Grain yield, yield components and WUE
In this study, the grain yield, spike number, grain number and 1,000-grain weight (TGW) of winter wheat were significantly affected by the year (Y) and irrigation time(I) (Table 3). The grain yield and TGW were also significantly affected by the interaction of year and irrigation time. The spike number first increased and then decreased with the delay in the irrigation time and peaked in the L4 treatment. Compared with L4, the average spike number in L3, L5 and L6 was reduced by 4.6%, 9.9% and 17.2%, respectively. However, the grain number decreased with the delay of irrigation time, showing that L4, L5, and L6 were 3.9%,7.8% and 16.5% lower than L3, respectively. The grain number in L6 was significantly lower than that in the other treatments, while the difference between L3, L4 and L5 was not significant. However, the TGW significantly increased with the delay of irrigation time. Compared with L3, for every postponement of irrigation for one leaf age, the average TGW increased by 1.9 g, 2.4 g, and 2.9 g.
Compared with 2020–2021, the grain yield of 2018–2019 and 2019–2020 was significantly reduced. The annual difference in grain yield was attributed to rainfall and other meteorological conditions. The lower precipitation after anthesis in 2018–2019 resulted in a lower TGW, and the low temperature in the spring of 2019–2020 resulted in a lower spike number (April 9–10). Despite the annual difference in GY, the highest yield during the three growth seasons was obtained for the L4 treatment. Except for the GY of L3 and L4 treatments in 2019–2020, there was no significant difference, and the other treatments were significantly lower than that of L4. Compared with L4, the average GY of L3, L5 and L6 decreased by 7.5%, 8.9% and 21.7%, respectively.
Table 3
Effects of different irrigation treatments on grain yield, yield components and water use efficiency in winter wheat. Different letters indicate significant difference treatments at P < 0.05 level. Y, year; I, irrigation time. *, **, and *** indicate significant effects at P < 0.05, P < 0.01 and P < 0.001, respectively; ns indicates no significant effect. Values are means ± standard error (n = 3). ANOVA, analysis of variance; WUE, water use efficiency.
Year | Treatment | Spike Number (×104 ha− 1) | Grain numbers per spike | 1000-grain Weight(g) | Grain Yield (kg ha− 1) | WUE (kg ha− 1 mm− 1) |
2018–2019 | L3 | 700.7 b | 34.3 a | 30.5 d | 6692.7 b | 16.5 c |
L4 | 740.3 a | 33.1 a | 32.7 c | 7644.2 a | 19.5 a |
L5 | 676.5 b | 31.5 ab | 34.1 b | 6601.6 b | 17.3 b |
L6 | 594.0 c | 28.8 b | 37.9 a | 5884.0 c | 16.1 c |
2019–2020 | L3 | 779.9 ab | 29.3 a | 35.5 d | 7436.9 ab | 16.6 b |
| L4 | 818.4 a | 27.5 a | 37.5 c | 7631.3 a | 17.6 a |
| L5 | 726.0 bc | 26.3 ab | 41.1 b | 7259.3 b | 17.0 b |
| L6 | 676.5 c | 23.3 b | 43.6 a | 6012.9 c | 14.4 c |
2020–2021 | L3 | 674.3 ab | 35.3 a | 34.7 d | 7626.6 b | 17.6 b |
| L4 | 699.6 a | 34.4 ab | 36.2 c | 8233.3 a | 19.5 a |
| L5 | 632.5 bc | 33.4 ab | 38.3 b | 7555.4 b | 18.1 b |
| L6 | 600.6 c | 30.5 b | 40.7 a | 6518.8 c | 16.5 c |
ANOVA | Year (Y) | *** | *** | *** | *** | *** |
Irrigation(I) | *** | *** | *** | *** | *** |
Y×I | ns | ns | *** | *** | *** |
Leaf area index (LAI), dry matter accumulation and harvest index (HI)
The LAI at the anthesis stage reached its maximum value at L3, and the three-year average was 6.8. Compared with L3, the LAI decreased with the delay in irrigation (L4, L5, and L6) by 7.4%, 15.2% and 30.1%, respectively (Fig. 3). After the anthesis period, the LAI of each treatment began to decrease, but the range of decrease differed. L3 and L6 decreased significantly from 15 days after anthesis (DAA) in 2018–2019 and 2020–2021, while those from 15 DAA responded similarly in 2019–2020. The LAI of L4 and L5 decreased from DAA 15–20, but the rate of decrease was lower than those of L3 and L6. After 20 DAA, the LAI of L4 and L5 were significantly higher than those of the other treatments. This indicated that a moderate delay in irrigation could effectively prolong the LAI of grain filling in the middle and late stages.
Dry matter accumulation at the anthesis stage (DMA) decreased with the delay of irrigation time (Fig. 4), and the DMA at L3 was significantly higher than those of the other treatments. Dry matter accumulation at maturity (DMM) reached its maximum value at L4, with a three-year average of 17,925.7 kg ha− 1, and there was no significant difference between L3 and L4. However, they were significantly higher than those of L5 and L6. Compared with L4, the DAM of L5 and L6 were 6.0% and 14.9% lower than that of L4, respectively.
There were differences in the HI values during the three-year period owing to the effects of the DMA and grain yield. However, L4 achieved the highest HI over three years with an average of 0.44 (Fig. 5), which was significantly higher than that of other treatments in 2018–2019 and 2021–2022. Compared with L4, L3, L5 and L6 were reduced by 6.0%, 4.5% and 6.5%, respectively.
Water consumption
Table 4
Effects of different irrigation time on total water consumption (ET), soil water consumption (ΔSW) and their water consumption ratios during the 2018–2021 wheat growing seasons.
Year | Treatment | ET (mm) | I (mm) | P (mm) | ΔSW (mm) | Ratio to total water consumption (%) |
Irrigation | Precipitation | Soil water |
2018–2019 | L3 | 405.4 a | 90 | 85.9 | 229.5 a | 22.2 c | 21.2 c | 56.6 a |
| L4 | 391.6 b | 90 | 85.9 | 215.7 b | 22.9 b | 21.9 b | 55.1 b |
| L5 | 381.8 b | 90 | 85.9 | 205.9 b | 23.6 b | 22.5 b | 53.9 b |
| L6 | 366.0 c | 90 | 85.9 | 190.1 c | 24.6 a | 23.5 a | 51.9 c |
2019–2020 | L3 | 446.8 a | 90 | 176.2 | 180.7 a | 18.2 c | 35.5 c | 36.4 b |
| L4 | 432.3 b | 90 | 176.2 | 166.1 b | 20.1 b | 39.4 b | 38.2 a |
| L5 | 427.6 b | 90 | 176.2 | 161.5 bc | 20.8 a | 40.8 b | 37.4 a |
| L6 | 417.2 c | 90 | 176.2 | 139.6 c | 21.1 a | 41.2 a | 32.6 c |
2020–2021 | L3 | 434.3 a | 90 | 103.4 | 240.9 a | 20.7 b | 23.8 b | 55.5 a |
| L4 | 423.1a | 90 | 103.4 | 229.7 a | 21.3 b | 24.5 b | 54.3 a |
| L5 | 416.5 ab | 90 | 103.4 | 223.1 ab | 21.6 ab | 24.8 ab | 53.8 ab |
| L6 | 396.0 b | 90 | 103.4 | 202.6 b | 22.7 a | 26.1 a | 51.2 b |
ANOVA | (Y) | *** | | | *** | *** | *** | *** |
(I) | *** | | | *** | *** | *** | *** |
Y×I | ** | | | *** | ** | *** | * |
Before the first irrigation in spring during the three growing seasons, the soil moisture in L4, L5, and L6 all reached the level of mild stress (Fig. 6). The duration of drought exposure increased with the delay of irrigation time, showing an average of 5, 13, and 18 days in L4, L5, and L6, respectively. On the anthesis stage, the soil moisture in L3 reached the moderate drought stress level in 2021–2022, while other treatments reached the mild drought stress level. The duration of post-anthesis drought stress gradually decreased with the delay in irrigation time, and there were differences in the influence of rainfall in different years. From the 3-leaf age irrigation to maturity, the plants in L4 were exposed to drought stress for the shortest duration (37 d on average) during the three-year period, while L3, L5, and L6 were 40 d, 42 d, and 43 d, respectively.
The proportion of irrigation water in the total water consumption was 20.3%, 20.1% and 21.6% on average during three wheat growing seasons in 2019, 2020, and 2021, respectively. Precipitation accounted for 22.3%, 39.2% and 24.8% of the total water consumption during these years, respectively. The proportion of soil water consumption to total water consumption was 54.4%, 36.2% and 53.7% on average during these years, respectively (Table 4). This shows that precipitation ratio only slightly varied under different annual types, while precipitation as a percentage of ET and soil water consumption as a percentage of ET varied substantially under different annual types, which was caused by different amounts of rainfall during the wheat season.
Year (Y), irrigation (I) treatment and their interaction significantly affected the total water consumption and stage water consumption. With the delay in the time of irrigation, the water consumption and soil water consumption gradually decreased. Compared with L3, delaying the irrigation time (L4, L5, and L6) reduced the consumption of water by 13.2 mm, 20.2 mm and 35.8 mm for L4, L5 and L6, respectively (Table 4). In 2018–2020, there was no significant difference in the consumption of total water between L4 and L5, but they were significantly lower than L3 but higher than L6. In 2020–2021, the difference in total water consumption between L3, L4 and L5 was not significant, but they were significantly higher than that of L6. The total soil water consumption in the three growth seasons was consistent with the trend of rainfall, which is expressed as the total water consumption between 2019–2020, which was the highest, followed by 2020–2021 and 2018–2019. This was owing to higher increases in rainfall, soil consumption and evapotranspiration.
Table 5
Water consumption characteristics in different wheat growth stages. CA, water consumption amount; CD, daily water consumption; CP, water consumption percentage; I, irrigation; Y, year.
Year | Treatment | Sowing to jointing | Jointing to anthesis | Anthesis to maturity |
CA (mm) | CD (mm d− 1) | CP (%) | CA (mm) | CD (mm d− 1) | CP (%) | CA (mm) | CD (mm d− 1) | CP (%) |
2018–2019 | L3 | 175.4 a | 1.0 a | 43.3 a | 115.3 a | 4.0 a | 28.7 ab | 114.7 b | 3.5 b | 28.7 c |
| L4 | 145.4 b | 0.8 b | 37.0 b | 119.3 a | 4.1 a | 30.7 ab | 126.9 a | 3.4 b | 32.3 b |
| L5 | 145.4 b | 0.8 b | 38.0 b | 112.3 a | 3.9 a | 29.3 a | 124.1 a | 3.4 b | 32.7 b |
| L6 | 145.4 b | 0.8 b | 39.7 b | 90.3 b | 3.2 b | 24.7 b | 130.3 a | 3.7 a | 35.7 a |
2019–2020 | L3 | 173 a | 1.0 a | 38.7 a | 115.0 a | 4.0 a | 25.7 bc | 158.8 b | 4.5 b | 35.7 c |
| L4 | 140.8 b | 0.8 b | 32.7 b | 125.9 a | 4.3 a | 29.3 a | 165.6 ab | 4.4 b | 38.3 bc |
| L5 | 140.8 b | 0.8 b | 33.0 b | 119.5 a | 4.1 a | 28.0 ab | 167.3 ab | 4.4 b | 39.3 ab |
| L6 | 140.8 b | 0.8 b | 34.0 b | 99.5 b | 3.5 b | 24.0 c | 176.9 a | 4.9 a | 42.3 a |
2020–2021 | L3 | 189.4 a | 1.0 a | 43.7 a | 123.6 ab | 4.3 ab | 28.7 ab | 121.4 b | 3.8 b | 27.7 c |
| L4 | 152.8 b | 0.9 b | 36.3 b | 130.8 a | 4.5 a | 31.0 a | 139.5 a | 3.9 b | 33.0 b |
| L5 | 152.8 b | 0.9 b | 36.7 b | 123.4 ab | 4.3 ab | 29.7 a | 140.3 a | 3.9 b | 33.7 b |
| L6 | 152.8 b | 0.9 b | 38.3 b | 101.4 b | 3.4 b | 24.0 b | 141.8 a | 4.2 a | 37 a |
ANOVA | Year(Y) | *** | ns | *** | *** | ns | ns | *** | *** | *** |
| Irrigation(I) | *** | *** | *** | *** | *** | *** | *** | *** | *** |
| Y×I | ns | ns | ns | * | ns | ns | ** | ns | ns |
Year and irrigation significantly affected the water consumption (CA), water consumption intensity (CD) and water consumption modulus coefficient (CP) in growth stages, and the interaction of year and irrigation significantly affected CA (Table 5). The CA, CD and CP from the sowing to jointing stages showed that the irrigated treatment (L3) was significantly higher than those of the non-irrigated treatments (L4, L5, and L6). CA and CD were significantly higher in L3, L4 and L5 than in L6 from the jointing to anthesis stages, while the CP from the jointing stage to anthesis stage was L4 > L5 > L3 > L6. The CA from anthesis to maturity in L3 was significantly lower than those of L4, L5, and L6. The results showed that delayed irrigation enables winter wheat to use more water during the grain-filling period.
Dynamics of the wheat population
The tiller numbers of wheat in different treatments showed a unimodal change, reaching a peak at the standing stage, and then decreasing with the growth process (Fig. 7) at the jointing stage. The total tiller number in the irrigation treatment (L3) was significantly higher than those in the non-irrigated treatments (L4, L5, and L6). From jointing to anthesis, the tiller death rate of L3, L4 and L5 increased with the delay in irrigation time. The average number of tillers at booting decreased by 6.4%, 13.6%, and 20.0%, respectively, compared with L3, and decreased by 2.7%, 9.8% and 16.8% at anthesis, respectively. L3 decreased rapidly compared with L4, L5 and L6 after anthesis, resulting in a lower spike number than L4, and the difference reached a significant level in 2018–2019 (Fig. 7a). These results indicated that delayed irrigation reduced the number of tillers before anthesis, but irrigating too early (L3) will cause a rapid decline in tiller numbers after anthesis.
The senescence process and photosynthetic characteristics
The photosynthetic rate (Pn) and chlorophyll content (Chl) showed a consistent trend of first increasing and then decreasing after anthesis, but the peak time of different treatments of limited irrigation differed (Fig. 8, 9). L3 and L6 reached their peaks at 5–10 DAA, while L4 and L5 reached their peaks at 10–15 DAA. Both Pn and Chl obtained their maximum values at L4, which were 27.3 and 4.5, respectively. The Pn peaks of L3, L5, and L6 were 5.8%, 3.2%, and 4.4% lower than those of L4, respectively. In addition, the Chl peaks of L3, L5, and L6 were 6.8%, 2.7%, and 3.0% lower than those of L4. This indicated that moderately delayed irrigation in L4 increased the maximum photosynthetic rate and chlorophyll content. After the peak value, the Pn and Chl trends were L4 > L5 > L6 > L3, and the Pn and Chl of L3 and L6 decreased rapidly compared with those of L4 and L5.
Compared to other treatments, L4 significantly increased the activity of the SOD in flag leaves after anthesis (Fig. 10a, b, c). The accumulation of MDA in L3 and L6 at the anthesis stage was significantly higher than that in L4, and the difference gradually increased with the progress of the filling period (Fig. 10d, e, f). The accumulation of MDA in L5 was significantly higher than that of L4 at 10 to 20 DAA. Overall, L4 enhanced the activity of leaf antioxidant enzyme activity and inhibited the excessive accumulation of MDA, which ultimately delayed leaf senescence.
Correlation analysis of yield factors, water consumption characteristics and physiological indicators
A significant positive correlation was observed between GY, SN and LAI, whereas LAI significantly positively correlated with DMM and CA-JA (Fig. 11). CA-JA also significantly positively correlated with GY. GN significantly positively correlated with DMM and MDA. There was a significant positive correlation between Pn and Chl. TGW significantly negatively correlated with DMM, GN and DMA. MDA significantly negatively correlated with Pn, Chl and SOD. No significant correlation was observed between HI, WUE, and the other indicators.