Agronomic performance of drought-resistance rice cultivars grown under alternate wetting and drying irrigation management in southeast China

Guang Chu,Tingting Chen,Song Chen,Chunmei Xu,Danying Wang,Xiufu Zhang

China National Rice Research Institute,Chinese Academy of Agricultural Sciences,Hangzhou 310006,Zhejiang,China

Keywords:Agronomic traits Alternate wetting and drying Drought-resistance rice cultivars Grain yield Water use efficiency

A B S T R A C T Compared to drought-susceptible rice cultivars(DSRs),drought-resistance rice cultivars(DRRs)could drastically reduce the amount of irrigation water input and simultaneously result in higher grain yield under water-saving irrigation conditions. However, the mechanisms underlying these properties are unclear.We investigated how improved agronomic traits contribute to higher yield and higher water use efficiency(WUE)in DRRs than in DSRs under alternate wetting and drying(AWD).Two DRRs and two DSRs were field-grown in 2015 and 2016 using two different irrigation regimes:continuous flooding(CF)and AWD.Under CF,no statistical differences in grain yield and WUE were observed between DRRs and DSRs.Irrigation water under the AWD regime was 275–349 mm,an amount 49.8%–56.2%of that(552–620 mm)applied under the CF regime.Compared to CF,AWD significantly decreased grain yield in both DRRs and DSRs,with a more significant reduction in DSRs,and WUE was increased in DRRs,but not in DSRs,by 9.9%–23.0%under AWD.Under AWD,DRRs showed a 20.2%–26.2%increase in grain yield and an 18.6%–24.5%increase in WUE compared to DSRs.Compared to DSRs,DRRs showed less redundant vegetative growth,greater sink capacity,higher grain filling efficiency,larger root biomass,and deeper root distribution under AWD.We conclude that these improved agronomic traits exert positive influences on WUE in DRRs under AWD.

1.Introduction

Global agriculture is currently facing two major challenges:an increase in food demand to feed the rapidly growing world population and a decline in global water resource availability[1].Rice(Oryza sativa L.)is a major food staple for ＞50%of the global population[2].It has been estimated that rice yield must increase by at least 1%annually to meet the challenge of food security for a rapidly growing population[3].However,with industrial development,rapid population growth,and urbanization,irrigation water is becoming increasingly scarce in China [1, 4], and the scarcity of freshwater resources threatens rice production[5].To help meet the food demand under limited water resources, researchers at Shanghai Agrobiological Gene Center(Shanghai,China)bred a new type of rice cultivar characterized by having a yield potential similar to that of the main irrigated lowland rice cultivars in use but with increased drought resistance[6].The development of drought-resistant rices(DRRs)to reduce irrigation water input during rice production has become a critical area of agricultural research in China[6].Although several field studies have demonstrated that DRRs are droughtresistance[7–10],the mechanisms underlying this trait are unclear.

AWD is an effective water-saving irrigation technology[11,12]that has been applied on ＞12 Mha of agricultural land in China each year and is being adopted in Asian countries such as Bangladesh,India,The Philippines,and Vietnam[13,14].There is no doubt that AWD can reduce irrigation water input,but it remains debatable whether this technology can increase or maintain grain yield[10,15–18].A meta-analysis of 56 studies with 528 side-by-side comparisons of AWD with CF found that AWD decreased rice grain yield by 5.4%;however,when soil water potential was higher than ?20 kPa or field water level did not drop below 15 cm from the soil surface,yields were not significantly reduced[15].Lampayan et al.[16]found in the Philippines,Vietnam,and Bangladesh that AWD could reduce irrigation water input by up to 38%with no yield reduction,and increase farmers'income by 17%–38%.Norton et al.[17]found that safe AWD could increase rice grain production compared to CF. In our previous studies, a moderate AWD(soil water potential=?15 kPa)could maintain or even increase grain yield in DRRs or newly bred“super”rice cultivars[10,18].Crops are irrigated 13–15 times under the moderate AWD regime during the entire growth period[19,20].However,in many areas in the lower reaches of Yangtze River,rice receives fewer than 10 irrigations during the entire growth period,owing to an incomplete irrigation system and an expensive irrigation cost (unpublished data). Further investigation into how DRRs perform in relation to grain yield and WUE under water-saving irrigation is thus warranted,particularly under AWD.

Understanding the agronomic traits of high-yielding and high-WUE rice cultivars could provide directions for future breeding and water-management efforts.Various agronomic traits are closely related to increased yield performance and high NUE or WUE in rice,including large sink capacity[21,22],LAI and LAD[23],large shoot biomass,high CGR[24]and NAR[25], and large root biomass[26]. However,our understanding of the agronomic performance of DRRs under AWD is limited.The purposes of this study were to(1)compare yield performance and WUE between DRRs and drought-susceptible rice cultivars (DSRs) under an AWD irrigation regime, (2) compare agronomic traits including percentage of productive tillers,LAD,CGR,NAR,root dry weight and shoot biomass,and NSC accumulation in the stem at heading stage and its remobilization during ripening,between DRRs and DSRs under AWD. Such a study was expected to shed light on the ways in which DRRs cope with water-limited conditions, as well as provide helpful information for breeding new cultivars with drought tolerance and rice water-saving cultivation methods.

2.Materials and methods

2.1.Plant materials and cultivation

The field experiment was conducted at the Fuyang Agricultural Experimental Station of the China National Rice Research Institute(30.30′N,120.2′E,11 m above sea level)in 2015–2016. Fig. S1 shows the daily air temperature and precipitation during the rice seasons of 2015 and 2016.Soil of the experimental field is classified as a Fec-Stagnic Anthrosol(Etisols,US classification).The original soil fertility parameters in 0–20 cm soil layer were as follows:organic matter 38.7 g kg?1,total N 2.02 g kg?1,available N 324.6 mg kg?1,available P 18.5 mg kg?1,and available K 72.1 mg kg?1,with pH 5.79.The gravimetric soil moisture content at field capacity was 0.187 kg kg?1,and the soil bulk density was 1.12 g cm?3.

Two newly bred DRRs and two DSRs were employed in the experiment.Hanyou-113(HY-113,a three-line indica hybrid from Huhan-11A×Hanhui-3)and Hanyou-8(HY-8,a three-line japonica hybrid from Huhan-2A×Xiangqing) were encoded as DRR-1 and DRR-2,respectively.Tianyouhuazhan(TYHZ,a three-line indica hybrid from Tianfeng-A×Huazhan) and Changyou-5(CY-5,a three-line japonica hybrid from Chang-01-11A×CR-27)were encoded as DSR-1 and DSR-2,respectively.All the cultivars are planted over large areas in the lower reaches of the Yangtze River.Seeds of HY-113 and HY-8 were provided by the Shanghai Agrobiological Gene Center(Shanghai,China),and seeds of TYHZ and CY-5 were provided by the Fuyang Seed Company(Fuyang,Zhejiang,China).

Pre-germinated seeds of all the rice cultivars were raised in the field.Sowing dates were May 20 in both 2015 and 2016.Twenty-day-old seedlings were transplanted at a hill spacing of 25 cm×16 cm with two seedlings per hill in both study years.Nitrogen(as urea)was applied at a total rate of 200 kg ha?1,and split-applied with 50% before transplanting, 20% at midtillering,and 30%at panicle initiation.Phosphorus,as single superphosphate,was applied at a rate of 45 kg ha?1before transplanting.Potassium,as potassium chloride,was applied at a total rate of 120 kg ha?1,and split-applied with 50%before transplanting and 50%at panicle initiation.HY-113 and TYHZ headed(half of total panicles)on August 15–17 and were harvested on October 1,HY-8 and CY-5 headed on August 30–31 and were harvested on October 20.

2.2.Treatments

Fig.1–Irrigation water for rice under various irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

In both study years,the field experiment was laid out in a randomized complete block design with three replicates.Each plot was 30 m2(6 m×5 m)in area and separated from the others by concrete walls(0.5 m in width and 1.0 m in depth).Two irrigation treatments,CF and AWD,were applied from the second week after transplanting until maturity. Field water was kept a depth of 2–3 cm under both irrigation regimes during the first week to allow the seedlings to recover and to control weeds.Under the AWD regime,plants were not re-watered until the soil water potential reached ?30 kPa at a depth range of 15–20 cm.With the exception of drainage at midseason,water depth under the CF regime was maintained at 2–3 cm from one week after transplanting to one week before harvest. Our preliminary experiments found that crops are irrigated fewer than 10 times under this soil water potential during the entire growth period and that this soil water potential reduced yield by about 25%–30%compared to the CF regime for some rice cultivars typically grown in this area.In each AWD regime plot,soil water potential was measured with a tensiometer with a 5-cm-length sensor at a soil depth range of 15–20 cm.In each AWD regime plot,four tensiometers were installed and readings were recorded at 11:30 a.m.each day.Plots were flooded 2–3 cm deep until thereading reached the threshold.The amount of water input in each plot was recorded with a water meter(model LXSGEBMYC-15, Hangzhou Water Meter Manufacturing Factory,Hangzhou,China)installed in the irrigation pipes.

Table 1–Number of tillers and the proportion of productive tillers of rice under two irrigation regimes.

Table 2–Leaf area index(LAI)of rice under two irrigation regimes.

2.3.Parameter measurements

In each plot,20 hills of plants were selected to record the number of tillers from transplanting until heading at intervals of five days and at maturity.The proportion of productive tillers was calculated as the ratio of the tillers at maturity to those at jointing.Shoot biomass and LAI were measured at jointing,heading and maturity stages.Plants from six hills were sampled and measured in each plot.After cleaning,plant samples were separated into leaves,stems,and panicles(at heading and maturity),and dry weight of each part was determined after oven drying at 70°C to constant weight.The concentrations of soluble sugars and starch were determined and calculated following Yoshida et al.[27]and Li et al.[28].Oven-dried plant samples were ground into fine powder and filtered through a 1-mm sieve.A ground sample of 0.1 g was added to 10 mL of ethanol and incubated in a 15 mL centrifuge tube at 80°C for 0.5 h and then centrifuged at 2000 r min?1for 20 min after cooling to room temperature.The supernatant was transferred to a 100 mL volumetric flask and the extraction was repeated three times.The three supernatants were pooled in the flask,followed by addition of distilled water to 100 mL.An aliquot of the extract was used for determination of soluble sugars with anthrone reagent.For starch determination,the residue after centrifugation in the tube was added to 2 mL of distilled water and then shaken in a boiling water bath for 15 min.Perchloric acid(9.36 mol L?1)of 2 μL was added to the tube,followed by immersion in an ice bath for 15 min for complete digestion of starch into glucose.The supernatant of the extract after centrifugation was collected in a 100 mL volumetric flask.The extraction was repeated by placing the residue in 2 mL of 4.68 mol L?1perchloric acid for 15 min for a second time.The supernatants were pooled and made to 100 mL with distilled water. For the colorimetric assay,optical density was measured at 620 nm in a spectrophotometer(UV-1800,Shimadzu,Tokyo,Japan).Glucose released in the extraction was estimated with anthrone reagent and converted to a starch value by multiplying by 0.9.The NSC concentration of a given plant part refers to the sum of the concentrations of soluble sugars and starch.NSC accumulation in stems(g m?2)was calculated as stem biomass multiplied by the NSC concentration.NPS was defined as the ratio of NSC accumulation in stems per m2to spikelets per m2. Leaf area was determined with a leaf area meter(LI-3000C,Li-Cor,Lincoln,NE,USA).LAD,CGR,and NAR were calculated as follows:

Fig.2–Leaf area duration of rice during vegetative,reproductive,and ripening stages under various irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

where L,M,and t refer to LAI,shoot biomass(g m?2),and growing days at measurement,respectively;the subscripts 1 and 2 refer to the first and second measurements.

Roots were sampled at heading.For sampling,roots were dug out by core sampling,including a soil volume around roots of 0.008 m3(0.25 m×0.16 m×0.20 m).This root sample contains about 95%of the total root dry weight[29].For each measurement,plants from six hills in each plot were pooled.Each root sample with soil was divided into 0–10 cm and 10–20 cm segments. After careful washing with a hydropneumatic elutriation device(Gillison's Cultivar Fabrication Inc.,Benzonia,MI,USA),roots were oven-dried at 70°C to constant weight and then weighed.

Grain yield and yield components were measured following Yoshida et al.[27].Grain yield was measured from a harvest area of 6.0 m2in the middle of each plot at maturity and adjusted to 14%moisture.Yield components(panicle number per square meter,spikelet number per panicle,grain filling rate,and 1000-grain weight) were derived from plants in1 m2,excluding border plants,randomly located in each plot.Grain filling rate was calculated as the ratio of filled grains(using a NaCl solution with specific gravity ≥1.06 g cm?3) to total spikelets.HI and WUE were calculated as follows:

Fig.3–Shoot biomass at stages of jointing,heading,and maturity of rice under various irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

2.4.Statistical analysis

Analysis of variance(ANOVA)was calculated in SPSS 20.0(SPSS,Chicago,IL,USA)with year and replicate as random effects and irrigation,cultivar,and their interactions as fixed effects.Data from each sampling date were analyzed separately.Multiple comparisons of means were made with the LSD test at the 0.05 probability level.

3.Results

3.1.Soil water potential

Although total rainfall during the rice growing season was greater in 2015(668 mm)than in 2016(498 mm),the difference in rainfall during the reproductive and ripening periods was relatively small between 2015 and 2016 in Fuyang(299 mm in 2015,278 mm in 2016)(Fig.S1).Irrigation water was 275–349 mm under the AWD regime,or 49.8%–56.2%of that(552–620 mm)applied under the CF regime(Fig.1).In the absence of rain, it took 8–10 days to reach a soil water potential of ?30 kPa under the AWD regime,depending on the plant growth stage(Fig.S2).The CF regime received 22–24 irrigations for the indica rice cultivars(HY-113 and TYHZ)and 24–26 irrigations for the japonica rice cultivars(HY-8 and CY-5),and the AWD regime received 7–8 irrigations for the indica rice cultivars (HY-113 and TYHZ) and 8–10 irrigations for the japonica rice cultivars(HY-8 and CY-5).The field was irrigated 1–2 time(s)fewer in 2015 than in 2016,owing to greater rainfall during the vegetative period in 2015 than in 2016.Under the same irrigation regime,the differences in soil water potential were not significant between different rice cultivars when the irrigation regime was the same(Fig.S2).

3.2.Tiller number,leaf area index,and duration

Fig.4–Crop growth rate of rice during vegetative,reproductive,and ripening stages under two irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

Under CF,the difference in tiller number at all the measurement periods and the percentage of productive tillers was not significant between DRR-1 and DSR-1 or between DRR-2 and DSR-2.Owing to greater precipitation during the vegetative period in 2015 (369 mm) than in 2016 (220 mm), tiller number at jointing stage was higher in 2015 than in 2016 for all the cultivars under AWD(Table 1 and Fig.S1).Under AWD, tiller number showed no significant difference at jointing stage between DRR-1 and DSR-1 or between DRR-2 and DSR-2,and tiller number was greater for DRR-1 than DSR-1 or DRR-2 than DSR-2 at heading and maturity stages,and the percentage of productive tillers was significantly greater for DRR-1 than for DSR-1 and for DRR-2 than for DSR-2(Table 1).

Compared to CF,AWD significantly reduced LAI at all measurement periods and reduced LAD during the full growth period(Table 2 and Fig.2).Coinciding with tiller number,LAI at jointing and LAD during the vegetative period were higher in 2015 than in 2016 for all cultivars under AWD(Table 2,Fig.2,and Fig.S1).LAI in all measurement periods and LAD during the full growth period did not significantly differ between DRR-1 and DSR-1 or between DRR-2 and DSR-2 under CF(Table 2 and Fig.2).LAI at jointing stage and LAD during the vegetative period did not significantly differ between DRR-1 and DSR-1 or between DRR-2 and DSR-2 under AWD,and total LAI at heading and maturity and LAD during reproductive and ripening periods were higher for DRR-1 than for DSR-1 and for DRR-2 than for DSR-2 under AWD (Table 2 and Fig. 2).Furthermore,compared to DSRs,DRRs showed a higher ratio of effective LAI to total LAI under AWD(Table 2).

3.3.Shoot biomass,crop growth duration,and net assimilation rate

As with tillers,shoot dry weight at jointing and CGR during the vegetative period were higher in 2015 than in 2016 under AWD owing to greater rainfall in 2015(Figs.3,4,and S1).Shoot biomass at all periods and CGR and NAR during the total growth periods did not significantly differ between DRR-1 and DSR-1 or between DRR-2 and DSR-2 under CF,and there was no significant difference in shoot biomass at jointing and CGR during the vegetative period between DRR-1 and DSR-1 or between DRR-2 and DSR-2 under AWD(Figs.3,4,and 5).Shoot biomass at heading and maturity,CGR,and NAR during the reproductive and ripening periods were significantly higher for DRR-1 than DSR-1 and for DRR-2 than for DSR-2 under AWD(Figs.3,4,and 5).

Fig.5–Net assimilation rate of rice during vegetative,reproductive,and ripening stages under two irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

3.4.Root biomass

Total root biomass under AWD markedly decreased by 18.9%–24.6% in DSRs compared to those under CF, whereas no significant difference in DRRs was found between CF and AWD (Table 3). Root biomass did not significantly differ between DRR-1 and DSR-1 or between DRR-2 and DSR-2 under CF,and root biomass was significantly higher for DRR-1 than for DSR-1 and for DRR-2 than for DSR-2 under AWD(Table 3).Compared to CF,AWD induced a significant increase in the root:shoot ratio of all rice cultivars at heading time,and a more significant increase in DRRs,indicating that AWD promotes root growth,particularly in DRRs(Table 3).Further observation showed that when total root weight was divided into 0–10 cm and 10–20 cm segments,DRR-1 or DRR-2 had larger root biomass in the 10–20 cm soil layer and a higher deep root proportion(the ratio of root dry weight at 10–20 cm depth to root dry weight at 0–20 cm depth)than DSR-1 or DSR-2 under AWD(Table 3),implying that DRRs has a deeper root distribution in soil than do DSRs under AWD.

NS means non-significant at P=0.05.

3.5.NSC in stems and NSC remobilization

AWD induced a significant reduction in NSC accumulation(27.2%–32.8%)in stems of DSRs at the heading stage and by 10.3%–13.9%for DRRs,compared to that under CF(Table 4).For DRRs,the amount of NPS at heading was 7.21–7.45 mg under AWD and 6.8%–10.0%higher than that under CF,indicating that DRRs have higher ratios of source to sink under AWD(Table 4).However,the NPS of DSRs did not significantly differ betweenCF and AWD.Furthermore,AWD induced a significant increase in NSC remobilization from stems,as well as in the contribution of NSC to grains in the four rice cultivars compared to that under CF,particularly in DRRs(Table 4).

Table 3–Root dry weight and root–shoot ratio of rice under two irrigation regimes.

3.6.Grain yield and water use efficiency

Compared to CF,grain yield under AWD was significantly lower for both DRRs and DSRs(Table 5).Furthermore,DRR-1 or DRR-2 showed significant higher grain yield(by 20.2%–26.2%)than DSR-1 or DSR-2 under AWD,although there was no significant difference in grain yield under CF(Table 5).For DSRs,a lower grain yield under AWD was attributable mainly to a 21.3%–25.0%reduction in the number of panicles and a 8.6%–11.4% reduction in spikelets per panicle (Table 5).However, AWD induced a 16.4%–19.0% reduction in the number of panicles in DRRs,whereas the number of spikelets per panicle in DRRs did not significantly differ between CF and AWD.Furthermore,AWD increased the grain-filling rate for DRRs(Table 5).Compared to that under CF,WUE for DRRs under AWD increased by 9.9%–13.1%in 2015 and 19.1%–23.0%in 2016.AWD had no effect on improving WUE in DSRs than CF,except that AWD significantly reduced WUE by 12.1%in CY-5 in 2015(Fig.6).

4.Discussion

Compared to DSRs,HY-113 and HY-8 showed similar yield potential under the CF regime,as well as stronger drought resistance and higher yield under AWD.Similar results have been reported previously:DRRs produces higher grain yield and simultaneously required less irrigation water input than common rice cultivars under water-saving irrigation[10,30].Agronomic traits of rice are closely associated with yield performance,NUE,or WUE[21–26];however,the agronomic performance of DRRs under AWD is little known.In this study,we observed that the main agronomic traits of DRRs that are related to higher yield and higher WUE under AWD were(1)greater sink capacity,as evidenced by a higher number of spikelets per panicle,more tillers at maturity,and a greater percentage of filled kernels;(2)less redundant vegetative growth as a result of a greater percentage of productive tillers and a higher ratio of effective to total LAI at the heading stage;(3)higher grain-filling efficiency as a result of larger LAD,CGR,and NAR during the ripening period,greater NSC remobilization from the stem to kernels during the ripening period and a greater amount of NPS at anthesis;and(4)larger root biomass and deeper root distribution.

Table 4–NSC accumulation in stems at heading and maturity,NPS at heading,and NSC remobilization during grain filling of rice under two irrigation treatments.

DRRs had higher yield and higher WUE under AWD,differences that could be attributed to a greater number of spikelets per panicle.The application of AWD resulted in 8.6%–11.4%reduction in the number of spikelets per panicle in DSRs,although no significant difference was observed between AWD and CF for DRRs.The mechanism underlying the maintenance of a high number of spikelets per panicle DRRs under AWD remains unclear.There was a higher CGR during the reproductive period of DRRs than during that of DSRs under AWD. In previous studies, greater CGR during the reproductive period could increase sink capacity by promoting spikelet differentiation,reducing spikelet degeneration,and increasing endosperm cells proliferation at the early seed development stage[31,32].We speculate that the observed greater CGR during the reproductive period is attributable to the higher number of spikelets per panicle,thereby leading to a higher grain yield and WUE in DRRs.

In previous studies, less redundant vegetative growth could improve canopy quality,in turn reducing water and nitrogen use in unproductive tillers and increasing radiation use efficiency[33–35].In the present study,DRRs showed less redundant vegetative growth during the vegetative period under AWD, as evidenced by the greater percentage of productive tillers and higher ratio of effective to total LAI at the heading stage.We suspect that less redundant vegetative growth during the vegetative period under AWD leads to higher yield and higher WUE in DRRs.

Grain filling is the final stage of growth when fertilized ovaries develop into caryopses[36].Compared to CF,AWD significantly increased grain filling rate in DRRs,whereas grain filling rate in DSRs did not significantly differ between CF and AWD. However, the mechanism in which DRRs shows higher grain filling rate than DSRs under either AWD is not fully understood.Based on the results of our present investigation,there are two possible explanations.First,DRRs have higher CGR,LAD,and NAR than DSRs under AWD during the ripening period.A greater ability to produce matter during the ripening period could enhance grain filling by increasing sink strength and source activity[37].Second,DRRs have higher rates of NSC accumulation in stems at heading and promote more pre-stored carbon remobilization from stems during the ripening period,which in turn could increase grainfilling rates under AWD.Furthermore,NPS did not significantly differ between CF and AWD in DSRs,whereas AWD significantly increased NPS in DRRs,indicating a higher ratio of source to sink in DRRs under AWD.In previous studies,NPS at heading and the amount of NSC in the stems are strongly related to sink strength and grain-filling rate[38],and an increase in carbon remobilization from the vegetative tissues to grains contributes to higher rice grain yield[36].It is thus possible that a larger CGR, LAD, NAR and an enhanced remobilization of accumulated NSC from stems to kernels during the ripening period contribute to an increase in grainfilling efficiency and harvest index under AWD in DRRs,resulting in better yield performance and higher WUE.

Table 5–Grain yield and yield components of rice under two irrigation regimes.

Fig.6–Water use efficiency for rice under two irrigation treatments in 2015 and 2016.HY-113 and HY-8 are DRRs;TYHZ and CY-5 are DSRs.Vertical bars represent±standard error of the mean where they exceed the size of the symbol.

Root biomass is regarded as the most important root morphological trait because root biomass is positively correlated with root N uptake and water uptake ability[39].In this study,root biomass did not significantly differ between DRRs and DSRs under CF. However, DRRs showed larger root biomass and deeper root distribution than DSRs under AWD.Given that Yang et al.[39]found that extensive root biomass could support large shoot dry matter accumulation, we speculate that the improved root growth contributed to higher grain yield and higher WUE in DRRs under AWD.In addition,we observed higher root biomass at the 10–20 cm soil depth and a higher ratio of root biomass at 10–20 cm to total root biomass in DRR than in DSRs under AWD.In our earlier studies[10,18],rice plants with higher percentage of root biomass in the deep soil layer(10–20 cm)to total root biomass could show higher yield and higher WUE.We speculate that larger root biomass and deeper root distribution in DRRs contribute to higher yield and WUE under AWD.

5.Conclusions

This study confirms previous findings that AWD could reduce irrigation water input compared to the CF regime.Compared to CF,AWD significantly reduced grain yield in both DRRs and DSRs,with a more significant reduction in DSRs.Under AWD,DRRs showed a 20.2%–26.2%increase in grain yield and an 18.6%–24.5%increase in WUE in comparison with DSRs.WUE was increased in DRRs but not in DSRs.by 9.9%–23.0%under AWD.Less redundant vegetative growth,greater sink capacity,higher grain-filling efficiency,larger root biomass,and deeper root distribution are important agronomic traits that are closely related to higher grain yield and WUE in DRRs subjected to AWD.

Acknowledgments

This study was supported by the National Key Research and Development Program of China (2016YFD0300507,2016YFD0300108),the National Natural Science Foundation of China (31671630, 31671638, 31501264), and the China Agriculture Research System(CARS-01).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2018.04.005.