Transfer characteristics of nitrogen fixed by leguminous green manure crops when intercropped with maize in northwestern China

LlU Rui,ZHOU Guo-peng,CHANG Dan-na,GAO Song-juan,HAN Mei,ZHANG Jiu-dong,SUN Xiao-feng,CAO Wei-dong

1 Institute of Soil and Fertilizer,Qinghai Academy of Agriculture and Forestry Sciences/The Graduate School,Qinghai University,Xining 810016,P.R.China

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

3 College of Resources and Environmental Sciences,Nanjing Agricultural University,Nanjing 210095,P.R.China

4 Institute of Soil and Fertilizer and Water-saving,Gansu Academy of Agricultural Sciences,Lanzhou 730070,P.R.China

Abstract To ascertain the possibility of cultivating maize using biological nitrogen fixation (BNF) by leguminous green manure crops in maize/leguminous green manure intercropping systems,BNF and nitrogen (N) transfer were studied in Xining and Wuwei,two typical northwestern Chinese cities.The experimental treatments included monocultured maize,monocultured green manures (hairy vetch and common vetch),and their intercropping systems.The proportions of N derived from the atmosphere (%Ndfa) in intercropping systems were not significantly different from that in monocultured green manure systems at either experimental site,except for that in hairy vetch (HV) in Xining.The amount of N derived from the atmosphere (Ndfa) of common vetch (CV) significantly decreased from 1.16 and 1.10 g/pot in monoculture to 0.77 and 0.55 g/pot when intercropped with maize,in Xining and Wuwei,respectively,and the Ndfa of HV when intercropped significantly decreased from 1.02 to 0.48 g/pot in Xining.In the intercropping systems in Xining and Wuwei,the amounts of N transferred (Ntransfer) from CV to maize were 21.54 and 26.81 mg/pot,accounting for 32.9 and 5.9% respectively of the N accumulation in maize,and the values of Ntransfer from HV to maize were 39.61 and 46.22 mg/pot,accounting for 37.0 and 23.3%,respectively,of the N accumulation in maize.Path analysis showed that soil nutrient and green manure biomass were mainly related to Ndfa,and that δ15N had a primary relationship with Ntransfer.We found that 5.9-37.0% of N accumulation in maize was transferred from green manures,and that the N transfer ability to maize of HV was higher than that of CV.In conclusion,intercropping with leguminous green manures provided a feasible way for maize to effectively utilize biologically fixed N.

Keywords:green manure,maize,intercropping,biological nitrogen fixation,nitrogen transfer

1.lntroduction

Soil nitrogen (N) is one of the main factors affecting crop growth in monocultured systems (Mcgrawet al.2008),and is very important for maintaining high crop yield(Thilakarathnaet al.2016).Biological nitrogen fixation(BNF) by legumes is an important strategy to replenish N in the soil.BNF uses N2from the atmosphere and is considered one way to develop sustainable ecological agriculture (Yaoet al.2019).Studies have shown that intercropping with legumes could reduce the use of chemical fertilizers,increase crop yield,and improve soil fertility through BNF (Xiaoet al.2004;Tanget al.2018;Duet al.2019).Because the fixed N from legumes can be transferred to cereals,the competition for soil N by cereals can improve the proportions of N derived from the atmosphere (%Ndfa) of legumes (Peopleset al.2015).However,the effects of intercropping systems of various cereals on BNF may differ (Chapagain and Riseman 2014;Fanet al.2019),and there may also be significant differences in N transfer from legumes to cereals (Zanget al.2015;Tsialtaset al.2018) depending on the crop variety,management and environment (Carlsson and Huss-Danell 2003;Chalket al.2014).

The15N natural abundance method is based on the comparison between the15N/14N ratio in an N2-fixing plant,a non-N2-fixing reference plant growing in the same soil,and the N2-fixing plant depending only on N2fixation for N supply (Zelleret al.1998).The effect of isotopic fractionation on the process of BNF in root nodules is minimal and is often neglected.The fractionation effect mainly occurs in the process of N transport from nodules to other plant tissues (Jinya 1984).Therefore,it can be ignored in the15N natural abundance method for measurement of BNF.Hence,the15N natural abundance method can be used to measure net N transfer from N2-fixing plants to associated non-N2-fixing plants under intercropping conditions (Moyer-Henryet al.2006;Sierra and Nygren 2006).However,with the15N natural abundance method,most studies on BNF and N transfer focus on major crop legumes in intercropping systems,such as the pea,soybean,and groundnut.Few studies have revealed the N transfer efficiency of BNF by leguminous green manure crops to the main crop in their intercropping systems.

In northwestern China,the traditional intercropping systems are maize or wheat intercropped with soybean.These traditional intercropping systems have been associated with problems such as high water consumption and soil fertility depletion.In recent years,China has been taking action to adjust the crop industry structure,to improve soil fertility and the ecological environment of farmland,and reduce the use of chemical fertilizers.In northwestern China,more sustainable production methods should be explored to fit local environmental needs.New intercropping systems have been proposed,such as leguminous green manures intercropped with maize (ZeamaysL.) (Caoet al.2017).Studies have shown that the common vetch (CV,ViciasativaL.) and hairy vetch (HV,ViciavillosaRoth) could replace part of the nitrogen fertilizer used,improve soil fertility,and increase cereal yields by BNF (Zhanget al.2011).

There is little information on the legume N fixation and N transfer to maize in these intercropping systems.The present study created a constrained soil environment so that the fixed N could transfer more effectively to the nearby grain crops.The main objectives of the study were to:(1) determine %Ndfaand the amount of N derived from the air (Ndfa) of CV and HV in monocultured and intercropped maize systems;(2) quantify the N transfer from CV and HV to maize in intercropped maize systems;and (3) reveal the influence of planting system,experimental site,soil chemical properties,plant nutrients,and shoot dry biomass on Ndfaand the amount of N transfer from the legumes to maize (Ntransfer).

2.Materials and methods

2.1.Experimental sites and materials

Pot experiments were conducted in 2019 in Xining(Qinghai Province) and Wuwei (Gansu Province),northwestern China.Xining (36°62′N,101°77′E) is located at an altitude of 2 300 m above sea level,where the annual average temperature,precipitation,and evaporation are 5.9°C,368 mm,and 1 730 mm,respectively.The soil used for the Xining study belongs to the Kastanozem group.Wuwei (37°43′N,102°35′E)is located at an altitude of 1 504 m above sea level,where the annual average temperature,precipitation,and evaporation are 7.8°C,222 mm,and 2 021 mm,respectively.The soil used in the Wuwei study was an irrigated desert soil.Soil for the experiment was obtained from the cultivated layer of farmland (0-20 cm) and then air-dried and screened (5 mm) for later use.The basic chemical properties of the soils before planting are shown in Table 1.

Table 1 Chemical properties of the soils used in the Xining,Qinghai and Wuwei,Gansu in pot trials in China1)

The maize and green manure varieties used were the main local cultivars.In Xining,the maize cultivar was Jinsui 3,and the CV cultivar was Qingmu 3333.In Wuwei,the maize cultivar was Wuke 2,and the CV cultivar was Longjian 2.The HV cultivar was Turkmen at both sites.

2.2.Experimental design and management

A completely randomized design was applied in the pot experiment,and it involved five treatments:monocultured maize (MM),monocultured common vetch (MC),monocultured hairy vetch (MH),maize intercropped with common vetch (IMC),and maize intercropped with hairy vetch (IMH).Five replicate pots were used for each experimental treatment.Each pot (length×width×height:48 cm×32 cm×16 cm) contained 20 kg soil (dry-weight basis).No N fertilizer was applied to any treatment.Application rates for the pot experiment of phosphorus (P)and potassium (K) fertilizers were double that of the most common levels applied in local fields (141 kg P2O5ha-1and 97.5 kg K2O ha-1and 150 kg P2O5ha-1and 97.5 kg K2O ha-1in Xining and Wuwei,respectively),i.e.,P and K fertilizers were applied as basal fertilizer at rates of 2.51 g P2O5and 1.73 g K2O/pot in Xining and 2.67 g P2O5and 1.73 g K2O/pot in Wuwei,respectively.Three maize seeds and six green manure seeds were planted in each hole,and one maize plant and three green manure plants were left after thinning at the seedling stage in each pot.The seeds were sown on April 15,2019.During the growth period,each pot was watered weekly with 5 L of well water.The planting pattern used in the pot trials is shown in Fig.1.

Fig.1 Planting pattern of maize and green manure crops in the pot trial. A,monocultured maize.B,monocultured leguminous green manure.C,maize intercropped with leguminous green manure.

2.3.Sample collection and analysis

Samples of plants and soils were collected in Xining on July 3,2019 and in Wuwei on June 29,2019.The green manures and maize shoots were inactivated at 105°C for 30 min and then dried to a constant weight at 65°C.After being weighed,plant samples were ground using a ball mill prior to determination of15N abundance and total N.At the same time,some fresh soil samples were collected from the pots for the determination of mineral N concentration,while the remaining soil samples were air-dried and screened for the determination of other soil physicochemical properties.

The total N concentration of the plants was determined using an elemental analyzer (Elementar Analysensysteme GmbH,Germany).The Vanadium-molybdenum yellow colorimetric method was used for the determination of plant total phosphorus (P) and flame photometer for the determination of plant total potassium (K) (Lu 2000).The15N abundance of plants and soil was determined using a mass spectrometer (Thermo Scientific,Waltham,MA,USA).Soil organic matter (SOM) was determined by the potassium dichromate-external heating method,and soil total N (TN) was determined by the Kjeldner method.Fresh soil was extracted with 2 mol L-1KCl,and the mineral N (Nmin) of the extract was detected using a flow analyzer (Seal AA3;Norderstedt,Germany).Available P was determined using the method described by Olsenet al.(1954),and available K was assayed using the ammonium acetate method (Helmke and Sparks 1996).The soil pH (soil:water=1:2.5) was determined using a compound electrode (LE438,Mettler-Toledo Instruments,Shanghai,China).

2.4.Data collection and analysis

A shoot dry biomass land equivalent ratio (LER)＞1 indicates that intercropped shoot dry biomass has an advantage over monoculture;value＜1 indicates a disadvantage.The LER of the intercropping systems was calculated using eq.(1):

where YIML-Mand YIML-Lrepresent the shoot dry biomass of intercropped maize and intercropped leguminous green manure,respectively,while YMMand YMLrepresent the shoot dry biomass of monocultured maize and monocultured leguminous green manure,respectively.

The relative15N (δ15N) abundance was calculated according to eq.(2) (Yoneyamaet al.1986):

The abundance of15N in the atmosphere is stable and uniform at 0.3663%,which is considered the standard isotope abundance of15N (Mariotti 1983).The atom%15N(sample) is the15N abundance of the samples to be tested.

The %Ndfaof green manure and the proportion of N transferred from green manure to maize (%Ntransfer) were calculated using eqs.(3) and (4),respectively (Shearer and Kohl 1986;Jensen 1996):

where δ15Nref.plantis the δ15N of the reference plant(monocultured maize).The δ15Nfixingplantis the δ15N of the intercropped leguminous green manure.The δ15Nreceiveris the δ15N of the intercropped maize.Bis the δ15N value of nodulated N2-fixing plants grown without N input;theBvalues of CV and HV are -1.56 and -0.93,respectively(Tadakatsuet al.1986).

2.5.Data statistics

Microsoft Excel 2019 was used for data processing.Significance of treatment effects was determined by oneway ANOVA,and the LSD (P＜0.05) andt-test procedures of the SPSS 22 statistical analysis system were used to determine significant differences between treatment means.Histograms and boxplots were drawn using OriginPro 2019 (OriginLab,Northampton,MA,USA).The“PLSPM”package in R 3.6.1 was used to explore the correlation between intercropping systems,experimental sites,soil chemical properties,plant nutrient accumulation,shoot dry biomass,and Ndfaand Ntransfer(Sanchez and Trinchera 2010).

3.Results

3.1.Effects of intercropping on the chemical properties of soil

The soil chemical properties for different treatments are presented in Table 2.For both experimental sites,the soil pH was significantly lower in the maize/HV intercropping system than in monocultured HV,and SOM was significantly lower in maize/green manure intercropping systems than in monocultured green manure.Nminand soil δ15N showed no significant differences between monocultured and intercropped green manures in Xining,but reduced significantly in the maize/CV intercropping system compared with that in monocultured CV in Wuwei.In Xining,available K reduced significantly in the maize/HV intercropping system compared with that in the monocultured HV system,but it increased significantly in the maize/CV intercropping system compared with that in the monocultured CV system.

3.2.Effects of intercropping on shoot dry biomass and N accumulation

The shoot dry biomass of intercropped maize at both experimental sites was significantly lower than in the monoculture (Table 3).Moreover,in the intercropping systems,the total shoot dry biomass of green manure and maize was not significantly different from that of the monocultured green manure and was higher than that of the monocultured maize.The LERs in the maize/CV intercropping systems were 1.07 and 1.05 and in the maize/HV intercropping systems were 1.02 and 1.13 in Xining and Wuwei,respectively,suggesting that the total shoot dry biomass in intercropping systems had an overall advantage over the monocultures.

Shoot N accumulation in intercropped maize was significantly lower than in monocultured maize.At the two experimental sites,shoot N accumulation of intercropped CV was significantly lower than that of monocultured CV by 23.4 and 30.2% in Xining and Wuwei,respectively.However,the total N accumulation per pot of green manure monoculture and maize/green manure intercropping systems was higher than that of monocultured maize (Table 4).

Table 2 Soil chemical properties in monoculture and intercropping treatments in pot trials at two experimental sites1)

Table 3 Shoot dry biomass and land equivalent ratio (LER) in monoculture and intercropping treatments in pot trials at two experimental sites

Table 4 Shoot nitrogen accumulation (g/pot) in monoculture and intercropping treatments in pot trials at two experimental sites

3.3.Effects of intercropping on %Ndfa and Ndfa of leguminous green manures

The %Ndfavalues of the intercropped green manures were not significantly different to the monocultures,except that the %Ndfaof intercropped HV was significantly lower than the monoculture in Xining (Fig.2-A).The %Ndfaof monocultured HV was significantly higher than that of CV (monocultured and intercropped) and intercropped HV in Xining,while the %Ndfaof CV was significantly higher than that of HV in Wuwei.In the monocultured and intercropped systems,the %Ndfavalues of CV were not significantly different between the two experimental sites.The %Ndfavalues of HV in the monocultured andintercropped systems were 78.7 and 54.7% in Xining,respectively,and were significantly higher than those of HV in Wuwei in both systems (35.4 and 29.0%,respectively) (Fig.2-A).

The Ndfaper pot was significantly lower in intercropped green manures than in monocultured green manures,except for in HV in Wuwei (Fig.2-B).However,the Ndfaper plant was not significantly different between the systems (Fig.2-C).The Ndfaper pot showed no significant difference between CV and HV,except that it was significantly higher in CV than in HV in monocultured systems in Wuwei.The amounts of Ndfavalues of intercropped CV and monocultured HV in Xining were 0.77 and 1.02 g/pot,respectively,which were significantly higher than in Wuwei (0.55 and 0.59 g/pot,respectively).

Fig.2 The proportions of N derived from the atmosphere(%Ndfa) per pot (A) and and the amount of nitrogen derived from the atmosphere (Ndfa) per pot (B) and per plant (C) of leguminous green manure plants in different monoculture and intercropping treatments.MC,monocultured common vetch;IMC-C,common vetch in maize/common vetch intercropping system;MH,monocultured hairy vetch;IMH-H,hairy vetch in maize/hairy vetch intercropping system.Different lowercase letters indicate significant differences among the treatments(n=5,P＜0.05,one-way ANOVA followed by LSD).＊ indicates significant differences among the experimental sites (n=5,P＜0.05,t-test).Bars are SE.

3.4.Effects of intercropping on %Ntransfer and Ntransfer to maize

In both intercropping systems,the N concentrations in leguminous green manure crops were significantly higher than that in maize (Fig.3).The N concentrations of CV and HV were between 2.2 and 3.1%,while and the N concentrations in intercropped maize were only 1.0-1.8%.The premise of N transfer between maize and legumes is the difference in nutrient concentration (Xiaoet al.2004).The N concentration gap between maize and legumes in this study was large enough to establish a source-pool relationship.

Fig.3 Plant nitrogen concentration (N%) in different intercropping systems in Xining,Qinghai (A) and Wuwei,Gansu (B). IMC-C,common vetch in maize/common vetch intercropping system;IMC-M,maize in maize/common vetch intercropping system;IMH-H,hairy vetch in maize/hairy vetch intercropping system;IMH-M,maize in maize/hairy vetch intercropping system.Different lowercase letters indicate significant differences among the treatments (n=5,P＜0.05,one-way ANOVA followed by LSD).Bars are SE.

The %Ntransferin the maize/HV intercropping system was significantly greater than in the maize/CV intercropping system.The %Ntransfervalues from CV and HV to maize were 32.9 and 37.0%,respectively,in Xining,and these proportions were significantly higher than the 5.9 and 23.3%,respectively,measured in Wuwei (Fig.4-A).The amounts of N transferred to maize intercropped with CV were 21.54 and 26.81 mg/pot,which were significantly lower than those of maize intercropped with HV (39.61 and 46.22 mg/pot) in Xining and Wuwei,respectively.However,there was no significant difference in Ntransferbetween the two experimental sites (Fig.4-B).The amounts of N transferred from green manures to maize were 7.17-15.41 mg/plant of maize (Fig.4-C).

Fig.4 The proportion of N transferred (%Ntransfer) per pot (A) and the amount of N transferred (Ntransfer) per pot (B) and per plant (C)in different intercropping treatments in Xining,Qinghai and Wuwei,Gansu. IMC-M,maize in the maize/common vetch intercropping system;IMH-M,maize in the maize/hairy vetch intercropping system.Different lowercase letters indicate significant difference among the treatments (n=5,P＜0.05,t-test);＊ indicates significant differences among the experimental sites (n=5,P＜0.05,t-test).Bas are SE.

3.5.Factors influencing the Ndfa and Ntransfer

The results of the path analysis of the factors influencing Ndfaand Ntransferare shown in Fig.5.Monocultured andintercropped green manure treatments and experimental site had significant influences on soil pH,soil nutrients and shoot dry biomass.Moreover,these treatments had a significant influence on soil δ15N and plant nutrient accumulation.Soil nutrient concentrations and shoot dry biomass were significantly correlated with Ndfa(Fig.5-A).The two intercropped treatments and experimental sites had a significant influence on soil pH,soil δ15N,plant nutrient accumulation,and shoot dry biomass.Moreover,treatments had a significant influence on soil nutrient concentrations,and soil δ15N correlated well with Ntransfer(Fig.5-B).

Fig.5 Path analysis of factors influencing the N accumulation derived from air (Ndfa) and nitrogen transfer accumulation (Ntransfer).A,monocultured common vetch,maize intercropped with common vetch,monocultured hairy vetch,and maize intercropped with hairy vetch.B,maize intercropped with common vetch and maize intercropped with hairy vetch.Experimental site indicates Xining,Qinghai and Wuwei,Gansu.Plant nutrient indicates total nitrogen accumulation (TN),total phosphorus accumulation (TP);and total potassium accumulation (TK).Biomass indicates shoot dry biomass.SOM,soil organic matter;TN,total nitrogen;Nmin,mineral nitrogen;AP,available phosphorus;AK,available potassium.＊,P＜0.05;＊＊,P＜0.01;＊＊＊,P＜0.001.

4.Discussion

4.1.Effects of intercropping systems on the BNF

In this study,the %Ndfaof green manure intercropped with maize was not significantly different from that in the monoculture,except for that in HV in Xining.Because no nitrogen fertilizer was applied in this experiment,soil N limitation at the early stage of growth may have led to a lack of N in the green manures,thus affecting the root development of the green manures and resulting in the reduction of N-fixation ability.Maet al.(2013) found that with an increase in N fertilizer application,the %Ndfaof legumes at first increased and then decreased,indicating that an appropriate application of exogenous N can increase the %Ndfa.Ingraffiaet al.(2019) found that the%Ndfavalues in wheat/faba bean intercropping systems were 21-108% greater than in monocultured faba bean.This was because most of the N in the soil was consumed by the wheat in the intercropping system,and the faba bean was driven to increase biologically fixed N to meet its own demand for N (Ingraffiaet al.2019).Dansoet al.(1987) also found that intercropping improved the %Ndfaof legumes,because the amount of N absorbed by legumes from the soil decreased in the intercropping system and induced an improvement in N-fixation efficiency.This effect becomes greater with an increase in the density of cereal (Fanet al.2019).In the present study,maize only grew to the early jointing stage,and its N-absorption ability was weak and could not drive the leguminous green manures to increase the %Ndfa.As shown in Table 5,the shoot dry biomass per maize plant was significantly higher in monoculture than in intercropping systems,whereas the shoot dry biomass per legume plant was significantly greater in intercropping systems,except for that of CV in Wuwei.This may be because the green manures had a relatively greater competitive ability than maize did for soil available N when they were intercropped together.Huet al.(2017) showed a similar result in which maize competition for N may not have contributed to the increase in %Ndfaof leguminous green manure.In practice,leguminous green manures are ploughed into the soil,and the negative impact on maize will be mitigated.

Table 5 Shoot dry biomass per leguminous green manure plant (g/plant) in monoculture and intercropping treatments in pot trials at two experimental sites

The %Ndfavalues at the two experimental sites were similar for the maize/CV intercropping system,while the %Ndfavalues of HV at the two sites differed,with the %Ndfaof HV in Xining being higher than that of HV in Wuwei.This difference may be attributed to different environments or legume species at the experimental sites.In this study,path analysis showed that the difference in Ndfawas mainly related to soil nutrient concentration and shoot dry biomass (Fig.5-A).Other studies have also shown that BNF is affected by soil nutrient levels (Obersonet al.2013;Huet al.2017).The N fixation potential of leguminous green manure crops in the intercropping system could be maximized by an appropriate combination of cereals and legumes,appropriate population density,and corresponding soil management measures (especially appropriate N supply and water regulation) (Saiaet al.2016;Fanet al.2019;Rodriguezet al.2020).

4.2.N transfer in intercropping systems

This study showed that the %Ntransfervalues of HV were significantly higher than CV at both experimental sites.In addition,5.9-32.9% and 23.3-37.0% of N in maize were derived from CV and HV,respectively,in the maize/green manure intercropping systems.Other reports have shown that 2-17% of N in maize was transferred from soybean roots (Ranells and Wagger 1997;Yonget al.2015),11%of N in barley was transferred from pea (Chapagain and Riseman 2014),and 19% of N in rye was transferred from HV (Poffenbargeret al.2015),and 7-42% of N in nonleguminous forage grass was transferred by legumes(Schipanski and Drinkwater 2012;Thilakarathnaet al.2012).In this study,the N transfer occurred before the leguminous green manure was ploughed into the soil.Unlike commercial crops of beans and forage,leguminous green manures are generally ploughed into the soil at the flowering stage,and the released N from the green manure can be absorbed by the intercropping crop or the following crop.Because legume varieties and environmental factors can affect N absorption,fixation,and transfer in legume and non-legume intercropping systems,further studies with different varieties are needed in different environments to verify the present results.

Through path analysis,it was found that soil δ15N was mainly related to Ntransfer.There was no significant difference in Ntransferbetween the two experimental sites,suggesting that consistent results for Ntransfermay be found in different environments.Significant and negative correlation was found between soil mineral nitrogen and δ15N (Table 6),indicating that Ntransferwas higher when soil mineral N was lower.Other studies have found that the value of δ15N in soil increased with increasing N cycling rate and14N loss (Xiaoet al.2004).In addition,it was found that Ntransferin intercropping systems mainly occurred through root contact (Xiaoet al.2004).In the present pot experiment,the maize and legume roots were in close contact,which may be one reason leading to high%Ntransfer.

Table 6 Correlation between the total nitrogen,mineral nitrogen,and δ15N of the soil in monoculture and intercropping treatments in pot trials at two experimental sites1)

5.Conclusion

Intercropping with maize had little effect on the %Ndfaand decreased the shoot N accumulation and Ndfaof green manure but increased the LER values.The%Ndfavalues of green manures varied among species and environments;the %Ndfavalues of CV under both monoculture and intercropping systems were significantly lower than that of monocultured HV in Xining and were significantly higher than that of HV in both systems in Wuwei.The Ndfavalues of HV and CV in monocultured and intercropped systems varied from 0.59-1.16 g/pot to0.48-0.77 g/pot,respectively.Moreover,a high Ntransferin the maize/HV system was found in this experiment;the Ntransferfrom HV to maize and from CV to maize amounted to 39.61-46.22 mg/pot and 21.54-26.81 mg/pot,respectively.Path analysis showed that soil nutrients and shoot dry biomass played important roles in Ndfa,whereas soil δ15N was more important for Ntransfer.The results of this study may support the establishment of maize/leguminous green manure intercropping systems in northwestern China.

Acknowledgements

This work was financially supported by the China Agriculture Research System of MOF and MARA (CARS-22),the National Natural Science Foundation of China(32072678),the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences (CAAS-ASTIP-2020),the Chinese Outstanding Talents Program in Agricultural Science,the Protection and Utilization of Crop Germplasm Resources of China Green Manure (19200393) and the Fund Project of Qinghai Academy of Agricultural Sciences (2019-NKY-06).

Declaration of competing interest

The authors declare that they have no conflict of interest.