Optimization of tillage rotation and fertilization increases the soil organic carbon pools and crop yields in a wheat-maize cropping system on China’s Loess Plateau
Xia Zhang1|Sixu Lu1|Chenguang Wang1|Afeng Zhang1|Xudong Wang1,2
1College of Natural Resources and Environment, Northwest A&F University, Yangling, Shaanxi 712100, PR China
2 Key Laboratory of Plant Nutrition and Agri-environment in Northwest China, Ministry of Agriculture, Yangling, Shaanxi 712100, PR China
Correspondence : Xudong Wang and Afeng Zhang; College of Natural Resources and Environment, Northwest A&F University, No. 3, Taicheng Road, Yangling 712100, Shaanxi, PR China.
E-mail: wangxudong01@126.com (X. Wang); zhangafeng@nwsuaf.edu.cn (A. Zhang).
ABSTRACT: Long-term application of high nitrogen and phosphorus fertilizer and mono-tillage practices can adversely affect soil health, carbon sequestration and crop growth. A 10-year field experiment was conducted in a wheat-maize cropping system on the Loess Plateau in China to explore fertilization and tillage methods that improve SOC sequestration and crop yields. We evaluated the effects of (1) fertilization (balanced fertilization (BF), low fertilization (LF), and conventional fertilization (CF)) and (2) alternating years of different tillage (no-tillage and subsoiling (NS), subsoiling and ploughing (SP), ploughing and no-tillage (PN)) or continuous ploughing tillage (PP) on input-C, SOC pool, and crop yields. BF and rotational tillage (NS, SP, and PN) increased the amount and stabilization rate of input-C, thereby increased SOC storage, and the highest effect was found in BF+NS treatment. Simultaneously, BF produced higher contents of SOC, readily oxidizable carbon (ROC), particulate organic carbon (POC) and dissolved organic carbon (DOC) and C pool management index (CMI) at 0-10 cm depth. For tillage, rotational tillage increased labile C contents and CMI at 0-10 cm, 20-35 cm and 35-50 cm depths, which improved soil quality. Crop yields showed an increase tendency with the increases of SOC content, labile C fraction contents, and CMI. Therefore, the higher yields of wheat and maize were found in BF and rotational tillage; the highest were in BF+NS treatment. Our finding suggested that NS combined with BF may be the best management to increase SOC storage, improve soil quality and productivity on China’s Loess Plateau.
KEYWORDS: rotational tillage; fertilization; soil organic carbon sequestration; organic carbon fractions; crop yields.

1. INTRODUCTION

Soil organic carbon (SOC) is not only a major indicator to evaluate soil fertility, but also a primary pool of global carbon storage (about 1580 Gt C) (Schimel, 1995). Efforts to increase SOC content are thought to significantly reduce atmospheric CO2 and enhance soil fertility and crop productivity (Sauerbeck, 2001). Soil tillage is one of the main human factors to cause the changes in the SOC pool. Long-term conventional tillage (CT) decreased labile C content (Cheng et al., 2012) and caused 25-75% SOC mineralization loss (Nicoloso et al., 2018). Meanwhile, conservation tillage reportedly increase the labile C contents and SOC sequestration (Somasundaram et al., 2016; Badagliacca et al., 2018). The influence of tillage is dependent on tillage depth. Carbonell et al. (2015) observed that the positive impacts of NT on SOC and labile C were limited to the topsoil (0-10 cm). The influence of tillage method on the SOC pool also is dependent on soil type and environmental factors. Also, different tillage practices significantly influence crop yields. Previous researchers found that NT and minimum tillage increase SOC sequestration and crop yields (Sun et al., 2018a; Márcio et al., 2018) and that subsoiling tillage (ST) could enhance root growth and improve crop yields (Sun et al., 2018b). However, other studies suggested that long-term mono-conservation tillage was noneffective to improve crop productivity (Stewart et al., 2017; Li et al., 2018) and some even found that NT negatively affected crop yields compared with CT (Zhao et al., 2017; Bogunovic et al., 2018). Therefore, the influences of single continuous conservational tillage on crop yields are unclear. Recent studies have reported that rotational tillage effectively prevented the disadvantages of mono-tillage, while increasing SOC content and crop productivity (Chu et al., 2016; Wang et al., 2018). Nevertheless, the data about the impacts of tillage rotation systems on SOC sequestration, labile C content, soil quality, and crop productivity on the China’s Loess Plateau is lacking.
Fertilization is an indispensable agricultural practice used to achieve high yields, and it also changes SOC content. Previous studies have indicated that inorganic fertilizer application significantly improved crop productivity and increased crop residues amount returned to the field and therefore increased SOC and labile C contents (Nayak et al., 2012; Ghimire et al., 2017). A study reported by Chaudhary et al. (2017) indicated that soils fertilized over the long term contained higher levels of SOC, readily oxidizable C (ROC), particulate organic C (POC), and dissolved organic C (DOC) than unfertilized soils. Some studies have indicated that the positive influences of fertilization on crop yields only occur with balanced fertilization and not with unbalanced fertilization (Kukal et al., 2009; Shahid et al., 2017). Whereas, Liang et al. (2012) observed that inorganic fertilizer application has no significant positive effects on SOC, labile C, and crop yields. These inconsistent results may be related to climate, soil type, crop system, experimental duration, and other factors (Samuel et al., 2018). Therefore, studies of the impacts of fertilization on SOC sequestration and crop growth should be accordance with local conditions.
The Loess Plateau is located in Northwest China and is a major production region for wheat and maize (Ren et al., 2016). Long-term intensive conventional tillage system has caused a series of troubles, such as increased soil erosion, loss of SOC due to mineralization, and degradation in soil quality (Chen et al., 2009). Therefore, conservational and rotational tillage practices with crop straw return have been viewed as potential options. Loess soil fertility is low because of low contents of SOM, total nitrogen (N), and available phosphorus (P), whereas total potassium (K) content is high due to the rich illite minerals in soil parent material (De et al., 2011). Thus, to pursue short-term high crop yields, farmers have always adopted a fertilizing mode of high N and P in this region, which is not conducive to sustainable soil production. Recently, some studies have shown that the balanced application of N, P, and K fertilization can strongly improve crop productivity (Lu et al., 2017). On the whole, to find the optimal mode of fertilization and tillage for the Northwest China region, it is essential to assess the impacts of fertilization and rotational tillage on SOC sequestration, soil quality and crop productivity in this region.
In this study, our objectives were to study the changes of SOC sequestration, soil labile C fractions, C pool management index (CMI), and crop yields under different fertilization and tillage methods. Thus, a 10-year (2007 to 2016) field experiment with different fertilization and tillage modes in a wheat-maize crop system in the Loess Plateau was established to test the following hypotheses: (i) balanced application of N, P, K fertilization with the rotation of different tillage practices would increase wheat and maize yields and increase the SOC stock by increasing the amount of crop residues returned to the field and (ii) annual rotation of no-tillage and subsoiling would increase the SOC content and soil labile C fractions, thus improving the CMI and crop productivity in the China’s Loess Plateau.

2 | MATERIALS AND METHODS

2.1 | Site

This work was conducted over ten years (2007 to 2016) in the Dryland Agricultural Research Station of Northwest A & F University, located in Heyang County (104°04′E, 35°19′N altitude 877 m), Shaanxi Province of Northwest China. The climate at the research station is temperate semi-arid continental monsoon. In this region, the mean annual frost-free period is 210 days. The annual mean precipitation, evaporation and temperature were 536.6 mm, 1,833 mm and 11.5 °C, respectively. The soil in this region is classified as a Chromic Cambisol (sand 34%, silt 39%, clay 27%) using the FAO/UNESCO Soil Classification (1993). The main soil characteristics (0-50 cm) at the beginning of study are listed in Table S1.

2.2 | Experimental design

We adopted a plit-plot design for this experiment. Fertilization treatment was kept in primary plot treatment and included: balanced fertilization (BF), low fertilization (LF), and conventional fertilization (CF). Tillage method was the sub-plot treatment and included: no-tillage rotated with subsoiling in alternating years (NS), subsoiling rotated with ploughing in alternating years (SP), ploughing rotated with no-tillage in alternating years (PN), and ploughing applied every year (PP). Five complete cycles of rotations were implemented during the ten years (2007 to 2016). The two factors were combined into 12 treatments.
Fertilization treatment: BF, N: 150 kg ha-1, P2O5: 120 kg ha-1, and K2O: 90 kg ha-1; LF, N: 75 kg ha-1, P2O5: 60 kg ha-1, and K2O: 45 kg ha-1; CF, N: 225 kg ha-1, P2O5: 180 kg ha-1, and no potassium fertilizer applied. Wheat and maize had the same amount of fertilizer application under the same fertilization treatment. The N, P2O5, and K2O were from urea (N: 46.4%), diammonium phosphate (N: 18%, P2O5: 44%), and potassium chloride (K2O: 60%), respectively. The full rates of P and K plus 50% of N was applied on the sowing date for maize, the remaining 50% of N was used at the 12th leaf stage (the middle of June each year). The full rates of N, P, and K fertilizers were used once on the planting date for wheat.
Tillage methods: In no-tillage (N), crop residue was chopped and covered on the topsoil, the soil was left undisturbed until sowing. In subsoiling (S), sub-soiling, the crop residue was chopped and covered on the topsoil. The topsoil remained undisturbed while the subsoil was subsoiled (30-35 cm depth, 60 cm width) using a subsoiler. In ploughing (P), the crop residue was chopped, and mixed with the soil (20-25 cm depth) using a moldboard plow. Fields were tilled according to the described methods after crop harvest each year.
Planting: The winter wheat/spring maize rotation system was carried out in this study. For winter wheat (cultivars: Jinmai 47 before 2014, Chang 6359 after 2014), the time of planting and harvesting were in the end of September and early June of the following year, respectively. The seeding rate and row spacing were 150 kg ha-1 and 20 cm, respectively. For spring maize (cultivars: Yuyu 22 before 2014, Zhengdan 958 after 2014), the time of planting and harvesting were in April and the end of September, respectively. The seeding rate and row spacing were 37.5 kg ha-1 and 60 cm, respectively. Table S2 describes the crop rotation schedule from 2007 to 2016.

2.3 | Soil sampling and processing

Soils were randomly sampled in 0-10 cm, 10-20 cm, 20-35 cm and 35-50 cm layers with three replications under different treatments in June 2016 after winter wheat harvest. After air-drying, samples were ground and sieved to 0.25-mm for analyzing the soil C fractions. SOC was obtained with the K2CrO7-H2SO4digestion method (Walkley and Black, 1934). Readily oxidizable C (ROC) was measured using the 333 mmol L-1 KMnO4 oxidation method described by Blair et al. (1995). Dissolved organic C (DOC) was determined as described by Jiang et al. (2006). The CHCl3 fumigation-extraction method was used to measure Microbial biomass C (MBC) (Vance et al., 1987). Particulate organic C (POC) was measured according to the description of Cambardella and Elliott (1992).
Total SOC stock was calculated as
\(\text{\ SOC}_{\text{stock}}=0.1\times\text{SOC}_{c}\times D\times T\)(1)
Where \(\text{SOC}_{\text{stock}}\) (Mg ha-1) is the stock of SOC, \(\text{SOC}_{c}\) (g/kg) is SOC content, D (g cm-3) and T (cm) represent soil bulk density and soil thickness, respectively.
\(\text{\ ΔSOC}_{\text{stock}}=\text{\ SOC}_{\text{stock}}-\text{\ SOC}_{stock0}\)(2)
Where\(\text{ΔSOC}_{\text{stock}}\) (Mg/ha) is the SOC stock accumulation and\(\text{\ SOC}_{stock0}\)(g kg-1) is the SOC stock before the experiment.
CMI was calculated according to the description of Blair et al. (1995). The CF+PP treatment is the reference in the equation:
\(L=\text{ROC}_{c}/\text{SOC}_{c}\) (3)
Where L is the lability of SOC, \(\text{ROC}_{c}\) is the content of readily oxidized carbon, and \(\text{SOC}_{c}\) (g kg-1) is SOC content.
\(\text{CPI}=\text{SOC}_{\text{cs}}/\text{SOC}_{\text{cr}}\) (4)
WhereCPIstands for carbon pool index, \(\text{SOC}_{\text{cs}}\) and\(\text{SOC}_{\text{cr}}\) (g/kg) are the soil organic carbon content of the sample and the reference (CF+PP treatment), respectively.
\(\text{LI}=L_{C}/L_{R}\) (5)
Where LI is the C lability index, \(L_{C}\) and \(L_{R}\) stand for the C lability of the sample and the reference (CF+PP treatment).
\(\text{CMI}=\text{CPI}\times\text{LI}\times 100\) (6)

2.4 | Estimation of plant biomass carbon inputs

In each treatment, wheat and maize samples were randomly collected from three 3 m2 and three 9 m2 areas at the harvest period. After air-drying, the samples were manually threshed and weighed to determine the yields and straw amount.
The input-C from straw, \(C_{\text{straw}}\), was calculated as:
\(C_{\text{straw}}=B_{\text{straw}}\times 0.40\) (7)
Where \(B_{\text{straw}}\) is crop straw biomass, plant biomass C input was calculated by assuming that C content in plant tissue is 40% according to the study of Johnson et al. (2006).
The input-C from crop stubble, \(C_{\text{stubble}}\), was calculated as:
\(C_{\text{stubble}}=B_{\text{straw}}\times r\times 0.40\) (8)
Where r is the stubble to straw ratio. For Maize, stubble represented 10% of straw biomass. For wheat, stubble represented 20% of straw biomass (Li et al., 2016).
The input-C from roots, \(C_{\text{root}}\), was calculated as:
\(C_{\text{root}}=B_{\text{straw}}\times r\times 0.40\) (9)
Wherer is the root to straw ratio. Root represented 23% (maize) and 22% (wheat) of straw biomass according to the description of Kong et al. (2005).
The input-C from rhizodeposition (\(C_{\text{rhizodep}}\)) was calculated as follows (Maillard et al., 2018):
\(C_{\text{rhizodep}}=C_{\text{root}}\times 0.65\) (10)
The stabilization rate (SR, %) of plant biomass carbon into SOC at the 0-50 cm depth was assessed as follows (Srinivasarao et al., 2012):
\(SR=\text{\ ΔSOC}_{\text{stock}}/C_{\text{input}}\) (11)

Where\(\mathbf{C}_{\mathbf{\text{input}}}\)is accumulative plant biomass C input.

2.5 | Statistical analysis

SAS 9.0 (SAS Systems, Cary, NC, USA) was used for data analysis. Origin 2016 was used for graphing. The Split-plot analysis of variance (ANOVA) was applied to assess the impacts of fertilization and tillage on input-C, SOC, soil labile C fractions (including ROC, DOC, MBC and POC), CMI and crop yields. The multiple comparisons were conducted on the basis of the Duncan Multiple Range Test (at < 0.05 level). And the relationships among the indexes were evaluated by regression equations.

3 RESULTS

3.1 | Accumulative plant biomass C input

The cumulative C input from crop residue (including crop straw, crop roots and rhizodeposition) in the 10-year experimental period (2007-2016) are shown in Table S3 and Figure 1. The fertilization treatments caused significant differences in the accumulative plant biomass C input, and the total C input under BF higher than that under LF and CF. Simultaneously, rotational tillage systems (SP, PN, and NS) significantly increased the C inputs from straw, root residue, stubble, and rhizodeposition. Compared with the PP, total C input were 18.10%, 11.04% and 14.72% higher in the SP, PN, and NS tillage than that in PP, respectively.

3.2 | SOC

The effects of the fertilization on SOC content at 0-20 cm depth were significant after ten years, but were not significant in the deeper layers (Figure 2, p < 0.05). Compared with the CF, the BF and LF significant increased the SOC content and stocks at the 0-10 cm and 10-20 cm depth. At the 0-50 depth, SOC stock under different fertilization treatments was followed by BF > CF > LF, and the statistical differences were found among three fertilization treatments (Table S4; Figure 3; p < 0.05). Significant effects on SOC content and stocks due to tillage system were detected at the 0-50 cm depth (p < 0.05). SOC stock was markedly higher in the rotational tillage (SP, PN, and NS) than that in PP tillage. More specifically, compared with PP tillage, the NS, SP, and PN tillage primarily increased SOC content and stocks at the 0-10 cm depth. Similarly, the rotational tillage systems also increased SOC content in > 10 cm soil layers. The difference was significant between NS and PP in10-20 cm and 35-50 cm layers, was significant between SP and PP in the 20-35 cm and 35-50cm layers, and was significant between PN and PP at the 20-35 cm depth (p < 0.05).
The SOC stocks of all treatments increased after the ten-year experiment, and the maximum increment by 6.83 Mg ha-1 was in the treatment of BF+NS, the minimum increment by 2.54 Mg ha-1 was in the treatment of CF+PP (Figure 3). The SOC stock accumulation at the 0-50 depth was higher under BF and LF than that under CF, and there was a statistical difference between BF and CF (p < 0.05). Changes in SOC stock accumulation under different tillage systems decreased in the order of NS > PN > SP> PP, and the statistical differences were found between rotational tillage and PP tillage (p < 0.05).

3.3 | Stabilization rate

The proportion of SOC stock accumulation to total C input is the stabilization rate (Figure 4). The stabilization rate in the treatment of BF+NS was 17.08%, which was the highest among all treatments. The lowest stabilization rate was recorded (7.66%) in the CF+PP treatment. For fertilization effect, the stabilization rate were increased by 39.00% in BF and by 33.79% in LF compared with CF. Meanwhile, the effect of tillage on the stabilization rate was significant. Compared with the PP, the stabilization rate were significantly increased by 61.54%, 20.83%, and 26.92% in the NS, SP, and PN, respectively (p <0.05).
Figure 5 showed that SOC stock accumulation was positively correlated with total C input, and the stabilization rate of plant biomass C (p <0.05).

3.4 | Soil labile C fractions

Table 1 showed the differences of soil labile C fractions (ROC, DOC, MBC, and POC) among treatments (p < 0.05). For fertilization effect, the order of ROC content under different fertilization treatments was BF > CF > LF in the topsoil (0-10 cm), and there were statistical differences between BF and LF (p < 0.05). However, the differences in ROC content among the three fertilization treatments were nonsignificant at > 10 cm depth. Tillage had significant effects on ROC in all soil layers (0-50 cm). Compared with PP tillage, NS and SP significantly increased ROC content at the 0-10 cm, 20-35 cm, and 35-50 cm depth; and PN significantly increased ROC content at the 0-10 cm and 20-35 cm depth (p < 0.05).
Significant effects of the fertilization treatments or tillage systems on DOC content were detected in all soil layers (0-50 cm) (Table 1). DOC content with BF was higher than that with CF at the 0-10 cm, 10-20 cm, and 20-35 cm depth, and DOC content was lower with LF than that with CF at the 0-10 cm and 35-50 cm depth (p < 0.05). For tillage effect, DOC content was significantly higher in rotational tillage systems (NS, SP, PN) than that in PP tillage in all soil layers (0-50 cm)(P <0.05). More specifically, compared with PP, DOC content increased by 40.92%, 10.89%, 15.64%, 30.95% under NS, increased by 13.56%, 5.03%, 8.00%, 20.48% under SP and increased by 24.94%, 12.29%, 10.55%, 13.33% under PN at 0-10 cm, 10-20 cm, 20-35 cm and 35-50 cm depth, respectively.
The content of MBC was followed by CF > BF > LF in all layers (0-50 cm), and there were statistical differences between LF and CF (Table 1). Tillage method affected MBC content in all layers (0-50 cm). Compared with PP tillage, rotational tillage systems (NS, SP and PN) significantly increased MBC content at the 0-10 cm, 20-35 cm, and 35-50 cm depth. Meanwhile, NS, SP and PN decreased MBC content compared to PP tillage at 10-20 cm depth (p < 0.05).
Fertilization treatment markedly affected POC content at the 0-10 cm and 10-20 cm depth, and POC content was higher with the BF treatment than that with the CF and LF treatments (Table 1). Tillage significantly affected POC content. Compared with PP, POC content increased by 58.79%, 15.73%, 67.39% under NS, increased by 26.06%, 41.57%, 36.96% under SP and increased by 42.42%, 28.09%, 30.43% under PN at 0-10 cm, 10-20 cm and 35-50 cm depths, respectively.