Changes in organic C stability within soil aggregates under different fertilization patterns in a greenhouse vegetable field

LUAN Hao-an ,YUAN Shuo ,GAO Wei ,TANG Ji-wei ,Ll Ruo-nan ,ZHANG Huai-zhiHUANG Shao-wen

1 Institute of Agricultural Resources and Regional Planning/Key Laboratory of Plant Nutrition and Fertilizer of Ministry of Agriculture and Rural Affairs,Chinese Academy of Agricultural Sciences,Beijing 100081,P.R.China

2 College of Forestry,Hebei Agricultural University,Baoding 071000,P.R.China

3 Tianjin Institute of Agricultural Resources and Environment,Tianjin 300192,P.R.China

4 Institute of Agricultural Resources and Environment,Hebei Academy of Agriculture and Forestry Sciences,Shijiazhuang 050051,P.R.China

Abstract Knowledge of the stability of soil organic C (SOC) is vital for assessing SOC dynamics and cycling in agroecosystems.Studies have documented the regulatory effect of fertilization on SOC stability in bulk soils.However,how fertilization alters organic C stability at the aggregate scale in agroecosystems remains largely unclear.This study aimed to appraise the changes of organic C stability within soil aggregates after eight years of fertilization (chemical vs.organic fertilization)in a greenhouse vegetable field in Tianjin,China.Changes in the stability of organic C in soil aggregates were evaluated by four methods,i.e.,the modified Walkley-Black method (chemical method),13C NMR spectroscopy (spectroscopic method),extracellular enzyme assay (biological method),and thermogravimetric analysis (thermogravimetric method).The aggregates were isolated and separated by a wet-sieving method into four fractions:large macroaggregates(>2 mm),small macroaggregates (0.25–2 mm),microaggregates (0.053–0.25 mm),and silt/clay fractions (<0.053 mm).The results showed that organic amendments increased the organic C content and reduced the chemical,spectroscopic,thermogravimetric,and biological stability of organic C within soil aggregates relative to chemical fertilization alone.Within soil aggregates,the content of organic C was the highest in microaggregates and decreased in the order microaggregates>macroaggregates>silt/clay fractions.Meanwhile,organic C spectroscopic,thermogravimetric,and biological stability were the highest in silt/clay fractions,followed by macroaggregates and microaggregates.Moreover,the modified Walkley-Black method was not suitable for interpreting organic C stability at the aggregate scale due to the weak correlation between organic C chemical properties and other stability characteristics within the soil aggregates.These findings provide scientific insights at the aggregate scale into the changes of organic C properties under fertilization in greenhouse vegetable fields in China.

Keywords:fertilization,organic C stability,soil aggregates,thermogravimetric analysis,13C NMR spectroscopy

1.lntroduction

Soil organic C (SOC) is a critical attribute for soil quality in agroecosystems (Smithet al.2013).Moreover,as an important sink and source of atmospheric CO2,SOC plays a crucial role in the global C balance (Karhuet al.2014).Therefore,sustaining a suitable level of SOC is conducive to improving soil quality and alleviating climate change.In this case,an understanding of SOC composition is needed,because the C-containing constituents in soil C differ largely in their stability against microbial degradation,which greatly affects SOC cycling and sequestration (Wang Het al.2017).It is commonly believed that SOC resistance to microbial degradation(i.e.,SOC stability) can be affected by physical occlusion within soil aggregates (physical protection;Mechanism 1),chemical absorption with minerals (chemical protection;Mechanism 2),and SOC inherent molecular structure’s“resistance”to microbial degradation (biochemical protection;Mechanism 3) (Schmidtet al.2011).The importance of Mechanisms 1 and 2 in soil C dynamics and SOC stability has been demonstrated by several researchers (Adhikari and Yang 2015;Huanget al.2016).A previous study on Mechanism 3 indicated that organic and chemical fertilization can reduce and increase SOC stability in bulk soils,respectively (Luanet al.2019),which,however,was not an in-depth investigation at the aggregate scale.

As the key units of soil structure,soil aggregates are involved in many ecosystem processes (e.g.,sequestration of SOC and nutrient cycling) (Liuet al.2019).Additionally,soil aggregates are heterogeneous assemblages of organic and mineral particles (Sixet al.2004) that provide physical protection of organic C and create microenvironments that differ in C resources among different aggregates (Guanet al.2019).For instance,soil macroaggregates (>0.25 mm) contain mainly active and labile organic C forms (e.g.,carbohydrates,lipids,and fatty acids) that originate principally from plant residues or organic resources (ORs) (Sixet al.2004).Conversely,microaggregates (<0.25 mm)consist of more stable SOC fractions (e.g.,phenols,lignin dimers,and aromatic compounds) formed by highly humified macromolecules (Gaoet al.2015).However,from long-term field experiments,Srivastavaet al.(2020)and Xuet al.(2021) found that labile C (e.g.,KMnO4oxidizable C and dissolved organic C) was higher in microaggregates than in macroaggregates,suggesting that C in microaggregates was less stable than C in other aggregates.This contradictory evidence indicates that organic C stability in soil aggregates remains highly uncertain and needs further exploration and study.

Greenhouse vegetable production (GVP) systems represent one of the most intensive agroecosystems in China (Huet al.2017).A survey conducted in 2013 showed that the areas of GVP systems had reached approximately 2.0 million ha in China (over 85% of total world production) (MOA 2014).However,GVP systems create both benefits and problems in the agricultural sector.For instance,in GVP systems,the overuse of N fertilizers (up to 1 000 kg N ha–1) leads to soil structure deterioration,acidification,and salinization,which,in turn,causes several issues,such as soil C loss and soil quality degradation (Guanet al.2019).To deal with these issues,organic manure and crop straw,as important ORs,are recommended as reasonable means of supplementary C sources for maintaining the balance of SOC accumulation and decomposition and aggregate formation in GVP systems (Liuet al.2018).Recently,the application of ORs to agricultural soils has been reported to alter SOC stability and composition in bulk soils (Luanet al.2019),but no such studies at the aggregate scale in GVP systems has been conducted.Thus,determining the influence of ORs on the stability of organic C within soil aggregates will provide a better understanding of SOC dynamics in GVP systems.

Numerous methods,including the modified Walkley-Black method (Liuet al.2018),thermal analysis techniques (Peltreet al.2017),and13C nuclear magnetic resonance spectroscopy (Wang Het al.2017),have been applied to estimate SOC stability and divide SOC into diverse C pools or fractions that differ in their lability (Luanet al.2019).However,researchers increasingly realized that information obtained by utilizing only one technique to evaluate SOC stability may have methodological defects or method-specific uncertainties (Angstet al.2019;Luanet al.2019).Further,other studies have demonstrated that the stability of organic C in bulk soils can be thoroughly evaluated from different angles (e.g.,spectroscopic,biological,and thermogravimetric angles)(Peltreet al.2017;Nandanet al.2019).

This study used four methods (the modified Walkley-Black method,13C nuclear magnetic resonance spectroscopy,extracellular enzyme assays,and thermogravimetric analysis) in combination to provide more integrated and reliable information regarding how fertilization alters organic C stability at the aggregate scale in a GVP field located in Tianjin,China.The major aims of this research were to:(1) comprehensively evaluate the effects of different fertilization patterns (organicvs.chemical fertilization) on soil aggregate distribution and aggregate-associated organic C stability;and (2) appraise the connections among the four methods,and identify whether these methods can be used to estimate organic C stability at the aggregate scale under these study conditions.

2.Materials and methods

2.1.Study site

An eight-year experiment (from October 2009 to January 2017),which featured the double cropping of leafy vegetables (celery) and fruit vegetables (tomato),was conducted on a solar greenhouse farm in Xiqing District,Tianjin,China (117°0´E,39°1´N).The experimental area is located in a typical warm sub-humid continental climate zone,with a mean annual temperature of 11.6°C,mean annual precipitation of 586 mm,and groundwater depth of 1 m.The experimental site has a loamy soil that is classified as medium-loam Chao (Aquic Cambisols) soil by the FAO soil classification.The initial soil properties of the plough horizon (0–20 cm) in October 2009 are shown in Table 1.

2.2.Experimental design

The experiment was arranged in a randomized block design with three replicates and four fertilization treatments.The area of one individual plot was 2.4 m wide and 6.0 m long (14.4 m2).Each plot was separated from the others by PVC plates to avoid nutrient and water cross-contamination between neighboring plots.Four treatments were included:(1) 100% chemical N addition(100CN),(2) 50% chemical N and 50% manure N addition(50CN/50MN),(3) 50% chemical N,25% manure N,and 25% straw N addition (50CN/25MN/25SN),and (4) 50%chemical N and 50% straw N addition (50CN/50SN).Notably,the study was designed to vary the quality and quantity of C resource inputs while maintaining a constant amount of added nutrients (N,P2O5,and K2O).The specific description of the nutrients and C addition and information about the fertilizers used are shown in Table 2 and Appendix A.

Table 1 The initial soil properties of the plough horizon (0–20 cm) sampled in October 20091)

Table 2 The amounts of nitrogen and carbon applied in the present study (kg ha–1)

2.3.Soil sampling and aggregate size preparation

Surface soil (0–20 cm depth) samples from each plot were collected in mid-January 2017.Five undisturbed soil block samples (20 cm×10 cm×5 cm) were collected randomly from each plot and mixed carefully to form one composite soil sample.The composite samples were stored in sterile containers (rigid plastic boxes) under cool conditions (4°C) and quickly transferred to the laboratory for sieving.All of the composite soil samples were passed through an 8-mm sieve by carefully breaking soil clods and avoiding soil deformation from mechanical compression.During this process,macrofauna,stones,and plant residue were removed.After mixing thoroughly,these samples were used for aggregate fractionation.

Soil aggregates were physically fractionated following the procedure described by Cambardella and Elliott(1993) with minor modifications.Briefly,the sieved fieldmoist soil subsamples (<8 mm) equivalent to 100 g of dry weight were placed at the top of three stacked sieves (2,0.25,and 0.053 mm in diameter),and carefully infiltrated in water for 10 min.The sieves were then manually moved (amplitude 3 cm) vertically 50 times within 2 min to separate the sieved soils into four aggregates,namely,large macroaggregates (>2 mm),small macroaggregates(2–0.25 mm),microaggregates (0.25–0.053 mm),and silt/clay fractions (<0.053 mm).Meanwhile,each separated aggregate fraction was divided into two subsamples.One subsample was stored at 4°C (for lessthan one week) until subsequent analysis of extracellular enzyme activities,whereas the other subsample was air-dried for the measurement of organic C pools and thermogravimetric and spectroscopic properties.

In the present study,a series of indices were used to assess the effects of fertilization (organicvs.chemical fertilization) on organic C composition and stability at the aggregate scale in the GVP system in Tianjin,China.Specifically,these indices were:chemical indices (2.4),spectroscopic indices (2.5),biological indices (2.6),and thermogravimetric indices (2.7) of organic C stability within soil aggregates.

2.4.Oxidizable organic C analysis (chemical method)

The oxidizable organic C within soil aggregates was measured according to the modified Walkley-Black method (Chanet al.2001).Briefly,0.5-g soil samples(<0.25 mm) were added to a 250-mL conical flask containing 10 mL of potassium permanganate (K2Cr2O7;0.167 mol L–1),and then 5,10,or 20 mL of concentrated sulfuric acid (H2SO4,18 mol L–1) was added to create gradient-oxidizing conditions.Thus,oxidizable organic C was divided into four types of C pools based on their oxidizable labilities.Specifically,organic C was categorized into very labile C (CVL,organic C oxidized under 5 mL H2SO4),labile C (CL,organic C oxidized in 10 mL H2SO4–organic C oxidized in 5 mL H2SO4),less labile C (CLL,organic C oxidized in 20 mL H2SO4–organic C oxidized in 10 mL H2SO4),and non-labile C (CNL,total organic C–organic C oxidized in 20 mL H2SO4).For easy interpretation,CVLand CLwere regarded as active C pools(Ca),while CLLand CNLwere considered passive C pools(Cp) (Nathet al.2016).

2.5.13C NMR spectroscopy (spectroscopic method)

Aggregate samples for solid-state13C cross-polarization with magic angle spinning (CP–MAS) NMR spectroscopy(Bruker Avance III 400 NMR spectrometer,Germany)were treated with hydrofluoric acid (HF;2%) to remove the mineral phase (Luanet al.2019),prior to13C NMR analysis using the procedure described by Wang Het al.(2017).Afterwards,the spectra obtained with13C NMR were divided into four C functional groups based on their chemical-shift areas (Luanet al.2019):(a) alkyl C:0–45 ppm,(b) O-alkyl C:45–110 ppm,(c) aromatic C:110–160 ppm,and (d) carbonyl C:160–190 ppm.

2.6.Soil extracellular enzyme assay (biological method)

Using fluorogenically labelled substrates according to the procedures of Zhanget al.(2015),this study determined two important extracellular enzyme activities to obtain the information about the enzyme-based indices involved in the soil C cycle and stability.The two activities were:(i) oxidase (phenol oxidase,PHOs),which oxidizes phenolic substrates,and utilizes oxygen to catalyze the oxidation of recalcitrant organic C (Bellet al.2013);and (ii) hydrolase (β-glucosidase,BG),an important indicator of C dynamics,which hydrolyzes cellulose and oligosaccharides and releases monosaccharides (e.g.,glucose) (Luoet al.2017).Briefly,two labelled fluorogenic substrates (4-methylumbelliferone-β-D-glucoside (4-MUB)and 1-3,4-dihydroxyphenylalanine (L-DOPA)) were used for determining the BG and PHOs activities,respectively.

The sieved moist aggregates (1.0 g in dry weight)were used to create a soil suspension by dissolving the aggregates in 50 mL of 50 mmol L–1Na-acetate buffer.Further details have been described in Zhanget al.(2015).For BG analysis,the buffer,sample suspension,10 μmol L–1references,and 200 μmol L–1substrates(4-MUB) were dispensed into the wells of a black 96-well microplate.The black 96-well microplates were incubated at 25°C for 4 h in the dark.For PHOs activity,the buffer,sample suspension,25 mmol L–1L-DOPA,and 0.3% (w/v) H2O2were dispensed into the wells of a clear 96-well microplate.The clear 96-well microplates were incubated at 20°C for 20 h in the dark.Afterwards,the activities of BG and PHOs were determined by a microplate fluorometer (Scientific Fluoroskan Ascent FL,Thermo Fisher Scientific,Waltham,MA,USA) with 365 nm excitation,450 nm emission filters (BG),and 450 nm absorbance (PHOs).All of the extracellular enzyme activities were expressed as nmol h–1g–1soil.

2.7.Thermogravimetric analysis (thermogravimetric method)

Thermogravimetric (TG) analysis can be used for evaluating the thermogravimetric composition of soil samples (Planteet al.2009).Briefly,TG analysis was performed with a Netzsch STA 409PC Luxx simultaneous thermal analyzer (Netzsch-Gerätebau GmbH,Selb,Germany).To eliminate the interference of inorganic C,soil samples were treated with 1 mol L–1HCl at room temperature for 24 h,and the HCl was washed out with distilled water.Then,the samples were air-dried and lightly ground in a mortar.Afterwards,10−15 mg HCltreated aggregate samples in a Pt/Rh crucible were progressively heated from ambient temperature (～20°C)to 800°C (10°C min–1) in a synthetic air (20% O2and 80% N2balance,30 mL min–1) and N2(10 mL min–1)atmosphere.Moreover,data recorded at <150°C were discarded to avoid the weight loss changes related to moisture loss (Covaledaet al.2011).

Based on several studies (Gaoet al.2015;Luanet al.2019),three phases were distinguished in the exothermic region of 150–600°C:(i) the first phase at 150–380°C was regarded as the mass loss of labile fractions (Exo1,thermally labile fraction);(ii) the second phase at 380–500°C was considered the mass loss of resistant fractions(Exo2,thermally resistant fraction);and (iii) the third phase at 500–600°C was considered the mass loss of refractory fractions (Exo3,thermally refractory fraction).The total mass loss (ExoT) was the sum of Exo1,Exo2,and Exo3.

2.8.Calculations

Soil aggregate stabilityThe mass proportions of the four aggregates were used to calculate soil aggregate stability,i.e.,mean weight diameter (MWD),as follows(Luanet al.2020):

whereXiis the average diameter of aggregates (mm);andAiis the mass proportions of aggregates (%).

Chemical indices of organic C stability within aggregatesThe lability index (LI) was used to reflect organic C chemical oxidizability (i.e.,stability) and calculated based on the values of CVL,CL,CLL,and organic C (Nandanet al.2019):

where CVL,CL,and CLLare very labile C,labile C,and less labile C,respectively.Moreover,the concentrations of organic C within aggregates were determined by the heated dichromate/titration method (Nelson and Sommers 1982).

Spectroscopic indices of organic C stability within aggregatesThree spectroscopic indices were calculated based on the proportions of different C functional groups(Wang Het al.2017):

Biological indices of organic C stability within aggregatesOne biological index for evaluating organic C stability within aggregates was calculated using the following equation (Sinsabaugh and Shah 2011;Luanet al.2019):

Thermogravimetric indices of organic C stability within aggregatesThe thermo-stability index (TSI),which can reflect the stability of organic C in soils,was calculated as Peltreet al.(2017):

Additionally,this study also calculated thermogravimetric indices,the temperature at which half of the thermogravi metric mass loss occurs (TG50),of organic C stability within aggregates (Planteet al.2009).

2.9.Data analysis

One-way analysis of variance with the Duncan test was used to test the significance (P<0.05) of soil variables among fertilization treatments or aggregates.Pearson’s correlation analysis and linear regression analysis were applied to evaluate the relationships among organic C chemical,spectroscopic,biological,and thermogravimetric indices within aggregates.All of the statistical analyses were conducted with SPSS 16.0 Software (SPSS Inc.,Chicago,IL,USA).

3.Results

3.1.Soil aggregate distribution and stability

Soil aggregate distribution in all fertilization treatments showed that the main fractions were macroaggregates(74.7–82.8%),followed by microaggregates (9.1–11.4%)and silt/clay fractions (7.9–13.9%) (Fig.1).Compared with the non-ORs-amended soil (100CN),the ORsamended soils (50CN/50MN,50CN/25MN/25SN,and 50CN/50SN) had more large macroaggregates (39.2–52.1%),less small macroaggregates (30.7–40.7%),and less silt/clay fractions (9.1–9.5%),and exhibited higher aggregate stability (as indicated by high values of MWD:2.43–2.97 mm) (Fig.1).

Fig.1 Changes in soil aggregate distribution and aggregate stability (MWD) under different fertilization patterns.100CN,100% chemical N;50CN/50MN,50% chemical N and 50%manure N;50CN/25MN/25SN,50% chemical N,25% manure N and 25% straw N;50CN/50SN,50% chemical N and 50%straw N;MWD,mean weight diameter.Different lowercase letters indicate significant differences (P<0.05) among different fertilization treatments within the same aggregates.Error bars indicate SE (n=3).

3.2.Organic C and oxidizable organic C contents within soil aggregates

Over the 8-year experimental period,the contents of organic C,Ca,and Cpin each aggregate showed an increasing trend with increasing C inputs (100CN<50CN/50MN<5 0CN/25MN/25SN<50CN/50SN) (Table 3).Additionally,compared to the 100CN treatment,organic amendments increased the LI values within large macroaggregates,small macroaggregates,microaggregates,and silt/clay fractions by 3.2−6.5,11.5−20.0,9.3−22.7,and 4.4−8.5%,respectively(Table 3).

Across the four aggregates,the contents of organic C and oxidizable organic C (Caand Cp) were the highest in microaggregates (24.4,13.8,and 10.6 g kg–1) and decreased in the order:microaggregates>macroaggregates>silt/clay fractions.The LI values for various aggregates varied from 1.54 to 1.80,which were the highest in silt/clay fractions (1.80),followed by macroaggregates (1.63–1.66)and microaggregates (1.54).

3.3.13C NMR spectra within soil aggregates

The relative proportions of different C functional groups at the aggregate scale were measured by13C NMR analysis (Table 4).In comparison with the non-ORsamended treatment,the ORs-amended treatments increased the relative proportions of O-alkyl C within large macroaggregates,small macroaggregates,microaggregates,and silt/clay fractions by 5.2–10.1,7.3–15.0,5.7–10.5,and 2.2–6.0%,respectively.Changes in the relative proportions of aromatic C within all aggregates showed the opposite trends of those of O-alkyl C under different fertilization treatments:50CN/50SN<50CN/25M N/25SN<50CN/50MN<100CN.Moreover,there were no significant differences (P<0.05) in the relative proportions of alkyl C and carbonyl C within aggregates among all the fertilization treatments.These results caused higher OA/A and lower AI and RI values within aggregates in the ORs-amended soils than in the non-ORs-amended soil(Table 4).

Table 3 Mean values of organic C contents,oxidizable organic C contents,as well as chemical indices of organic C stability within different aggregates under different fertilization patterns

For all of the fertilization treatments,the relative proportions of alkyl C in the aggregates decreased significantly in the following order:silt/clay fractions>macroaggregates>microaggregates.The relative proportions of O-alkyl C were substantially greater(3.6–5.4%) in microaggregates than in other aggregates.Meanwhile,the relative proportions of aromatic C increased significantly by 17.1,18.7,and 14.6% in large macroaggregates,small macroaggregates,and microaggregates,respectively,compared to those in silt/clay fractions.The OA/A values were the highest in microaggregates (2.10),intermediate in macroaggregates(1.87),and the lowest in silt/clay fractions (1.69) among the four aggregates.The AI values were significantly lower (14.1–17.7%) in silt/clay fractions than in other larger aggregates.The RI values decreased in the order:macroaggregates (0.93–0.96)>silt/clay fractions(0.89)>microaggregates (0.85).

3.4.Extracellular enzyme activities within soil aggregates

All of the measured extracellular enzyme activities(BG and PHOs),as well as one enzyme-based index(BP),differed strongly among the different fertilization patterns (Table 5).Compared with the non-ORsamended treatment,the ORs-amended treatments increased the hydrolase (BG) activity within all of the aggregates (50CN/50SN>50CN/25MN/25SN>50C N/50MN>100CN).Oxidase (PHOs) activity in each aggregate was the highest in 50CN/25MN/25SN and decreased in the order:50CN/25MN/25SN>50CN/50SN and 50CN/50MN>100CN.Moreover,the ORsamended treatments increased the values of BP within large macroaggregates,small macroaggregates,microaggregates,and silt/clay fractions by 82.5–235.1,7.1–168.7,31.7–242.2,and 27.9–111.9%,respectively,compared with the non-ORs-amended treatment.

Within the four aggregates,the highest hydrolase(BG) activity across treatments was observed in microaggregates (498.2 nmol h–1g–1soil),which was significantly higher (33.9−118.3%) than those in other aggregates.The oxidase (PHOs) activity showed a decreasing trend with increasing aggregate size,with silt/clay fractions exhibiting 8.0−33.1% higher oxidative activity across treatments than other aggregates.In addition,the BP values were significantly higher (60.5–119.3%) in microaggregates than in other aggregates.

3.5.Thermogravimetric fractions of organic C within soil aggregates

The amounts of ExoTwithin all the aggregates,which were positively correlated with organic C content(Appendix B),were significantly (P<0.05) higher in the ORs-amended soils than in the non-ORsamended soil (Table 6).The proportions of Exo1in each aggregate decreased in the order of 50CN/50SN>50CN/25MN/25SN>50CN/50MN>100CN,while the proportions of Exo2and Exo3showed the opposite trend (100CN>50CN/50MN>50CN/25MN/25SN>50CN/50SN).In addition,the ORs-amended treatments significantly increased the thermogravimetric indices of organic C (TSI and TG50) within all the four aggregates compared with the non-ORs-amended treatment.

Among the four aggregates,microaggregates exhibited the highest proportions of Exo1across treatments(54.6%) and the lowest proportions of Exo2and Exo3(29.4 and 16.1%,respectively),whereas silt/clay fractionsshowed the lowest proportions of Exo1(50.0%) and the highest proportions of Exo2and Exo3(31.4 and 18.7%,respectively).TSI and TG50,the thermogravimetric indices of organic C in aggregates,varied from 0.84 to 1.01 and from 368.4 to 380.4°C for the GVP system,respectively(Table 6).The changes in the thermogravimetric indices (TSI and TG50) showed similar trends with the proportions of Exo2and Exo3among aggregates:silt/clay fractions>macroaggregates>microaggregates.

Table 4 Mean values of relative proportions of different C functional groups and spectroscopic indices of organic C stability within different aggregates under different fertilization patterns

3.6.Linkages among the chemical,spectroscopic,biological,and thermogravimetric indices of organic C stability within aggregates

Pearson’s correlation (Table 7) revealed that the spectroscopic (OA/A and RI),biological (BP),and thermogravimetric (TG50and TSI) indices were all positively or negatively associated with each other(P<0.01).Similarly,the AI values were significantlycorrelated with LI (P<0.01),BP (P<0.05),TSI (P<0.05),and TG50(P<0.05).However,the spectroscopic,biological,and thermogravimetric indices,including OA/A,BP,TSI,and TG50,were not significantly correlated(P>0.05) with the LI values.More interestingly,the proportions of thermally stable (or labile) fractions (P<0.01)and Cp(or Ca) (P<0.05) were positively correlated with the proportions of resistant (or labile) C functional groups,but no significant correlations were found between the proportions of thermally stable (or labile) fractions and Cp(or Ca) (Appendix C).Together,these results suggest that the spectroscopic,thermogravimetric,and biological properties of organic C within aggregates were closely associated with each other.

Table 5 Mean values of enzyme activities within different aggregates under different fertilization patterns

4.Discussion

4.1.lncreasing organic C stability in the non-ORsamended soil relative to the ORs-amended soils at the aggregate scale

Substantial evidence obtained in this study (as shown by the higher LI,BP,and OA/A,and lower RI,TSI,and TG50values) indicated that ORs application reduced the chemical,biological,spectroscopic,and thermogravimetric stability of organic C in soil aggregates relative to chemical fertilization alone,which were similar to the findings of a previous study on SOC stability in bulk soils (Luanet al.2019).These results could be explained by two possible reasons.First,large amounts of chemical N inputs,as well as high temperature and humidity in GVP fields,promoted the decomposition of organic C (especially labile C fractions) in the non-ORsamended soil (Bootet al.2016).Second,the manure and straw used in the ORs-amended treatments can provide sufficient labile C resources in the soils (Luanet al.2020).Therefore,the higher labile C inputs related to ORs,as well as the consumption of labile C fractions in the non-ORsamended soil,resulted in lower organic C stability in the ORs-amended soils than in the non-ORs-amended soil.

More interestingly,the trends of organic C stability in silt/clay fractions induced by ORs application were weaker than those in the other aggregates (Appendix D).Guanet al.(2015) and Wang Yet al.(2017) revealed that ORs-derived labile C is preferentially preserved in macroaggregates and difficult to reach in small-size aggregates (i.e.,silt/clay fractions) unless over long-term microbial metabolism.Moreover,silt/clay fractions,which are physically protected by their mineral surface as well as iron and aluminum oxides,are the most stable fractions and their characteristics are hardly affected by fertilization(Chunget al.2008).These findings could explain the present results (Appendix D) that ORs application had different effects on organic C stability among different aggregates (i.e.,weak effects on silt/clay fractions and strong effects on other large aggregate fractions).

4.2.Variation in organic C stability among aggregates:silt/clay fractions>macroaggregates>microaggregates

As proposed by previous studies,organic C is unevenlydistributed among different aggregates and may be variably sensitive to soil environmental changes (Chenget al.2011;Liuet al.2019).Sixet al.(2004) indicated that microaggregates are formed by primary particles(i.e.,silt and clay particles) and stable binding agents(e.g.,humified organic C).These microaggregates are further combined by labile organic binding agents(e.g.,microbial-and plant-derived polysaccharides) to form macroaggregates,which subsequently results in the accumulation of organic C in the larger aggregates(Ayoubiet al.2012).However,in this study,the content of organic C was the highest in microaggregates among all the aggregates (Table 3).Another study found that microaggregates are physically more stable than macroaggregates and preserve greater quantities of organic C (Shahbazet al.2016).Smithet al.(2014)also revealed that microaggregates may be more effective in stabilizing C,which could be attributed to their micropore-and nanopore-dominated structure that is inaccessible to microbes or enzymes.These findings help explain the present results (Table 3) and suggest that microaggregates can be considered as a suitable microhabitat for the preservation of organic C.

Table 6 Mean values of ExoT,Exo1,Exo2,Exo3,and thermal indices of soil organic C stability within different aggregates under different fertilization patterns

Table 7 Pearson’s correlation coefficients (r) among the chemical,spectroscopic,biological,and thermal stability indices ofsoil organic C (n=48)1)

Given that ORs inputs,presumably enriched in labile C resources,should preferentially be incorporated into large-size aggregates (e.g.,macroaggregates) (Liet al.2016),it was expected to observe higher proportions of new (labile) organic C and higher organic C lability (i.e.,lower organic C stability) in macroaggregates,and higher proportions of old (stable) organic C and lower organic C lability in smaller aggregates (<0.053 mm).In the present study,organic C stability (as measured by the lower BP and OA/A values and higher RI,TSI,and TG50values)was the highest in silt/clay fractions among the four aggregates (Tables 4–6).This result was consistent with expectations and the prior findings of Xuet al.(2020),who reported that organic C mainly exists in silt/clay fractions as old and recalcitrant C forms.The possible explanation for these findings is that the C resources are limited in silt/clay fractions (Table 3),which leads to the consumption of labile C by microbes and allows recalcitrant C to be stored in silt/clay fractions (Smithet al.2014).Therefore,organic C in silt/clay fractions would be more stable than that in other aggregates.

Moreover,this sutdy found that organic C in microaggregates,rather than in macroaggregates,was the most labile within the aggregates (Tables 4–6),which was contradictory to the expectations of this study and the prior finding reported by Gaoet al.(2015) that organic C stability declined with increasing aggregate size.Penget al.(2017) suggested that new (labile)C resources (e.g.,exogenous ORs) preferentially accumulated in macroaggregates,but these C resources in macroaggregates persisted for a shorter time than those in microaggregates.Other studies have reported that water and oxygen conditions,microbial characteristics,and physical protection are the major factors affecting the variability in organic C decomposition in various aggregates (Shahbazet al.2016;Chenet al.2019).Microaggregates are characterized by a low porosity and pore connectivity (Rabbiet al.2016),low O2availability and diffusion rate,and relatively stable water potential (Konget al.2011),which is beneficial for creating a microenvironment where microbes cannot access substrates (e.g.,labile C fractions),and in turn,preserve the labile C fractions and enhance the organic C ‘inherent’ lability (i.e.,reduce organic C stability) within microaggregates.Taken together,the stability of organic C in microaggregates was the lowest,while that in silt/clay fractions was the highest among the four soil aggregates.

4.3.Unsuitability of the modified Walkley-Black method for evaluating organic C stability at the aggregate scale under these study conditions

The spectroscopic (RI and OA/A),biological (BP),and thermogravimetric indices (TSI and TG50) seem to be more useful and accurate than the chemical indices(LI),because these indices are more closely associated with one or more mechanisms of organic C stabilization(Mastrolonardoet al.2014) and were significantly (P<0.05)correlated with each other (Table 7).Moreover,the strong relationships among these indices were supported by the results from linear regression (thermally stable (or labile) fractions (%)vs.resistant (or labile) C functional groups (%);R2=0.48＊＊) (Appendix C).These findings demonstrated that organic C thermogravimetric properties are closely related to their spectroscopic characteristics at the aggregate scale.This was consistent with the findings of Merinoet al.(2014) and Peltreet al.(2017),who demonstrated that major classes of organic ingredients in SOC characterized by thermogravimetric analysis technology were intimately associated with chemical structures (i.e.,C functional groups) identified by13C NMR analysis.In comparison with13C NMR analysis,thermogravimetric analysis techniques are more economical,more convenient,and faster (Fernándezet al.2011).More importantly,thermogravimetric analysis can adequately characterize the complete quality continuum that organic C comprises and integrate multiple organic C stability mechanisms.Specifically,it can integrate organic C’s inherent biochemical“resistance”(Mechanism 3)and chemical absorption between SOC and minerals(Mechanism 2) (Mastrolonardoet al.2014).Therefore,thermogravimetric analysis can be recommended as an important analytical tool for predicting organic C quantity(Appendix B;ExoTwas closely associated with organic C contents,R2=0.74＊＊) and stability at the aggregate scale.

Nandanet al.(2019) found that the modified Walkley-Black method could divide organic C into different fractions based on their oxidizable labilities and then be used to measure organic C stability (as indicated by LI values) to some extent.However,the present study found no significant correlations between chemical (LI) and other indices (except for AI and RI) (Table 7),as well as different results on organic C stability among soil aggregates (chemical stability:microaggregates>macroaggregates>silt/clay fractionsvs.thermogravimetric and biological stability:silt/clay fractions>macroaggregates>microaggregates).These inconsistent results could be attributed to differences in the characteristics of soil aggregates and the means used for determining organic C stability (Luanet al.2019;Zhanget al.2019).Hanet al.(2016) reported that silt/clay fractions,the most stable fractions among aggregates,are characterized by plenty of minerals (e.g.,Fe-hydroxides and other protective mineral phases) that provide abundant reactive surface sites by which organic C can be strongly adsorbed.Ussiriet al.(2014) noted that the presence of various mineral phases (e.g.,Fe-and Aloxides) in soils can reduce the efficiency of old (stable) C removal from soils by chemical oxidants (e.g.,NaOCl and K2Cr2O7).These findings suggest that minerals in silt/clay fractions can reduce the old (stable) organic C extraction efficiency by chemical reagents (e.g.,K2Cr2O7),which was proved by the lower Cpcontents in silt/clay fractions than in other aggregates (Table 3).However,other studies have found that soil minerals have only minimal interference with the application (e.g.,measurement of organic C stability) of thermogravimetric and13C NMR analysis (Planteet al.2009;Ussiriet al.2014).Taken together,this study suggests that the modified Walkley-Black method is not suitable for characterizing organic C stability at the aggregate scale under these study conditions.

5.Conclusion

The integration of multiple methods allowed the evaluation of organic C stability within aggregates at a level previously not achieved.This study showed that eight years of organic amendments improved the soil structure in a GVP system by inceasing the proportions of large macroaggregates and enhancing soil aggregate stability (MWD).Relative to the non-ORs-amended treatment,organic amendments strongly increased the organic C content and reduced its stability(as indicated by spectroscopic,thermogravimetric,and biological stability indices) within aggregates,mainly due to the increase of labile C fractions.The content of organic C in microaggregates was significantly higher than that in other aggregates,while the stability of organic C followed the opposite trend (silt/clay fractions>macroaggregates>microaggregates).Moreover,thermogravimetric analysis can be considered an important means for evaluating organic C quantity and stability within soil aggregates.However,the modified Walkley-Black method is not suitable for characterizing organic C stability within soil aggregates,especially in silt/clay fractions.This study lays the foundation for future research exploring the mechanisms of organic C stabilization at the aggregate scale.

Acknowledgements

The authors sincerely acknowledge the financial support provided by the China Agriculture Research System of MOF and MARA (CARS-23-B02),the National Key Research and Development Program of China(2016YFD0201001),and the scientific research projects for talents introduce in Hebei Agricultural University(YJ2020054).Moreover,we thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

Declaration of competing interest

The authors declare that they have no conflict of interest.

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