Bingshen JIANGJianlin SHEN∗Minghong SUNYajun HUWenqian JIANGJuan WANGYong LI and Jinshui WU
1Key Laboratory of Agro-ecological Processes in Subtropical Region and Changsha Research Station for Agricultural & Environmental Monitoring,Institute of Subtropical Agriculture,Chinese Academy of Sciences,Changsha 410125(China)
2University of Chinese Academy of Sciences,Beijing 100049(China)
3College of Biological and Food Engineering,Huanghuai University,Zhumadian 463000(China)
4College of Resources & Environment,Hunan Agricultural University,Changsha 410128(China)
(Received June 27,2019;revised September 29,2019)
ABSTRACT Agronomic practices affect soil phosphorus(P)availability,P uptake by plants,and subsequently the efficiency of P use.A field experiment was carried out to investigate the effects of various agronomic practices(straw incorporation,paddy water management,nitrogen(N)fertilizer dose,manure application,and biochar addition)on soil P availability(e.g.,soil total P(STP),soil available P(SAP),soil microbial biomass P(SMBP),and rice P uptake as well as P use efficiency(PUE))over four cropping seasons in a rice-rice cropping system,in subtropical central China.Compared to the non-straw treatment(control,using full dose of chemical N fertilizer),straw incorporation increased SAP and SMBP by 9.3%-18.5% and 15.5%-35.4%,respectively;substituting half the chemical N fertilizer dose with pig manure and the biochar application increased STP,SAP,and SMBP by 10.5%-48.3%,30.2%-236.0%,and 19.8%-72.4%,respectively,mainly owing to increased soil P and organic carbon inputs;adding a half dose of N and no N input(reduced N treatments)increased STP and SAP by 2.6%-7.5% and 19.8%-33.7%,respectively,due to decreased soil P outputs.Thus,soil P availability was greatly affected by soil P input and use.The continuous flooding water regime without straw addition significantly decreased SMBP by 11.4% compared to corresponding treatments under a mid-season drainage water regime.Total P uptake by rice grains and straws at the harvest stage increased under straw incorporation and under pig manure application,but decreased under the reduced N treatments and under biochar application at a rate of 48 t ha−1,compared to the control.Rice P uptake was significantly positively correlated with rice biomass,and both were positively correlated with N fertilizer application rates,SAP,SMBP,and STP.Phosphorus use efficiency generally increased under straw incorporation but decreased under the reduced N treatments and under the manure application(with excessive P input),compared to the control.These results showed that straw incorporation can be used to increase soil P availability and PUE while decreasing the use of chemical P fertilizers.When substituting chemical fertilizers with pig manure,excess P inputs should be avoided in order to reduce P accumulation in the soil as well as the environmental risks from non-point source pollution.
Key Words:biochar,manure fertilizer,nutrient management,P use efficiency,paddy soil,straw incorporation,water regime
Phosphorus(P)is an essential plant nutrient,along with nitrogen(N)and potassium(K),for crop growth(Schachtmanet al.,1998).Soil P deficiencies may directly affect crop metabolism,which in turn can inhibit growth and yield.However,excessive application of P fertilizers(e.g.,calcium superphosphate and calcium-magnesia)can cause P accumulation in the soil(Csathóand Radimszky,2009),and can lead to P losses from leaching of the surface water and subsequent eutrophication of water bodies(MacDonaldet al.,2011;Townsend and Porder,2012).Phosphorous is a non-renewable resource,and with continuously unsustainable consumption,it is predicted that P fertilizers will soon become limited and more expensive(Zhanget al.,2008).Only a small amount of soluble P exists in soils because P can form various minerals,particularly with Ca2+,Al3+,and Fe3+ions(Hinsinger,2001;Scalengheet al.,2014),and also because P has a high affinity with the surfaces of many soil constituents(Scalengheet al.,2014;Gérard,2016).Therefore,the in-season P use efficiency(PUE)is usually low,at approximately 10%-15%(Roberts and Johnston,2015).
In the recent years,improving PUE through different agronomic practices has become a priority as high PUE has been shown to confer agronomic,economic,and environmental benefits.Phosphorus use efficiency depends largely on soil P availability as well as on crop P uptake(Simpsonet al.,2011;Veneklaaset al.,2012).Agronomic practices can change soil physicochemical properties and biological characteristics,and thus,could also affect soil P availability and crop P uptake.For example,long-term balanced chemical fertilization(N,P,and K fertilizers used in a suitable ratio)has been confirmed to enhance soil P availability,rice P uptake,and yield(Zhanget al.,2009;Bhattacharyyaet al.,2015;Suet al.,2015).Straw incorporation in croplands has been described as a simple and effective way to increase soil nutrient utilization efficiency and subsequent crop yields(Guptaet al.,2007;Zhuet al.,2010).This is because straw incorporation can increase soil P availability by enhancing microbial biomass Pviathe cellulose component,and by competitive adsorption of Fe and Al oxides by lignin functional groups(Sunet al.,2019).In addition,straw incorporation can increase soil organic carbon(SOC)to improve microbial biomass,and thus,soil microbial biomass P(SMBP)may increase accordingly based on the principle of stoichiometric homeostasis(Mastoet al.,2013).Changes in soil pH and redox potentials(Eh)under certain water regimes(e.g.,continuously flooded and intermittently flooded)in paddy fields can also affect P availability as well as P adsorption and desorption characteristics in the soil(Krairapanondet al.,1993;Salequeet al.,1996;Senget al.,1999).Scalengheet al.(2014)found that under flooding conditions,calcium plays a major role in controlling P dynamics,even in acidic soils.In recent years,biochar addition has been recommended as a method of enhancing soil fertility as well as soil P availability.Biochar is produced through pyrolysis of organic materials under conditions of little or no oxygen(Lusibaet al.,2017).Biochar is porous with a large surface area,and adding it to soils has been approved as a measure to improve infiltration,aggregate stability,porosity,acid buffer capacity,and water and nutrient holding capacities(Lehmannet al.,2011).It has been reported that biochar enhances soil available P due to its own high available P content,and that biochar decreases the adsorption of P in the soil(Cuiet al.,2011;Zhanget al.,2016).
The rice-rice cropping system is an important cropping system in China.The system consists of two rice seasons per year and is used across approximately 10.5 million hectares in Southern China(Xin and Li,2009).In recent years,the intensive and excessive use of chemical P fertilizers has resulted in P accumulation in the soil,and increases in yield or crop P uptake due to P fertilizer application have decreased,resulting in a low PUE(Ramaekerset al.,2010;Xuet al.,2016).For example,with P fertilizer application rates as high as 43.7 kg P ha−1,the apparent PUE was reported to be only 3% to 6% in areas using the rice-rice cropping system(Jiet al.,2011).To reduce the negative impacts of excessive chemical fertilizer application,as well as to enhance soil fertility in paddy rice production systems,a number of agronomic practices(e.g.,straw incorporation,substitution of some chemical fertilizers with animal manure,reducing N fertilizer application,and biochar amendment)have been encouraged.Nevertheless,little is known about the responses of soil available P and rice P uptake to different agronomic practices in the rice-rice cropping system.A field experiment was conducted to examine the influences of various agronomic practices(including straw incorporation under different water regimes,different N fertilizer doses,chemical fertilizers combined with pig manure,and biochar addition)on soil P availability(based on soil total P,available P,and microbial biomass P),rice P uptake,and yield in a typical rice-rice cropping system with moderate soil fertility in Southern China.The purposes of the study were i)to examine the differences in soil P availability and rice P uptake under different agronomic practices and ii)to identify effective agronomic practices to improve soil P availability and PUE in the rice-rice cropping system.
The field experiment was carried out in a typical paddy field with rice-rice cropping system in Hunan Province,China(28°33′04′′N,113°19′52′′E).The study site has a typical subtropical climate,where the annual average temperature is 17.5°C and the annual average rainfall is approximate 1 500 mm.The paddy soil has a parent material of granite red soil,and is classified as a Stagnic Anthrosol based on the Chinese soil taxonomic system(Gonget al.,2007)or as a Hydragric Anthrosol based on the World Reference Base for Soil Resources(IUSS Working Group WRB,2015).The basic soil properties of the topsoil(0-20 cm)at the start of the experiment can be found in Table I.
The field experiment began in 2012,and nine treatments were designed,which were divided into two sets.The first set of treatments was a combination of two factors:water regimes(mid-season drainagevs.continuous flooding)and straw incorporation(withvs.without rice straw applied at a rate of 6 t ha−1).The four treatments in this set were:i)mid-season drainage without rice straw incorporation(MD−S,control),ii)mid-season drainage with rice straw incorporation(MD+S),iii)continuous flooding without rice straw incorporation(CF−S),and iv)continuous flooding with rice straw incorporation(CF+S).The basic properties of the rice straw applied to this experiment can be found in Table I.A mid-season drainage water regime was implemented according to the locally-adopted paddy water regime.In the mid-season drainage water regime,the paddy field was kept at a ponding water depth of 5-8 cm for about 30 d and was then drained for 10 d to reduce ineffective tillering,before being reflooded until one week before rice harvest.In the continuous flooding water regime,the paddyfield was kept flooded,and no drainage was conducted until one week before rice harvest.The quantity of the rice straw incorporated into the soil was close to the quantity of rice straw yield per cropping season.The rice straw was cut into 10 cm pieces and was evenly distributed over the paddy field before straw was incorporated into the soil.

TABLE IProperties of the soil,amended rice straw,pig manure,and biochar
The second set of treatments consisted of six treatments,which were:no N fertilizer(N0),full dose of chemical N fertilizer(N;as same as the MD−S treatment),half dose of chemical N fertilizer(1/2N),half dose of chemical N fertilizer with pig manure(1/2N+M),full dose of chemical N fertilizer with 24 t biochar ha−1(N+LB),and full dose of chemical N fertilizer with 48 t biochar ha−1(N+HB).The N fertilizer application rate was the same across the N,1/2N+M,N+LB,and N+HB treatments.No N fertilizer was applied in the N0 treatment,while a half dose of chemical N fertilizer(compared to the dose applied in the N treatment)was applied in the 1/2N fertilizer treatment.The amount of pig manure used in the 1/2N+M treatment was determined based on the idea that pig manure was substituted for half dose of the chemical N fertilizer(60 and 75 kg ha−1in the early(first)and late(second)rice seasons,respectively),giving 6.45 and 9.74 t ha−1of manure for early and late rice,respectively.The composed pig manure was all used as basal fertilizer in each rice season.The biochar used in this study was provided by Sanli New Energy Ltd.(Shangqiu,China),and was produced from wheat straw at a pyrolysis temperature of 500°C.The biochar amendment rates of 24 and 48 t ha−1were equivalent to 1% and 2% of the dry weight of the topsoil(0-20 cm),respectively.The biochar was passed through a 5-mm sieve before being applied,and was only applied once,at the beginning of the experiment,in the 2012 early rice season,because biochar has high biochemical stability and can be stored in soil for a long time(Glaseret al.,2000;Kuzyakovet al.,2014).The basic properties of the biochar applied in this experiment can be found in Table I.The water regime was mid-season drainage in the second set of experiments.
Total N fertilizer was applied at rates of 120 and 150 kg N ha−1in the early and late rice seasons,respectively,in all treatments except N0 and 1/2N,and the amount of N fertilizer applied was determined by the expected grain yields for the early and late rice crops(Maet al.,2014;Shenet al.,2014).Urea was used as the chemical N fertilizer;50%of the N fertilizer was applied as a base fertilizer,30% as topdressing at the tillering stage,and the remaining 20% as topdressing at the heading stage.Phosphorous in the form of calcium superphosphate was applied at a rate of 17.5 kg P ha−1in the early rice seasons and at a rate of 25.1 kg P ha−1in the late rice seasons in 2016 and 2017,and potassium in the form of potassium sulfate was applied at a rate of 100 kg K2O ha−1once as the base fertilizer in each rice season(Maet al.,2014;Shenet al.,2014).The amounts of N,P,and K contained in the incorporated straw were subtracted from the total N,P,and K fertilizer application rates in the MD+S and CF+S treatments.For the 1/2N+M treatment,no chemical P fertilizer was used because of the P input from the pig manure,which provided averages of 104 and 128 kg P ha−1in the early rice season and the late rice season,respectively.The basic properties of the pig manure used in this experiment can be found in Table I.Total N and P fertilizer application rates,chemical N and P fertilizer application rates,and total N and P application rates from organic materials in the treatments can also be found in Table II.
All treatments had three replications in a randomized block design.Each field plot was 35 m2(7 m×5 m).The rice cultivar wasOryza sativaL.‘Xiangzaoxian No.45’in the early rice season and wasO.sativaL.‘T-you 207’in the late rice season.The rice seedlings in the early seasons were planted with inter-plant intervals of 16.7 cm and with an inter-row spacing of 20 cm,in the late seasons,seedlings were plated with inter-plant intervals of 20 cm and with an inter-row spacing of 20 cm.Each hill contained three to four rice seedlings.Other agricultural practices were in accordance with local norms.
At the beginning of the experiment,as only the soil total organic carbon increase caused by biochar addition was considered,the extra nutrients caused by biochar addition were not subtracted from the total fertilizer input(Table I),which could have led to artificially higher N and P inputs in the biochar treatments in the first crop season(the earlyrice season in 2012).The grain and straw N uptake under the biochar treatments were almost equal to that under the control in 2015,as shown in Table SI(see Supplementary Material for Table SI),and the grain and straw P uptake under the biochar treatments were almost equal to under the control from 2012-2015,as shown in Table SII(see Supplementary Material for Table SII).Therefore,it may be that there was no additional available N and P input in the biochar treatments in the fourth year after biochar addition.Thus,the analyses described in this study were conducted during 2016 and 2017,the fifth and sixth years of the experiment,respectively.

TABLE IINitrogen and P fertilizer application rates for all treatments in the 2016 and 2017 early and late rice seasons
Soil samples were collected during the tillering,heading,and harvesting stages of the early and late rice seasons in 2016 and 2017,respectively.Five soil cores(0-20 cm)per plot were randomly obtained using a soil auger.The samples were mixed,and then taken back to the laboratory.Half of each sample was stored at 4°C for late soil microbial biomass phosphorus(SMBP)analysis.The remaining part was airdried at room temperature,passed through a 2-mm sieve for soil available P(SAP)and pH analysis,and ground to pass through a 0.15-mm sieve for STP,soil total N(STN),and soil total organic carbon(STOC)analyses.In this study,the samples collected at the harvest stage were used to measure SAP,STP,STN,and STOC for each season,as these variables are relatively stable,while the soil samples collected at the three growth stages were used to measure SMBP during the 2016 and 2017 early and late rice seasons.
The plant samples for yield and P concentration measurements were also collected at the harvest stage.The rice grain and straw yields were determined by harvesting rice plants from five 1-meter-squared areas of each plot.To measure rice P concentrations,rice samples were randomly collected from five hills per plot,avoiding unhealthy plants and those lacking neighbors(Tirol-Padreet al.,1996).These rice samples were then cleaned,divided,and categorized as root,straw,or grain,and then placed into respective sample bags.Samples were dried in an oven at 80°C until the weight was constant(Kalra,1998).After drying,the samples were ground to pass through a 0.425-mm sieve and were then stored in sample bags at room temperature for later chemical analyses.
The SAP was determined by extracting the soil with 0.5 mol L−1NaHCO3at pH 8.5(Olsen and Sommers,1982).The SMBP was calculated as the difference between the amount of P extracted(using 0.5 mol L−1NaHCO3at pH 8.5)from fumigated and unfumigated soils using a conversion factor of 0.4(Brookeset al.,1982).Soil basic properties(STOC,STN,STP)and N and P concentrations in rice plants,pig mature,and biochar were determined using the methods described in Pageet al.(1982)and Kalra(1998).Soil total P content was determined by fusing the soil samples with sodium hydroxide at 720°C(Pageet al.,1982).The P concentrations in the rice plants,pig manure,and biochar were measured by digesting the samples using H2SO4and H2O2solutions.For both plant and soil samples,P concentrations in the extracts or the digestion solutions were determined using a UV-2600 ultraviolet light and visible spectrophotometer(Shimadzu Co.Ltd.,Hong Kong),according to the ammonium molybdate-ascorbic acid method.The available P content in the biochar was measured using the same method as that used to determine SAP.Soil total organic carbon content was measured by wet digestion using H2SO4-K2CrO7.Soil total nitrogen content was measured by digesting soil samples using H2SO4solution,with a mixture of 1:10 copper sulfate and potassium sulfate as the catalyst.After the soil was digested,the N concentration of the digestive solution was determined using a continuous flow analytical system(Tecator FIAstar 5000 Analyzer,Foss Tecator,Sweden).The rice plant,pig manure,and biochar total N concentrations were also determined using a flowinjection auto-analyzer after samples were digested with H2SO4-H2O2(Kalra,1998).
The P activation coefficient(PAC)is the ratio of SAP to STP and is one of the most important indexes of soil fertility.It is calculated using the following formula(Wuet al.,2017):

The formula for total rice P uptake(TPU,kg ha−1)is as follows(Dobermann,2007):

wherePsis the P concentration of the straw(g kg−1),Pgis the P concentration of the grain(g kg−1),Ysis the straw yield(kg ha−1),andYgis the grain yield(kg ha−1).The biomass(straw+grain)yield was calculated from the actual straw(excluding stubble)and grain yields in each plot at the harvest stage.
The differences among the treatments were analyzed based on Duncan’s multiple range test using the SPSS 19.0 software package(IBM Corp.,USA),andP<0.05 was considered as the significance level.Graphs were made using SigmaPlot 12.5(Systat Software Inc.,USA).Redundancy analysis(RDA)was conducted to analyze the effects on soil fertility properties(STOC,STN,STP,SAP,SMBP,and soil available N as indicated by the N fertilization application rate)in CANOCO 5.0(Microcomputer Power,USA).
The average STP contents at the harvesting stage of the 2016 and 2017 early and late rice seasons ranged from 548 to 561 mg kg−1across the different water regime and straw incorporation treatments(Fig.1a).The STP contents did not differ significantly among the four treatments.Average SAP content at the harvesting stage of the 2016 and 2017 early and late rice seasons ranged from 8.1 to 9.6 mg kg−1across the treatments(Fig.2a),and SAP contents were significantly higher(P<0.05)under the MD+S and CF+S treatments than under the MD−S and CF−S treatments.SAP levels were increased by 9.3% to 18.8% under the MD+S and CF+S treatments,respectively,in comparison to the corresponding treatments without straw incorporation.SAP did not differ significantly between treatments with different water regimes(e.g.,MD−Svs.CF−S;MD+Svs.CF+S).The PAC values ranged from 1.5% to 1.7% across the different water regime and straw incorporation treatments(Fig.3a).The pattern of PAC values among the treatments was consistent with those of the SAP values.The PAC values were higher under straw incorporation treatments(MD+S and CF+S)than under non-straw treatments(MD−S and CF−S).

Fig.1 Average soil total P(STP)contents under different water regime and straw treatments(a)and under different N fertilizer and biochar treatments(b)in the 2016 and 2017 early and late rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate significant differences among treatments at P<0.05.See Table II for the description of all treatments.
The average SMBP contents under the different treatments ranged from 20.6 to 27.9 mg kg−1(Fig.4a).The SMBP contents increased significantly by 15.5% to 35.6%(P<0.05)under the MD+S and CF+S treatments,respectively,compared to the corresponding treatments without straw incorporation(23.2 and 20.6 mg kg−1).The SMBP contents did not differ significantly between the MD+S and CF+S treatments.However,the SMBP content was significantly less(P<0.05)under the CF−S treatment compared to under the MD−S treatment.

Fig.2 Average soil available P(SAP)contents under different water regime and straw treatments(a)and under different N fertilizer and biochar treatments(b)in the 2016 and 2017 early and late rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate significant differences among treatments at P<0.05.See Table II for the description of all treatments.

Fig.3 Average P activation coefficients(PAC)under different water regime and straw treatments(a)and under different N fertilizer and biochar treatments(b)in the 2016 and 2017 early and late rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate mean significant differences among treatments at P<0.05.See Table II for the description of all treatments.

Fig.4 Average soil microbial biomass P(SMBP)contents under different water regime and straw treatments(a)and under different N fertilizer and biochar treatments(b)in the 2016 and 2017 early and late rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate significant differences among treatments at P<0.05.See Table II for the description of all treatments.
The average STP contents under the N fertilizer and biochar treatments in 2016 and 2017 ranged from 558 to 828 mg kg−1,with the highest values under the 1/2N+M treatment,followed by the N+HB,N0,N+LB,1/2N,and N treatments,in descending order(Fig.1b).However,the differences among the N0,N+LB,and N+HB treatments were not significant.Average SAP contents ranged from 8.6 to 28.9 mg kg−1and followed a similar pattern to that of STP across the different treatments(Fig.2b).For example,the SAP content was the lowest in the N treatment and was the highest in the 1/2N+M treatment.The SAP content under the N0,1/2N,N+LB,and N+HB treatments did not different significantly.The PAC values ranged from 1.6% to 3.5% across the N fertilizer and biochar treatments(Fig.3b).The pattern observed in the PAC values among the treatments was consistent with that of the SAP values.For example,the highest PAC value was observed under the 1/2N+M treatment,while the lowest PAC value was observed under the N treatment.
The average SMBP ranged from 20.6 to 40.0 mg kg−1under the different fertilizer and biochar treatments,and was the greatest overall under the 1/2 N+M treatment(Fig.4b).The SMBP was also dramatically higher(P<0.05)under the N+LB and N+HB treatments compared to the N and 1/2N treatments,while it was the lowest under the N0 treatment.There were no significant differences in SMBP between the N+LB and N+HB treatments or between the 1/2N and N treatments.
Grain yields as affected by straw incorporation and water regime followed different trends in the early and late rice seasons(Fig.5a,b).At the mature stage,grain yields ranged from 3.5 to 4.3 t ha−1in the early rice season and increased to from 5.2 to 5.6 t ha−1in the late rice season.Rice grain yields under the straw treatments(MD+S and CF+S)were significantly lower(P<0.05)than those under the nonstraw treatments(MD−S and CF−S)in the early rice season(Fig.5a),but in the late rice season,the rice grain yields did not differ significantly between straw and non-straw treatments(P>0.05)(Fig.5b).Grain yields under the mid-season drainage treatments(MD−S and MD+S)were not significantly different from those under the continuous flooding treatments(CF−S and CF+S)(Fig.5b).Straw yields at the mature stage ranged from 3.1 to 3.6 t ha−1during the early rice season and increased to from 4.6 to 5.4 t ha−1during the late rice season.Straw yields under the different treatments followed the same trend as those of grain yields in the early rice season(Fig.5a).However,in the late rice season,straw yields were significantly higher(P<0.05)under straw treatments compared to under nonstraw treatments,and were significantly higher under midseason drainage treatments compared to under the continuous flooding treatments(Fig.5b).

Fig.5 Average grain and straw yields under different water regime and straw treatments(a and b)and under different N fertilizer and biochar treatments(c and d)in the 2016 and 2017 early(a and c)and late(b and d)rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate significant differences among treatments at P<0.05.See Table II for the description of all treatments.
Among the fertilizer and biochar treatments,grain and straw yields ranged from 2.2 to 4.4 t ha−1and from 1.9 to 3.9 t ha−1in the early rice season,respectively,and increased to from 4.0 to 5.3 t ha−1and from 3.5 to 4.9 t ha−1in the late rice season,respectively.Grain and straw yields generally increased with increased N fertilizer application rate in the early and late rice seasons(Fig.5c and d).The grain yields under the 1/2N+M,N+LB,and N+HB treatments were not significantly different from those under the N treatment,but the straw yield under the 1/2N+M treatment was significantly lower(P<0.05)than under the N treatment in both early and late rice seasons.Straw yields were significantly higher under the biochar treatments(N+LB and N+HB)(P<0.05)than under the N treatment in the early rice season.
Rice P uptake by grains and straw as affected by different water regimes and straw incorporation followed different trends at the mature stage during both the early and late rice seasons(Fig.6a,b).Phosphorus uptake did not differ significantly between straw and non-straw treatments(except for decreased grain P uptake in the CF+S treatment compared to the CF−S treatment)in the early rice season,but uptake was greater(P<0.05)under straw treatments in the late rice season.The MD−S treatment increased straw P uptake compared to the CF−S treatment(P<0.05)in the late rice season.Compared to the MD+S treatment,the CF+S treatment decreased grain P uptake in the early rice season and increased straw P uptake(P<0.05)in the late rice season.
The grain and straw P uptake generally increased with increased N fertilizer application rate in both the early and late rice seasons(Fig.6c,d).Compared to the N treatment,under the 1/2N+M,N+LB,and N+HB treatments,P uptake did not differ significantly in the early rice season.On the other hand,in the late rice season,the straw and grain P uptake was significantly higher(P<0.05)under the 1/2N+M treatment,and was significantly lower(P<0.05)under the N+LB and N+HB treatments.
For all treatments,RDA showed that the first and second ordination axes explained 90.0% and 5.6% of the variation in average annual rice P uptake,respectively(Fig.7a).The RDA ordination plot showed that average annual rice P uptake values from treatments with the same N fertilizer application rate were clustered together.The rice P uptake along axis 1 showed a positive correlation with rice biomass and P concentration and was also positively correlated with soil available N(indicated by N fertilizer application rate),SMBP,STN,and STOC.The results of the Pearson correlation analysis of soil properties and rice P uptake(Table III)were consistent with those of the RDA,showing that the rice P uptake was positively correlated(P<0.01)with N fertilizer application rate,SMBP,and STN.For treatments with the same N fertilizer application rate,RDA showed that the first and second ordination axes explained 73.6% and 5.4%,respectively,of the variation of the rice P uptake(Fig.7b).Rice P uptake along axis 1 was positively correlated with rice P content and was negatively correlated with rice biomass.In addition,rice P uptake was positively correlated with SMBP,SAP,and STP,and was negatively correlated with STOC.

Fig.6 Average grain and straw P uptake under different water regime and straw treatments(a and b)and under different N fertilizer and biochar treatments(c and d)in the 2016 and 2017 early(a and c)and late(b and d)rice seasons.Bars are the standard deviations of means(n=3).Different letters above the columns indicate significant differences among treatments at P<0.05.See Table II for the description of all treatments.

Fig.7 Redundancy analysis(RDA)of the effects of soil fertility properties on average annual rice P uptake in the 2016 and 2017 early and late rice seasons across all the treatments(a)and under treatments with the same N fertilizer application rate(b).The arrows with dotted lines represent samples that were distributed and sorted spatially.1(or 8)=MD−S(or N);2=MD+S;3=CF−S;4=CF+S;5=N0;6=1/2N;7=1/2N+M;9=N+LB;10=N+HB;Biomass=average annual rice straw and grain yield;P-uptake=average annual rice straw and grain P uptake;P-concen=average annual rice straw and grain P concentration;FNR=annual total N fertilizer application rate;STN=soil total N;STP=soil total P;SAP=soil available P;STOC=soil total organic carbon;SMBP=soil microbial biomass P.See Table II for the description of all treatments.

TABLE IIIPearson correlation coefficients between soil fertility properties and average annual rice P uptake across all treatments and under treatments with the same total N application rate
To evaluate PUE under the various agronomic practices,the ratio of P output to P input was compared across treatments(Roberts and Johnston,2015).Among the treatments,the ratios were remarkably lower(P<0.05)under the N0,1/2N,and 1/2N+M treatments compared to the N treatment,as shown in Table SIII(see Supplementary Material for Table SIII),which indicates that both N shortage and P overuse can cause PUE to decrease.The ratio also increased significantly(P<0.05)under the straw treatments compared to the corresponding non-straw treatments,indicating that straw incorporation can increase PUE.There were no significant differences in the ratios between water regime(MD−Svs.CF−S;MD+Svs.CF+S)treatments and among the N,N+LB,and N+HB treatments,and thus water regime and biochar amendment had limited effects on PUE.In addition,considering that chemical P fertilizer was reduced by 51.2%and that there were no yield decreases in the straw incorporation treatments,PUE dramatically increased by 210% under the straw treatments compared to under the N treatment,based on the partial factor productivity of applied P,which was calculated as the ratio of rice grain yield to chemical P fertilizer application(Roberts and Johnston,2015).
In this study,straw incorporation increased SAP and SMBP,but did not significantly affect STP.Straw incorporation can provide carbon sources for soil microbes,which enhances microbial growth,and thus promotes the mineralization of soil organic P by enhancing phosphatase activity(Sunet al.,2019).In addition,straw degradation releases organic anions,which could reduce the immobilization of fertilizer P or induce the mobilization of adsorbed Pviacompetition for P adsorption sites,thereby improving soil available P(Nziguhebaet al.,1998).Recent studies have also shown that the main components of straw(cellulose and lignin)improve soil P availability by different mechanisms(Sunet al.,2019).Sunet al.(2019)reported that cellulose increased soil P availability by enhancing microbial biomass P,while lignin increased soil P availabilityviacompetitive adsorption of Fe and Al oxides by lignin functional groups.Straw incorporation increases soil organic carbon and increases microbial biomass,and thus,SMBP increases accordingly based on the principle of stoichiometric homeostasis(Mastoet al.,2013).
The observed decrease in SMBP under the CF−S treatment compared to the MD−S treatment may be due to that the activities of some aerobic microbes are restrained under continuous flooding(Devêvre and Horwáth,2000).Additionally,it has been reported that there are more leaching losses of soil dissolved N and P in paddy fields under continuous flooding compared to under a controlled irrigation water regime(Penget al.,2011).Decreases in soil dissolved C,N,and P under continuous flooding may reduce the soil microbial biomass,and thus SMBP decreases accordingly.Former studies have shown that SAP increases in anaerobic conditions compared to in aerobic conditions due to the fact that Al oxides,Fe oxides,and especially Ca2+,become more soluble under reduced oxygen,and thus the P associated with them is released into the soil solution(Scalengheet al.,2014).However,there were no observed increases in SAP under the continuous flooding water regime compared to the mid-season drainage water regime in this study,which also indicates that there may be more P leaching losses in paddy fields under continuous flooding.Thus,it is more favorable to adopt mid-season drainage than continuous flooding in paddy fields,not only to increase soil P availability,but also to decrease P leaching.
In this study,the addition of biochar increased STP,SAP,and SMBP.The increase in STP under the biochar treatments may largely be due to the P input from the amended biochar.Considering that soil amendment with straw-derived biochar reuses resources(Woolfet al.,2010),biochar amendment is therefore a good method of increasing STP stock.Adding biochar to acidic soils can increase soil pH(Lehmannet al.,2011;Wanget al.,2018),which reduces P immobilization by Fe and Al oxides(Loneet al.,2015;Singhet al.,2016).Moreover,increased soil pH could promote alkaline phosphatase activity,which enhances the hydrolysis of SAP(Gul and Whalen,2016).Thus,biochar amendment could increase SAP.Adding biochar to paddy soils has been found to increase soil microbial biomass C due to increasing soil dissolved carbon and available nutrients(e.g.,available N,P and K)(Wanget al.,2018),which can also increase SMBP,based on the principle of stoichiometric homeostasis(Mastoet al.,2013).
The increases in STP,SAP,and SMBP under the pig manure treatment observed in the present study could be due to high P input from the pig manure.In this treatment,half of the chemical N fertilizer dose was replaced with pig manure.However,due to the high P concentration in the pig manure,the P fertilizer application rate in the 1/2N+M treatment(104 and 128 kg P ha−1in the early and late rice seasons,respectively)was much higher than in other treatments.The P concentration in pig manure is highly plant-available as approximately 48%-76% of the total P may be labile P,and approximately 25%-48% of the P may be organic(Pagliariet al.,2012;Liet al.,2014,2015),which is hard for soil ions to immobilize(e.g.,Ca2+,Fe3+,Al3+).This was reflected in the high P activation coefficient measured(Fig.3b).However,the increased STP and SAP values in the pig manure treatment may increase the risk of P losses by runoffand leaching(McDowell and Sharpley,2001;Allenet al.,2006;Karimiet al.,2018).Former studies have shown that when SAP is greater than a certain value(or“change point”),the P leaching risk is dramatically increased(Hesketh and Brookes,2000;Zhaoet al.,2007).For example,the soil Olsen-P change points for a high P leaching risk ranged from 30 to 160 mg kg−1in 22 soil samples collected from 11 provinces in China(Zhaoet al.,2007).The averaged Olsen-P values were 27.2 to 30.1 mg kg−1in the 1/2N+M treatment,which indicated that high P leaching losses may occur under this treatment.
Because of the decrease in P uptake by rice plants,soil P surplus(the difference between fertilizer P input and P output by rice grain and straw P uptake)in the reduced N treatments(N0 and 1/2N)was as much as 8.2-15.6 kg ha−1year−1(Table SIII).This may be the main reason behind the enhanced STP and SAP levels in the N0 and 1/2N treatments.Therefore,soil P surplus in fields could be one of the major factors affecting soil P availability.It is important to apply P fertilizers at a reasonable level to avoid soil P accumulation as well P leaching losses.
In the rice-rice cropping system,the results of the present study showed that rice P uptake generally increased under straw treatments(MD+S and CF+S)and under the 1/2 N+M treatment,and these increases all occurred in the late rice season,although straw incorporation decreased rice P uptake in the early rice season.The RDA analysis showed that when treatments shared the same N application rate,rice P uptake was positively correlated with rice P concentration,and increases in rice P concentration were mainly caused by increases in SAP,SMBP,and STP(Fig.7b).This indicates that agronomic practices that increase SAP,SMBP,and STP have the potential to increase rice P uptake.There were decreases in grain and straw yields in the straw incorporation and pig manure treatments in the early rice season,which may be caused by insufficient N supplied to rice crops(Andrénet al.,1993;Agostini and Sholefield,2006).This is because the release of N from organic materials decreases under low temperatures(Zhanget al.,2015;Tanget al.,2016).In addition,increases in soil microbial biomass caused by straw incorporation or manure application may also increase the immobilization of fertilizer N by microbes(Devêvre and Horwáth,2000),and thus also reduce N supply to rice crops.Decreased rice yields may then result in decreased or plateaued rice P uptake under the straw and pig manure treatments in the early rice season.Therefore,straw incorporation in the late rice season should be encouraged to increase soil P availability and to promote rice P uptake,thus enhancing PUE.
Although the average rice P uptake under the 1/2N+M treatment was the highest of all the treatments,soil P surplus was also the highest under this treatment.This is because with increased P fertilizer input,rice P uptake will reach a point at which it can no longer increase(Pantenet al.,2016).The P fertilizers are also hard to remove from soils,leading to soil P surplus.Higher rates of P input tend to increase the soil P pool,leading to a higher risk of non-point source pollution(Sharpley and Wang 2014).Therefore,when partly substituting organic pig manure for a chemical N fertilizer,it is important to consider the N input from pig manure as well as the P input,in order to avoid excessive P input and soil P accumulation(Shiet al.,2015).
In this study,decreased rice P uptake was found in treatments with reduced N fertilizer application rates(N0 and 1/2N),as well as in the CF−S and N+HB treatments.Nitrogen is one of the essential nutrients for crop growth.A reduction in N application rate causes a decrease in rice biomass,leading to decreased rice P uptake(Fig.7a).Under the CF−S treatment,SMBP content decreased,which could have caused the decrease in rice P concentration,as shown in Fig.S1(see Supplementary Material for Fig.S1),and consequently the decrease in rice P uptake(Fig.7b).Due to the large reduction in rice straw P concentration observed in the N+HB treatment,as shown in Fig.S2(see Supplementary Material for Fig.S2),biochar addition at a high rate(48 t ha−1)remarkably reduced rice straw P uptake.As SAP and SMBP both increased under the biochar treatments,the decreased straw P uptake may be caused by reductions in root P uptake capacity.This is because the root structure,root P uptake kinetics,and root exudates influence plant P uptake in the aboveground parts(Schachtmanet al.,1998;Shenet al.,2011).Further studies need to be conducted to understand how biochar application in paddy fields affects root P uptake as well as P transportation and distribution in plant different tissues.The results of this study show that straw-derived biochar application at a rate of 24 t ha−1can promote soil P availability and will not result in decreased rice P uptake.
This study mainly focused on data obtained in the fifth and sixth years of the field experiment.Data from the first four years of the field experiment is not shown to avoid reporting the effects of the additionally-available N and P from biochar additions on soil P availability and rice P uptake.Therefore,the data obtained in this study can provide stable information on how soil P availability and rice P uptake respond to biochar amendment in a rice-rice cropping system.However,long-term studies are still needed to identify the long-term effect of biochar amendment.There was much higher P input in the 1/2N+M treatment,because the manure P input was much higher than the total P input designed for the treatments when substituting half the chemical N fertilizer dose with pig manure.Thus,comparing this treatment with the other treatments may not be meaningful.However,the 1/2N+M treatment also provided useful information on the effects of soil P surplus on soil P availability,and on the risks of P leaching due to high P input in paddy fields.In this study,there was no zero P fertilizer treatment from which to calculate soil P availability,rice P uptake,and P fertilizer use efficiency.This is because this major purpose of this experiment was to compare soil P availability and rice P uptake under various agronomic practices within normal conventional practices,and the zero P treatment was not designed at the beginning of the field experiment.Considering that zero P treatments are very important when studying PUE,a zero P treatment was added in the late rice season of 2016.However,the zero P treatment was compromised due to the strong residual effect of P fertilizers up to the late rice season of 2017,therefore the soil P availability and rice P uptake calculated for the three rice seasons without P input do not truly reflect the soil P availability and the soil P supply under no P input.Thus,the results of the zero P treatment were not presented in this study.
This study found that,in a rice-rice cropping system:i)straw incorporation enhanced soil P availability by increasing SAP and SMBP,ii)substituting half the chemical N fertilizer dose with pig manure and the biochar application augmented soil P availability by increasing STP,SAP,and SMBP,mainly owing to increased soil P and organic C inputs,and iii)applying half a dose of chemical N or no N increased soil P availability by increasing STP and SAP due to increased soil P surplus.Thus,the balance between soil P input and output had important effects on soil P availability.The continuous flooding water regime reduced soil P availability by decreasing SMBP compared to a mid-season drainage regime.The rice P uptake increased with increases in soil P availability when N supply was not restrained and decreased when N supply was reduced.Biochar application at a high rate also decreased rice P uptake,which may be related to a reduced root P uptake capacity.Phosphorus use efficiency generally increased under straw treatments,and decreased under the reduced N treatments and under the manure treatment with excessive P input.
In the rice-rice cropping system,straw incorporation in the late rice season can be a good method of increasing soil P availability as well as PUE.Replacing half the chemical N fertilizer dose with pig manure caused a large P surplus in the soil and may,thus,increase P lossesvialeaching or runoff.In future,a suitable application rate for pig manure in the rice-rice cropping system should be established to improve nutrient use efficiency while reducing the environmental risks associated with non-point source pollution.Biochar application at a rate of 24 t ha−1can promote soil P availability and will not result in decreased rice P uptake.
ACKNOWLEDGEMENT
The research is supported by the National Key Research and Development Program of China(Nos.2016YFD0200307 and 2018YFC0213302)and the Youth Innovation Promotion Association of Chinese Academy of Sciences(No.2017418).
SUPPLEMENTARY MATERIAL
Supplementary material for this article can be found in the online version.