Conservation Tillage Increases Water Use Efficiency of Spring Wheat by Optimizing Water Transfer in a Semi-Arid Environment

Water availability is a major constraint for crop production in semiarid environments. The impact of tillage practices on water potential gradient, water transfer resistance, yield, and water use efficiency (WUEg) of spring wheat was determined on the western Loess Plateau. Six tillage practices implemented in 2001 and their effects were determined in 2016 and 2017 including conventional tillage with no straw (T), no-till with straw cover (NTS), no-till with no straw (NT), conventional tillage with straw incorporated (TS), conventional tillage with plastic mulch (TP), and no-till with plastic mulch (NTP). No-till with straw cover, TP, and NTP significantly improved soil water potential at the seedling stage by 42, 47, and 57%, respectively; root water potential at the seedling stage by 34, 35, and 51%, respectively; leaf water potential at the seedling stage by 37, 48, and 42%, respectively; tillering stage by 21, 24, and 30%, respectively; jointing stage by 28, 32, and 36%, respectively; and flowering stage by 10, 26, and 16%, respectively, compared to T. These treatments also significantly reduced the soil–leaf water potential gradient at the 0–10 cm soil depth at the seedling stage by 35, 48, and 35%, respectively, and at the 30–50 cm soil depth at flowering by 62, 46, and 65%, respectively, compared to T. Thus, NTS, TP, and NTP reduced soil–leaf water transfer resistance and enhanced transpiration. Compared to T, the NTS, TP, and NTP practices increased biomass yield by 18, 36, and 40%; grain yield by 28, 22, and 24%; and WUEg by 24, 26, and 24%, respectively. These results demonstrate that no-till with straw mulch and plastic mulching with either no-till or conventional tillage decrease the soil–leaf water potential gradient and soil–leaf water transfer resistance and enhance sustainable intensification of wheat production in semi-arid areas.


Introduction
Wheat (Triticum aestivum L.) is a major food crop in the world, which plays an important role in ensuring food security [1]. The western Loess Plateau of China is characterized by harsh climatic conditions including frequent spring drought, severe wind erosion, and water erosion [2,3]. Spring drop to -22 ℃ in January. Average annual temperature 6.4 ℃. Long-term climatic records show that annual cumulative air temperature >10 ℃ is 2240 ℃ and annual radiation is 5930 MJ/m 2 , with 2480 h of sunshine per year. Average annual evaporation is 1531 mm (coefficient of variation: 24.3%), which is three-to four-fold greater than precipitation.

Experimental Design and Agronomic Management
The experimental design was a randomized complete block with four replications. Plots were 4 m wide × 17 m long in block 1, 21 m long in blocks 2 and 3, and 20 m long in block 4. The long-term experiment included six tillage practice treatments in a two-year spring wheat/pea (Pisum sativum L.) rotation, with both phases of the rotation present in each year. All measurements in this study were made from plots planted with wheat. The conventional tillage with no straw Agronomy 2019, 9, 583 4 of 18 (T) treatment included the removal of all aboveground crop residues at the time of grain harvest before moldboard plowing to a depth of 20 cm. The conventional tillage with straw incorporated (TS) treatment was the same as T, except that all residues from the previous crops were retained and incorporated into the soil with tillage. The no-till with no straw (NT) treatment had all aboveground crop residues removed at the time of grain harvest and no tillage operations. The no-till with straw cover (NTS) treatment was the same as NT, except that all residues from the previous crops were retained. The conventional tillage with plastic mulch (TP) treatment was the same as T, except that alternating ridges (10 cm high × 40 cm wide) and furrows (10 cm wide) were made after harrowing with a ridging implement and all ridges and furrows were covered with colorless plastic film mulch using a plastic mulch laying machine prior to sowing crops in the furrows. The no-till with plastic mulch (NTP) treatment was the same as NT, except that the entire plot area was covered with colorless plastic film mulch using a plastic mulch laying machine. There were the same ridges and furrows with TP.
The spring wheat and pea cultivars were Dingxi 40 and Lvnong 2, respectively. Wheat was sown at a rate of 187.5 kg ha −1 in rows spaced 20 cm apart and pea was seeded at 180 kg ha −1 in rows spaced 24 cm apart. Immediately prior to the time of plastic mulch laying in the treatments with plastic mulch, all treatments were fertilized with calcium superphosphate (105 kg P 2 O 5 ha −1 for wheat and pea) and urea (105 and 20 kg N ha −1 for wheat and pea, respectively) that was broadcast uniformly over the entire plot area. Wheat was sown on 27 March 2016 and 26 March 2017, and harvested on 25 July 2016 and 20 July 2017. Weeds were removed by hand during the growing season and controlled with herbicides during the fallow period.

Precipitation and Drought Index
Daily precipitation was measured with a rainfall canister at the experimental site and DI was calculated as follows [34]: where Ar is annual rainfall, M is average annual rainfall, and δ is the standard deviation for annual rainfall. Drought index can be used to distinguish among wet (DI > 0.35), normal (−0.35 ≤ DI ≤ 0.35), and dry (DI < −0.35) soil water conditions for various time periods including on an annual basis, for a growing season, and for a fallow period [34]. Therefore, rainfall during the growing season and fallow period were used to also calculate the DI for these periods in the two study years.

Water Potential and Soil-Leaf Resistance
Water potential indexes were measured at four growth stages of wheat including the seedling stage . Three representative plants were randomly selected per plot, their leaves were removed with scissors, and placed into the leaf sample box. Next, a root and soil sample for the selected plants was taken using a soil corer (9-cm inner diameter) from the 0-10 cm soil depths at the seedling stage; at the 0-10 and 10-30 cm soil depths at tillering and jointing; and 0-10, 10-30, and 30-50 cm soil depth at flowering, respectively. Sampled root systems were gently shaken to let the rhizosphere soil fall into the soil sample box, then the root system was placed into the root sample box. Leaf water potential, root water potential, and soil water potential were measured immediately after each were sampled using a dew point water potential meter (WP4C Dewpoint PotentiaMeter, METER Group, Pullman, WA, USA) [35,36].
Transpiration rate and net photosynthetic rate was measured at 9:00 to 11:00 on the morning of the flowering stage (15 June 2016 and 27 June 2017) of wheat with a portable photosynthesis system (model GFS3000, Heinz Walz GmbH, Effeltrich, Germany). Three wheat plants were randomly selected Agronomy 2019, 9, 583 5 of 18 in each plot, the flag leaves of each plant were measured, and the average value of the three plants was obtained as the transpiration rate and net photosynthetic rate of the plot. Soil-leaf water transfer resistance (R sl ) was calculated using following equation [37]: where R sl is the soil-leaf water transfer resistance; Ψ s is soil water potential; Ψ l is leaf water potential; and CT is also transpiration rate.

Soil Water Content, Evapotranspiration, and Evaporation
Soil water content was measured to a depth of 2 m before sowing and after harvest in 2016 and 2017 using the oven-dry method [38] for the 0-5 and 5-10 cm soil depths, and using a time domain reflectometry soil moisture sensor (TRIME-PICO IPH/T3, IMKO GmbH, Ettlingen, BW, Germany) for the 10-30, 30-50, 50-80, 80-110, 110-140, 140-170, and 170-200 cm soil depths. The volumetric moisture content for the 0-5 and 5-10 cm soil depths was calculated by weight moisture content multiplied by corresponding soil bulk density. Evapotranspiration (ET) was calculated using the following equation [9]: where ET is evapotranspiration during the growing season; P is precipitation during the growing season; and W 1 and W 2 are water storage in the 0-200 m soil layer before sowing and after harvest, respectively. Soil evaporation was measured with a micro-evaporator made from polyvinylchloride tubing with the length of 150 mm, internal diameter of 110 mm, and external diameter of 115 mm [39]. On the sampling day, the soil mass of the micro-evaporator was weighed using an electronic balance with a sensitivity of 0.01 g, returned back to its original location in the field, and measured again at 07:00 h the following day. The loss in mass was the amount of evaporation the day before (equivalent to 0.1051 mm g −1 ). Soil inside the micro-evaporator was changed every three days and after precipitation, the tube emptied of soil, and placed in a new location in the field, which ensured that soil moisture inside the micro-evaporator was consistent with the surrounding soil. The calculation of evaporation in a growth period is based on the daily average evaporation measured during the growth stage multiplied by the number of days during the growth period without precipitation. The amount of transpiration during a growing season is the sum of that for all growth periods in the growing season using the following equation [40]: where T is transpiration during growing season; ET is evapotranspiration during growing season; and E is soil evaporation during growing season.

Yield and Water Use Efficiency
The whole plot was harvested manually using sickles at 5 cm above ground. The edges (0.5 m) of the plot were trimmed and discarded. Biological yield (BY) was measured by natural drying and before threshing. The grain moisture content after threshing was measured by the PM-8188 grain moisture meter (Kett Electric Lab., Tokyo, Japan), repeated five times, and the mean was taken. In addition, grain yield (GY) at 13% water content was calculated. All straw and chaff from stubble incorporated treatments were returned to the original plots immediately after threshing. Water use efficiency was calculated using following equations [9]: where WUE g and WUE b are water use efficiency of the grain and biomass yield, respectively.

Statistical Analysis
All data were checked for normality of distribution using the SPSS 19.0 software (IBM Corp., Chicago, IL, USA) and the Shapiro-Wilk test, and for homoscedasticity using the Levene's test with the general linear model. Data were transformed using either square root or natural log transformation to achieve normality when assumptions could not be met. All data were normal after testing. Analysis of variance was conducted for all dependent variables. Year and tillage practice were considered as fixed effects, and replication was considered a random effect. Differences among means were determined using Tukey's honestly significant different test (p ≤ 0.05). The linear relationship of water potential indexes with transpiration, BY, GY, WUE g , and WUE b were assessed using Pearson's correlation coefficient.

Effect of Tillage Practices on Water Potential at Different Growth Stages
Soil water potential varied with year, tillage practice, soil depth, and growth stage of wheat (Table 3). In 2016, soil water potential with NTS and TP were significantly greater in the 0-10 cm soil layer at the seedling and jointing stages compared to T. In 2017, soil water potential with the different treatments had a similar pattern to that in 2016. On average, compared with T, soil water potential with NTS was significantly greater in the 0-10 cm soil depth at the seedling and jointing stages. Soil water potential with TP was significantly greater than that with T in the 0-10 cm soil depth at the seedling stage and in the 0-10 and 10-30 cm soil depths at the jointing stage. Compared to T, soil water potential with NTP was significantly increased in the 0-10 cm soil depth at the seedling stage, in the 10-30 cm soil depth at tillering stage, and in the 10-30 cm soil depth at the jointing stage.
Year, tillage practice, soil depth, and growth stage of wheat influenced root water potential (Table 4). In general, compared to T, root water potential was significantly increased with NTS and NT in the 0-10 cm soil depth at the seedling and jointing stages, and with NTS in the 30-50 cm soil depth at flowering. Root water potential was not significantly different between TS and T in all soil layers at every growth stage. Root water potential with TP was significantly greater than that with T in the 0-10 cm soil depth at the seedling, tillering, and jointing stages, and in the 0-10 and 30-50 cm soil depths at flowering. Root water potential with NTP was significantly greater than that with T in the 0-10 cm soil depth at the seedling stage, in the 0-10 and 10-30 cm soil depths at tillering and jointing, and in the 0-10 and 30-50 cm soil depths at flowering.
a Within a column for a given year, means followed by different letters are significantly different (p ≤ 0.05); b T, conventional tillage with no straw; NTS, no-till with straw cover; NT, no-till with no straw; TS, conventional tillage with straw incorporated; TP, conventional tillage with plastic mulch; NTP, no-till with plastic mulch.
Leaf water potential differed with year, tillage practice, soil depth, and growth stage of wheat (Table 5). In 2016, compared to T, leaf water potential with NTS was significantly increased at the seedling stage, and not significantly different with NT and TS at any growth stage. Leaf water potential in 2016 was significantly greater with NTP and TP at the seedling stage, and with TP at flowering, compared to T. In 2017, compared to T, leaf water potential with NTS was significantly increased at the seedling and tillering stages; however, leaf water potential with NT was not significantly increased at any growth stage. Leaf water potential was significantly greater with TS than T at the seedling and tillering stages, and with TP than T increased at the seedling, tillering, and jointing stages. On average, leaf water potential with NTS and NTP was significantly greater than that with T at the seedling, tillering, and jointing stages. Leaf water potential with NT and TP was not significantly different when compared to that with T at any growth stage. However, leaf water potential with TS was significantly greater than that with T at the seedling stage.

Effect of Tillage Practices on Water Potential Gradient at Different Growth Stages
The soil-root water potential gradient was affected by year, tillage practice, soil layer, and growth stage of wheat (Table 6). In 2016, the soil-root water potential gradient was not significantly different among tillage practices at all soil layers at all growth stages. In 2017, the soil-root water potential gradient was significantly reduced with NTS and NTP compared to the other tillage practices in the 0-10 cm soil depth at the jointing stage and in the 0-10 and 30-50 cm soil depths at the flowering stage.
The root-leaf water potential gradient varied with year, tillage practice, soil depth, and growth stage of wheat (Table 7). On average, compared to T, the root-leaf water potential gradient with NTS was significantly reduced at the 0-10 cm soil depth at the seedling stage, 10-30 cm soil depth at jointing stage, and 30-50 cm soil depth at the flowering stage; however, the root-leaf water potential gradient with NT was significantly increased at the 0-10 cm soil depth at the tillering stage. The root-leaf water potential gradient was significantly decreased with TS at the 0-10 cm soil depth at the seedling stage, and with TP at the 0-10 cm soil depth at the seedling stage and 30-50 cm soil depth at flowering, compared to T. The root-leaf water potential gradient with NTP was significantly reduced at the 0-10 cm soil depth at the seedling stage and 30-50 cm soil depth at flowering, compared to T.
The soil-leaf water potential gradient varied with year, tillage practice, soil layer, and growth stage of wheat (Table 8). On average, the soil-leaf water potential gradient with NTS was significantly less than that with T at the 0-10 cm soil depth at the seedling stage and 30-50 cm soil depth at flowering. The soil-leaf water potential gradient with NT and TS was not significantly different from that with T at all soil depths and growth stages. Compared to T, the soil-leaf water potential gradient was significantly decreased with TP at the 0-10 cm soil depth at the seedling stage and at the 30-50 cm soil depth at flowering, and with NTP at the 0-10 cm soil depth at the seedling and jointing stages and at the 30-50 cm soil depth at flowering.

Effects of Tillage Practices on Transpiration Rate and Soil-Leaf Water Transfer Resistance at Flowering
Transpiration rate of wheat at flowering varied with tillage practice (Figure 2). Compared with T, transpiration rate was significantly increased with NTS, TP, and NTP, but not significantly different with NT and TS in all years (Figure 2A,B); on average, NTS, TP, and NTP significantly increased transpiration rate by 103, 143, and 91%, respectively, compared with T. Net photosynthetic rate and soil-leaf water transfer resistance at flowering were impacted by tillage practices (Figures 2 and 3). Net photosynthetic rate was significantly increased with NTS, TP, and NTP, but was not significantly different with NT and TS ( Figure 2C,D); over the two years, NTS, TP, and NTP significantly increased the net photosynthetic rate by 20, 19, and 19%, respectively, when compared to T. Compared to T, soil-leaf water transfer resistance at all soil layers was significantly reduced with NTS, TP, and NTP, but not significantly different with NT and TS ( Figure 3). Averaged across years and soil layers, compared to T, the soil-leaf water transfer resistance with NTS, TP, and NTP was significantly decreased by 66, 70, and 63%, respectively.

Effect of Tillage Practices on Yield and Water Use Efficiency
Tillage practice significantly affected transpiration at flowering, BY, WUE b , GY, and WUE g (Table 9). Over the two years, compared with T, transpiration with NTS, TP, and NTP was significantly increased by 40, 64, and 76%, respectively; however, transpiration was not significantly different with NT and TS. Compared to T, BY was significantly increased with NTS, TP, and NTP by 18, 36, and 40%, respectively; however, it was not significantly different with NT and TS. Water use efficiency of BY was significantly increased with TP and NTP by 25 and 22%, respectively, but was not significantly different with NTS and TS, compared to T. Grain yield with NTS, TP, and NTP was significantly increased by 28, 22 and 24%, respectively, compared to T; however, it was not significantly different among NT, TS, and T. Water use efficiency of GY with NTS, TP and NTP was significantly increased by 24, 26, and 24%, respectively, but not significantly different with NT and TS, compared to T. Table 9. Transpiration at the growing season, biomass and grain yields, and water use efficiency of grain yield and biomass yield (WUE b and WUE g , respectively) of wheat as affected by tillage practice in 2016 and 2017.

Correlations of Water Potential Indexes with Transpiration, Biomass and Grain Yields, and Water Use Efficiency of Grain and Biomass Yields
Significant correlations among the water potential indexes, transpiration at growing season, BY, WUE b , GY, and WUE g of wheat were observed (Table 10). Soil water potential, root water potential, and leaf water potential at the seedling stage was highly significant and positively associated with transpiration, BY, WUE b , GY, and WUE g . Soil water potential, root water potential, and leaf water potential at other growth stages showed different patterns. The root-leaf water potential gradient and soil-leaf water potential gradient at the seedling stage had a significant negative correlation with transpiration, BY, WUE b , GY, and WUE g . The soil-root water potential gradient, root-leaf water potential gradient, and soil-leaf water potential gradient at the 30-50 cm soil depth at flowering had a significant negative correlation with transpiration, BY, WUE b , GY, and WUE g . The soil-root water potential gradient, root-leaf water potential gradient, and soil-leaf water potential gradient at other growth stages showed different patterns. Table 10. Pearson's correlation coefficient for correlations of water potential indexes with transpiration, biomass and grain yields, and water use efficiency of biomass and grain yields (WUE b and WUE g , respectively) across years for different growth stages of wheat and soil layers. Correlation coefficients followed by * and ** are significant at P ≤ 0.05 and 0.01, respectively; b S, soil water potential; R, root water potential; L, leaf water potential; S-R, soil-root water potential gradient; R-L, root-leaf water potential gradient; S-L, soil-leaf water potential gradient.

Effects of Tillage Practices on Water Potential in the Soil-Plant System
Soil, roots, and leaves are important indicators of whether plants are subject to drought stress [41][42][43], and have been employed in the selection of appropriate tillage practices. Tillage practices can affect soil, root, and leaf water potential [44,45]. In this study, NTS significantly increased soil water potential in the 0-10 cm soil depth at the seedling and jointing stages of wheat compared to T because NTS increased topsoil moisture at the seedling stage. However, with wheat growth and development, canopy coverage increased, transpiration dominated evapotranspiration, and the positive effect of straw mulching on topsoil moisture gradually weakened [26,46], thus NTS did not significantly increase the soil water potential at flowering. Conventional tillage and no-till improved soil water potential compared to T in the 0-30 cm soil depths at all growth stages, mainly because plastic film mulching reduced soil evaporation, which led to greater soil water moisture throughout the growing season [47]. No-till with straw cover, TP, and NTP increased leaf water potential compared to T at all growth stages, in agreement with results from previous studies [44,48]. However, Zhang et al. [49] found that NTS reduced leaf water potential by 11% compared to T. This discrepancy is likely to be due to differences in soils and early rainfall prior to measurement. The study reported by Zhang et al. (1999) was conducted on a quaternary red clay soil with high viscosity, and long-term no-till led to subsurface soil compaction and shallow root systems. The present study was conducted on a deep loess soil with deep uniform texture and high water storage capacity [50], which is favorable for the growth and development of crop root systems.
Water potential gradients drive water transport from soil to plants, with a greater water potential gradient resulting in faster water absorption [51]. In this study, NTS, TP, and NTP reduced the soil-root water potential gradient in the 30-50 cm soil depth at the flowering of wheat. No-till with straw cover, TP, and NTP significantly decreased the root-leaf water potential gradient compared to T at the 0-10 cm soil depth at the seedling stage and 30-50 cm soil depth at flowering. These treatments also significantly reduced the soil-leaf water potential gradient at the 0-10 cm soil depth at the seedling stage and 30-50 cm soil depth at flowering, most likely because they stored more water from the fallow period. Moreover, wheat canopy coverage reaches a maximum at flowering, thereby limiting evaporation after this stage.
Water transfer resistance exists in the process of water transport from soil to plants [52]. In this study, NTS, TP, and NTP reduced the soil-leaf water transfer resistance at flowering of wheat compared to T. This could be due to NTS, TP, and NTP having increased root length and root surface area, and more favorable spatial distribution of roots for water uptake [53]. This was demonstrated in this study, as NTS, TP, and NTP had greater soil water absorption by plants than T.
In this study, NTS, TP, and NTP significantly increased the transpiration and net photosynthetic rate of wheat at flowering compared to T, as shown in previous studies [54][55][56]. The net photosynthetic rate of wheat flag leaves has been reported to be 24 to 39% higher with NTS compared to conventional tillage, and also has a significantly higher transpiration rate [54,57]. In contrast, Jiang et al. [58] found that NTS reduced the photosynthetic rate of wheat, likely because their straw cover was applied after sowing, resulting in less soil moisture stored during the fallow season. Straw coverage in this study occurred after harvest, leading to more soil moisture stored during the fallow season, thereby enabling an increase in photosynthetic rate. Transpiration is fundamental to understanding crop water use efficiency [10]. In this study, transpiration with NTS, TP, and NTP was significantly increased compared to T, mainly because NTS, TP, and NTP increased precipitation infiltration and reduced soil evaporation [23,47,59].
Biomass yield of wheat was significantly greater with NTS, TP, and NTP compared to T. Garofalo and Rinaldi [60] found that a greater rate of transpiration was associated with greater BY. However, Dam et al. [61] found that the long-term BY of maize did not differ between NTS and T. This may be attributable to differences in soil texture at the experimental sites, which was sandy loam in their study and loess in the present study. In agreement with our results, Zhang et al. [62] found that plastic mulching increased the BY of maize. This could be due to enhanced crop growth resulting from greater soil temperature [63,64], soil moisture [62], and radiation capture [65] with plastic mulching.

Effects of Tillage Practices on Grain Yield and Water Use Efficiency
Conservation tillage practices have been shown to increase soil water storage, wheat yield, and WUE on the semi-arid Loess Plateau of China [27,66]. However, Pittelkow et al. [15] found that conservation tillage practices did not increase the GY of cereals in moist regions. This is likely to be because the impact of conservation tillage on yield varies among climatic zones. The improvement of wheat GY and WUE g with NTS, TP, and NTP compared to T in this study can be attributed to increased water potential and decreased water potential gradient and water transfer resistance, thus enhancing transpiration and BY.

Conclusions
This study demonstrated that NTS, TP, and NTP significantly increased grain yield and WUE g as a result of increased water potential, decreased water potential gradient, and water transfer resistance, and led to increases in transpiration rate, transpiration, and biomass yield. These results demonstrate that no-till with straw cover, conventional tillage with plastic mulch, and no-till with plastic mulch are suitable tillage practices for the sustainable intensification of wheat production in semi-arid areas.