Effects of long-term straw return on soil organic carbon fractions and enzyme activities in a double-cropped rice paddy in South China

Key Laboratory of Crop Physiology,Ecology and Genetic Breeding,Ministry of Education/Collaborative Innovation Center for the Modernization Production of Double Cropping Rice/College of Agronomy,Jiangxi Agricultural University,Nanchang 330045,P.R.China

Abstract Long-term straw return is an important carbon source for improving soil organic carbon (SOC) stocks in croplands,and straw removal through burning is also a common practice in open fields in South China. However,the specific effects of long-term rice straw management on SOC fractions,the related enzyme activities and their relationships,and whether these effects differ between crop growing seasons remain unknown. Three treatments with equal nitrogen,phosphorus,and potassium nutrient inputs,including straw/ash and chemical nutrients,were established to compare the effects of straw removal (CK),straw return (SR),and straw burned return (SBR). Compared to CK,long-term SR tended to improve the yield of early season rice (P=0.057),and significantly increased total organic carbon (TOC) and microbial biomass carbon (MBC) in double-cropped rice paddies. While SBR had no effect on TOC,it decreased light fraction organic carbon (LFOC) in early rice and easily oxidizable organic carbon (EOC) in late rice,significantly increased dissolved organic carbon (DOC),and significantly decreased soil pH. These results showed that MBC was the most sensitive indicator for assessing changes of SOC in the double-cropped rice system due to long-term straw return. In addition,the different effects on SOC fraction sizes between SR and SBR were attributed to the divergent trends in most of the soil enzyme activities in the early and late rice that mainly altered DOC,while DOC was positively affected by β-xylosidase in both early and late rice. We concluded that straw return was superior to straw burned return for improving SOC fractions,but the negative effects on soil enzyme activities in late rice require further research.

Keywords:double-cropped rice paddy system,straw return,straw burned return,SOC fractions,soil enzyme activities

1.lntroduction

Soil organic carbon (SOC) is an important determinant of the physical,chemical,and biological properties of soil (Gonget al.2009). Soil fertility is affected by many factors,including crop residue return,fertilizer application,and tillage management (Yanet al.2013; Zhuet al.2015).However,while SOC is not sensitive to short-term changes in straw return (Zhaoet al.2016),some labile organic C fractions,including light fraction organic C (LFOC),easily oxidizable organic C (EOC),dissolved organic C (DOC),particulate organic C (POC),and microbial biomass C(MBC),are sensitive to and respond rapidly to changes in soil management. Xuet al.(2011) found that total organic C (TOC) did not change after 2 years of straw return,but MBC,POC and EOC were significantly increased. Wang Wet al.(2015) reported that straw return significantly increased MBC,DOC and LFOC in a double-cropped rice paddy after 3 years of straw return. Therefore,these SOC fractions are commonly used as efficient indicators for identifying optimized agricultural practices that increase the stock and quality of soil C (Murata and Goh 1997; Malhiet al.2011;Plaza-Bonillaet al.2014; Benbiet al.2015).

Soil biological components are another element that responds rapidly to changing soil conditions (Powlson and Jenkinson 1981; Anderson and Gray 1990). Soil enzymes synthesized and secreted by soil microorganisms are responsible for the formation and decomposition of organic matter. Salazaret al.(2011) reported that an enzyme catalyzed biochemical reaction can be influenced by its specific substrate. For example,amylase and cellulase activity can be influenced by the addition of organic matter to the soil (Pancholy and Rice 1973); β-glucosidase is involved in soil metabolic and carbon cycle activity (Stottet al.2010); β-xylosidase activity increases after adding peat (Vepsäläinenet al.2004); and Geet al.(2010) found a significant correlation between SOC and sucrase activity.Soil enzyme activity has been used as a primary early indicator of soil quality (Ceccantiet al.1993; Badianeet al.2001). Thus,studying the effects of straw return on the formation of SOC fractions and C-cycle enzyme activities is necessary,and understanding their relationship can reveal the underlying mechanisms of SOC transformation under long-term straw return.

China is rich in crop straw resources,with 598 million tons of food crop straw produced in 2017 (Liet al.2017).Straw return is widely considered as an effective method for increasing SOC and reducing chemical fertilizer application(Lehtinenet al.2014; Liuet al.2014). Zhanget al.(2017)found that the SOC of rice paddy fields increased 34-56%under the practice of straw return compared with fertilizer application alone. Liuet al.(2014) also observed that SOC and soil active C fractions significantly increased with straw return. In addition,changes in meteorological conditions,such as temperature (Kirschbaum 1995) and rainfall (Tianet al.2013),between the early and late rice cropping seasons can also affect SOC sequestration,because soil moisture and temperature affect the decomposition and transformation of organic carbon by microorganisms(Schimelet al.1999; Davidsonet al.2000). The effect of straw return on crop yield also depends on meteorological conditions (Pittelkowet al.2015). Straw return has been shown to increase rice yield in subtropical regions(Huanget al.2013),because warm conditions stimulated the decomposition of straw and the release of nutrients(Yadvinderet al.2005). In contrast,Blanco-Canqui and Lal (2009) found that straw return mulching reduced crop yields in humid temperate regions,since the cold and wet soil conditions were prolonged by straw return in the early growing season. In a study focusing on China,Wang Jet al.(2015) also observed a slightly negative effect on crop yield by straw return. However,in South China,rice agriculture is dominated by the double-cropped rice systems. In these systems,the meteorological conditions of early and late season are quite different,but research on the effects of seasonal meteorological changes on crop yield and SOC fractions in double-cropped rice paddy system is lacking.

The double-cropped rice system in South China is an important cropping system,accounting for 45% of the total rice planting area and 40% of total rice yields (Xueet al.2015). As rice production has increased over the years,so has the amount of available straw. Moreover,open burning of crop straw produces harmful air pollutants and is therefore banned in China. However,this practice has become more common in recent decades,especially during the period after harvesting of the early season rice crops in South China. Open burning of straw is also a convenient and efficient way for farmers to remove crop straw (Zhanget al.2014). Previous studies have shown that long-term burning of rice stubble in paddy fields does not affect total SOC (Powlsonet al.1987; Carter and Mele 1992; Hoyleet al.2006). However,few studies have investigated the comparative effects of long-term SR on SOC fractions and soil enzyme activities compared with SBR in the doublecropped rice system,especially regarding the mechanism of the SOC fraction changes and their differences in earlyseason and late-season rice paddy. Hence,the aims of the present study were to:1) investigate the effects of longterm rice straw return on SOC fractions and soil enzymes activities; 2) determine the sensitivity of SOC fractions under long-term straw return in double-cropped rice fields; and 3) distinguish the seasonal differences in SOC fractions between the early-season and late-season rice paddy fields in the middle and lower Yangtze River Basin,South China.We hypothesized that long-term straw return would increase the SOC fractions and affect the soil enzymes activities,and that there would be seasonal differences in double-cropped rice paddy. This study provides insight into the effect of straw return on the formation of SOC fractions,and may inform straw return management practices in double-cropped rice systems in South China.

2.Materials and methods

2.1.Experimental site

The experiments were conducted in a rice paddy field at the National Soil Fertility Monitoring (NSFM) site,located in Yangxi Village,Wenzhen Town,Jinxian County,Jiangxi Province,China (28°20´7.14´´N,116°5´29.73´´E).There are two main seasons for rice cultivation in Jiangxi Province,namely early rice (March to July) and late rice (June to October). The site is located in a typical subtropical monsoon humid climate,with an annual average temperature of 17.5°C and annual precipitation that ranges from 1 600-1 800 mm. The field site contains quaternary red clay developmental paddy soil. The fields were subjected to a series of long-term straw return treatments from 2010 to 2017. Prior to this treatment,the soil contained SOC (19.8 g kg-1),total nitrogen (N) (2.3 g kg-1),available N (126.0 mg kg-1),available phosphorus(P2O5) (31.4 mg kg-1),available potassium (K2O) (97.9 mg kg-1),and pH (1:5,soil:water) 5.5.

2.2.Experimental design

Three straw return treatments were established with an area for each treatment of 326.7 m2(19.70 m×16.58 m).The three treatments included straw removal (CK),straw return (SR),and straw burned return (SBR) with three replications for each treatment. The same amounts of nitrogen,phosphorus,and potassium nutrients were applied to all treatments. The nutrient content of rice straw and straw ash was determined by direct measurement(Appendix A) and any deficient nutrient was supplemented with chemical fertilizers which contained N,P2O5,and K2O. The applications of N,P2O5,and K2O for early rice were 165,75 and 150 kg ha-1; and for late rice were 195,87.75 and 175.5 kg ha-1,respectively. Chemical N fertilizer was applied as base fertilizer,tillering fertilizer,and panicle fertilizer at rates of 5:2:3 for early rice and 4:2:4 for late rice,respectively. Chemical K2O fertilizer was applied as tillering fertilizer and panicle fertilizer at a rate of 7:3 for both early and late rice. All chemical P2O5fertilizer was applied as basal fertilizer. The sources of chemical nitrogen,phosphorus,and potassium fertilizers were urea,calcium magnesium phosphate,and potassium chloride,respectively. Rice varieties Luliangyou996 and WufengyouT025 were selected and planted in the early-and late-season rice paddy in 2010-2016,respectively. In 2017,Wufengyou 286 and HYou 518 were selected and planted in the early-and late-season rice paddy,respectively. The rice paddy fields were rotary tilled by rotary cultivator machines and rice seedlings were transplanted manually. Seedlings were transplanted with row spacing of 13.33 cm×23.31 cm for early-season rice and 13.33 cm×26.64 cm for late-season rice. Minced (by harvester) or ashed rice strawin situwas immediately returned to the paddy field after harvesting the early and late rice crops on July 23-27 and October 28-31 of each year from 2010 to 2017,respectively; and then the straw/ash material was plowed into the soil with rotary tillage. The estimated amounts of straw/ash are shown in Appendix B,assuming a ratio of yield to rice straw of 1:1 and a combustion factor of straw of 0.8572,according to the burning habits of local farmers. Other management measures were carried out according to typical farming practices in the region.

The data of precipitation and temperature in 2017 were provided by the Meteorological Bureau of Jinxian County in Jiangxi Province,China.

2.3.Grain yield measurement

Grain yield was determined from a 7 m2sampling area in each treatment,with three replicates,before harvest in early-season and late-season rice,and it was then adjusted to 13.5% safe moisture content for weighing.

2.4.Soil sampling and analysis

Soil samples from the early and late rice paddies were collected after the rice harvests on July 14 and October 25,2017,respectively. Five soil cores were taken from each plot (0-20 cm in depth and 3 cm in diameter) with three replicates for each treatment. The fresh soil samples were homogenized,sieved through a 2-mm mesh and divided into two subsamples. One of the fresh subsamples was stored at 4°C and the other was air-dried for the analysis of the SOC fractions. Soil pH was determined by a pH meter in a 1:5 (w/w) soil to water mixture.

SOC fraction analysisTOC was measured using a standard potassium dichromate digestion method (Baiet al.2005).

EOC was measured according to the 333 mmol L-1KMnO4digestion method with a slight modification (Blairet al.1995). Air-dried soil (containing 15 mg of C) was shaken with 20 mL of 333 mmol L-1potassium permanganate for 1 h at 200 r min-1and then centrifuged for 10 min at 2 000×g. The supernatant solution was diluted by 1:250 with distilled water and the absorbance of the diluted solution was measured at 565 nm.

MBC was measured using the chloroform fumigationextraction method adapted from Lu (1999). Soil samples were extracted with 0.5 mol L-1potassium sulfate for 30 min at 300 r min-1and MBC was measured using a TOC analyzer.

DOC was determined by extracting fresh soils with distilled water (1:5 ratio) for 1 h at 250 r min-1,centrifuging for 10 min at 4000×g,and measuring the C concentration with a TOC analyzer (Jianget al.2006).

The light fraction (LF) and LFOC were determined according to the NaI separation method (Lu 1999). Briefly,a 1:3 ratio of air-dried soil and sodium iodide (density 1.8 g cm-3) was placed in a 100-mL centrifuge tube,shaken for 60 min at 200 r min-1,and centrifuged at 4 000×g for 10 min.The floating LF was collected with a vacuum unit with a 0.45-μm filter,and then washed three times with 0.01 mol L-1calcium chloride and distilled water. The floating LF was dried at 60°C for 48 h and then weighed. The LFOC content was then measured using the same method as for TOC.

POC was determined using the method of Cambardella and Elliott (1992) with some modifications. A 10 g air-dried soil sample was placed in a 50-mL centrifuge tube,30 mL of sodium hexametaphosphate (5 g L-1) was added,and the mixture was oscillated at 200 r min-1for 12 h. The mixture was passed through a 53-μm sieve and washed until the water in the sieve was colorless. The sieve was collected,dried at 60°C for 48 h,the particulate fraction (PF) was weighed,and the POC content was determined using the same method as for TOC.

Soil enzyme activity determinationSoil enzyme activities were determined using an analysis kit (Suzhou Branch Ming Biotechnology Co.,Ltd.,China). Sucrase and amylase activity were determined using 3,5-dinitrosalicylic acid colorimetry. Cellulase activity was determined using the anthrone colorimetric method. The activity of β-glucosidase was assayed using thep-nitrobenzene-β-D glucopyranoside colorimetry method. The activity of β-xylosidase was assayed using thep-nitrophenol-β-D-xylose colorimetry method.

2.5.Statistical analysis

The data were analyzed with ANOVA using the general linear model procedure in the SPSS 24.0 (SPSS Inc.,Chicago,IL,USA). The mean values for the treatments were compared using the least significant difference (LSD) test at a significance level ofP＜0.05. Pearson’s correlation analysis was used to determine the relationships between SOC fractions and soil enzyme activities. RDA was performed to determine the relationships between soil enzyme activities and SOC fractions using CANOCO 5.0 (Microcomputer Power Inc.,Ithaca,NY,USA).

3.Results

3.1.Effects of straw return on grain yield in a doublecropped rice system

Averaged across eight years,the grain yield in the doublecropped rice system showed no significant differences in either SR or SBR compared with CK,although treatments with either SR or SBR significantly increased the grain yields(P＜0.05) (Fig.1) in certain years. SR tended to improve the yields of early season rice (P=0.057),but not the yields of late season rice.

3.2.Differences in precipitation and average temperature in early-and late-season

Fig.1 Effects of straw return modes on rice yield in a double-cropped rice paddy (2010-2017). CK,SBR,and SR are straw removal,straw burned return,and straw return,respectively. Different letters indicate significant differences between treatments(P＜0.05). Error bars indicate SD (n=3).

In 2017,the average temperature in late rice was higher than that in early rice by 1.1°C (Fig.2). The average temperature during the fallow period between early rice and late rice was 31.85°C. In addition,46 and 29 rainfall events were observed in the early rice and late rice,with the average precipitation levels of 14.77 and 10.81 mm,respectively.

3.3.Effects of straw return on SOC fractions in the double-cropped rice system

Compared to the CK treatment,SR significantly increased soil TOC by 5.6-7.4% and MBC by 14.1-17.3% in both early and late rice (Table 1),whereas SBR did not significantly affect TOC or MBC. SR significantly reduced DOC by 22.4%in early rice and by 5.4% in late rice,whereas SBR increased DOC. SBR significantly reduced EOC in late rice.

In the double-cropped rice paddy fields,SR increased LF and LFOC compared to either SBR or CK (Table 2),with significant differences in LF for CK and in LFOC for SBR. There were no significant differences in PF among the treatments. The POC was only significantly decreased by CK in the late rice.

In the early rice,DOC was negatively and significantly correlated with TOC,LFOC,and MBC (Table 3). In the late rice,EOC was negatively correlated with DOC,but LFOC was positively correlated with TOC and MBC.

3.4.Effects of straw return on soil enzyme activities in the double-cropped rice system

The activities of sucrase,amylase,cellulase,and β-glucosidase were significantly increased by 18.6,343.7,61.8,and 20.0%,respectively,by SR compared to CK in the early rice season,whereas the opposite trends were observed in the late rice (Table 4). Amylase and cellulase activities were significantly increased in the SBR early rice compared to CK,but were significantly decreased in the late rice. In both the early and late rice seasons,β-xylosidase was significantly decreased by SR,but it was increased by SBR. Moreover,for the five soil enzyme activities,the treatment effects for SR and SBR were not consistent between early and late rice.

Fig.2 Average daily temperature and precipitation from transplanting to harvesting of double-cropped rice (2017).The data are provided by the Meteorological Bureau of Jinxian County in Jiangxi Province,China.

Table 1 Effects of straw return modes on TOC,EOC,DOC,and MBC in a double-cropped rice paddy1)

Table 2 Effects of straw return modes on soil LF,LFOC,soil PF,and POC in a double-cropped rice paddy1)

3.5.Relationships between SOC fractions and soil enzyme activities

In the early rice,TOC was positively correlated with sucrase and β-glucosidase activity but negatively correlated with β-xylosidase,whereas the opposite trends were seen for DOC (Table 5). There were also positive relationships between LFOC and β-glucosidase activity,and between MBC and both sucrase and cellulase activities. In the late rice,TOC,LFOC,and MBC were all negatively correlated with sucrase activity. There were also negative relationships between POC and amylase,cellulase,and β-glucosidase activities,and between DOC and both β-xylosidase and amylase activities. EOC was positively correlated with amylase and cellulase activities. Overall,DOC was significantly correlated with most enzymes in both early and late rice,suggesting that the DOC gave the most sensitive response to the changes of soil enzyme activities in the double-cropped rice paddies.

The coordinate from the first two ordination axes of the overall variances each explained 99.5 and 0.03% ofthe variance in the early rice,and 82.2 and 0.23% of the variance in the late rice,respectively (Fig.3). The sucrase and β-glucosidase had the longest projections,suggesting the greatest effects on SOC fractions on the first axis in the early rice. However,in the late rice,β-xylosidase had the greatest effect on SOC fractions on the first axis. In both early and late rice,the angle between β-xylosidase and DOC was the smallest,which means that DOC is mainly influenced by β-xylosidase.

Table 3 Pearson’s correlation coefficients among different soil organic carbon fractions in a double-cropped rice paddy1)

Table 4 Effects of straw return modes on soil enzyme activities in a double-cropped rice paddy

Table 5 Pearson’s correlation analysis between soil enzyme activities and soil organic carbon fractions1)

3.6.Effects of straw return on soil pH in the doublecropped rice system

All of the treatments lowered soil pH by 7.6-9.6% from the initial value of 5.5 in both the early and late rice. Among the treatments,SBR resulted in the lowest soil pH with a significant difference,followed by SR (Fig.4).

4.Discussion

Many studies have reported that straw return increased crop yields (Liuet al.2014; Wang Jet al.2015). Our study showed similar results during eight-year field experiments where the same nutrient inputs were applied among the treatments,especially for early rice (Fig.1),although the effects did not reach a significant level (P=0.057). This might have been due to the increases in SOC and the improvements in soil bio-physical properties (Blanco-Canqui and Lal 2009). In the present study,the trends of improving early rice yields indicated that the changes of SOC and soil bio-physical properties showed seasonal differences,because straw derived from the early rice harvest had less time to decompose for the late rice in a given year (Liaoet al.2018).

Fig.3 Redundancy analysis (RDA) of soil enzyme activities and soil organic carbon fractions. TOC,total organic carbon; LFOC,light fraction organic carbon; EOC,easily oxidizable organic carbon; POC,particulate organic carbon; DOC,dissolved organic carbon; MBC,microbial biomass carbon.

In this study,SR significantly increased TOC by 7.4% in early rice and by 5.6% in late rice,compared to CK (Table 1).This level was similar to the results of previous studies that found straw return improved SOC sequestration (Yanet al.2013; Zhuet al.2015; Zhaoet al.2016). However,in other studies straw burning was found to have no effect on soil organic carbon compared with straw removal (Rietlet al.2012). In this study,the same trend was also observed,because most C input from straw in the SBR plots escaped as CO2during straw burning (Powlsonet al.1987). In the field experiment,soil C inputs are higher in SR than in SBR.As a result,SR had relatively high TOC,especially for the early rice,suggesting that straw burned return could not improve SOC sequestration,and seemed to simply waste biomass resources (Liet al.2017).

Fig.4 Effects of straw return modes on soil pH in a doublecropped rice paddy. CK,SBR,and SR are straw removal,straw burned return,and straw return,respectively. Different letters indicate significant differences between the treatments(P＜0.05). Error bars indicate SD (n=3).

SOC fractions are sensitive to changes in C supply(Haynes 2000). In this study,SR significantly increased the MBC concentration in double-cropped paddy fields.This increase might be due to the decomposition of straw leading to the release of nutrients such as dissolved organic matter (DOM) and stimulating microbial growth. DOM,which contains DOC,is a great source of carbon and can supply other nutrients for microorganisms (Chenet al.2010).Meanwhile,DOC is an indicator of potential organic matter decomposition (Xuet al.2011). Long-term straw return can significantly increase the DOC concentration (Xuet al.2011; Benbiet al.2015; Wang Wet al.2015),although our study showed the opposite trends for SR (Table 1).There are several possible explanations for the decrease in DOC with straw return in our study. Straw return can stimulate soil microbial growth whereas the DOM has higher biodegradability,resulting in the DOM being easily consumed by microorganisms as their preferred substrate(Chenet al.2010). However,SR inputs a large amount of C but does not input more N compared with other treatments,so the microbial decomposers may adapt to the high-C(and relatively low-N) conditions by reducing their microbial carbon use efficiency (Manzoniet al.2008,2010). This would cause the microorganisms to preferentially consume a large amount of DOM with a lower microbial carbon use efficiency (Kalbitzet al.2010; Mooshammeret al.2014). In addition,the effects of environmental changes on the soil should be also considered,such as precipitation leaching during the winter fallow period in South China. Because DOC is also affected by hydrology (Neff and Asner 2001),the DOC leaching could be increased by straw return (Hua and Zhu 2018). Moreover,the early rice season had more rainfall than the late rice season (Fig.2),which may result in increased DOC leaching. Furthermore,the negative relationships between DOC and TOC,LFOC and MBC in early rice indicated that DOC could not contribute to an improvement in SOC (Table 3). Our study showed longterm straw return did not change the EOC concentration relative to straw removal,which confirmed the results of Benbiet al.(2015). Straw burned return can decrease soil labile C pools (Muqaddaset al.2015). In our study,straw burned return also significantly reduced LFOC in early rice and EOC in late rice. These results further suggest that burning straw residue has negative effects on the stability of SOC fractions.

Straw return has been found to significantly increase soil POC and LFOC (Yanet al.2007; Nayaket al.2012; Liuet al.2013). In this study,we found similar results in the late rice (Table 2). Because LFOC and POC concentrations in soil mainly depend on a balance between straw input and decomposition (Malhiet al.2011; Zhuet al.2015),straw return provides higher carbon input than straw removal.However,SR did not increase either POC or LFOC in early rice. This is most likely due to the different ambient temperatures between the early and late rice seasons.Increasing temperature will promote quick rice straw decomposition (Tanget al.2016),whereas the average temperature for late rice was higher than that for early rice by 1.1°C (Fig.2). In addition,straw added after the end of the late season has much more time to decompose (Liaoet al.2018). SR had no effect on POC compared to SBR(Table 2). These results are in agreement with Chanet al.(2002). SBR did not affect LF in the late rice season,but significantly decreased LF in the early rice season. Soil LF is comprised of fresh or semi-decomposed plant residues(Wander 2004) and also contains charcoal,seeds,and animal debris (Gregorich and Janzen 1996). Therefore,we infer that straw ash accounts for only a certain proportion of LF in the late rice. However,because the soil continuously assimilates straw ash,especially after tillage,the LF in the early rice had been decreased by SBR during the previous year. Soil LF is a component of the PF,but soil PF is much higher than soil LF,so it is not easily changed (Wander 2004; Bartuškaet al.2015). Therefore,our results agree with previous studies in showing no significant difference in PF (Chenet al.2017).

Straw return increased microbial biomass and MBC,thereby supplying enough energy and creating a suitable environment for the production of soil enzymes (Jiaoet al.2011). Long-term straw return has been shown to significantly increase soil enzyme activities (Gianfredaet al.2005). Similarly,we found that SR significantly increased the activities of sucrase,amylase,cellulase,and β-glucosidase in the early rice compared with CK (Table 4). However,SBR significantly reduced the majority of soil enzyme activities in the late rice,in contrast to the early rice. This could be due to the high temperature of straw burning causing the incineration and volatilization of C compounds,thus weakening microbial activity,and resulting in reduced soil enzyme activities (Dicket al.1988). However,β-xylosidase activities increased with increasing temperature,showing maximal activity at 90°C (Maneliuset al.1994). So,SBR increased β-xylosidase in double-cropped rice paddies,and β-xylosidase was also positively correlated with DOC in double-cropped rice paddies (Table 5). We conclude that the increase in DOC concentration in SBR was caused by an increase in β-xylosidase,because this enzyme is involved in the process of DOC production,and the increase of β-xylosidase activity can further accelerate the degradation process of carbon substances such as cellulose,hemicellulose and lignin in soil (Waldropet al.1994).

Meanwhile,soil enzyme activities are also affected by abiotic conditions (e.g.,temperature and soil pH) and root activity (Biederbecket al.1994; Liuet al.2008). Straw shows a decomposition mode of “first fast and then slow”,and the rate of decomposition increases as the temperature increases (Pal and Broadbent 1975; Nakajimaet al.2016).In this study,the straw from early rice paddy fields that is returned to late rice paddy fields may be rapidly decomposed at an appropriate temperature to generate a large amount of organic acids (Kuwatsuka and Shindo 1973; Rao and Mikkelsen 1977). These acids could inhibit the growth of late rice seedlings,hinder the respiration of rice roots,and reduce the absorption of various nutrient elements(Tanaka 1967). Furthermore,the toxic effect of organic acids could accelerate the decrease in soil pH (Rao and Mikkelsen 1977; Xieet al.2018). Liuet al.(2010) reported that long-term straw return could decrease soil pH,and we observed similar results in our study (Fig.4). All soil enzyme activities in SR were decreased by 9.8-34.9%,and the negative relationships with SOC fractions increased in the late rice (Table 5). In other words,C input from straw in the early rice might reduce the soil enzyme activities related to the C-cycle in late rice. Therefore,this might be one of the reasons that the yield increasing effect of straw return was weakened in late rice. Overall,there is a need to strengthen paddy management strategies for improving the enzyme activities which are reduced in SR in late rice.However,the reduction of enzyme activities in SBR could not be improved,which were caused by the high temperature associated with burning straw.

The RDA showed differences in the enzymes affecting the changes of SOC fractions in early and late rice,which were mainly influenced by sucrase and β-glucosidase in the early rice,and β-xylosidase in the late rice (Fig.3). Soil enzymes catalyze the decomposition and synthesis of SOC(Liet al.2016),and each enzyme has specific substrates and the ability to catalyze specific biochemical reactions(Salazaret al.2011). Our results showed that DOC was mainly influenced by β-xylosidase in double-cropped rice paddies; and while sucrase and β-glucosidase mainly affected MBC in early rice,amylase mainly affected EOC in late rice. These observations lead to the conclusion that SOC fractions and microbial communities in response to straw return were different between the early and late rice.Whereas,the effect of the microbial community on the soil enzymes and SOC fractions in double-cropped rice paddy system remains to be further studied.

5.Conclusion

In this study,compared to CK,long-term SR tended to improve grain yield of early rice (P=0.057),and significantly increase TOC with the improvement of MBC in the doublecropped rice paddy. DOC was lower in SR than in CK in early rice,but POC and LFOC were higher in late rice,and EOC was similar between SR and CK in both early rice and late rice. These results indicated that MBC was the most sensitive parameter to long-term straw return in double-cropped rice paddy. However,SBR could have no effect on TOC due to the significantly decreased LFOC in early rice and EOC in late rice,in spite of the significantly increased DOC,suggesting that long-term straw burned return was detrimental to SOC sequestration. In addition,SR significantly increased most of the soil enzymes activities in the early rice,but negative impacts were found for SR and SBR in the late rice,due to the decreased pH. Therefore,we propose that straw return should be conducted,rather than straw burned return,based on the SOC fractions in the double-cropped rice paddy system in South China.Furthermore,our results highlight the potential negative impacts of straw return on SOC sequestration and yield in the late rice season.

Acknowledgements

Thanks to those who are at the forefront of the fight against the COVID-19,so that we can write this paper with peace of mind. This work was supported by the National Key Research and Development Program of China (2017YFD0301601),the China Postdoctoral Science Foundation (2016M600512),the Open Project Program of State Key Laboratory of Rice Biology,Ministry of Science and Technology,China (20190401),and the Jiangxi Province Postdoctoral Research Project Preferential Grant,China(2017KY16).

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