. In contrast, the amount of organic C derived from maize increased in the soil with a very slow annual increment of 0.17 g C kg⁻¹, probably because of the removal of all the plant residues from the plots. Based on the soil organic matter dynamics observed in the study plots, intentional recycling/maintenance of plant residues was proposed as a way of increasing soil fertility in maize or soybean cultivation.

Key words:

• continuous cropping
• δ¹³C
• maize
• soil organic matter
• soybean

INTRODUCTION

Maize (Zea mays L.), soybean (Glycine max Merr.) and rice (Oryza sativa L.) are major staple crops in China. The area cultivated with maize in China covered 1.4, 2.0 and 2.4 × 10⁷ ha in 1960, 1980 and 2003, respectively, whereas soybean continuously covered approximately 9.3 × 10⁶ ha during this period (http://zzys.agri.gov.cn/nongqing/nongqing.asp). Thus, although the increase in the cultivated area from 1960 to 2003 was only 1.7-fold for maize and negligible for soybean, the production of maize and soybean increased drastically during this period: 1.6 × 10⁷ and 6.4 × 10⁶ tons in 1960, 6.3 × 10⁷ and 7.9 × 10⁶ tons in 1980, and 1.2 × 10⁸ and 1.5 × 10⁷ tons in 2003 for maize and soybean, respectively (http://zzys.agri.gov.cn/nongqing/nongqing.asp).

Northeast China including Liaoning, Jiling, Heilongjiang and Neimenggu Provinces was the main production area of maize and soybean, and accounted for 35% and 39% in 1960, 27% and 42% in 1980 and 29% and 50% in 2003, respectively, of the total production area in China (http://zzys.agri.gov.cn/nongqing/nongqing.asp). These statistical records require that the soil fertility of Northeast China be carefully monitored under maize and soybean cultivation to enable stable and sustainable production in the future.

It is generally recognized that soil fertility, which is evaluated based on the organic matter content in soil, is decreasing in farmlands worldwide. Changes in soil fertility can be estimated from the balance between the annual input of plant residues to soil (e.g. crops, weeds, compost) and the decomposition rate of soil organic matter (SOM). Plant residues derived from C3 and C4 plants display δ¹³C values of −12 and −27, respectively (Boutton et al. 1998; Diels et al. 2001; Smith and Epstein 1971), because of the difference in their photosynthetic mechanisms (Bender 1971; Smith and Epstein 1971). Therefore, the annual input of SOM derived from C4 plants and the decomposition rate of the original SOM accumulated under C3 vegetation can be evaluated independently in farmlands where C3 plants were replaced with C4 plants.

The determination of the natural ¹³C abundance of SOM is a suitable method for evaluating the effect of land-use changes and field management on the decomposition and stabilization rates of soil organic carbon (SOC) (Balesdent et al. 1990; Flessa et al. 2000; Ludwig et al. 2003; McDonagh et al. 2001; Römkens et al. 1999) and for quantifying the C turnover rates of SOC in several particle-size fractions (Besnard et al. 1996; Gerzabek et al. 2001; Gleixner et al. 1999; Jolivet et al. 2003).

The SOC content was 30% and 45% lower in farmlands with continuous maize cultivation for 12 and 27 years, respectively, than in nearby forests in France, and 18% and 23% of the SOC was derived from maize residues (Jolivet et al. 2003). The SOC content decreased more drastically from 84–103 g C kg⁻¹ to 38 g C kg⁻¹ for 25–30 years after deforestation for maize cultivation in the sub-humid Ethiopian highlands, where 20–36% of SOC was derived from maize (Solomon et al. 2002). Land-use change from grasslands to farmlands (maize) was less serious in terms of soil degradation, and no decrease in the SOC level was recognized in Canada and France (Bolinder et al. 1999; Puget et al. 1995), where 9–21% and 45% of SOC were derived from maize residues after 15 and 23 years in the farmlands, respectively.

It is well known that field management influences the C turnover rate based on ¹³C analysis, and 3.70, 3.95 and 4.00 kg C m⁻² of SOC that originated from grasses remained in soil after 17 years of maize cultivation under conventional, no-tillage and superficial tillage management, respectively (Balesdent et al. 1990). In contrast, the effect of fertilization on the turnover rate of SOC was not recognized from ¹³C analysis after 32 years of maize cultivation in Canada (Gregorich et al. 1996).

Thus, although information about the rates of degradation and replacement of SOC is accumulating worldwide, information about the SOC dynamics in the farmlands of Northeast China, the major grain belt in China, is not available. The purpose of the present study was to estimate the rates of decomposition and replacement of SOC by using the soil inventory collected from a long-term maize cultivation plot in Heilongjiang Province, Northeast China, to estimate the sustainability of the present maize cultivation system.

MATERIALS AND METHODS

Soil samples and field history

We used air-dried soil samples that had been collected periodically from maize and soybean monoculture plots at Hailun Experimental Station for Agricultural Ecology, Northeast Institute of Geography and Agricultural Ecology, Chinese Academy of Sciences, Heilongjiang Province (47°26′N, 126°38′E, 240 m a.s.l.) since 1991. Urea and (NH₄)₂HPO₄ amounting to 62 kg N ha⁻¹ and 69 kg P₂O₅ ha⁻¹ were applied to maize fields as basal fertilizer, and an additional 69 kg N ha⁻¹ from urea was applied at the booting stage. For soybean, 27 kg N ha⁻¹ and 69 kg P₂O₅ ha⁻¹ from (NH₄)₂HPO₄ were applied as basal fertilizers. Seeding and harvesting were carried out in early May and early October, respectively. The collection from 9 holes in each plot of the 0–10 cm soil layer, which were mixed to obtain a composite sample, was carried out in April or May except for 1994, 2001 (October) and 2004 (June). Although the cultivation history before 1991 was not fully documented, the δ¹³C value of the soil samples collected in 1991 (−23.89) indicated that C3 plants had been cultivated as major crops in the past.

The mean annual precipitation was 530 mm and the mean annual temperature was 1.5°C. The soil was a Black soil (Mollisol) derived from loamy loess. Particle-size distribution determined using the pipette method was as follows: 40% sand, 26% silt and 34% clay for the 0–20 cm soil layer. The mean grain yields (± annual deviation expressed as standard deviation) of maize and soybean, 5.7 ± 0.8 and 2.1 ± 0.4 ton ha⁻¹, did not vary appreciably between 1991 and 2004, respectively.

The soil samples were passed through a 2 mm mesh screen and any plant residues were carefully discarded. All the sieved samples were further crushed to pass through a 0.5 mm mesh screen.

The above-ground parts of maize were collected in June 2004 from four replicates. The materials were oven-dried and pulverized using a vibration mill (TI-100, HEIKO, Japan).

Determination of total C content and δ¹³C values of soil samples

The presence of a significant amount of inorganic C in the soil samples was checked by the addition of 2 mol L⁻¹ HCl under submerged conditions. Because the addition of the acid did not result in CO₂ ebullition, the total C content and δ¹³C values of the soil samples were directly measured without any pretreatment (even if 100 discernible bubbles with a diameter of 1 mm evolved with the addition of HCl to a 1 g soil sample, the amount of inorganic C was 28 µg, which was approximately one thousandth of the total C content in the soil sample [approximately 28 mg C g⁻¹ soil]). The total C content in soil and δ¹³C values of SOM (maize plot only) in plant samples were measured using a NC analyzer (Micro Corder JM1000CN, J. Science Laboratory, Kyoto, Japan) and a stable isotopic ratio mass spectrometer connected to an elemental analyzer (DELTA Plus–NC2500, ThermoFinnigan, San Jose, CA, USA), respectively.

Estimation of the proportion of SOC derived from maize C

The proportion of SOC derived from maize C (a; %) in the plot with continuous maize cultivation was estimated using the following equation:

RESULTS AND DISCUSSION

Total C content in soil samples and δ¹³C values of SOM

As shown in Fig. 1, the total C content in soil was stable without any appreciable changes under continuous maize (28.9 ± 0.9 g C kg⁻¹) and soybean (28.3 ± 1.2 g C kg⁻¹) cultivation. These findings are not in agreement with those obtained in a wheat–soybean rotation in farmers’ fields in the Heilongjiang Province by Liu et al. (2003), in which the SOC content decreased by 15%, 25% and 48% compared with an uncultivated field (54.4 g C kg⁻¹) after 5, 14 and 50 years of cultivation, respectively. The drastic decrease in SOC content in the field under the wheat–soybean rotation was attributed, in part, to the determination of SOC changes during the first 50 years

Figure 1  Total carbon (C) content in soil under continuous maize and soybean cultivation. (•), maize cultivation field; (▴), soybean cultivation field.

of cultivation after the virgin lands were changed to farmlands. In contrast, in the present study, the SOC changes were estimated in the past 13 years in the farmland. In addition, the lower yield in farmers’ fields seemed to have resulted in a lower input of crop residues to the fields. The average yields of maize and soybean in Heilongjiang Province were 3.84 and 1.49 ton ha⁻¹ in 2001, respectively.

The δ¹³C values of SOM under continuous maize cultivation had increased linearly since 1991 (Fig. 2), and the increase in the δ¹³C value was best fitted to the following equation:

in which t is the year (AD) of measurement.

Estimation of temporal decrease in the initial SOC pool and accumulation rate of maize-derived C in soil

The carbon remaining after partial mineralization may differ in terms of the δ¹³C value from the initial carbon, with generally ¹³C enrichment, because isotopic discrimination against the heavier ¹³C isotope during microorganism-mediated respiration and decomposition can enrich the remaining organic matter with ¹³C (Agren et al. 1996). Natelhoffer and Fry (1988) showed in a laboratory experiment that because of overall isotopic fractionation during litter decomposition, the residual SOC was enriched with ¹³C, although they did not observe changes in the ¹³C content of SOC during the humification process. In contrast, it was reported in another study that organic matter from both wheat and maize underwent decomposition without significant ¹³C enrichment (Balesdent et al. 1990). In the following data analyses and discussion we assumed that there would

Figure 2  Changes in the δ13C value of soil organic carbon (SOC) under continuous maize cultivation. δ13C () = 0.07 × (y − 1990) − 23.93 (significant at 1% level; y, the year of measurement).

Figure 3  Changes in the content of carbon (C) derived from initial soil organic matter (SOM) under continuous maize cultivation.

Figure 4  Changes in the content of carbon (C) derived from maize residues under continuous maize cultivation.

not be any differences in the δ¹³C values between the initial and residual C.

Temporal decrease in the initial SOC pool of the previous vegetation (Fig. 3) and temporal increase in the SOC pool derived from maize (Fig. 4) since 1991 were determined based on variations in the total C content and the proportion of maize-derived C estimated using Eq. 1. The average δ¹³C values of the initial SOM and maize plants were −23.89 (standard deviation 0.01) and −12.50 (standard deviation 0.28). Judging from the pattern of decrease in the initial SOC pool with time (t), a simple exponential model (C _t  = A₀ × e ^−kt ) was used (Jolivet et al. 1997):

The equation indicated that only approximately 13% of the initial SOC was decomposed during the 13-year period of maize cultivation in the study field. The half-life (t_hl) and turnover rate (t_tr) of the initial SOC expressed as 0.693/k and 1/k, respectively, were 65 (t_hl) and 94 (t_tr) years.

The amount of organic C derived from maize (C_m) increased with the duration of cultivation as shown in Fig. 4. Again a simple exponential model [C_m = A_m × (1 − e ^−kt )] was used to estimate the C_m chronosequence (Jolivet et al. 1997):

It was deduced from Eqs 3, 4 that half of the SOC in the field would be derived from maize residues (C_{initial SOC} = C_maize) in approximately AD 2077 under the present field management.

Half-life of the initial SOC and accumulation rate of maize-derivd C in soil

Table 1 shows a comparison of the decrease in the SOC content and the half-life of the initial SOC determined in the present study using data reported in the literature under continuous maize cultivation. As in many studies soil samples were not available at the time of the onset of the experiment (Gregorich et al. 1996; Jolivet et al. 2003; McDonagh et al. 2001; Solomon et al. 2002), reference soil samples that were located near the experimental plots were assumed to show the same C contents and δ¹³C values as those of the original soils in the literature. Thus, systematic soil collection was found to be indispensable for evaluating the changes in soil fertility associated with agricultural activities. The half-life of SOC at the present study site was longer than that recorded at sites described by other researchers (ranging from 15 to 33 years). Exceptions included the sites in Germany (Flessa et al. 2000; Ludwig et al. 2003) and Tanzania (McDonagh et al. 2001), where the half-life was longer than that in our study, ranging from 97 to 152 years. Because the SOC content at those sites was low (6.7–13.1 g C kg⁻¹), SOC at the onset of the experiment might consist mainly of recalcitrant organic C remaining after readily and slowly decomposable C was consumed by soil microorganisms. Although the Ethiopian and French soil samples (27.2–38.2 g C kg⁻¹) contained similar amounts of SOC to that of the present soil samples, SOC half-life was very short, 15–18 years. The longer half-life of SOC at the present study site, notwithstanding the large SOC content (approximately 30 g C kg⁻¹), probably resulted from the long freezing period of soil (170–180 days year⁻¹) and low annual temperature (1.5°C), as well as the lower annual rainfall (530 mm).

Although all the maize residues were removed from the plot fields every year for cattle feeding, the SOC content had remained nearly constant since 1997 (Fig. 1). The annual decreasing rate of initial SOC [(A₀ − A₀ × e ^−k) / A₀ × 100] was estimated from Eq. 3 to be approximately 1.1%, corresponding to 0.31 (= 29.6 – 29.6e ^−0.0106) g C kg⁻¹ or approximately 310 kg C ha⁻¹ (in the case where the bulk density = 1 and plow layer = 10 cm). In contrast, the accumulation rate of C derived from maize (most probably from maize roots) was 0.17 [= 33.1 × (1 − e ^−0.0052)] g C kg⁻¹ year⁻¹, based

Table 1 Comparison of the decrease in soil organic carbon (SOC) content and half-life of the initial SOC after maize cultivation

Country	Site	Mean Ann. Temp (°C)^1)	Rainfall (mm year^−1)	Soil texture^2)	Original vegetation	Fertilizer	Period (year)	SOC (g kg^−1)		% of maize C	Soil depth (cm)	Half-life of SOC (year)3)	Reference
								Original	Farmland
Canada	Ontario	8.7	876	28%Cl, 35%Si	Sod	+	32	51.3*	21.8	28	0–10	19	Gregorich et al. (1996)
	Ontario	8.7	876	28%Cl, 35%Si	Sod	−	32	51.3*	19.5	18	0–10	19	Gregorich et al. (1996)
	Québec	4	1200	18%Cl, 55%Si	Grassland	+	15	33.0**	23	9	0–20	23	Bolinder et al. (1999)
Ethiopia	Wushwush	18	1800	clayey	Forest	+	25	84.4*	38.1	20	0–10	17	Solomon et al. (2002)
	Munesa	19	1250	clayey	Forest	+	30	103.1*	38.2	36	0–10	15	Solomon et al. (2002)
France	Boigneville	10.5	600	22%C	C3 plants	+	17	10.0**	9.6	30	0–10	29	Balesdent et al. (1990)
	Boigneville	10.5	600	25%Cl, 69%Si	C3 plants	+	23	10.0**	10.2	45	0–30	28	Puget et al. (1995)
	Boigneville	12.7	900	29%Cl, 39%Si	Forest	+	15	31.1*	20	16	0–25	17	Jolivet et al. (2003)
	Boigneville	12.7	900	27%Cl, 42%Si	Forest	+	28	31.1*	18.6	21	0–25	26	Jolivet et al. (2003)
	Boigneville	12.7	900	30%Cl, 16%Si	Forest	+	11	24.7*	20	8	0–25	25	Jolivet et al. (2003)
	Boigneville	12.7	900	20%Cl, 13%Si	Forest	+	32	24.7*	15.4	19	0–25	33	Jolivet et al. (2003)
	Boigneville	12.7	900	31%Cl, 23%Si	Forest	+	12	39.5*	27.7	11	0–25	18	Jolivet et al. (2003)
	Boigneville	12.7	900	26%Cl, 58%Si	Forest	+	27	39.5*	21.6	19	0–25	23	Jolivet et al. (2003)
Germany	Halle	9.2	465	10%Cl, 20%Si	Rye	+	37	14.9*	13.1	15	0–20	87	Flessa et al. (2000)
	Halle	9.2	465	10%Cl, 20%Si	Rye	+	39	13.7*	12.8	14	0–25	123	Ludwig et al. (2003)
	Halle	9.2	465	10%Cl, 20%Si	Rye	−	39	10.6*	9.8	9	0–25	152	Ludwig et al. (2003)
The Netherlands	Noord-Brabant			3%Cl, 5%Si	Pasture	+	26	23*	16	23	0–20	29	Römkens et al. (1999)
Tanzania	Vitono		500–800		Forest/bushland	+	25	6.7*	6.3	11	0–20	97	McDonagh et al. (2001)
	Ikuwala		500–800		Forest/bushland	+	25	21.9*	10.2	22	0–20	17	McDonagh et al. (2001)
	Mlulula^4)		500–800		Forest/bushland	+	25	10.1*	8.2	30	0–20	31	McDonagh et al. (2001)
China	Hailun	1.5	530	30%Cl, 26%Si	Farmland	+	13	30.7**	31.1	8^5)	0–10	65^6)	Present study
^1)Mean annual temperature.
^2)Cl and Si: clay and silt, respectively.
^3)Half-life of initial SOC.
^4)Rotation with sorghum.
^5)The value was calculated from Eqs 2 and 3.
^6)The value was calculated from Eq. 2.
*Reference soil located nearby.
**Soil samples at the time when the experiment started.

on Eq. 4. As the dry weight ratios among grain, straw and roots were roughly 2:2.5:1 for maize (Wang G, unpubl. data, 2002) and the mean grain yield of maize was 5.7 ton ha⁻¹ in the study plots, the annual inputs of root residues and root C to the soil were estimated to exceed 2.9 ton ha⁻¹ and 1.2 ton C ha⁻¹, respectively, assuming that the C content in the root biomass was 42%. Approximately 15% of the added root residues appeared to remain and compensate for the decomposition of the initial SOC. Therefore, the SOC content of the farmlands in the study areas may increase by intentional recycling/maintenance of crop straw in the fields. As the SOC content in the forest areas was approximately 50 g C kg⁻¹, according to the survey conducted at Zhaoguang, Heilongjiang Province, by Liu et al. (2003), 50 g C kg⁻¹ of SOC may become one of the targeted values for farmlands in Heilongjiang Province.

ACKNOWLEDGMENTS

We thank Mr Noriyuki Ikai, Nagoya University, for his technical assistance in the determination of the total C content. This research was conducted under the support of the Nakajima Foundation and the Chinese Academy of Sciences (Innovation Project; grant number KZCX1-SW-19-2).