Guoqiang Zhang,Dongping Shen,Bo Ming,Ruizhi Xie,Xiuliang Jin,

Chaowei Liua,Peng Houb,c,Jun Xueb,c,Jianglu Chene,Wanxu Zhanga,b,c,Wanmao Liua,b,c,Keru Wangb,c,*,Shaokun Lia,b,c,*

a The Key Laboratory of Oasis Eco-agriculture,Xinjiang Production and Construction Group,College of Agronomy,Shihezi University,Shihezi 832000,Xinjiang,China

b Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,Beijing 100081,China

c Key Laboratory of Crop Physiology and Ecology,Ministry of Agriculture,Beijing 100081,China

d UMREMMAH,INRA,UAPV,84914 Avignon,France

e Research Institute of Agricultural Sciences,Division 6 of Xinjiang Production and Construction Group,Wujiaqu 831300,Xinjiang,China

Keywords:Irrigation frequency Soil m oisture Maize High yield(＞15 Mg ha-1)Water use efficiency

A B S T R A C T Worldwide,scarce water resources and substantial food demands require efficient water use and high yield.This study investigated w hether irrigation frequency can be used to adjust soil m oisture to increase grain yield and w ater use efficiency(WUE)of high-yield m aize under conditions of m ulching and drip irrigation.A field experim ent w as conducted using three irrigation intervals in 2016:6,9,and 12 days(labeled D6,D9,and D12)and five irrigation intervals in 2017:3,6,9,12,and 15 days(D3,D6,D9,D12,and D15).In Xinjiang,an optim al irrigation quota is 540 m m for high-yield m aize.The D3,D6,D9,D12,and D15 irrigation intervals gave grain yields of 19.7,19.1-21.0,18.8-20.0,18.2-19.2,and 17.2 Mg ha-1 and a WUEof 2.48,2.53-2.80,2.47-2.63,2.34-2.45,and 2.08 kg m-3,respectively.Treatm ent D6 led to the highest soil w ater storage,but evapotranspiration and soil-w ater evaporation w ere low er than other treatm ents.These results show that irrigation interval D6 can help m aintain a favorable soil-moisture environment in the upper-60-cm soil layer,reduce soilw ater evaporation and evapotranspiration,and produce the highest yield and WUE.In this arid region and in other regions with sim ilar soil and clim ate conditions,a sim ilar irrigation interval w ould thus be beneficial for adjusting soil m oisture to increase m aize yield and WUEunder conditions of m ulching and drip irrigation.

1.Introduction

China's food production is highly depend on irrigation[1],especially in north and northw est China,w hich have only 18%of China's w ater resources.Northern China accounts for as m uch as 65%of China's arable land[2]and 60%of the population,and groundw ater in this region has been severely over-extracted[3].In particular,in Xinjiang,w hich is a typical arid region of China,agricultural w ater accounts for over 90%of total w ater use.In arid and sem iarid areas,w ater shortage is the m ain factor lim iting crop yield.Deng et al.[4]reported an agricultural w ater-use efficiency(WUE)of 0.46 kg m-3in the north and northw est of China.Irrigation w ater faces the double lim itations of w ater shortage and low WUE.Thus,the developm ent of w ater-efficient agriculture and the im provem ent of crop yield and WUE have high potential as effective m easures to develop sustainable agriculture in irrigated areas.

Maize(Zea mays L.)is the m ost w idely grow n crop in China and plays an im portant role in ensuring China's food security[5].Improving crop yield per unit land area is a key to solving the problem of food security.In m odern m aize production,increased yields per unit area com e from increasing the optim um planting density[6,7].Also,in arid regions,w ater is the m ajor factor lim iting agricultural yield.Drip irrigation and plastic film m ulching are new agricultural w ater conservation technologies that have been w idely used in crop production in recent years.

The tim e interval betw een irrigation applications is a crucial factor for drip-irrigation m anagem ent because its affects soil-m oisture distribution,root distribution,w ater uptake by roots,and water percolation under the root zone[8-12].For these reasons,WUE and crop yield depend on irrigation interval and can thus differ even for the sam e total am ount of irrigation.At each irrigation,excessive or inadequate w ater application can influence both WUE and grain yield.Although som e studies[9,13]have show n that high irrigation frequency increases crop yield and WUE,others[14-16]have found that crop yield under low-frequency irrigation does not significantly differ from that obtained under highfrequency irrigation.This discrepancy m ay reflect the choice of different clim atic conditions and different crops for study.

Irrigation frequency can change the spatial distribution of soil moisture and soil-w ater storage[17].High-frequency drip irrigation (once every three days)produced higher soil m oisture in the 0-20 cm soil layer than in the deep soil layer,w hereas low-frequency drip irrigation(once every 10 days)favored w ater infiltration and lateral infiltration:deep soil m oisture w as higher,but the w ater supply w as not tim ely and surface soil m oisture w as low er.Overall,m ediumfrequency drip irrigation(once every seven days)w as beneficial to w ater infiltration and lateral infiltration.Mediumfrequency drip irrigation is conducive to a uniform distribution of w ater in the soil profile[18].Low irrigation frequency corresponds to excessively long irrigation intervals and may cause w ater stress,especially in sandy soils.It can also lead to substantial percolation below the root zone during irrigation because the irrigation am ount at each irrigation m ay exceed the soil water-storage capacity.In contrast,an excessively high irrigation frequency m ight lead to desirable conditions for w ater uptake by roots,but at the price of increased energy and labor costs.If the irrigation am ount is too sm all,soil w ater is distributed m ainly on the surface and is insufficient to m aintain crop grow th,thereby causing w ater stress and increasing soil-w ater evaporation[19].A suitable irrigation frequency can establish a balance betw een soil m oisture and oxygen conditions in the crop root zone,reduce root soaking,and m aintain a high soil m atric potential in the rhizosphere to reduce plant w ater stress throughout the grow ing season[20].Thus,optim izing irrigation frequency and w ater-application rate could help m axim ize crop yield and WUE[21].The observation that several previous studies[22-25]have show n that high yield and WUE result from suitable irrigation frequency indicates that a frequent and uniform w ater supply is im portant for m eeting the w ater requirem ents of plants and m axim izing crop yield and WUE.

Drip-irrigation frequency affects soil-w ater distribution,and high-frequency irrigation increases potato tuber grow th and WUE [12].In salt-affected soil,a once-in-five-days frequency of drip irrigation under m ulch leads to the m ost suitable soil-m oisture range for cotton,w hereas high-frequency(5 days)irrigation prom otes soil-salt leaching in the root zone[26].How ever,in sandy drip-irrigated soil,m aize yield and WUE increase w ith increasing irrigation frequency and rate[27].Subsurface drip-irrigation frequency does not affect m aize production w ith deeper silt loam soils[14,28].In norm al years,irrigation frequency has no effect on grain yield under subsurface drip-irrigation;however,in dry years(w ith a seasonal precipitation of 415 m m or less)a high irrigation frequency can result in greater grain yield[21].The differences in these results probably reflect differences in crops,irrigation technology,irrigation rate,climatic conditions,and/or soil texture.Thus,optim um irrigation frequency is not a constant,but is affected by soil conditions,clim ate conditions,irrigation conditions,and crop factors.

Our previous study[29]in Xinjiang investigated high-yield m aize-irrigation technology and yielded an optim um irrigation quota of 540 m m for a high yield of 17.4 Mg ha-1of m aize w ith an irrigation interval of 9 days.To further explore and perfect w ays to increase m aize yield and WUE based on this irrigation quota,w e hypothesized that increasing the irrigation frequency w ould further increase m aize yield and WUE.The objectives of the present study were(1)to clarify how irrigation frequency affects soil m oisture and(2)to determ ine w hether irrigation frequency affects grain yield and WUE for m aize.The results provide new insights into crop yield and WUEimprovement in sem i-arid and arid regions that should be helpful for im proving irrigation technology and designing irrigation system s.

2.Materials and m ethods

2.1.Experimental region and site

A tw o-year field experim ent w as conducted from April to Septem ber in 2016 and 2017 at the Qitai Farm Experimental Station of the Chinese Academ y of Agricultural Sciences(Xinjiang,China,43°50′N,89°46′E,altitude:1020 m).The soil in the field is sandy loam w hose m ain properties in the top layer(0-100 cm)are presented in Table 1.The average p H of the soil is 7.8 and the w ilting point is 8.6%(g g-1).The upper-60-cm soil profile contained 14.9 g kg-1of organic matter,1.46 g kg-1total N,99.7 m g kg-1available K,and 49.7 m g kg-1available P.These physical and chem ical properties of the soil w ere m easured at the beginning of each field experim ent.

Table 1-Soil p hysical and hyd raulic param eters in the ex perim ental field in 2016 and 2017.

The climate in this region is characterized by m inim al rainfall and m any hours of sunshine.From 1997 to 2017,during the entire spring m aize grow ing season(April-Septem ber),the m ean precipitation w as 158.6 m m,the m ean daily temperature w as 18.8°C,the mean reference crop evapotranspiration w as 1386.0 m m,and the m ean total annual hours of sunshine w as 1693.3 h.The precipitation w as 208.2 m m(2016)and 166.0 m m(2017)during the grow th period of maize.Fig.1 shows climatic variables,including m ean m onthly precipitation distribution(Fig.1-a),m ean daily air tem perature,and daily reference evapotranspiration(ETo)(Fig.1-b)during the seasons from 1997 to 2017.The daily reference evapotranspiration ETow as determ ined using the FAO Penm an-Monteith m ethod[30].Table 2 gives the precipitation,average air tem perature,and sunshine hours during the 2016 and 2017 m aize-grow ing period.Meteorological data w ere obtained from meteorological stations located at the farm experim ental station.

2.2.Experiment design

The experim ent included three irrigation-frequency treatm ents in 2016 and five in 2017.The local irrigation interval w as used as the control interval(D9).The three irrigation intervals for 2016 w ere 6,9,and 12 days(labeled D6,D9,and D12).The five irrigation intervals for 2017 w ere 3,6,9,12,and 15 days(D3,D6,D9,D12,and D15).The total irrigation am ount w as the optimal amount(540 mm),as determined in a previous study [29]for drip irrigation w ith plastic-film m ulching system s in arid regions.One day after sow ing,15 m m of w ater w as applied to assure uniform,rapid germ ination.To prevent late lodging and to harden seedlings,no irrigation w as applied from sow ing to 60 days after sow ing.Table 3 describes the irrigation strategies.

2.3.Irrigation system and agronomic practices

Zhengdan 958(ZD958)and Xianyu 335(XY335)are tw o comm only planted,high-yield,density-tolerant maize hybrids in China.ZD958 and XY335 w ere used in 2016,and XY335 w as used in 2017.Maize w as sow n on April 18,2016 and April 21,2017 and harvested on October 18 in both 2016 and 2017.The planting density w as 12×104plants ha-1.Plants w ere seeded in alternating w ide and narrow row s at an alternating row spacing of 40-70-40 cm,and the spacing betw een plants w ithin a row w as 15 cm.Surface drip irrigation and plastic-film mulching were applied,and a com bination planter[29]w as used to apply drip tape and plastic film,punch holes,and m anually sow.The area of each plot w as 66 m2(10 m by 6.6 m).Each irrigation treatm ent included three replications.Water movement between plots w as prevented by w aterproof m em branes buried at a depth of 1 m below the soil surface betw een each plot and by 1-m-w ide buffer zones betw een plots.

Fig.1-Changes in precipitation,temp erature and reference evapotranspiration during spring maize grow ing seasons in Qitai for the period 1997-2017.(a)Mean m onthly precip itation;(b)m ean daily tem p erature and d aily reference evap otranspiration.

Maize w as drip-irrigated using w ater pum ped from groundw ater[29].The drip irrigation system included singlew ing drip tape(Tianye Inc.,Shihezi,China)placed in the m iddle of each narrow row.The em itter spacing w as 30 cm and the flow rate w as 3.2 L h-1at an operating pressure of 0.1 MPa.Careful design and management led to stable discharge and pressure.Each plot w as connected to a highprecision w ater m eter(LXS-25F,Ningbo,China)and control valve.

Table 2-Precip itation,average tem perature,and sunshine hours during 2016 and 2017 m aize-grow th period.

Before sow ing,base fertilizers w ere applied at concentrations of 150 kg ha-1N(as urea),225 kg ha-1(NH4)2HPO4(am m onium phosphate),and 75 kg ha-1K2O(potassium sulfate).An additional 600 kg ha-1urea w as applied during the grow ing stage to ensure an adequate supply of nutrients.Chem ical control(DA-6 Ethephon,China Agrotech,Shanxi,China)w as applied at 600 m L ha-1in the V8-V10 period of maize.All weeds,diseases,and pests in the experimental plots w ere controlled.

2.4.Sampling and measurements

2.4.1.Measurement of soil-moisture content

Soil-moisture content in 20-cm-thick soil layers(0-100 cm)w as m easured using the oven-drying m ethod and a tim edom ain reflector(TDR,TRIME-T3,Germ any).Five 100-cm-long tubes w ere deployed under the drip tape in all treatm ents after sow ing and in each season.Samples were collected before sow ing and physiological m aturity,after rainfall,and one day before and after irrigation.Before sow ing and physiological m aturity,soil-m oisture content w as m easured using the oven-drying method.

2.4.2.Evapotranspiration

Maize actual evapotranspiration ETc(m m)w as calculated during the growing season using the soil w ater balance equation[29]:

w here ETcis evapotranspiration(m m)during the grow ing season,I is the am ount of irrigation w ater applied(m m),P is precipitation(mm),Cris capillary rise(m m),Dpis percolation(m m),Rfis runoff(m m),andΔS is the change in soil-w ater storage(m m).

In Eq.(1),Cris considered to be zero because the groundw ater table w as 70 to 80 m below the surface;runoff is also assum ed to be insignificant because the field w as flat,and Dpis considered negligible because the soil-w ater content below 100 cm did not reach field capacity(FC)on any sampling date.

2.4.3.Soil-water evaporation

Soil-w ater evaporation Esw as m easured w ith a m icrolysim eter(MLS)[31-33].The MLS consisted of tw o parts:an outer cylinder and an inner cylinder.The inner cylinder was m ade of steel pipe 10 cm in inner diam eter and 15 cm long w ith a w all thickness of 1 m m.The outer cylinder w as m ade of a PVC cylinder w ith inner diam eter of 12 cm and length 15 cm.An electronic balance(ES6000,D&T,Tianjin,China)w ith a precision of 0.01 g w as used for m easuring m ass.Three MLSs w ere placed in each plot and w eighing w as perform ed every day around sunset.The difference in w eight over 2 days is the am ount of evaporation,and a 1-g change in w eight in the MLS corresponded to an evaporation of 0.127 m m.For each m easurem ent,the inner cylinder w as forced vertically into the soil and w ithdraw n to rem ove the soil.The bottom w as sealed w ith alum inum foil and the soil w as w eighed,after w hich the inner cylinder w as placed in the outer cylinder fixed betw een the w idely spaced m aize row s.To m aintain the same soil moisture conditions in each plot,the original soil in the MLS w as replaced every tw o days.The soil w as replaced after precipitation or irrigation events.

The soil-w ater evaporation Esper unit tim e can be calculated follow ing[34,35]

where Esis soil-w ater evaporation(mm),αis a conversion factor(0.127 mm),andΔm(g)is the mass difference between MLS in one unit of tim e.

2.4.4.Water-use efficiency

Water-use efficiency w as calculated as the ratio of grain yield to the total evapotranspiration for the w hole season[36,37]:w here GY is grain yield(kg ha-1)and ETcis total evapotranspiration(m m)calculated from Eq.(1).

Table 3-Irrigation schedule for irrigation interval treatm ents in 2016 and 2017.

2.4.5.Grain yield

At physiological maturity,an area of 13.2 m2(central six row s of each plot,4 m long)from three plots w as harvested m anually and the grain m ass w as m easured.The plants and ears w ere counted and the number of ears per plant w as determ ined.Kernel num ber per ear and 1000-kernel w eight w ere m easured for 20 representative ears per plot.Grain m oisture content w as determ ined w ith a portable m oisture m eter(PM8188,Kett Electric Laboratory,Japan).Grain yield w as expressed at 14%m oisture content and used for statistical analyses and calculation of WUE.

Fig.2-Soil-w ater storage as a function of d ays after sow ing during the m aize grow ing season in 2016 and 2017.

Fig.3-Effect of irrigation intervals on soil-w ater storage.The three broken lines represent the soil-w ater storage for different soil horizons corresp onding to the 70%field capacity(SS70%,m ed ium-d ashed line),60%(SS60%,short-dashed line),and w ilting point(SSw p,d otted line).

2.5.Statistical analysis

Calculations w ere perform ed and charts w ere prepared using Microsoft Excel 2013(Microsoft Corporation,Redmond,Washington,USA)and SigmaPlot 12.5(Systat Softw are Inc.,San Jose,California,USA).Analysis of variance w as used to test for differences in yield,WUE,and ETcas a function of irrigation frequency.Correlation analysis w as perform ed w ith SPSS 18.0(SPSS Inc.,Chicago,Illinois,USA)to determ ine the relationships betw een WUE and irrigation frequency and between WUE and ETc.Means were compared using Fisher's least significant difference tests w ith P＜0.05(LSD0.05).

3.Results

3.1.Soil-w ater storage

Irrigation frequency significantly affected soil-w ater storage(SS)during the irrigation period(Fig.2).Irrigation began at 61 days after sowing(DAS),and the SS of the 0-60 cm soil layer for each treatm ent varied great before and after irrigation.The range of variation increased w ith irrigation interval length.The SS in the 60-100 cm soil layer was relatively stable.Average SS values in the 0-60 cm soil layer for D3,D6,D9,D12,and D15 w ere 121.0,134.2-148.9,127.6-138.1,125.4-137.7,and 117 m m,respectively.In the 60-100 cm soil layer,the average SSvalues for intervals w ere 80.3,89.6-94.0,83.7-89.0,79.9-86.2,and 74.7 m m.SSw as higher in 2016 than in 2017,and no great difference in SS occurred betw een XY335 and ZD958.In the 0-60-cm soil layer,the D6 irrigation treatment average SSvaried from 134.2 to 148.9 mm throughout the irrigation period.

3.2.Effect of irrigation interval on soil w ater storage

During the irrigation period(61-142 DAS),SS varied in the 0-60 cm soil layer during the different grow th stages(Fig.3).Soil-w ater storage under treatm ent D6 exceeded that under the other treatm ents at all grow th stages.Low er irrigation frequency corresponded to greater fluctuation in SS.We used SS70%as the ideal low er lim it and SS60%as the low er lim it for m ild w ater stress,soil w ater storage status can also reflect how irrigation frequency affects soil-m oisture status in different grow th stages.Table 4 show s,for each irrigation treatment,the percent of days during the irrigation period that the SS attained the given levels.

The SSwpw as 66.3 m m.The num ber of days w ith SS exceeding SS70%for treatm ent D6 w as significantly greater than that under the other treatm ents.At SS60%,SS for treatm ents D12 and D15 declined significantly in the later period in 2017,indicating that these treatm ents led to som e w ater deficit.

3.3.Soil-w ater evaporation

Irrigation frequency significantly affected soil-w ater evaporation Es(Fig.4).During 61-76 DAS,the average daily Esof D6,D9,D12 w ere 1.12,1.16,and 1.18 m m day-1in 2016,and in2017 the average daily Esfor treatments D3,D6,D9,D12,and D15 w ere 1.21,1.1,1.15,1.23,and 1.29 m m day-1.No significant difference in Esbetw een XY335 and ZD958 w as observed in 2016.

Table 4-Percent of d ays w hen soil w ater storage w as at a d ifferent w ater storage horizon d uring irrigation period.

3.4.Total soil-water storage

From 61 to 160 DAS,ETc,SS(Fig.5),and Es(Fig.6)differed significantly as a function of irrigation frequency.Esfor treatm ent D6 w as great low er than for the other irrigation treatm ents,but SS w as higher than for the other irrigation treatm ents.The results show that irrigation treatm ent D6 reduced Esand thus ETc.A suitable irrigation frequency can m aintain SSat a relatively high level and m aintain a favorable soil-m oisture environm ent.

3.5.Grain yield,water-use efficiency,and evapotranspiration

Treatment D6 achieved high yield(19.1-21.0 Mg ha-1)and WUE(2.53-2.80 kg m-3)over both grow ing seasons(Table 5).The WUEfor D6 w as 8.3%,4.8%,11.7%,and 28.7%greater than that for D3,D9,D12,and D15,respectively.Maize yield and WUE are quadratically related to irrigation interval:ygrainyield=-0.0316x2+0.3391x+19.153, R2=0.701**; yWUE=-0.0082x2+0.1096x+2.2743,R2=0.808**(w here x is irrigation interval).Evapotranspiration ETcw as significantly greater for treatm ent D15 than for the other treatm ents,and ETcw as significantly lower for treatment D6 than for the other treatm ents.Maize yield did not increase w ith ETc,but WUE decreased as ETcincreased.Maize yield and WUE w ere exponentially related to ETc:ygrainyield=80.83e-0.002x,R2=0.508*;yWUE=28.107e-0.003x,R2=0.746**(x is ETc).

Com paring the sam e treatm ents,grain yield w as 1.9%higher in 2017 than in 2016.In 2016,XY335 yield w as significantly greater(5.2%)than ZD958 yield.This difference in yield m ay reflect the m ore num erous rainy days in 2016 and reduced sunshine tim e during the grain-filling stage.Differences in yield betw een cultivars reflect m ainly the cultivar attributes.

4.Discussion

Irrigation frequency changed the spatial distribution of soil m oisture and SS,w hen irrigation frequency increases,the upper-layer soil m oisture increases w ithin a certain range,and the up per soil layer storage increases.The appropriate irrigation frequency may thus increase the SS capacity.We reached conclusions similar to that of studies[12,17].In the p resent study,the soil m oisture w as concentrated m ainly in the 0-60 cm soil layer,and increasing the irrigation frequency reduced soil-m oisture fluctuations in this upper soil layer.Treatm ent D6 m aintained a high soil-m oisture content throughout the irrigation period.Soil-w ater storage in late 2016 w as higher than that in 2017,m ainly because of higher SS before sow ing in 2016,higher rainfall in July and August,m ore cloudy days and less sunshine.These factors inhibited soil-w ater evaporation and evap otranspiration and reduced the w ater consum ption of m aize,resulting in higher SS.

Fig.4-Soil-w ater evap oration under different irrigation intervals from 61 to 76 DAS in 2016 and 2017.

Different irrigation frequencies also lead to different SS.Studies[38]have show n that,for m ulching plus drip irrigation in arid areas,the low er lim it of irrigation is 65%FC,and obtained high yield.Previous studies[39-41]show ed that 70%FC is the soil moisture content suitable for the key w ater-requirement period for m aize.The classification of soil w ater stress and the determ ination of the low er lim it of irrigation are affected by m any factors,including soil environm ent,climate,m ulching,irrigation amount,and irrigation frequency.Irrigation treatm ent D6 m aintained a high soil-m oisture regim e,w hereas the other treatm ents caused m ild w ater stress.Thus,in future research,soil-w ater stress should be further analyzed and tested at different grow th stages.

Fig.5-Evapotransp iration and soil w ater storage in the 0-100 cm soil layer at 61-160 days after sow ing and for d ifferent irrigation intervals.ETc,crop evapotranspiration;SS,soil w ater storage at 160 DAS.Means follow ed by different low ercase letters are significantly different at P＜0.05.

Treatm ent D6 led to the highest SS,but evapotranspiration and Esw ere low er than other treatm ents.Thus,this irrigation treatm ent schedule balanced irrigation am ount and the physiological and ecological w ater consum ption of m aize,thereby m aintaining a favorable soil w ater environm ent.The negative correlation betw een soil-w ater storage and evapotranspiration indicates that a suitable irrigation frequency can increase the total am ount of w ater in the soil[17,20].

Reducing Esis im portant for im proving WUE and saving w ater.The average Esw as low est for treatm ent D6 from 61 to 76 DAS,and ranking the treatments in term s of Esgives D15＞D12＞D3＞D9＞D6.Thus,Esis one possible reason for the difference in soil m oisture.The m ain factor is irrigation frequency:a single,high-volum e irrigation leads to high Es.Although film and m ulching reduce Esby 55%,uncovered soil(betw een w idely spaced row s)still experiences copious evaporation.Soil-w ater evaporation and irrigation am ount correlate positively and increase exponentially w ith surface soil m oisture content[42,43].A high irrigation frequency leads to more evaporation because of the high long-term w ater content at the soil surface,resulting in low er w ater storage in the soil layer[17].Our findings follow ed these rules.An excessively high irrigation frequency causes a high Es,as show n in the schem atic diagram of Fig.7.Thus,a suitable irrigation frequency reduces Esand increases WUE.These results show that an irrigation frequency that is too high or too low leads to ineffective soil-w ater evaporation.By reducing ineffective Es,a suitable irrigation frequency thus helps maximize the use of soil moisture by crops.

Fig.6-Soil-w ater evap oration at 61-160 days after sow ing and for different irrigation intervals.

Table 5-Grain yield,crop evapotransp iration,and w ater-use efficiency of m aize for different irrigation treatm ents in 2016 and 2017.

Drip-irrigation plastic-film m ulching is an effective w ay to save w ater,increase production,and im prove WUE[44,45].The grain yield of D6 w as 20.7%greater than that(17.4 Mg ha--1)of Zhang et al.[29]for the sam e area and conditions in Xinjiang.The previous study used nine-day irrigation intervals(D9).Thus,reducing the irrigation interval to six days increased the m aize yield.Other w ork that supports this conclusion includes that of Irm ak et al.[21],w ho show ed that irrigation frequency affected grain yield significantly in dry years,and that high irrigation frequency led to higher grain yield(14.7 Mg ha-1).Yazar et al.[22]reported that m aize yields varied from 7.9 to 11.3 Mg ha-1and 7.3 to 11.9 Mg ha-1for three-and six-day irrigation intervals,respectively.Thus,a suitable irrigation frequency leads to high yield,in agreem ent w ith the results of previous studies[22,46].How ever,a suitable irrigation frequency is affected by soil texture,w eather condition,rainfall and irrigation rate.Thus,our findings m ay be applied in other arid regions w ith sim ilar soil and clim ate conditions.

Fig.7-Schem atic diagram of soil-m oisture environm ent.

High WUE(a m ean of 2.62 kg m-3)w as obtained in this study by application of a suitable irrigation am ount(540 m m),w hich reduces evapotranspiration for m ulching plus drip irrigation[29].In this arid region,evapotranspiration is high,and irrigation intervals in local production are 9-15 days.How ever,the current irrigation interval is too long,low ering the yield of m aize and the WUE.To further im prove the yield and WUEof m aize,w e ran field trials for tw o years to identify the highest WUE(2.80 kg m-3)for a suitable irrigation interval(6 days)and am ount(540 m m).The WUE im proved by 6.9%com pared w ith previous w ork.In the present study,the WUE for treatm ent D6 w as 6.5%and 12.9%greater than that for D9 and D3,respectively.We conclude that a suitable irrigation frequency leads to a high WUE,as in previous studies[9,22,25].

Low WUEis a comm on problem.How ell et al.[14]reported m aize WUE from 1.08 to 1.62 kg m-3,and Yazar et al.[22]reported that a six-day irrigation interval gave a high WUE(2.27 kg m-3).Bozkurt et al.[47]obtained a WUEof 1.4 kg m-3,and Kuscu et al.[48]reported a WUEfrom 1.40 to 1.93 kg m-3.Ham m ad et al.[49]obtained a m aize WUE of 1.04 to 1.55 kg m-3,and Zhao et al.[50]obtained a WUEof 1.84 kg m--3.How ever,com pared w ith these studies,w e obtained a higher WUE.The main reasons behind this result are the use of dense planting,plastic film m ulching,drip irrigation,and an irrigation frequency increasing grain yield and reduce evapotranspiration.Thus,a suitable irrigation frequency leads to both high yield and high WUE.

Increased WUE can be achieved by coordinating m aize yield and evapotranspiration.Maize yield can be optim ized by use of high-yield hybrids,dense planting,mulching,drip irrigation,and w ater and fertilizer integration technologies.Reducing evapotranspiration can also increase WUE.Previous studies have show n that soil-w ater evaporation can be reduced by m ulching[51-53],straw m ulching[54],deficit irrigation[22,36,55],optimizing irrigation[29],and optimizing irrigation frequency[56,57]to increase WUE.We also conclude from the present study that a suitable irrigation frequency(D6 in this case)is conducive to reducing soil-w ater evaporation and im proving WUE.

A shortcom ing of the present study w as that soil w ater stress w as not recorded at different grow th stages.Future studies should investigate how irrigation frequency affects dry m atter production and photosynthesis in m aize.Soil w ater stress and grading should also be recorded at different grow th stages.Further study m ay show that tuning the irrigation frequency and irrigation am ount at different grow th stages can further increase grain yield and WUE.

5.Conclusions

Soil m oisture,soil-w ater evaporation,evapotranspiration,yield,and WUE w ere investigated as a function of irrigation frequency and am ount for drip irrigation w ith plastic-covered m ulch.Given an irrigation quota of 540 m m,the optim um irrigation interval(6 days)helped the upper 60 cm soil layer m aintain high w ater storage,producing a soil m oisture environm ent favoring m aize grow th,reducing soil-w ater evaporation and ETc,and thereby increasing m aize yield and WUE.Grain yield reached 19.1 to 21.0 Mg ha-1and WUE reached 2.53 to 2.80 kg m-3.A suitable irrigation frequency thus helps optim ize soil m oisture and thereby increase m aize yield and WUE.Adjustm ent of an irrigation frequency m atched to the regional evapotranspiration is thus conducive to im proving WUE.Sim ilar m anagem ent practices m ay be applied in other arid regions w ith sim ilar soil and clim ate conditions.

Conflict of interest

The authors have declared that no conflict of interest.

Acknow ledgm ents

We are grateful for research support from the National Key Research and Developm ent Program of China(2016YFD0300110,2016YFD0300101);the National Basic Research Program of China(2015CB150401);the National Natural Science Foundation of China(31360302);the Science and Technology Program of the Sixth Division of Xinjiang Construction Corps in China(1703);and the Agricultural Science and Technology Innovation Program for financial support of this study.We w ould also like to thank the review ers for helping us improve our original manuscript.