The first factor affecting dryland winter wheat grain yield under various mulching measures: Spike number

Yingxia Dou ,Hubing Zhao # ,Huimin Yang ,Tao Wang,Guanfei Liu,Zhaohui Wang,Sukhdev Malhi

1 College of Natural Resources and Environment, Northwest A&F University, Yangling 712100, China

2 Key Laboratory of Plant Nutrition and the Agri-Environment in Northwest China, Ministry of Agriculture and Rural Affairs, Yangling 712100, China

3 Department of Renewable Resources, University of Alberta, Edmonton T6G 2H1, Canada

Abstract Water is the key factor limiting dryland wheat grain yield. Mulching affects crop yield and yield components by affecting soil moisture. Further research is needed to determine the relationships between yield components and soil moisture with yield,and to identify the most important factor affecting grain yield under various mulching measures.A long-term 9-year field experiment in the Loess Plateau of Northwest China was carried out with three treatments:no mulch (CK),plastic mulch (MP) and straw mulch (MS). Yield factors and soil moisture were measured,and the relationships between them were explored by correlation analysis,structural equation modeling and significance analysis. The results showed that compared with CK,the average grain yields of MP and MS increased by 13.0 and 10.6%,respectively. The average annual grain yield of the MP treatment was 134 kg ha–1 higher than the MS treatment. There were no significant differences in yield components among the three treatments (P<0.05).Soil water storage of the MS treatment was greater than the MP treatment,although the differences were not statistically significant. Soil water storage during the summer fallow period (SWSSF) and soil water storage before sowing (SWSS) of MS were significantly higher than in CK,which increased by 38.5 and 13.6%,respectively. The relationship between MP and CK was not statistically significant for SWSSF,but the SWSS in MP was significantly higher than in CK. In terms of soil water storage after harvest (SWSH) and water consumption in the growth period(ET),there were no significant differences among the three treatments. Based on the three analysis methods,we found that spike number and ET were positively correlated with grain yield. However,the relative importance of spike number to yield was the greatest in the MP and MS treatments,while that of ET was the greatest in CK. Sufficient SWSSF could indirectly increase spike number and ET in the three treatments. Based on these results,mulch can improve yield and soil water storage. The most important factor affecting the grain yield of dryland wheat was spike number under mulching,and ET with CK. These findings may help us to understand the main factors influencing dryland wheat grain yield under mulching conditions compared to CK.

Keywords: dryland winter wheat,plastic mulch,spike number,straw mulch

1.Introduction

Wheat is the world’s most important food crop (Yuet al.2021;Zhaoet al.2023),providing about 21% of the food for human consumption (FAO 2019). The northwest dryland is the main wheat growing region in China,with about 4.3 million hectares of land devoted to wheat cultivation (Zhao H Bet al.2019). However,drought is the main factor restricting crop yield on the Loess Plateau,China (Tanget al.2019). The main water source for wheat growth is natural precipitation (Li Ret al.2013).However,the period of intensive rainfall is seriously out of synch with the wheat growth period. This has led to a decline in wheat production,which is failing to meet the needs of a rapidly growing economy (Fanget al.2022).Therefore,some measures need to be taken to ensure food security (Rempeloset al.2020).

Mulching to improve the soil water storage capacity is an effective measure for achieving high yield in wheat (Sunet al.2020). Mulching with plastic or straw can reduce soil surface evaporation and increase leaf productive transpiration compared to no mulch (Dinget al.2017). The increase in transpiration means greater carbon dioxide assimilation and nutrient uptake (Senayet al.2011;Bastiaanssenet al.2012). Therefore,mulching with plastic or straw provides more water for leaf transpiration,thereby increasing grain yield. Plastic mulch can also raise the soil temperature (Hanet al.2013;Moet al.2016;Prosdocimiet al.2016),increase soil water storage (Chakrabortyet al.2008;Kaderet al.2017),reduce soil water evaporation (Li S Xet al.2013)and improve the water use efficiency of crops (Almeidaet al.2015). At present,plastic mulch combined with hole planting and soil covering has become the main form of plastic mulch usage in northern China,which can lead to increases in wheat yield (Chaiet al.2022b).Straw mulch is a conservation tillage measure for saving water and increasing yield. Straw mulch increases evapotranspiration and reduces soil evaporation (Penget al.2020),loosens the soil surface to facilitate rainwater infiltration (Jordánet al.2010;Parhizkaret al.2021),and enhances the soil water storage capacity (Cooket al.2006;Wanget al.2018). Straw mulch can prevent direct solar radiation from reaching the ground,reduce the heat supply driving water evaporation (Liuet al.2014;Akhtaret al.2019),and effectively improve water storage efficiency (Donget al.2018).

Mulching affects soil temperature (Yang Jet al.2018;Yuet al.2018),with higher temperatures increasing evaporation and lower temperatures decreasing evaporation intensity,thus affecting soil moisture. Due to the blocking effect of the plastic film,the heat of the soil is dissipated very slowly,which can effectively convert solar energy into heat (Ganet al.2013). Moreover,the use of plastic film also increases greenhouse gas emissions (Tsiropouloset al.2015),resulting in higher soil temperatures. However,straw has a strong shading property,which reduces the absorption and conversion of solar radiation as well as heat conduction (Chenet al.2007). The air permeability of the straw makes it easier for heat to escape from the soil surface,therefore it enhances the cooling effect. In a study on the effect of plastic mulch on ecophysiology and yield,Wanget al.(1998) emphasized that plastic mulch increased the soil temperature,and deep soil moisture moved upward through the temperature gradients to supply the wheat. Liet al.(2021) found that straw mulch reduced the temperature of the 0–25 cm soil layer and reduced soil evaporation,and their combined effects promoted the formation of panicle number in winter wheat. Mulch first affects soil moisture and temperature,which in turn affects the yield components. Plastic mulch increased the soil water content and soil temperature,thus promoting increases in wheat spike number and grain number per spike (Zhanget al.2022). Yang Y Het al.(2018) showed that compared with no mulch,straw mulch increased spikelet number and grain number per spike,and reduced thousand-kernel weight in 2014–2015. These three components of yield have competitive relations,so a change in one factor is often accompanied by changes in the other two factors (Slaferet al.2014). However,further exploration is needed to determine the relationships between soil water and yield components with grain yield,and to identify the factor responsible for the grain yield increase under mulching conditions. Therefore,the three different methods of correlation analysis,structural equation models,and importance ranking were used in this study to comprehensively explore the key factor affecting dryland wheat grain yield.

In this study,a long-term positioning experiment with different mulching measures was conducted in the rainfed region of the Loess Plateau for 9 years. The main objectives of this work were: (i) to compare the effects of plastic mulch and straw mulch on winter wheat grain yield,yield components and soil moisture,and (ii) to identify the main factor leading to grain yield variations by comprehensively considering the relationships between grain yield,its components and soil moisture through three analytical methods.

2.Materials and methods

2.1.Site description

The long-term field experiment was conducted from 2012 to 2021 in Yongshou County (35°3′N,108°3′E),Shaanxi Province,which is situated in the dryland winter wheat planting area of the Loess Plateau of Northwest China.The site is 995 mm above sea level and has a semi-humid continental monsoon climate in the warm temperate zone,with an average annual temperature of 10.9°C,annual rainfall of 496 mm and potential evaporation of 807 mm. With no irrigation source,the agriculture in this area is typically rain-fed. The soil at the experimental site was silty loam,as classified by the USDA soil textural classification system. At the beginning of the experiment,the basic properties of the 0–20 cm surface soil were as follows: pH,8.2;organic matter,11.7 g kg–1;nitrate nitrogen,14.5 mg N kg–1;ammonium nitrogen,2.7 mg N kg–1;total nitrogen,0.87 g N kg–1;available phosphorus,10.7 mg P kg–1;and available potassium,99.9 mg K kg–1.Taking increments of 20 cm as layers,the soil bulk density was determined in the 0–200 cm soil profile as 1.25,1.37,1.34,1.38,1.36,1.49,1.43,1.44,1.44 and 1.35 g cm–3,respectively.

2.2.Experimental design

The experiment was a 9-year long-term positioning experiment,and three treatments were set up: no mulch(CK),plastic mulch (MP) and straw mulch (MS). A random block arrangement was adopted with four replicates,and the plot area was 48 m2(12 m×4 m). The same fertilizer amount and field management measures were applied to each treatment in the experiment. During the growth of wheat,pesticides were used to control weeds and pests as needed. Weeds and insect pests did not significantly affect the growth of the winter wheat.

The winter wheat cultivar was Luohan 6,which was sown manually at a sowing rate of 150 kg ha–1. The same amount of fertilizer was applied to all treatments as 150 kg N ha–1in the form of urea and 127.5 kg P2O5ha–1in the form of superphosphate. Fertilizer was applied as base fertilizer in one application before sowing the winter wheat. The CK treatment was conventional flat planting,with row spacing of 20 cm. In the MPtreatment,the soil surface was covered with 1 m wide agricultural common white film (0.008 mm thickness),which was then covered with 1 cm of soil. About 10 seeds were sown in holes at a spacing of 12 cm,with rows spaced 20 cm apart. The MStreatment was a traditional flat cropping,and had row spacing of 20 cm. After sowing,the surface was evenly covered with 10,000 kg ha–1wheat straw.

2.3.Sampling and measurements

Fresh soil samples were collected from the 0–200 cm depth in layers of 20 cm intervals before sowing and after harvest of the winter wheat,and soil water content was measured by the drying method (Yanet al.2018).At the wheat harvest stage,4 m2of plants with uniform growth were randomly selected from each plot,and the aboveground parts were harvested and used to calculate grain yield per hectare (ha).

2.4.Precipitation

The annual precipitation in the experimental area is mainly concentrated in July–September,i.e.,the summer fallow period (Table 1). From 2012 to 2021,the rainfall levels in the summer fallow periods were 292.5,205.3,296.5,224.2,157.9,152.0,277.0,303.0,and 335.3 mm,respectively. The rainfall amounts during the growth period were 146.3,332.1,342.2,190.4,307.7,327.4,145.9,262.6,and 127.8 mm,respectively.

Table 1 Precipitation (mm) from 2012 to 2021 at the experimental site

2.5.Calculations

Determination of soil moistureSoil water content (ω,%)was measured in 10 soil layers (0–20 cm,20–40 cm,40–60 cm,60–80 cm,80–100 cm,100–120 cm,120–140 cm,140–160 cm,160–180 cm,and 180–200 cm) before sowing and after harvesting. Soil samples were randomly collected from each plot using a steel-core auger with a diameter of 4.5 cm. Soil moisture content was calculated by drying at 105°C for 48 h as follows (Yanet al.2018):

where W2is the weight of the aluminum box and fresh soil(g),W1is the weight of the aluminum box and dry soil (g),and W0is the weight of the aluminum box (g).

Soil water storageSoil water storage before sowing(SWSS,mm) of the 0–200 cm soil was calculated as follows (Zhanget al.2017):

where ωiis the soil water content (%),hiis the depth of soil layer (cm),ρiis the soil bulk density (g cm–3),andiis the soil depth (0–20,20–40,40–60,60–80,80–100,100–120,120–140,140–160,160–180,and 180–200 cm).

Soil water storage after harvest (SWSH,mm) of the 0–200 cm soil was calculated as follows (Zhanget al.2017):

Soil water storage during the summer fallow period(SWSSF,mm) of the 0–200 cm soil was calculated as follows (Zhao H Bet al.2019):

where SWSSais the soil water storage before sowing(mm),SWSHa–1is the soil water storage after harvest(mm),a is the current growth period,and a–1 is the previous growth period.

Soil water consumptionThis field experiment was located in the Loess Plateau without irrigation conditions.The groundwater level in that area is deep and is not easy to elevate or recharge. At the same time,the experimental site was located in a flat area with little surface runoff. Therefore,water consumption in the growth period (ET,mm) was calculated as follows (Liet al.2004):

where SWSS is the soil water storage before sowing (mm),SWSH is the soil water storage after harvest (mm),and P is the precipitation during the growth period (mm).

2.6.Statistical analyses

Excel 2010 software was used for data processing.Statistical analyses were performed using DPS 7.05.The LSD method was used to test the significance of differences between treatments,and the significance level was set as α=0.05. All figures were drawn using Origin Pro 2021.

Pearson correlation analysis was performed using SPSS 26 (SPSS Inc.,Chicago,IL,USA). A greater absolute value of the correlation coefficient in the Pearson correlation analysis means a greater correlation between the two factors. Structural equation modeling (SEM) was conducted using the lavaan package in the R software(v4.1.1,R Foundation for Statistical Computing,Vienna,Austria) (Breiman 2001). The SEM is a statistical method based on the covariance matrix of variables that is used to analyze the relationships between variables,and it can consider and deal with multiple factors simultaneously. Random forest model construction and variable importance ranking were carried out using the randomForest package in the R software (v4.1.1,R Foundation for Statistical Computing,Vienna,Austria).The %IncMSE was used as the criterion for evaluating the relative importance of factors affecting wheat grain yield.The %IncMSE refers to the degree to which a variable is changed into a random number to reduce the mean square error of the random forest model. The greater the change in value of the standardized mean error,the greater the importance of the variable in the establishment of the random forest model.

3.Results

3.1.Grain yield and yield components

The mulching measures had a significant effect on wheat grain yield (Fig.1-A). The average grain yields of MPand MSwere significantly higher than CK. The average grain yields of the CK,MPand MStreatments were 5,679,6,416 and 6,283 kg ha–1,respectively. Compared with CK,the MPtreatment increased the yield by 13.0% and the MStreatment increased it by 10.6%. There was no significant difference between the MPand MStreatments.The average annual grain yield of the MPtreatment was 134 kg ha–1higher than that of the MStreatment.

Fig. 1 Average wheat grain yield (A),spike number (SN) (B),grain number per spike (GS) (C),and thousand-kernel weight (GW)(D) in the nine growing seasons,under the no mulch (CK),plastic mulch (MP) and straw mulch (MS) treatments. Different lowercase letters indicate significant (P<0.05) differences between treatment means according to the LSD test.

Compared with CK,the spike number,grain number per spike and thousand-kernel weight were increased in the MPand MStreatments,but the differences were not statistically significant (Fig.1-B–D). The spike numbers in the MPand MStreatments increased by 8.1 and 8.5%compared with CK,respectively. The spike number of the MStreatment was 0.4% higher than that of the MPtreatment. Compared with CK,the MPand MStreatments increased grain number per spike by 3.0 and 0.2%,respectively. The MPtreatment increased grain number per spike by 2.7% compared with the MStreatment. The thousand-kernel weights of the MPand MStreatments were 2.3 and 0.1% higher than that of the CK treatment,respectively. The MPtreatment increased the thousandkernel weight by 2.2% compared with the MStreatment.

3.2.Soil water storage

The average SWSSF,SWSS,and SWSH as well as ET tended to be higher under the MPand MStreatments than under CK. The values for the MStreatment tended to be greater than the MPtreatment in all cases (Fig.2). The average SWSSF of MSwas significantly higher than CK,but the difference was not significant between MPand CK.Compared with CK,the average SWSSF in the MPand MStreatments increased by 23.9 and 38.5%,respectively.Compared with MP,the MStreatment increased the average SWSSF by 11.8%,but the difference was not significant.

The average SWSS was significantly higher in MPand MSthan CK. Compared with CK,the average SWSS in MPand MSincreased by 8.6 and 13.6%,respectively.However,there was no difference in SWSS between MPand MSin the 9-year mean (P>0.05),and MSwas only 21.8 mm higher than MP.

The ranges of variation in SWSH of the CK,MPand MStreatments were 255.0–393.7 mm,264.0–404.1 mm and 248.0–441.1 mm,respectively. But there was no significant difference among the treatments. Compared with CK,the MPand MStreatments increased the average soil water storage by 12.1 and 16.5 mm,respectively,and the MStreatment was 4.4 mm higher than MP.

Fig. 2 Average soil water storage during the summer fallow period (SWSSF) (A),before sowing (SWSS) (B) and after harvest(SWSH) (C),and water consumption in the growth period (ET) (D) in the nine growing seasons,under the no mulch (CK),plastic mulch (MP) and straw mulch (MS) treatments. Different lower-case letters indicate significant (P<0.05) differences between treatment means according to the LSD test.

The ranges of variation in ET of the CK,MPand MStreatments were 173.4–506.7 mm,166.7–546.0 mm and 193.6–533.6 mm,respectively,but the differences were not significant among the treatments. Mean water consumption levels in the growth period in all treatments were essentially similar (P<0.05). Compared with CK,the average water consumption of MPand MSincreased by 25.0 and 31.1 mm,respectively. The MStreatment was 6.1 mm higher than MP.

3.3.Relationships between grain yield,yield components and soil water storage

Pearson correlation analysisCorrelation analysis demonstrated that the wheat grain yield was significantly correlated with yield components and soil water storage(Fig.3). In the CK treatment,grain yield was significantly positively correlated with spike number,grain number per spike,SWSS,SWSSF and ET. The ET and spike number were the two factors most correlated with yield.There was no significant correlation between grain yield and thousand-kernel weight and SWSH,and there was a negative correlation between grain yield and thousandkernel weight. SWSSF was positively correlated with SWSS. Similarly,the correlation between SWSS and grain number per spike was very significant.

In the MPtreatment,grain yield was significantly positively correlated with spike number,SWSS,SWSH,SWSSF,and ET. Spike number and ET were the factors most closely related to wheat yield. There was little correlation between grain number per spike,thousandkernel weight and grain yield. The relationships between thousand-kernel weight and grain yield were negative.SWSSF was positively correlated with SWSS. SWSH also had a significant effect on spike number.

In the MStreatment,grain yield and spike number,grain number per spike,SWSS,SWSH,SWSSF,and ET had significant positive correlations (Fig.3). Spike number and SWSS were the two factors most closely related to grain yield. The correlation coefficient between thousand-kernel weight and grain yield was only 0.055.SWSSF had a positive effect on SWSS. There was also a significant positive correlation between SWSS and grainnumber per spike.

Fig. 3 Correlation coefficients of grain yield,yield components and soil water storage,under the no mulch (CK),plastic mulch(MP) and straw mulch (MS) treatments. SN,spike number;GS,grain number per spike;GW,thousand-kernel weight;SWSS,soil water storage before sowing;SWSH,soil water storage after harvest;SWSSF,soil water storage in the summer fallow period;ET,water consumption in the growth period. Asterisks indicate significant correlations. ＊,＊＊ and ＊＊＊ indicate differences at P<0.05,P<0.01,and P<0.001,respectively.

In the three treatments,spike number,SWSS,SWSSF,and ET were significantly and positively correlated with grain yield. In the CK treatment,ET had the greatest correlation coefficient with grain yield. Spike number was the most strongly correlated factor with yield in MPand MS.There was no significant correlation between thousandkernel weight and yield among the three treatments. The relationship between SWSSF and SWSS was extremely significant in all three treatments,which indicated that SWSSF provided more support for SWSS. Compared with the CK treatment,the MPand MStreatments showed little correlation between SWSS and spike number. In the CK treatment,the correlation between SWSH and spike number was not significant. In contrast,there was a significant positive correlation between the MPand MStreatments.

Structural equation modelsStructural equation modeling showed that in the CK treatment,ET,spike number,grain number per spike and thousand-kernel weight had direct positive effects on grain yield (Fig.4). Among them,ET contributed the most,followed by spike number. The ET was affected by SWSH and SWSS. Spike number was mainly affected by SWSS. However,there was no significant direct effect on ET and SWSH (P<0.05).SWSSF significantly affected SWSS. In the MPtreatment,ET,spike number,grain number per spike and thousandkernel weight had direct positive effects on wheat grain yield. Similarly,the direct contribution of spike number was the greatest. SWSS and SWSH affected spike number.However,spike number was mainly affected by SWSH.ET also significantly affected grain yield,and the main source of ET was SWSS. In the MStreatment,the grain yield was affected by ET,spike number,grain number per spike and thousand-kernel weight. Spike number had the greatest effect on the grain yield variation. Spike number was significantly and directly affected by SWSH. ET had little effect on grain yield. SWSSF had an indirect effect on both.

In the three treatments,grain yield was significantly affected by ET,spike number and grain number per spike.Among them,the most direct contributing factor to grain yield was ET in the CK,and spike number in the MPand MStreatments. The ET in the CK treatment was mainly affected by SWSS. In the MPand MStreatments,spike number was mainly affected by SWSH. SWSSF indirectly increased SWSH by increasing SWSS,which affected ET and spike number,and ultimately increased grain yield.

Relative importance of factors affecting wheat grain yieldThe analysis shows that yield components and soil water content had an impact on wheat grain yield,but the order of importance of each factor was different (Fig.5).In the CK treatment,the primary factor influencing wheat grain yield was ET,followed by spike number. The most important factor was spike number in the MPtreatment,followed by ET. Spike number and SWSS were the top two factors influencing grain yield in the MStreatment.In conclusion,the most important factor affecting wheat yield was ET in CK,and spike number in the MPand MStreatments.

4.Discussion

4.1.Effects of mulching on grain yield and its components

Our study found that mulching could improve the grain yield of winter wheat compared with CK (Fig.1-A). After extracting and analyzing the data from 74 articles,Hanet al.(2022) also found that film mulch and MSsignificantly increased wheat grain yield on the Loess Plateau. This is mainly because MPreduces heat loss and maintains the soil temperature through a blocking effect (Yang Y Het al.2018). Other studies have reported that MPimproved soil moisture and nutrient use efficiency (Abduwaitiet al.2021),and promoted photosynthesis and crop transpiration(Farahet al.2021). In addition,MSinhibits weed growth toavoid competition with crops for nutrients,and maintains soil water,which facilitates the dissolution and transport of root nutrients (Wuet al.2020). In our study,we also found that the average annual grain yield under the MStreatment was 134 kg ha–1lower than that of the MPtreatment. This was mainly due to the fact that MSsignificantly reduced the soil temperature compared with MP(Appendix A). Chenet al.(2015) also showed that MSgreatly reduced the soil temperature at the early stage of crop growth,which led to a reduction in the wheat emergence rate,inhibited the growth and development of winter wheat,and finally led to a reduction in grain yield.

Fig. 4 Structural equation models of grain yield,yield components and soil water storage,under the no mulch (CK),plastic mulch (MP) and straw mulch (MS) treatments. SN,spike number;GS,grain number per spike;GW,thousand-kernel weight;SWSS,soil water storage before sowing;SWSH,soil water storage after harvest;SWSSF,soil water storage during the summer fallow period;ET,water consumption in the growth period. Solid arrows indicate positive effects. Dotted arrows indicate no significant effect;the width of the lines indicates the degree of influence;the numbers represent standardized path coefficients. ＊,＊＊ and ＊＊＊ indicate differences at P<0.05,P<0.01,and P<0.001,respectively.

Mulching tended to increase spike number,grain number per spike and thousand-kernel weight of wheat compared with CK,but the results of MPand MSwere different in these three aspects. In our findings,the MStreatment tended to have a higher spike number than the MPtreatment (Fig.1-B). This may be because MSretains more soil moisture (Fig.2). The straw decomposition process releases more N,and soil water affects the crop root distribution,which is conducive to the absorption of N in the soil (Zhao S Cet al.2019;Yanget al.2020). Soil water and N can affect the tillering of wheat (Tedoneet al.2018),which is closely related to the increase in spike number in the late growth stage (Appendix B). Compared with the MStreatment,the grain number per spike and thousand-kernel weight of the MPtreatment increased by 2.7 and 2.2%,respectively (Fig.1-C and D). The straw cooling effect may not be conducive to the formation of grain number per spike (Appendix A;Chenet al.2014).The grain-filling stage is the key period that determines the thousand-kernel weight. Chaiet al.(2022a) showed that during the winter wheat filling stage,the mean soil temperature for six years under MSwas 1.43°C lower than that under MP. Low temperature caused a delay in the wheat growing season,which affected grain filling (Donget al.2007),and resulted in reduced thousand-kernel weight.

Fig. 5 Relative importance of factors impacting the wheat grain yield,under the no mulch (CK),plastic mulch (MP) and straw mulch (MS) treatments. SN,spike number;GS,grain number per spike;GW,thousand-kernel weight;SWSS,soil water storage before sowing;SWSH,soil water storage after harvest;SWSSF,soil water storage during the summer fallow period;ET,water consumption in the growth period. %IncMSE,increase in mean SE.

4.2.Effects of mulching on soil water storage

Plastic and MSgreatly reduce soil water evaporation(Linet al.2016). Compared with the CK,the mulching treatments increased SWSSF,and tended to increase SWSS and SWSH,in which the effect of the MStreatment was usually greater than the MPtreatment (Fig.2-A–C). There may be several reasons for this. First,the MSon the surface significantly lowers the accumulated soil temperature,as suggested by Zhaoet al.(2022),thus reducing ineffective water evaporation. Second,the barrier of plastic film led to the direct evaporation of precipitation instead of allowing it to enter the soil,which reduced the water storage of soil. In the MPtreatment,the soil surface was covered with plastic film. There were holes in the mulch,which was how the wheat grows.Precipitation could only enter the soil through the holes due to the plastic film barrier. Part of the rainfall on top of the plastic film became ineffective precipitation and then evaporated directly. In contrast,the soil was evenly covered with straw in the MStreatment,and the straw could not prevent rainfall infiltration into the soil column.Third,the amount of MSin the MStreatment was 104kg ha–1. Wanget al.(2018) found that higher straw coverage could improve the soil water storage capacity.

Compared with CK,the mulch treatments increased ET (Fig.2-D). The reason is that ET is obtained from the difference between (SWSS) and SWSH plus precipitation during the growth period. The precipitation during the growth period was consistent. SWSS under mulching was significantly higher than with CK (Fig.2-B),but there was no significant difference in SWSH (Fig.2-C).Therefore,the ET of the mulching treatments was higher than the CK. Plastic and MSretain more water and provide support for crop growth and development,which is the main reason for the higher yield of mulch treatments(Fig.1-A).

4.3.Relationships of soil water,yield related components and grain yield of wheat

In the three treatments of our long-term field experiment,we found that spike number and ET were positively correlated with grain yield (Fig.3). This is consistent with the previous findings of other researchers (Xieet al.2005;Liuet al.2018;Yuet al.2021). In addition,both our study and that of Liet al.(2018) showed that there was a significant positive correlation between SWSSF and SWSS.

Structural equation models can estimate the relationships between factors and test whether they fit the data(Hair and Alamer 2022). Thus,based on our current understanding and experience,we constructed structural equation models to explore the direct or indirect effects of yield components and soil moisture on the grain yield of each treatment. In this analysis we found that spike number and ET showed significant direct positive effects on grain yield in the three treatments (Fig.4). Increasing SWSS and SWSH could increase spike number. The reason could be that our experiment was set in the Loess Plateau where water was scarce,and the only water source was precipitation. SWSS and precipitation during crop growth determine the amount of SWSH. Huanget al.(2003) showed that SWSS accounted for about 32% of water consumption. As shown,SWSS affected spike number and ET. SWSSF was the most important source of SWSS (Fig.4). Accordingly,spike number and ET can be improved by increasing SWSSF. This result also verified our hypothesis. In the MPand MStreatments,SWSH showed a significant positive effect on spike number,while it had little effect in CK. The main reason for this result is that the exposed surface of the Loess Plateau has high evaporation and low precipitation(Liet al.2002,2022),so there is not enough water to support crop growth. Compared with MP,SWSS had a significant contribution to SWSH in the MStreatment. This may indicate that straw can better preserve soil moisture(Zhaoet al.2021) and reduce water loss (Liet al.2016)compared with plastic film,allowing SWSH to increase.

In the relative importance analysis (Fig.5),by comparison,spike number was the most important factor affecting grain yield in the MPand MStreatments.Other studies have found a similar result (McMasteret al.1994;Yanget al.2020). As we mentioned above,this is mainly attributed to mulch increasing soil water storage (Zhanget al.2007),which was beneficial to the growth of wheat seedlings,thus improving the formation of tillers and spike number (Liet al.2017). Sufficient soil moisture increased spike number,leading to the spike number being a key factor in determining grain yield. However,the relative importance of ET in the CK treatment was the greatest,and spike number was the second. The reason for this may be that under the condition of a bare land surface,water is the most important factor limiting grain yield due to the presence of less soil water (Liet al.2000;Wanget al.2016).Thus,grain yield is more dependent on ET. The relative importance of grain number per spike in CK was higher than in the other two treatments (Fig.5). This is mainly because grain yield is determined by spike number,the number of grains per spike and thousand-kernel weight.Thousand-kernel weight is a relatively stable component(Chaiet al.2022b;Yanget al.2022),and the importance of spike number was lower in CK than in the MPand MStreatments. Therefore,the grain number per spike ranked relatively high in importance. Combining the data in Figs.3–5,we found that ET was not high in the relative importance for grain yield in MScompared to MP.It may be that ET of the MStreatment was higher than that of the MPtreatment (Fig.2-D),so the limitation of ET on grain yield was relatively small.

5.Conclusion

In this 9-year long-term experiment,the MPand MStreatments increased grain yield,yield components and soil water storage compared to CK. Among them,the grain yield of MPwas higher than that of MS. The soil moisture of MSwas higher than MP. By exploring the relationships between soil moisture,yield components and grain yield,we found that spike number and ET were significantly correlated with grain yield in the three treatments. Sufficient SWSSF was beneficial for increasing spike number and ET,which was an important condition for high grain yield. The most important factor affecting wheat grain yield was spike number in the MPand MStreatments,and ET in the CK treatment. Therefore,our study has important implications for understanding the most important factors affecting dryland winter wheat grain yield under plastic and straw mulching measures in the Loess Plateau of Northwest China and elsewhere.

Acknowledgements

This study was supported financially by the National Key Research and Development Program of China(2021YFD1900703) and the National Natural Science Foundation of China (31272250).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on https://doi.org/10.1016/j.jia.2023.05.034