Effects of supplemental irrigation based on soil moisture levels on photosynthesis, dry matter accumulation, and remobilization in winter wheat (Triticum aestivum L.) cultivars

Abstract Two winter wheat (Triticum aestivum L.) cultivars, namely Jimai22 (JM22) and Zhouyuan9369 (ZY9369), were used to study the effects of a new irrigation policy, supplemental irrigation (SI) based on soil moisture levels, photosynthesis, dry matter accumulation, and remobilization from 2009 to 2011 in Northern China. Two SI treatments were designed based on relative soil moisture contents in the 0–140 cm soil layer: (1) the target soil relative water contents were 75% of field capacity (FC) at jointing and 65% of FC at anthesis (W1), 75% and 70% (W2) in 2009–2010, and (2) the target soil relative water contents were 75% at jointing and 75% at anthesis (W1′), 75% and 80% (W2′) in 2010–2011. Rain-fed treatment (W0) was used as control. Results showed that SI significantly improved the biomass, grain yield and water use efficiency (WUE) of both wheat cultivars. The biomass and grain yield of W1 and W1’ treatments were higher than those of others. The net photosynthetic rate, the actual photochemical efficiency of flag leaf, the accumulation of dry matter, and its remobilization from the vegetative parts to the grains after anthesis in W1 and W1’ treatments were significantly higher than in the other treatments. By contrast, the WUE and irrigation efficiency of W2 and W2’ were significantly lower than those of W1 and W1’. Under the experimental conditions, ‘JM22’ showed higher photosynthetic rate in the last stage of grain filling, more spike number per ha, more kernels per spike, higher 1000-kernels weight and eventually higher WUE than ‘ZY9369’.


Experimental design and irrigation management
The winter wheat cultivars 'Jimai 22' (JM22) and 'Zhouyuan 9369' (ZY9369), which are widely planted in the NCP, were studied. The experiment was conducted using a split block design with three replicates, with the wheat cultivars in the main block and irrigation treatments in the secondary block.
The irrigation treatments were designed on the basis of the soil moisture content of the 0-140 cm soil layers. Using a previously reported irrigation formula (Jalal et al., 2012), we increased the relative water content of the 0-140 cm soil layers to 75% of field capacity (FC) at the jointing stages for all the irrigation treatments, to 65% (W1) and 70% (W2) of FC at the anthesis stages in 2009-2010, and With limited irrigation in the critical stages of crop growth, the WUE and yield of crop in dry areas have been improved substantially (Oweis & Hachum, 2006). Nakamura et al. (2003) concluded that grain yield under deficient soil moisture condition was higher by about 15-40% with greater dry matter production compared to that with adequate soil moisture condition. However, previous studies seldom considered soil moisture and crop water requirement when determining irrigation water.
Supplemental irrigation (SI) is an effective water-saving irrigation technique based on considerations of precipitation, soil water storage, and crop requirements in different growth periods. The SI amount is calculated on the basis of actual relative soil moisture content before irrigation and the target relative soil moisture in the 0-140 cm soil layer, which can prevent waste of water resources caused by blind flooding Wang, Yu, & White, 2013). However, limited information is available regarding yield and water use of various winter wheat cultivars under different SI levels based on soil moisture. Moreover, wheat cultivars vary in their sensitivities to water stress-total dry matter, WUE, relative water loss (transpiration efficiency), and initial water content of flag leaves (uptake efficiency) all differ significantly among the seven bread wheat cultivars (Alireza & Farshad, 2013). Under drought stress, the yield parameters of sensitive wheat cultivars decrease more than those of tolerant cultivars (Guóth et al., 2009).
Our objectives were to: (i) determine the effects of varying irrigation levels on the basis of soil moisture at anthesis stage on the photosynthate formation, accumulation, translation, and distribution in different winter wheat cultivars; (ii) evaluate the effects of different SI regimes on grain yield and WUE of different winter wheat cultivars; and (iii) select a winter wheat cultivar that exhibits high yield and WUE, and determine its optimal SI regime based on soil moisture level.

Experimental site
Field experiments were conducted from October 2009 to June 2011 in the experimental station of Shandong Agricultural University (36′17″ N, 117′15″ E), which is located in the center of the NCP. The environment at the study site is typical and representative of the plain. The organic matter, total nitrogen, available nitrogen, available phosphorus, and available potassium in the topsoil (0-20 cm) of the experimental plots were 16.9-18.6 g kg −1 , 1.2-1.3 g kg −1 , 93.3-105.0 mg kg −1 , 30.2-31.8 mg kg −1 , and 105.8-110.5 mg kg −1 , respectively. The soil moisture content of the top 0-200 cm of the soil (in 20 cm increments) is shown in Figure 1. The amounts of precipitation during the  to 75% (W1′) and 80% (W2′) of FC at the anthesis stages in 2010-2011. Rain-fed treatment (W0) with no manual irrigation was set as control in both growing seasons. The size of the experiment plot was 2 m × 5 m, with a 1.5 m-wide unirrigated zone between two adjacent irrigation plots to minimize the interaction of different treatments.
The irrigation was supplied at the jointing and anthesis stages. The amount of SI was calculated using the following equation (Jalal et al., 2012): where CIR (mm) is the amount of SI, γbd (g cm −3 ) is the soil bulk density, D h (cm) is the soil layer depth (140 cm in this study), θt (mg water g −1 dry soil) is the target soil moisture content on a weight-basis after SI, and θn (mg water g −1 dry soil) is the soil moisture content on a weight-basis before irrigation. θt was calculated as follows: where θmax (mg water g −1 dry soil) is the field capacity and θr (%) is the target relative soil moisture content.
The SI amounts of different treatments are shown in Figure 3. Water was sprayed evenly on the experimental plots under pressure, and the irrigation amount was measured using a flow meter (Model: DN50-R80, trademark: Juxingwang Inc., Ningbo, China). When soil irrigation water reached relative equilibrium 3 days later, changes in soil moisture in the 0-140 cm soil layers after SI were measured. The relative soil moisture contents are shown in Figure 4.

Crop management
All plots were supplied with 240 kg N ha −1 , 112.5 kg P 2 O 5 ha −1 , and 112.5 kg K 2 O ha −1 . All P and K fertilizers and 105 kg ha −1 N fertilizer were applied pre-sowing, and the remaining N fertilizer was top-dressed at jointing. The fertilizers used were urea (46% N), ammonium dibasic phosphate (46% P 2 O 5 and 18% N), and potassium sulfate (52% K 2 O). Wheat seeds were sown at a density of 180 plants m −2 on 7 October 2009 and 7 October 2010. Wheat was harvested on 13 June 2010 and 15 June 2011. Other management practices were similar to conventional practices for field wheat production, including pest control and tilling practices.

Photosynthetic characteristic parameter
The net photosynthesis rate (P n ) and the actual photochemical efficiency (ФPSII) of flag leaf for each treatment were measured at the early (10 days after anthesis, DAA) and late (30 DAA) grain-filling stages using a portable photosynthesis system (CIRAS-2, PP-Systems, Hitchin, Herts, (2) t = max × r, UK) and a Pulse Modulated Fluorometer (FMS) chlorophyll fluorometer acquisition system (FMS-2 Hansatech, Britain), respectively. Both parameters were measured between 9:00 am and 11:00 am with 10 replicates under natural light.

Dry matter
At anthesis and maturity, the aboveground parts of the plants were sampled and divided into three parts, namely, leaf + stem + leaf sheath, cob + glume shell organs, and grains, which were weighed after drying to constant weight at 80 °C. The following parameters related to dry matter (DM) accumulation and remobilization within the wheat plant were calculated following Man et al. (2015): (1) Amount of dry matter accumulation during grain filling (ADMA) = dry matter content of the whole plant at anthesis − dry weight of aboveground vegetative organs at maturity (DWAVO).

Grain-filling rate and grain weight
Emerging flowering spikes were tagged on the same day.  Oven drying was used to measure the soil moisture content every 20 cm up to a depth of 200 cm and ultimately to estimate soil water consumption (ΔS). Soil core samples were randomly collected using a 50 mm-diameter steel sampling tube manually driven into the soil. The soil cores were weighed to obtain the current weight, ovendried at 105 °C for 48 h, and then re-weighed to determine the gravimetric water content. Measurements were obtained at sowing, jointing, anthesis, and maturity (Gan et al., 2000).

Statistical analysis
Data are presented as means of all replicates. ANOVA was used to detect significant differences among the mean values for different treatment conditions at the 0.05 probability level. Statistical analysis was performed using SPSS Version 13.0 for Windows (SPSS, Chicago, Illinois, USA).

Photosynthetic rate in flag leaf
Irrigated plants exhibited higher P n of flag leaves at 10, and 30 day after anthesis (DAA) than rain-fed plants ( Figure 5), and the differences in P n between rain-fed and SI-treated plants increased as grain filling advanced. The P n of flag leaves of Jimai22 ('JM22') did not differ significantly between the two SI treatment groups in both growing seasons; the same result was observed in Zhouyuan9369 ('ZY9369') in 2010/2011. In 2009/2010, P n of W1 of 'ZY9369' was lower than that of W2 at 10 and 30 DAA. In both growing seasons, P n of 'JM22' was significantly higher than that of 'ZY9369' at 30 DAA. sampled at 7-day intervals from anthesis to maturity. The grain-filling rate was estimated from accumulated dry weight. At each sampling date, grains were separated from the glumes and then dried at 105 °C for 30 min and at 80 °C until constant weight was reached. The total number of grains was determined, and their dry weight was recorded.

Grain yield and WUE
Grain yields were determined by manually harvesting grains from a 2 m 2 area in each plot and expressed at 12.5% moisture content (Xue et al., 2006). WUE was calculated using Equation (3) Sun et al., 2006) as follows: where WUE (kg ha −1 mm −1 ) is the WUE for grain yield, Y (kg ha −1 ) is the grain yield, and ET (mm) is the total evapotranspiration.
ET during the growth stage of winter wheat was calculated using the water balance Equation (4)  where P is the growing seasonal precipitation (mm), I is irrigation (mm), S G is the groundwater contribution to plant available water (mm), D is downward drainage out of the root zone (mm), R is surface runoff (mm), and ΔS is change in soil water stored in the upper 200 cm of the soil between sowing and maturity (mm). In this experiment, the groundwater was 15 m below the surface, and runoff was prevented from the experimental plots. Therefore, S G , D, and R were negligible. in the flag leaves did not differ significantly between the two cultivars in their early filling stages. In the middle and late filling stages, ФPSII in the flag leaves of 'JM22' was higher than that of 'ZY9369' receiving the same treatment. Figure 6 shows that ФPSII in the flag leaves of rainfed plants were lower than that in the flag leaves of SI-treated plants in the filling stage, and the difference increased as grain filling progressed. ФPSII did not differ significantly between the two SI treatments of both  plants. By contrast, the amount of accumulated dry matter was higher in SI-treated 'ZY9369' than in rain-fed 'ZY9369' ( Table 1). As shown in Table 1, the dry matter accumulation of total, grain, stem + sheath + leaf, and spike axis + glume were lower in 'ZY9369' than in 'JM22' of the same treatment group.

Contribution of dry matter redistribution and accumulation during grain filling to grain
The amount of dry matter remobilization during grain filling (ADMR) and its contribution rate to grain in SI-treated 'JM22' in the two growing seasons and in 'ZY9369' in 2009/2010 were lower than those in the rain-fed plants. By contrast, the amount of dry matter accumulation during grain filling (ADMA) and its contribution rate to grain were remarkably higher in the SI-treated plants than in the rainfed plants (Table 2).
There were no significant differences in ADMR between W1 (W1′) and W2 (W2′) in 'JM22'; however, its contribution rate to grain was lower in W1 (W1′)  The dry matter in the grains of rain-fed winter wheat was mainly contributed by ADMR; conversely, the dry matter in the grains of SI-treated winter wheat was mainly contributed by ADMA. SI evidently increased the ADMA and its contribution rate to grain in the filling stages. W1 or W1' treatment yielded higher ADMA and had higher contribution rate to grain than W2 or W2' treatment in the two growing seasons for both wheat cultivars. Table 1 shows that the amount of accumulated dry matter at anthesis was not different between the irrigated treatment group and the rain-fed treatment group of JM22 in 2009/2010. The amount of accumulated dry matter at anthesis was significantly higher in the irrigated treatment group than the rain-fed treatment group of JM22 in 2010/2011 and ZY9369 in the two growing seasons.

Dry matter accumulation and distribution at anthesis and maturity
Table 1 also shows that the amount of accumulated dry matter at maturity was significantly higher in the irrigated treatment group than the rain-fed treatment group. The amounts of dry matter accumulated in W1 and W1' treatment groups were higher than those in W2 and W2' treatment groups, respectively. Moreover, the amount of accumulated dry matter in 'JM22' was higher than in 'ZY9369' of the same treatment group. The amount of accumulated dry matter differed significantly between the two cultivars under non-irrigated conditions, and the gap between the two cultivars was 1512 kg hm −2 in 2009/2010 and 3743 kg hm −2 in 2010/2011. With irrigation, the differences between the two cultivars decreased to 442-592 kg hm −2 in 2009/2010 and 2054-2658 kg hm −2 in 2010/2011. With excessive irrigation of W2' , the amount of accumulated dry matter decreased faster in 'ZY9369' than in 'JM22' . In addition, the amount of accumulated dry matter at maturity was higher in 'ZY9369' than in 'JM22' of the same treatment group. The amounts of accumulated dry matter in different organs exhibited the following pattern: grain > stem + sheath + leaf > spike axis + glume. The amounts of accumulated dry matter in grains of both cultivars exhibited the following pattern: W1 > W2 > W0 in 2009/2010 and W1' > W2' > W0' in 2010/2011. The amounts of accumulated dry matter in the aboveground vegetative organs did not differ significantly among the treated 'JM22'  2375ab  10024a  12400a  7313d  1615d  6677b  15605c  JM22W1  2022c  9904a  11927a  9465a  1871c  7399a  18735a  JM22W2  2022c  9904a  11927a  8890b  1943bc  7304a  18138ab  ZY9369W0  2338b  8714b  11052b  7116d  1264e  5714c  14094d  ZY9369W1  2496a  9490a  11986a  8856b  2158a  7279a  18293ab  ZY9369W2  2496a  9490a  11986a  8409c  2028b  7108a  17545b  2010/2011 JM22W0'  1848b  7628c  9476b  7412c  1649c  4914c  13975d  JM22W1'  2014a  8930a  10944a  9386a  1865a  6587a  17838a  JM22W2'  2014a  8930a  10944a  9002a  1852a  6508a  17362a  ZY9369W0'  1079c  4230d  5310c  5794d  1010d  3429d Compared with those of rain-fed plants, the spike number per ha and kernels per spike of SI-treated plants increased in both growing seasons (Table 3). The kernels per spike and 1000-kernels weight of W1 (or W1′) treatment were higher than W2 (or W2′) treatment. WUE and irrigation benefits were also significantly affected by SI treatments. The WUE, and IB in W1 (or W1′) treatment were higher than those in W2 (or W2′) treatment for both cultivars. The spike number per ha and 1000-kernel weight were significantly higher in 'JM22' than in 'ZY9369' in both growing seasons, while kernels per spike was higher in 'ZY9369' than in 'JM22' . Under the same SI treatment, the grain yield, WUE, and irrigation benefit of 'JM22' were higher than those of 'ZY9369' . Table 3 also shows that different growing seasons had no significant effect on spike number per ha, while the impact on kernels per spike, 1000-kernels weight, grain yield, WUE, irrigation benefit were extremely significant. Different cultivars had no obvious effect on irrigation benefit, but the influences on other items reached extremely significant level. Water treatment extremely significantly affected other items except 1000-kernels weight. Different growing seasons and cultivars showed interactions and exerted extremely significant effects on 1000-kernels weight and grain yield, and had significant impact on WUE and irrigation benefit. Cultivars and water treatments had extremely significant effect on kernels per spike.

Discussion
Photosynthesis is one of the most important metabolic processes determining grain yield in wheat (Foulkes et al., 2007;Guóth et al., 2009). Leaf photosynthetic rate has been an important index of plant physiology to analyze the limiting factors of crop yield (Yang et al., 2014). Crop photosynthetic physiological characteristics may change with external environmental conditions. The photosynthetic capacity of wheat flag leaf after flowering can be significantly influenced by irrigation regimes. Water deficiency during grain filling stage reduces the photosynthetic rate and chlorophyll content of flag leaves and accelerates leaf senescence (Ahmadi & Siosemarde, 2005;Saeedipour & Moradi, 2011a). The effects of light drought stress on ΦPSII are not significant, but moderate and severe drought stresses can gradually decrease PSII electron transport (Zivcak et al., 2013). In this study, a precise irrigation method was proposed to control soil moisture content to a specific value in jointing and anthesis stages (Figure 3). Compared with the control, the photosynthetic rate of the flag leaves of the irrigated treatments was significantly improved, especially in the late grain filling stage, by maintaining 75% soil moisture content in the jointing stage and 65-75% soil moisture content in the anthesis

Grain filling rate
The grain filling rate of the cultivars showed slow-fastslow trend in both growing seasons. However, a faster filling rate was observed at 14-28 DAA (Figure 7).
The grain filling rate of the SI-treated plants was lower than that of the rain-fed plants at the beginning of the grain filling stage. The grain filling rates of the SI-treated plants and the rain-fed plants were similar in the mid-filling stages. By contrast, the grain filling rate of the SI-treated plants was higher than that of the rain-fed plants in the late filling stage. In 2010/2011, the grain filling rate of W1' treatment was higher than that of W2' treatment at 14 and 35 DAA.
The grain filling rate was significantly higher in 'JM22' than in 'ZY9369' under the same treatment in both growing seasons.

Grain yield and WUE
The grain yield differed significantly among different treatment conditions (Table 3). The highest yields in 2009/2010 were 9387.3 and 8362.4 kg hm −2 in W1 of 'JM22' and 'ZY9369' , respectively. By comparison, the highest yields in 2010/2011 were 9076.5 and 7546 kg hm −2 in W1' of 'JM22' and 'ZY9369' , respectively, followed by W2 (or W2′) treatment. The yield was significantly higher in SI-treated plants than in rain-fed plants. of 14 wheat genotypes under irrigated and rain-fed conditions and found that the leaf chlorophyll content and stomatal conductance of these wheat varieties grown under drought stress varied significantly. The new varieties maintained significantly higher values of P n , effective PSII quantum yield (ΦPSII), apparent rate of photosynthetic electron transport (ETR), coefficient of photochemical quenching (qP) and coefficient of non-photochemical quenching (NPQ) than the old varieties at low soil-water content (Ma et al., 2013). Wu and Bao (2011) investigated responses of wet climate (CY17) and dry climate (XN889) stage ( Figure 4). Under the same condition, the P n of flag leaves of 'JM22' at 20 and 30 DAA was higher than that of rain-fed plants, and ΦPSII increased by 60.6-78.2% (Figures 5 and 6). However, the P n and ΦPSII of the flag leaves did not increase more when the soil moisture content increased from 65% (W1) to 70% (W2) in 2009/2010 and from 75% (W1′) to 80% (W2′) in 2010/2011, by irrigating more (26-30 mm) water. The effects of irrigation on photosynthetic capacity of different wheat varieties were different. Khamssi and Najaphy (2012) studied the photosynthetic components  respectively, were 748-1080 kg hm −2 or 5.1-5.3%, in 'ZY9369' (Table 2). Thus, 'ZY9369' was more sensitive to excessive irrigation than 'JM22' . Grain number and single grain weight are reduced under drought stress or waterlogging stress in grain filling stage (de San Celedonio et al., 2014;Zhang et al., 2014), which results in significant reduction in yield (Dorostkar et al., 2015). Irrigation can increase the duration of grain filling by prolonging the functional period of leaves, which can in turn increase grain weight (Verma et al., 2004). However, compared with well-irrigated treatment, intermediate drying treatment is likely to increase the grain filling rate in superior and inferior spikelets (Zhang et al., 2015). Our study found that W0 treatment promoted grain filling in the early filling stage. By contrast, SI treatments kept a higher grain filling rate in the late filling stage, because the relatively high soil moisture content may delay plant senescence ( Figure 7). However, with excessive SI, especially when the relative soil moisture content was higher than 75%, the grain filling rate increased in the early filling stage but decreased quickly in later stages. The results also showed that SI treatments remarkably increased the spike number per ha and the kernels per spike than W0 (Table  3). However, SI did not significantly influence the single grain weight because rain-fed plants exhibited less spike number per ha and kernels per spike, and the sink was low, so each kernel could be filled more easily.
Water deficiency during cell division and grain filling stages would reduce the 1000-kernels weight and grain yield significantly, especially in the sensitive cultivars (Saeidi et al., 2012). Abdoli et al. (2013) also reported that the 'Marvdasht' cultivar is more sensitive to water deficiency in the grain filling stage than the 'DN-11' cultivar. Our study showed that the spike number per ha, 1000kernels weight, grain filling rate, and yield potential were higher in 'JM22' than in 'ZY9369' under the same treatment condition in both growing seasons (Table 3). The sensitivity of the yield to relative soil moisture content at anthesis stage differed between the cultivars. Compared with W1, more irrigation at anthesis reduced the kernels per spike by 4.9% in 'JM22' and reduced the 1000-kernels weight by 4.3% in 'ZY9369' in the 2009/2010 growing season.
Moderately limited irrigation or water deficiency can improve WUE (Panda et al., 2003;Sun et al., 2006). In traditional quantitative irrigation, grain yields ranging from 5130 to 6770 kg hm −2 and WUE ranging from 1.33 to 1.81 can be obtained in 16 cultivars (Zhang et al., 2010). Studies to date indicate that wheat produces maximum yield under the condition of 70% relative soil moisture content at the jointing stage (Pan et al., 2014) and 60-70% relative soil moisture content after anthesis (Ma et al., 2007). Wang et al. (2011) showed that the highest grain yield and WUE were obtained at 75% and 70% relative soil T. aestivum L. cultivars under 85, 55 and 25% field water capacity (FC), and found that at 55% field water capacity, CY17 showed a lower net P n than XN889 mainly due to stomatal closure. In our study, the photosynthetic ability of the flag leaves of 'JM22' was higher than that of 'ZY9369' in middle and late grain filling stages, which could be beneficial for accumulation and translocation of photosynthetic products ( Figure 5). The P n of the flag leaves of 'ZY9639' at 30 DAA decreased more sharply with excessive water than that of 'JM22' , which indicated that 'ZY9369' was more sensitive than 'JM22' .
The grain filling substance partly comes from the stored matter before anthesis and partly from photosynthetic product in wheat leaves after anthesis, and approximately 70-90% comes from the latter (Zhou et al., 2012). Soil moisture content remarkably influences dry matter accumulation and distribution in wheat (Ziaei & Sepaskhah, 2003). Drought or waterlogging after anthesis can significantly reduce the photosynthetic capacity, plant dry matter accumulation and the amount of dry matter distriuted to the grain (van Herwaarden et al., 1998;Jiang et al., 2004). However, moderate water deficit can accelerate dry matter transportation to the grain and improve harvest index . Our study showed that a target relative soil moisture content in the 0-140 cm soil layer of 75% at jointing stage, and 65-75% at anthesis stage (W1 or W1′) was the best treatment for high dry matter accumulation at maturity in the two seasons ( Figure 4). Moreover, the ADMA and its contribution rate to the grains in W1 or W1' treatment were 5972-6894 kg hm −2 and 71.2-73.5%, respectively (Table 2). These values decreased significantly in W2 and W2' treatment groups although they were irrigated with more water. The high grain yield of the W1 and W1' treatment groups is likely the result of high translocation rate of dry matter after anthesis. In addition, our results showed that the maximum remobilization amount of dry matter in the aboveground vegetative organs before anthesis and the contribution rate to grains were achieved in the rain-fed treatment groups in the 2009/2010 growing season. However, ADMA during grain filling in rainfed treatments decreased significantly, while moderate SI accelerated dry matter transport to the grains post-anthesis, which likely accounts for the observed high yield.
The amount of accumulated dry matter in post-anthesis was higher in the new cultivar under water-limited conditions than in old cultivars. Nevertheless, no significant difference was found when water was adequate (Fang et al., 2011). Our study found that the ADMA and ADMR of 'JM22' were significantly higher than that of 'ZY9369' under the same water conditions. Compared with W1 and W1' , the ADMA and the CADMA of W2 and W2' , respectively, decreased to 476-598 kg hm −2 and 2.0-2.2% in 'JM22' . By contrast, the ADMA and the CADMA of W2 and W2' , conditions in different rainfall years. Yield gap between the two varieties in the dry season was significantly greater than that in the normal year, indicating that 'JM 22' exhibited a high and stable yield under different irrigation and precipitation conditions. The P n and ФPSII of 'JM 22' was higher than those of 'ZY9369' in the late grain filling stage. Furthermore, the grain yield of 'ZY9369' decreased sharply because of drought and excessive irrigation. 'ZY9369' displayed higher sensitivity to soil moisture than 'JM 22' . (3) Results from the two growing seasons indicated that W1 or W1' treatment is the preferred treatment and 'JM 22' is the recommended winter wheat cultivar to save water and obtain high yields in the NCP.

Disclosure statement
No potential conflict of interest was reported by the authors. moisture contents at jointing and anthesis stages, respectively. Our study was performed under natural conditions, and the grain yield and WUE differed between the two years (Table 3). In the 2009/2010 growing season, total precipitation was 149.1 mm and was evenly distributed. The W1 treatment with 75% and 65% soil moisture contents in jointing and anthesis stages produced the highest grain yield at 9387.3 kg hm −2 and the highest WUE at 21.55 kg hm −2 mm −1 of 'JM22' . In the 2010/2011 growing season, total precipitation was 172 mm, but 109 mm of this was mainly recorded in the anthesis stage. In this case, the W1' treatment reached 75% soil moisture content in the anthesis stage without irrigation, resulting in adequate soil moisture, particularly during the grain-filling period. Furthermore, W1' treatment produced the highest grain yield, WUE, and irrigation efficiency. More irrigation water decreased the yield, WUE and irrigation benefits. Our results showed that treatment with SI at 65% (with a moderate amount of water) or 75% (with no irrigation because the soil moisture content had reached 75% from precipitation alone) at anthesis stage produced the highest yield, WUE and irrigation efficiency, on the basis of soil moisture content control to 75% at jointing stage, which was suitable for irrigation schemes under this experimental condition.
Various cultivars exhibited varying sensitivities to soil water content, which is related to their varying resistance capabilities to drought (Saeedipour & Moradi, 2011b). Under the various experimental soil moisture conditions described here, 'JM 22' produced higher yield and WUE, and exhibited higher yield stability compared to 'ZY9369' at both the rain-fed and irrigated conditions.

Conclusion
This study investigated the effects of SI under varying soil water contents on the photosynthetic characteristics of flag leaves, dry matter accumulation and remobilization, grain filling rate, yield and its component, and WUE of two different winter wheat cultivars. The following results were obtained: (1) the highest grain yield, WUE, amount of translocated dry matter after anthesis, and contribution rate to kernel were obtained when the treatment was administered in the anthesis stage through SI at 65% (with a moderate amount of water) or 75% (with no irrigation because the soil moisture content reached 75% from precipitation alone) at anthesis stage with soil moisture content set at 75% at jointing stage. This finding may be attributed to the significantly improved P n and ФPSII of flag leaves.
(2) The grain yield, spike numbers per ha, 1000-kernels weight, amount of translocated dry matter after anthesis, and contribution rate to kernel of 'JM 22' were significantly higher than those of 'ZY9369' under the same irrigation