Dynamics of organic carbon and nitrogen in deep soil profile and crop yields under long-term fertilization in wheat-maize cropping system

Muhammad QASWAR,Ll Dong-chu,HUANG Jing,HAN Tian-fuWaqas AHMED,Sehrish ALlMuhammad Numan KHANZulqarnain Haider KHAN,XU Yong-mei,Ll QianZHANG Hui-min,WANG Bo-ren,Ahmad TAUQEER

1 National Engineering Laboratory for Improving Quality of Arable Land/Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,Beijing 100081,P.R.China

2 Key Laboratory of Industrial Ecology and Environmental Engineering,School of Environmental Science and Technology,Dalian University of Technology,Dalian 116024,P.R.China

3 National Observation Station of Qiyang Agri-Ecology System/Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,Qiyang 426182,P.R.China

4 Guangdong Provincial Key Laboratory for Radionuclides Pollution Control and Resources,School of Environmental Science and Engineering,Guangzhou University,Guangzhou 510006,P.R.China

5 Agro-Environmental Protection Institute,Ministry of Agriculture and Rural Affairs,Tianjin 300191,P.R.China

6 Institute of Soil,Fertilizer and Agricultural Water Conservation,Xinjiang Academy of Agricultural Sciences,Urumqi 830091,P.R.China

7 College of Agriculture,Henan University of Science and Technology,Luoyang 471000,P.R.China

8 School of Chemical Engineering,Dalian University of Technology,Dalian 116024,P.R.China

Abstract Soil organic carbon (SOC) and nitrogen (N) are two of the most important indicators for agricultural productivity. The primary objective of this study was to investigate the changes in SOC and N in the deep soil profile (up to 100 cm)and their relationships with crop productivity under the influence of long-term (since 1990) fertilization in the wheat-maize cropping system. Treatments included CK (control),NP (inorganic N and phosphorus (P) fertilizers),NPK (inorganic N,P and potassium fertilizers),NPKM (NPK plus manure),and M (manure). Crop yield and the properties of topsoil were measured yearly from 2001 to 2009. C and N contents were measured at five different depths in 2001 and 2009. The results showed that wheat and maize yields decreased between 2001 and 2009 under the inorganic fertilizer (NP and NPK) treatments.The average yield between 2001 and 2009 under the NP,NPK,NPKM,and M treatments (compared with the CK treatment)increased by 38,115,383,and 381%,respectively,for wheat and 348,891,2 738,and 1 845%,respectively,for maize.Different long-term fertilization treatments significantly changed coarse free particulate (cfPOC),fine free particulate (ffPOC),intramicroaggregate particulate (iPOC),and mineral-associated (mSOC) organic carbon fractions. In the experimental years of 2001 and 2009,soil fractions occurred in the following order for all treatments:mSOC>cfPOC>iPOC>ffPOC. All fractions were higher under the manure application treatments than under the inorganic fertilization treatments. Compared to the inorganic fertilization treatments,manure input enhanced the stocks of SOC and total N in the surface layer (0–20 cm)but decreased SOC and N in the deep soil layer (80–100 cm). This reveals the efficiency of manure in increasing yield productivity and decreasing risk of vertical loss of nutrients,especially N,compared to inorganic fertilization treatments.The findings provide opportunities for understanding deep soil C and N dynamics,which could help mitigate climate change impact on agricultural production and maintain soil health.

Keywords:carbon stock,nitrogen stock,carbon fraction,soil profile,organic amendments,long-term experiment

1.lntroduction

Soil is the main reservoir of carbon (C) and nitrogen (N) which determine the soil health and agroecosystem sustainability(Bünemannet al.2018). C and N dynamics in the soil are mainly regulated by climate change,the soil environment,and human activities such as field management practices(Brevik 2013). Despite of exceptional research efforts to understand the C and total N stocks in soils and their controlling factors,there remain many unknowns,especially the changes in deep soil C and N contents in croplands(Doetterlet al.2015;Angstet al.2018). Understanding the C and N dynamics in soils is important not only for improving crop productivity but also for mitigating climate change and improving ecosystem management practices(Lawet al.2018). Management practices that increase the organic C input and sequestration in soil have been suggested as a possible approach to offset 5–10% of global fossil fuel emissions (Lal 2004). Consequently,the effects of soil management practices on C stocks especially has become a main focus of research. Understanding the influence of continuous management practices on SOC and N dynamics in the soil is thus important for sustainable agricultural systems.

Fertilization management practices are a major strategy to attain high crop productivity. Based on 153 field trials,Chenet al.(2014) reported that the application of organic manure enhanced crop yield by 40% compared with inorganic fertilizer inputs. Contrastingly,Hijbeeket al.(2017)performed several field experiments and found that manure application did not increase the crop yield significantly,while synthetic fertilizer application increased the crop yield by 2.0 t ha?1. These contrasting results suggest a further need to research and investigate the mechanism of crop yield associated with soil fertility under different fertilization management practices. Fertilization techniques influence crop yield by affecting soil properties,with soil organic C and N being the two main fertility indices (Hijbeeket al.2017;Qaswaret al.2020a). The nutrient stocks in the soil can be increased by continuous fertilization. Previous studies mainly investigated the C and total N stocks in surface soil layer (0–30 cm) because it is commonly thought that this layer is mostly influenced by plant roots and fertilization(Minasnyet al.2017). Limited studies have investigated the C and total N stocks at the soil depth below 30 cm (Poeplau and Don 2015;Tautgeset al.2019). Radiocarbon data show that C content in deeper soil depths is more resistant to degradation (Chabbiet al.2009). Generally,subsoils have a high reactive surface area and are less affected by surface agricultural management practices and SOC in deeper soil layers is in the form of organo-mineral complexes,which are important for belowground C stabilization (Rumpelet al.2015). Understanding these subsoil dynamics can help improve field management practices and strategies,and achieve a sustainable copping system.

Fractions of SOC are characterized by their stability and turnover rates under the influence of different management practices;therefore,defining these fractions cannot be excluded from a study focusing on long-term C and total N stocks in the soil. Sixet al.(2002) proposed a method and concept to categorize the physical fractions of bulk SOC into four different aggregate pool fractions according to their specific mechanisms of SOC protection. Identifying these fractions of SOC can be helpful in understanding the influence of continuous fertilization practices on the dynamics of SOC stocks (Sixet al.2002;Heet al.2015). For example,it has been shown that organic manure application and crop residue incorporation enhance SOC content and a higher proportion of the SOC stock (up to 72%) is stored in the mineral-associated soil organic carbon (mSOC) fraction(Courtier-Muriaset al.2013). Few studies have investigated the physical fractions of SOC,especially under long-term fertilization practices,in Chinese croplands (Tonget al.2014). The present study is one of the few long-term studies that investigate the dynamics of C and N stocks in deep soil profile in agricultural systems and highlight the significance of deep soil C and N dynamics under long-term fertilization management practices. Few studies have investigated C and N changes at a soil depth below 30 cm (Poeplau and Don 2015),even though it is known that C in subsurface soil is more protected from biotic and abiotic losses than that in surface soil (Prieset al.2018). This experiment was conducted to analyze the long-term impact of different fertilization strategies on crop productivity,the efficacy of resource utilization,and environmental impacts. The principal objective of this study was to investigate:(1) the pattern of changes in the soil organic C and total N stocks across the soil profiles (up to 1.0 m depth) under long-term fertilization inputs,and (2) the changes in the soil organic carbon fractions in the topsoil layer (0–20 cm) and their influencing factors.

2.Materials and methods

2.1.Experimental location

The field trial started in 1990 at the National Observation and Research Station for Farmland Ecosystems (26°45′42′′N,111°52′32′′E) in southern China. This region has a monsoon climate with an average temperature of 17.7°C and an annual precipitation of 1 290 mm (Appendix A).In the region,the rainfall period of the year is from April to June. The climate data for this study were collected regularly from the regional county weather station. The major cropping rotations in this region are rice–rice and wheat–maize systems. The soil is a red soil type (Baxter 2007) and classified as Eutric Cambisol according to the World Reference Base classification of soils (IUSS Working Group WRB 2015). The soil is 43.86% clay,31.86% silt,and 24.28% sand in the 0–20 cm depth. The initial (1990)soil characteristics of the topsoil were pH of 5.7,SOC of 7.9 g kg?1,total N of 1.07 g kg?1,available N of 79 mg kg?1,total P of 0.45 g kg?1,available P of 14.0 mg kg?1,total K of 13.7 g kg?1,and available K of 104 mg kg?1.

2.2.Experimental design and crop management

All treatments were arranged randomly with two replications under the wheat–maize rotation. Every replicated plot (size 20 m×10 m) was separated from its neighboring plot by a 20-cm cement barrier to avoid mixing of water and nutrients.The field was kept fallow for three years before starting the experiment to ensure homogenous soil properties.The treatments for this study included:(1) CK (control);(2) NP (inorganic N and P fertilizers);(3) NPK (inorganic N,P and K fertilizers);(4) NPKM (NPK plus manure);and(5) M (manure). Fertilizer input rates were the same in all fertilization treatments at 300 kg ha?1of N,120 kg of P2O5,and 120 kg ha?1of K2O. Inorganic fertilizer N and P were applied as urea and calcium superphosphate,respectively,while mineral K was applied as potassium chloride. In the NPKM and M treatments,pig manure application rates were 42 000 and 60 000 kg ha?1,respectively. All manure treatments were applied as a basal application during land preparation before crop sowing. Of the total inorganic fertilizer inputs,70% was applied as a basal application for the maize crop,with the remaining 30% applied to the wheat crop. Every year,crop yield and straw were removed from the field,and stubble residues were allowed to remain in the field. The fresh pig manure consisted of SOC at 413 g kg?1,N at 20.1 g kg?1,P at 12.9 g kg?1,and K at 12.5 g kg?1.Moisture content in fresh pig manure was 70%.

Every year,winter wheat (Xiangmai cultivar) was sown and cultivated at a rate of 63 kg ha?1(160 seeds m?2),and summer maize (Yedan 13 cultivar) was cultivated at a rate of 6.0 seeds m?2. No irrigation management was applied to either crop due to sufficient rainfall and high precipitation during the cropping season. Insecticides (omethoate and carbofuran) were applied to control aphids on wheat and maize borers on maize. After the crop was harvested manually,stubbles (approximately 6 cm in height) and roots remained in the field.

2.3.Sample collection and laboratory analysis

Every year,in the first week after the maize crop harvest,soil samples were collected at a depth of 0–20 cm from five randomly selected points in each treatment plot. Soil samples were collected from five different depths between 0 and 100 cm for soil C and total N stock determination in 2001 and 2009. Soil samples were collected using a stainless-steel core sampler with an inner diameter of 50.46 mm. Composite samples were mixed thoroughly and transported to the laboratory in clean polythene bags for further analysis after air-drying. Fractions of the composite samples were ground and passed through a 0.25-mm sieve to determine soil chemical properties. The SOC was measured according to the Walkley-Black wet oxidation method (Pageset al.1982). The soil total and available N was determined in accordance with Lu (2000) and the soil total and available P were measured according to Murphy and Riley (1964) and Olsen (1954),respectively. Soil pH was determined in a 2.5:1 (water:soil) suspension. The soil bulk density (BD) was estimated with the cutting ring method(with an inner diameter of 50.46 mm;a volume of 100 cm3,and a sampling depth of 50 mm) with three replications (Lu 2000). In the 2001 and 2009 fertilization periods,physical fractions of SOC were separated into different fractions according to the modified fractional method by Sixet al.(2002) and Sleutelet al.(2006);the pools were as follows:the coarse-free particulate organic carbon (cfPOC),the interaggregate fine-free particulate organic carbon (ffPOC),the intramicroaggregate particulate organic carbon (iPOC),and the mineral associated carbon (mSOC) fractions. Then,10 g of the soil sample was briefly sieved through a 250-mm sieve followed by an overnight prewetting treatment at 4°C and gently and manually shaken with 20 glass beads (with a 4-mm diameter) with a constant flow of water through the sieve column. The disturbance of the microaggregates by the beads was avoided,and the process continued until all macroaggregates were completely broken and the water ran clean. The material remaining on the 25-mm sieve was classified as cfPOC intermacroaggregates+coarse sand.The water with the suspended soil passing through the 250-mm sieve was transferred onto a 53-mm sieve and placed into a container. The sieve was then oscillated up and down 50 times in the container. The free silt and clay that passed through the 53-mm sieve was then centrifuged for 10 min with an additional 1.0 mL of 0.2 mol L–1clay flocculation material (CaCl2) and then collected. The microaggregatesized fraction that remained on the 53-mm sieve was dried at 60°C. These microaggregate-sized fractions consisted of the fine sand ffPOC,the iPOC,and the microaggregated silt and clay,and they were separated through the density flotation technique according to Sixet al.(2014). The microaggregate fraction was centrifuged with 30 mL of sodium iodide (NaI) at 1 250×g at 25°C for 60 min. The light fraction was considered as ffPOC and was extracted onto a 20-mm nylon filter and rinsed with deionized water. NaI was thoroughly removed from the remaining heavy fractions by rinsing five times with deionized water. The heavy fractions were dispersed by shaking at 200 oscillations min–1for 18 h in a 5 g L–1sodium hexametaphosphate solution.The dispersed microaggregates were sieved through a 53-mm sieve and rinsed with deionized water to collect the iPOC (53 to 250 mm) and fine sand. The remaining intramicroaggregated silt+clay (<53-mm) fractions were collected with the same method as the free silt+clay,mixed thoroughly and then considered as mSOC.

2.4.Calculation

The annual C input includes C from crop residues plus manure applied to the field and was calculated according to Caiet al.(2019). It was measured from the belowground root C (Cbelowground,t ha–1),the C from stubbles (Cstubble,t ha–1),and the C input from the pig manure application (Cmanure,t ha–1). The following equations were used to estimate the C input (Cinput,t ha?1):

where the Rbgis the ratio of the underground to the aboveground C from crops and is assessed as 30%according to Kunduet al.(2007). Rstubblesis the ratio of incorporated stubbles to aboveground biomass.

Soil organic C stock (Cstock,t ha?1) or total N stock (Nstock,t ha?1) were determined by the following equation:

where SOC,BD,and H indicate the soil organic carbon (g kg?1),bulk density (g cm?3),and soil depth (cm),respectively.TN was used instead of SOC for the total N stock in the above equation.

2.5.Statistical analysis

Significant differences among treatments at different soil profile depths were analyzed by one-way ANOVA followed by Tukey’s HSD test atP=0.05 level of significance. The interactive relationships between treatments and years were analyzed by two-way ANOVA. A redundancy analysis correlation was performed with Canoco Software (Windows version 5.0). The three-dimensional surface plots were made to observe the effects of the soil profile SOC and N contents,and regression analysis were performed using Sigma plot (Windows version 14).

3.Results

3.1.Crop yield,baseline soil properties,and carbon input

Different fertilizer input treatments significantly affected the yield (Figs.1 and 2) and soil properties (Fig.2). Crop yield decreased over the years under inorganic fertilization treatments. Organic manure input boosted the crop yield under the NPKM and M treatments compared to the inorganic fertilizer application treatments. The average yield across the years,compared to the CK treatment,increased by 38,115,383,and 381% for wheat,respectively,and increased by 348,891,2 738,and 1 845% for maize,respectively,under the NP,NPK,NPKM,and M treatments (Table 1).Soil pH and SOC decreased under the CK,NP,and NPK treatments when compared to the NPKM and M treatments.Soil P content showed a sharper increasing trend over the years under organic manure treatments when compared to the synthetic fertilization treatments. On average,over the years,compared to the CK treatment,manure applications increased soil pH,while inorganic fertilizers decreased soil pH significantly. Among all treatments,soil pH ranged between 6.63 (under the M treatment) and 4.47 (under the NP treatment). Compared with the baseline CK treatment,the NP,NPK,NPKM,and M treatments increased the SOC content by 19.1,23.4,75.3,and 79.4%,increased the TN content by 7.4,25.7,59.5,and 59.4%,and increased the available N content by 35.6,34.9,87.9,and 81.6%,respectively. Soil total and available P contents respectively increased by 64 and 750% under NP,86 and 774% under NPK,202 and 3 165% under NPKM,and 170 and 2 389%under M,compared to those under CK. The annual C inputs did not show significant differences between the years for a given treatment,but on average across the years,the fertilization treatments did significantly change the annual C inputs (Fig.3). Compared to the control,fertilization increased annual C input significantly,with the manure treatment causing a steeper increase than the inorganic fertilization treatments.

Fig.3 Boxplot showing long-term C inputs under long-term application of manure and inorganic fertilizers. CK,no fertilization;NP,inorganic nitrogen (N) and phosphorus (P)fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure.

Table 1 Mean crop yield of each treatment from 2001 to 2010 under long-term application of manure and inorganic fertilizers

Fig.1 Yield of wheat and maize crop from 2001 to 2010 under long-term application of manure and inorganic fertilizers. CK,no fertilization;NP,inorganic nitrogen (N) and phosphorus (P)fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure. Error bars indicate the standard deviation based on mean data of yield every two years.

Fig.2 Changes in soil pH (A),soil organic carbon (B),total nitrogen (N) (C),available N (D),total phosphorus (P) (E) and available P (F) contents in topsoil (0–20 cm) from 2001 to 2009 under long-term application of manure and inorganic fertilizers. CK,no fertilization;NP,inorganic and P fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure.

3.2.Physical fractions of the soil organic carbon and their relationship with soil properties

Fertilization treatments significantly influenced SOC fractions in 2001 and 2009 (Fig.4). In the experimental years 2001 and 2009,SOC fractions were present in the following order:mSOC>cfPOC>iPOC>ffPOC for every treatment. With the CK treatment as a baseline,all fractions except the mSOC were increased in all other treatments in both 2001 and 2009. The mSOC fraction decreased under the inorganic fertilization treatment in 2001. The mSOC fraction also increased in all the treatments in 2009,compared to the CK treatment. Compared to 2001,the cfPOC significantly decreased under the CK,NP,NPKM,and M treatments but did not show a significant difference under the NPK treatment in 2009. The ffPOC fraction also decreased between 2001 and 2009 under the CK,NP,and M treatments. The ffPOC fraction in 2009 increased by 24.9 and 2.2% under the NPK and NPKM treatments,respectively,compared to 2001. The iPOC fraction increased in all the treatments,and the mSOC fraction decreased in all the treatments,between 2001 and 2009.The redundancy analysis (RDA) revealed that the pH,SOC,and nutrient contents significantly and positively correlated with the physical fractions of SOC (Fig.5). Simple term effects by RDA showed that the overall variation explained by the individual factors AP,SOC,TN,AN,TP,and pH were 67.1,66.4,59.1,53.4,46.9,and 34.5%,respectively. RDA 1 accounted for 69.1% of the total variation,and RDA 2 accounted for 15.3% of the total variation.

Fig.4 Physical fractions of soil organic carbon in 2001 (A) and 2009 (B) under long-term application of manure and inorganic fertilizers.cfPOC,coarse free particulate organic C;ffPOC,fine free particulate organic carbon;iPOC,intra-microaggregate particulate organic carbon;mSOC,mineral-associated soil organic C fraction. CK,no fertilization;NP,inorganic nitrogen (N) and phosphorus (P)fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure.

Fig.5 Redundancy analysis correlation between soil chemical properties and physical soil organic carbon fraction under longterm application of manure and inorganic fertilizers. Red arrows indicate the explanatory variables and blue arrows indicate the response variables. SOC,soil organic carbon;AP,available phosphorus;AN,available nitrogen;TP,total phosphorus;TN,total nitrogen;cfPOC,coarse-free particulate organic C;ffPOC,fine free particulate organic carbon;iPOC,intra-microaggregate particulate organic carbon;mSOC,mineral-associated soil organic C fraction.

3.3.Stocks and contents of soil organic carbon and total nitrogen within the soil profile

Continuous fertilizer inputs changed the SOC and N contents and stocks across the soil profile (Figs.6 and 7). Overall,compared to 2001,the SOC content and stock decreased in 2009 for all the fertilization treatments. Compared to the top soil layer (0–20 cm),the SOC at the 80–100 cm soil depth decreased under the CK,NP,NPK,NPKM,and M treatments,respectively,by 82.3,15.3,65.1,71.8,and 68% in 2001 and by 74.7,35.9,63.9,80.8,and 73.1% in 2009. Under the CK treatment in 2001,the SOC stock in the 80–100 cm soil depth decreased by 11.4% compared to that in the topsoil layer (0–20 cm). Compared to the topsoil layer (0–20 cm),the SOC stock at the 80–100 cm soil depth increased by 323.7,74.4,40.9,and 60.1% under the NP,NPK,NPKM and M treatments,respectively. In 2009,compared to the topsoil layer (0–20 cm),the SOC stock at the 80–100 cm soil depth increased by 26.3,220.5,80.3,and 34.6% under the CK,NP,NPK and M treatments,respectively. The NPKM and M treatments also increased the soil total N content and stock in the topsoil layer (0–20 cm)and decreased the total N content in the deep soil profile (at 80–100 cm) compared to the NPK treatment. Compared to the topsoil layer (0–20 cm),TN at the soil depth of 80–100 cm was decreased under the CK,NP,NPK,NPKM,and M treatments by 50.5,42.2,39.7,67.7,and 54.6%,respectively,in 2001 and decreased by 66.3,35.3,46.9,62.4,and 64.1% respectively,in 2009. In 2001,the soil TN stock at the 80–10 cm soil depth increased by 147.5,187.7,201.7,61.6,and 126.9% under the CK,NP,NPK,NPKM,and M treatments,respectively. In 2009,compared to the topsoil layer (0–20 cm),the TN stock increased at the soil depth of 80–100 cm under the CK,NP,NPK,NPKM,and M treatments by 68.7,223.6,165.5,87.8,and 79.5%,respectively. The average SOC content across the soil profile between 2001 and 2009 decreased by 38.8,43.4,51.3,41.6 and 23.7%,respectively,and the SOC stock decreased by 35.3,47.4,58.5,50.9 and 25.3%,respectively,under the CK,NP,NPK,NPKM,and M treatments. The average soil total N content across the soil profile between 2001 and 2009 decreased under the CK,NPK,and NPKM treatments by 18.8,9.6,and 18.2%,respectively,but increased under the NP and M treatments by 5.2 and 10.6%,respectively. Compared to 2001,the soil total N stock in 2009 was decreased under the CK,NPK,and NPKM treatments by 19.9,22.5,and 27.7%,respectively.The average SOC stock across the soil profile under the NP,NPK,NPKM,and M treatments respectively increased by 156,101,99,and 93% in 2001 and by 108,29,51,and 122% in 2009 with the CK treatment as the baseline. The total N stock under the NP,NPK,NPKM,and M treatments respectively increased by 13.4,33.7,62.3 and 27.8% in 2001 and by 49.2,29.3,46.5 and 65.4% in 2009 compared to that under the CK treatment.

Fig.6 Soil organic carbon (A and B) and total nitrogen (N) contents (C and D) in the soil profile (0–100 cm) during 2001 and 2009 under long-term application of manure and inorganic fertilizers. CK,no fertilization;NP,inorganic N and phosphorus (P) fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure. Data are mean of numbers of replication (n=2) and error bars represent the standard deviation.

Fig.7 Soil organic carbon (A and B) and total nitrogen (N) stocks (C and D) in the soil profile (0–100 cm) during 2001 and 2009 under long-term application of manure and inorganic fertilizers. CK,no fertilization;NP,inorganic N and phosphorus (P) fertilization;NPK,inorganic N,P and potassium (K) fertilization;NPKM,inorganic NPK and manure addition;M,manure. Data are mean of numbers of replication (n=2) and error bars represent the standard deviation.

3.4.Relationships between soil characteristics,nutrient inputs,and crop yield

Regression analysis showed that the annual C input,C stock,and N stock within the topsoil were strongly and positively correlated with crop yields (Fig.8). Threedimensional surface analyses give insight into the effects of deep soil organic C stock and total N stock on annual crop yields (Fig.9). The three-dimensional surface plots showed that the SOC and the total N stocks in the topsoil layer (0–20 cm) had more influence on the crop yield than SOC and total N at deeper depths of the soil profile.

Fig.8 Relationships between annual carbon inputs,organic carbon stock and total nitrogen stock in topsoil and annual crop yield under long-term application of manure and inorganic fertilizers.

Fig.9 Three dimensional surface plots of crop yield affected by soil organic carbon (SOC;A),total nitrogen (TN;B),SOC (C),and total nitrogen stock (D) under long-term fertilization.

4.Discussion

The results indicated that continuous manure input,alone or combined with inorganic fertilizer,significantly improved wheat and maize crop yields when compared to the inorganic fertilizer application alone (Figs.1 and 2). These outcomes are supported by previous studies(Yanget al.2015;Caiet al.2019),indicating that manure application restores nutrients,consistently resulting in high crop yields. In the present study,a significant decline in crop yield over the years under the NPK and NP fertilizer treatments might be attributed to significant acidification through continuous inorganic fertilization. Previously,Hollandet al.(2019) observed a positive correlation between soil pH and crop yield under a 35-year long-term field experiment. Our results are also supported by the finding of Choudharyet al.(2018),which showed that long-term inorganic fertilization under the wheat–soybean cropping system significantly decreased crop yield over time. Choudharyet al.(2018) showed that net H+ions are normally released from plants,but plants release OH–or HCO3–if net anion uptake is significantly increased (Tanget al.2011). Inorganic fertilization,especially with urea,reduces the net base cations in the soil,which adversely impacts the soil pH. It has also been found that inorganic N fertilizer inputs substantially shift the soil’s Al3+buffering phase,which releases the net Al from Al-hydroxides on clay minerals in acidic soil,thus reducing the saturation of base cations and promoting soil acidity (Stevenset al.2009).Moreover,manure inputs provide retained nutrients to the crop for many years (Demelashet al.2014). It has also been shown that the alkaline nature of manure increases the pH of acidic soils (Rukshanaet al.2013). Manure input increases the soil pH by neutralizing protons within the soil.Therefore,as could be expected,in the present study,the soil pH was the highest under the M treatment among all the treatments (Fig.2). Furthermore,continuous application of the manure treatment also improved the availability of nutrients,such as N and P,which has also been supported by Miet al.(2016) and Qaswaret al.(2020b). The higher pH of the manure-amended soil was also attributed to buffering from the bicarbonates and organic acids within the manure.Another study demonstrated that mineral N (NH4-N+NO3-N),available P,K,Ca,and Mg increased in the soil immediately after manure application,which consequently increases the soil pH (Whalenet al.2000).

The increase in the crop yield under the NPKM and M treatments was also associated with the high C inputviathe manure,which increased the SOC content (Fig.3).Tautgeset al.(2019) reported that a 9 t ha–1estimated C input through continuous poultry manure input improves SOC content. Long-term applications of manure can increase the SOC and soil nutrients. Manure application increases the C input and crop residues within the soil,which enhances the soil C sequestration rate and increases soil productivity (Caiet al.2019). Manure can also increase microbial biomass and activity in the soil,which provides a better environment for crop growth (Peacocket al.2001).In this study,manure application treatments significantly increased crop yield compared to the inorganic fertilization treatments. Soil organic carbon protection and stability play a substantial role in the storage and the sequestration potential of SOC in agricultural soil. Tonget al.(2014)reported that an input of manure alone or in combination with inorganic fertilizer increased all fractions of SOC,compared to inorganic fertilizer inputs alone,which is consistent with our results (Fig.4). Organic fractions are associated with the composition and type of organic manure applied to the soil(Yadvinder-Singhet al.2005;Tonget al.2014). Therefore,manure inputs supply different organic compounds that contain all fractions of the SOC. Some studies proposed that the ffPOC and coarse ffPOC fractions establish a conceptual“unprotected”pool of SOC (i.e.,ffPOC),which is mainly biomass derived from microorganisms (Sixet al.2002). Therefore,the ffPOC fraction is also known as a significant indicator for C dynamics under continuous agricultural practices (Carteret al.2003;Wanget al.2019).Among the different treatments,this study observed that all fractions were increased by the addition of manure (Fig.4).Moreover,soil pH,SOC,and total and available nutrient contents were positively correlated with all fractions of SOC. Organic C storage in the soil increased under longterm manure application,mainlyviastabilization within the microaggregates of the soil. It has been shown that iPOC accounts for only 15–20% of the total SOC sequestration under long-term fertilization (Tonget al.2014). Mineralassociated C is important for the enhancement of the C sequestration potential in croplands,which increases the organic C stock in the soil,especially under long-term manure application (Sixet al.2002). In the present study,mineral-associated C fractions were the highest among all fractions in 2001 and 2009.

Changes in the stocks of SOC and total N across the soil depth profiles impact production and sustainability within the agroecosystem. Compared to the topsoil layer,organic C and total N contents decreased with soil depth,but their stocks were particularly increased at a depth of 80–100 cm in all the fertilizer treatments. Wanget al.(2016) reported that the SOC and total N contents in cropland between 0 and 60 cm soil depth decreased with depth regardless of the fertilization treatment. Fresh organic carbon inputs on the surface of the soil and the pattern of biomass allocation strongly influence the below-ground soil carbon (Fontaineet al.2007). In the present study,lower SOC contents in the subsoil than in the surface soil could be explained by the slow decomposition of organic matter on the surface of the soil due to acidic soil conditions. A lower soil pH decreases the microbial activity and slows the decomposition of soil organic matter (Motavalliet al.1995). The SOC content under all fertilization treatments,particularly under the inorganic fertilization treatment,decreased between 2001 and 2009 (Fig.2-B). This could be due to the acidic soil conditions,because SOC is easily depleted in acidic soil(Kalbitzet al.2000). Another reason could be that the annual precipitation in 2009 was lower than that in 2001 (Appendix A). Previously,Chenet al.(2020) found a decrease in mSOC under high N input,mainly due to soil acidification.Interestingly,in the study,soil organic C contents were higher in the topsoil and lower in the subsurface,especially at the 80–100 cm depth,under the long-term manure application treatment compared to the inorganic fertilization treatment(Figs.6 and 7). Several factors influence the soil C and N contents in the soil profile such as soil texture,irrigation,type of vegetation,and soil amendments (Tuoet al.2018).Meanwhile,soil total N stock was also higher at the surface and lower at the soil depth of 80–100 cm under the manure treatments when compared to the NP treatment. Soil bulk density was mostly decreased in the soil surface layers due to the continuous addition of manure,which includes particulate organic matter;however,some studies have found an inverse relationship between the SOC content and soil bulk density (Perie and Ouimet 2008). The high soil organic C and total N stocks in the deep soil profile under inorganic fertilization compared to the manure application could be due to the high vertical flow of N and C under inorganic fertilization (Yadvinder-Singhet al.2005;Hobbie and Ouimette 2009;Kindleret al.2011). Additionally,the higher soil bulk density under inorganic fertilization treatments compared to the manure application treatments might be a reason for the high C and total N stocks in the deep soil profile (Tautgeset al.2019). However,the study did not measure the vertical nutrient losses directly,but an analysis of the deep soil N content indicated that manure application reduces the vertical nutrient loss compared to the inorganic fertilization treatments. Moreover,the study observed a positive relationship between the annual C input,the soil organic C,and the total N stock and the annual crop yield (Fig.8),indicating the beneficial impact of manure input on sustaining long-term productivity.

5.Conclusion

The present study concluded that compared to the inorganic fertilizer treatment,long-term manure application increases crop yield by increasing the soil organic C inputs,the SOC stock,the total N stock,and the soil pH,especially in the topsoil layer (0–20 cm in depth),under the wheat–maize cropping system. Among the different fractions,ffPOC was the highest under all the fertilization treatments and showed the highest sensitivity to different fertilization treatments,indicating a difference in the SOC content among different fertilization treatments. The complete soil profile (0 to 100 cm in depth) study showed that continuous manure application alone or in combination with inorganic fertilizers increases the stock of SOC and the total N within the surface layer(0–20 cm) but decreases the SOC and total N in the deep soil layer (80–100 cm) compared to the inorganic fertilization treatments. These results suggest that manure inputs can reduce the vertical nutrient flow and therefore reduce vertical losses of nutrients.

Acknowledgements

This research was financially supported by the National Key Research and Development Program of China(2016YFD0300901 and 2017YFD0800101),and the Fundamental Research Funds for Central Non-profit Scientific Institution,China (161032019035,1610132020022 and 1610132020023).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendixassociated with this paper is available on http://www.ChinaAgriSci.com/V2/En/appendix.htm