DU Xiang-bei ,XI Min ,WEI Zhi ,CHEN Xiao-fei ,WU Wen-ge# ,KONG Ling-cong#

1 Crop Research Institute, Anhui Academy of Agricultural Sciences, Hefei 230031, P.R.China

2 Rice Research Institute, Anhui Academy of Agricultural Sciences, Hefei 230031, P.R.China

Abstract The yield of wheat in wheat–rice rotation cropping systems in the Yangtze River Plain,China,is adversely impacted by waterlogging.A raised bed planting (RBP) pattern may reduce waterlogging and increase the wheat yield after rice cultivation by improving the grain number per spike.However,the physiological basis for grain formation under RBP conditions remains poorly understood.The present study was performed over two growing seasons (2018/2019 and 2019/2020) to examine the effects of the planting pattern (i.e.,RBP and flat planting (FP)) on the floret and grain formation features and leaf photosynthetic source characteristics of wheat.The results indicated that implementation of the RBP pattern improved the soil–plant nitrogen (N) supply during floret development,which facilitated balanced floret development,resulting in a 9.5% increase in the number of fertile florets per spike.Moreover,the RBP pattern delayed wheat leaf senescence and increased the photosynthetic source capacity by 13.9%,which produced more assimilates for grain filling.Delayed leaf senescence was attributed to the resultant high leaf N content and enhanced antioxidant metabolism.Correspondingly,under RBP conditions,7.6–8.6% more grains per spike were recorded,and the grain yield was ultimately enhanced by 10.4–12.7%.These results demonstrate that the improvement of the spike differentiation process and the enhancement of the leaf photosynthetic capacity were the main reasons for the increased grain number per spike of wheat under the RBP pattern,and additional improvements in this technique should be achievable through further investigation.

Keywords: grain number,floret development,photosynthetic capacity,wheat grown after rice

1.Introduction

Wheat (TriticumaestivumL.) is grown worldwide and provides the global population with more than 20% of the total dietary calorie requirements (Shiferawet al.2013).Waterlogging significantly restrains wheat production globally,particularly in the rice–wheat rotation regions in South and Southeast Asia (Arakiet al.2012;Tiryakio?luet al.2015;Wuet al.2015).For example,the Yangtze River Plain,China,is one of the main food grain production areas (Dinget al.2013),where the problem of impacts on food security attributed to waterlogging is very serious.Heavy and unevenly distributed rainfall in these areas leads to frequent waterlogging during the wheat-growing season (Duet al.2021).Waterlogging considerably affects the survival,growth and development of wheat and results in a severe yield decline (Shaoet al.2013;Wuet al.2018;Ciancioet al.2021).

To overcome waterlogging during the growth period of wheat cultivated after rice,the raised bed planting pattern (RBP) (Fig.1) was developed (Duet al.2021,2022).This technique modifies the field microtopography,improves the water drainage efficiency,and decreases the soil water content.This production practice has been partially implemented in Australia and China (Bakkeret al.2007,2010;Maniket al.2019;Duet al.2022).Duet al.(2021) indicated that in the Yangtze River Plain,the wheat grain yield increased by 11.3–14.1% under the RBP pattern,and the increase in yield was mostly attributed to an improvement in the grain number per spike of 7.9–8.7%over that achieved under conventional flat planting (FP)conditions.However,the physiological mechanism driving the increase in the number of grains per spike under the RBP pattern remains unclear.

Fig.1 Layout of the different wheat planting patterns.A,wheat grown under the raised bed planting (RBP) pattern during the rejuvenation period.B,wheat grown under the flat planting (FP) pattern during the rejuvenation period.C,wheat grown under the RBP pattern during the grain-filling period.D,wheat grown under the FP pattern during the grain-filling period.

The number of grains per spike of wheat is the final outcome of a series of physiological processes,such as floret differentiation,development,degeneration,seed setting and grain filling (Serragoet al.2008).The number of grains per spike is determined by the number of spikelets per spike,the number of florets per spikelet,the seed-setting rate of the florets and the occurrence of a sufficient supply of photosynthetic sources at the grainfilling stage (Arduiniet al.2016).Increasing the number of grains per spike has become an important goal of highyield wheat cultivation and breeding (Foulkeset al.2011;Reynoldset al.2012).Previous studies have found that the formation of grains per spike is mainly regulated by climatic conditions (Gonzálezet al.2003;Wanget al.2010),the nutrient supply (Demotes-Mainard 2004;Sinclair and Jamieson 2006) and many other factors.Among these factors,waterlogging significantly affects grain establishment and formation (Cotrozziet al.2021;Becheranet al.2022).

More assimilates are produced for floret survival and grain development by improving plant growing conditions and the nutrient supply and by achieving prolonged photosynthetic function,which maximizes the crop potential productivity (Demotes-Mainard 2004;Wanget al.2010;Derkxet al.2012).Therefore,we hypothesized that the RBP pattern can increase the number of grains per spike of wheat,which is related to spike differentiation and leaf photosynthetic sources.To test this hypothesis,this study evaluated the dynamic features of RBP regarding the spike differentiation and leaf photosynthetic source characteristics of wheat grown after rice in order to better understand the physiological basis for the improved formation of grains per spike under the RBP pattern.The results of this study could provide a new theoretical basis and technical support for achieving high-yield cultivation in wheat after rice production systems.

2.Materials and methods

2.1.Site description

Field experiments were conducted during the 2018/2019 and 2019/2020 growing seasons at the Lujiang Experimental Station (31°46′N,117°26′E) of the Anhui Academy of Agricultural Sciences,Lujiang County,Anhui Province,China.A rice–wheat rotation system was adopted at the experimental site.The annual mean precipitation and temperature were 1 277 mm and 16.2°C,respectively.The groundwater depth reached approximately 0.8 m.The soil was a typical clay loam soil with a soil bulk density of 1.38 g cm–3in the 0–20 cm soil layer.The soil before the start of the experiment in 2018 exhibited an organic matter content of 17.9 g kg–1,total nitrogen (N) content of 1.69 g kg–1,Olsen phosphorus (P)content of 57.4 mg kg–1,and exchangeable potassium (K)content of 70.6 mg kg–1in the 0–20 cm soil layer.

2.2.Experimental design and management

The experiment involved a completely randomized block design with three replications.Two planting patterns were followed,namely,the FP and RBP patterns (Fig.1).Beds under the RBP pattern were 0.25 m in height and 1.80 m in width,with a furrow gap of 0.25 m,and wheat was sown in nine rows across the beds at a 0.20-m spacing (Duet al.2021).All the beds were prepared with an all-inone tractor,which could realize rice stubble incorporation,tilling,bed formation,fertilization,and sowing within a single operation (Duet al.2021).The beds were generally destroyed after the wheat harvest prior to puddling for the purpose of rice planting.Regarding the FP pattern,the stubble of the formerly planted rice was first crushed,followed by separate tilling,fertilization,and sowing operations.Under the FP pattern,wheat was sown in rows at a 0.20-m spacing.The plots were 8.2 m×25.0 m in size,and each plot comprised four beds under the RBP pattern.The Yangfumai 6 wheat cultivar was selected in this study.Sowing was mechanically performed on November 1 and 9 during the 2018/2019 and 2019/2020 growing seasons,respectively,at a seeding density of 22.5 g m–2.The head ditch was manually dredged after sowing to reduce water damage and improve drainage.

Recommended doses of 240 kg N ha–1(urea),90 kg P ha–1(diammonium phosphate) and 120 kg K ha–1(potassium sulfate) were consistently applied to each plot.As a basal fertilizer,the required P and K and 60%of the required N were applied during land preparation.The remainder of the required N fertilizer was employed as topdressing at jointing.Agronomic practices were executed following local cultural practices in all experimental fields and were maintained consistently across all of the plots.All plants were harvested at maturity on May 26,2019 and May 28,2020.

2.3.Meteorological information

During the wheat-growing season,the daily meteorological data,including the daily temperature,rainfall amount,and relative humidity,were obtained from a weather station located 700 m from the experimental fields (Fig.2).

Fig.2 Daily maximum (Tmax) and minimum (Tmin) temperatures,relative humidity and rainfall during the 2018/2019 (A) and 2019/2020(B) wheat-growing seasons.

2.4.Sampling and measurements

Soil NO3–-N contentMeasurements of the soil NO3–-N content were obtained with a discontinuous-type flow analyzer (Bran+Luebbe TRAACS Model 2000 Analyzer,Ltd.,Hamburg,Germany) at a depth from 0–20 cm twice per month from wheat overwintering until harvest.

Spike differentiation characteristicsMeasurements were performed at 3-day intervals from the panicle differentiation and grain-filling stages to the full kernel stage,and 10 plants were randomly selected from each plot to observe the panicle differentiation process.The numbers of spikelets,florets,fertile flowers and young embryos were recorded (Fischer and Stockman 1986).

Plant growth and physiological measurementsFifty successive wheat stems randomly selected from each plot were assessed and averaged to determine the various developmental stages (Zadokset al.1974).

During the wheat growth period,wheat plants within an area of 0.6 m2(three rows with a length of 1 m) in each plot were randomly sampled at 15-day intervals and separated into leaves,stems,ears,and grains.The green leaf areas of the samples were measured with a leaf area meter (Li-3000C,Li-COR Inc.,Lincoln,NE,USA),and the leaf area index (LAI) was calculated as the sum of the green leaf area per unit area of planted land(Awal and Ikeda 2003).All samples were deactivated at 105°C for 30 min to achieve enzyme sterilization and then dried at 80°C to a constant weight in a forced-air oven.

Wheat flag leaves were randomly collected from each plot starting at full expansion (10 days before anthesis)and anthesis,and then continuously at 7-day intervals until physiological maturity was reached.Half of these leaves were used to determine the chlorophyll content according to the method of Wellburn and Lichtenthaler(1984),and the other half was reserved to measure the flag leaf green area with a leaf area meter.The green and yellow parts of the collected flag leaf samples were manually separated based on visual observation.A subsample of 20 flag leaves was obtained from each plot and milled through a 0.5-mm sieve to determine the N contentviathe Kjeldahl method (Ogg 1960).

The photosynthetic rate (Pn) of the flag leaves of wheat was determined with an Li-6400 portable photosynthesis system (LI-COR Inc.,Lincoln,NE,USA).Five wheat flag leaves were selected in each plot for measuring the photosynthetic rate starting at full expansion and anthesis,and then continuously at 7-day intervals until physiological maturity was attained.All measurements were performed from 9:00 to 11:00 a.m.under a constant CO2concentration of 400 μmol mol–1and a photosynthetic photon flux density (PPFD) of 1 200 μmol m–2s–1.

Antioxidant enzyme activityWheat leaf senescence is related to oxidative stress and antioxidant defense mechanisms during reproductive development (Srivalli and Khanna-Chopra 2001;Cotrozziet al.2021).The same flag leaves employed for the photosynthetic rate measurement were sampled to evaluate the malondialdehyde (MDA)content and the activities of superoxide dismutase (SOD),catalase (CAT),and peroxidase (POD).The activities of POD and CAT were measuredviathe ultraviolet absorption method,and the SOD activity was quantifiedviathe nitrogen blue tetrazolium photochemical reduction method (Li 2000).The MDA content was determinedviathe thiobarbituric acid method (Li 2000).These measurements were repeated three times,and the average values were calculated.

2.5.Yield measurement

At maturity,the wheat grain yield,number of grains per spike,spike density and 1 000-grain weight were measured by harvesting an area of 2.05 m2(one bed and one furrow with a length of 1 m) in each plot.

2.6.Calculations and data analysis

Wheat canopy photosynthetic characteristics,i.e.,the decay rate of the leaf area (DLA) and leaf area duration(LAD),were calculated using eqs.(1) and (2): (Evans 1972;Hunt 1978):

where LAI and t are the total leaf area and measurement time,respectively.

Microsoft Excel 2010 and Origin 2022 were used for data processing and graph construction,respectively.All data are presented as the average values of three replications.SPSS 22.0 statistical software was employed for statistical data analysis.

3.Results

3.1.Growing conditions

Excessive rainfall occurred during the wheat-growing seasons in both years,and the air relative humidity was high (Fig.2).The total rainfall amounts reached 518.1 and 545.4 mm during the 2018/2019 and 2019/2020 wheatgrowing seasons,respectively.The total precipitation amounts per month during the 2018/2019 wheat-growing season were 102.6,88.7,34.2,81.2,50.6,82.3 and 78.5 mm in November,December,January,February,March,April and May,respectively.Total monthly precipitation amounts of 60.2,64.7,102.2,44.5,144.6,63.6 and 65.6 mm were recorded during the 2019/2020 wheat-growing season in November,December,January,February,March,April and May,respectively.

3.2.Spike differentiation characteristics and yield

The number of fertile florets per spike increased dramatically,peaked at 18 days after floret initiation and declined rapidly in all planting patterns (Fig.3).The planting pattern significantly affected the floret development and grain set characteristics of the wheat(Table 1).The number of fertile florets per spike,survival rate of florets and floret setting rate were significantly improved under the RBP pattern,but no notable effects were observed on either the maximum number of florets per spike or the fertile floret setting rate.Compared to the FP pattern,the number of fertile florets per spike,survival rate of florets and floret setting rate under the RBP pattern were considerably improved by 4.5 and 3.4 florets/spike,by 6.4 and 5.3 percentage points,and by 4.8 and 4.5 percentage points,respectively,during the 2018/2019 and 2019/2020 wheat-growing seasons,respectively.Finally,the RBP pattern significantly increased the grain numberand yield by 8.6 and 12.7% in the 2018/2019 wheatgrowing season,and by 7.6 and 10.4% in the 2019/2020 wheat-growing season,respectively,compared to the FP pattern.

Table 1 Floret development,set grains and yield composition of wheat as affected by the different planting patterns during the 2018/2019 and 2019/2020 wheat-growing seasons

Fig.3 Number of fertile florets per spike under the different planting patterns during the 2018/2019 and 2019/2020 wheat-growing seasons.FP,flat planting;RBP,raised bed planting.The error bars indicate standard errors (n=3).＊ indicates significance at the 0.05 level.

3.3.Wheat flag leaf photosynthetic characteristics and nitrogen content

Chlorophyll contentThe chlorophyll content in the flag leaves first increased and then decreased,reaching the highest value at anthesis under all planting patterns(Fig.4).Under the RBP pattern,the chlorophyll content in the flag leaves at the later growth stages of wheat was considerably higher than under the FP pattern.Implementation of the RBP pattern significantly improved the chlorophyll contents at–10,0,7,14,21 and 28 days after anthesis (DAA) by 13.5,16.1,14.7,19.2,39.7 and 125.9% in the 2018/2019 wheat-growing season,and by 9.2,6.2,9.1,20.9,47.0 and 121.1% in the 2019/2020 wheat-growing season,respectively,over the FP pattern.

Fig.4 Chlorophyll content,green flag leaf area,photosynthetic rate (Pn) and nitrogen content in the flag leaves under the different planting patterns during the 2018/2019 and 2019/2020 wheat-growing seasons.FP,flat planting;RBP,raised bed planting.The error bars indicate standard errors (n=3).＊ and ＊＊ indicate significances at the 0.05 and 0.01 levels,respectively.The values followed by different letters on each date (days after anthesis) indicate significant differences at the 0.05 level.

Green flag leaf areaThe green flag leaf area of wheat plants decreased from heading to maturity under both planting patterns (Fig.4).No significant differences in the green flag leaf areas were found between the RBP and FP patterns up to 7 DAA.From 14 to 28 DAA,a sharp decline in the green flag leaf area was found under the FP pattern,and notable differences were observed between the two planting patterns.Implementation of the RBP pattern significantly improved the green flag leaf areas at 14,21 and 28 DAA by 13.0,28.7 and 170.2% in the 2018/2019 wheatgrowing season,and by 11.4,37.1 and 384.8% in the 2019/2020 wheat-growing season,respectively,when compared to the FP pattern.

Photosynthetic rate (Pn)Pngradually decreased with the progression of flag leaf senescence (Fig.4).Significant differences inPnwere observed from 7 to 28 DAA between the RBP and FP patterns.Under the RBP pattern,thePnvalues at 7,14,21 and 28 DAA were significantly higher than under the FP pattern,by 10.4,28.3,57.4 and 264.1% in the 2018/2019 wheatgrowing season,and by 6.1,33.3,38.7 and 336.4% in the 2019/2020 wheat-growing season,respectively.

Leaf nitrogen contentThe flag leaf N content in the sampled wheat plants decreased from heading to maturity under all planting patterns and all years (Fig.4).Under the RBP pattern,a significantly higher flag leaf N content(P<0.05) was achieved during both growing seasons.The leaf N content levels at–10,0,7,14,21 and 28 DAA in the RBP pattern were significantly raised by 4.2,6.0,5.4,10.8,24.9 and 37.6% in the 2018/2019 wheat-growing season over the FP pattern,and in the 2019/2020 wheatgrowing season,the values were increased by 2.5,2.7,8.5,15.8,21.2 and 31.5%,respectively.

Oxidative stress and antioxidant enzymesWith the advancement of the growth process,the SOD activity of the collected wheat flag leaves gradually decreased with the progression of leaf senescence under all planting patterns.The CAT and POD activities of the flag leaves first increased and then declined under all planting patterns (Fig.5).Under the RBP pattern,the activities of enzymes SOD,CAT and POD in the flag leaves increased significantly throughout the entire growth period relative to the FP pattern.Averaged across each growing season,the SOD,CAT and POD activities under the RBP pattern showed 17.2,14.4 and 19.7% higher levels than those under the FP pattern in the 2018/2019 wheat-growing season,while they were 15.4,21.4 and 22.1% higher in the 2019/2020 wheat-growing season,respectively.

Fig.5 Superoxide dismutase (SOD),catalase (CAT),peroxidase (POD) and malondialdehyde (MDA) contents in the wheat flag leaves under the different planting patterns during the 2018/2019 and 2019/2020 wheat-growing seasons.FP,flat planting;RBP,raised bed planting.The error bars indicate standard errors (n=3).＊ and ＊＊ indicate significances at the 0.05 and 0.01 levels,respectively.

The MDA content in the sampled leaves followed the same pattern under both planting patterns,and the implementation of the RBP pattern resulted in a reduced MDA content during both seasons.Compared to the FP pattern,the average MDA contents in the sampled flag leaves under the RBP pattern were reduced by 27.3 and 22.8% in the 2018/2019 and 2019/2020 wheat-growing seasons,respectively (Fig.5).

3.4.Wheat canopy photosynthetic characteristics and biomass

The planting pattern significantly affected the DLA and LAD values (Table 2).Compared to the FP pattern,application of the RBP pattern resulted in a significant increase in LAD (by 15.9% in 2018/2019 and 12.3% in 2019/2020) and reduction in DLA (by 7.3% in 2018/2019and 12.0% in 2019/2020).Changes in the wheat canopy photosynthetic characteristics can significantly influence the biomass production of wheat.Implementation of the RBP pattern distinctly increased biomass production.The average biomass after anthesis and the total biomass under RBP conditions were 17.3 and 11.7% higher than those under the FP conditions in 2018/2019 wheatgrowing season,while in the 2019/2020 wheat-growing season,they were higher by 16.1 and 10.7%.

Table 2 Decay rate of the leaf area (DLA),leaf area duration (LAD),biomass accumulation and wheat yield as affected by the different planting patterns during the 2018/2019 and 2019/2020 wheat-growing seasons

3.5.Temporal dynamics of the soil NO3-N content

NO3–-N is the main form of N that is absorbed and used by crops,and owing to its high mobility,leaching represents the main pathway of N loss.The effect of the planting pattern on the soil NO3–-N content in the 0–20 cm layer is shown in Fig.6.The soil NO3–-N content in the 0–20 cm soil layer varied temporally and was significantly affected by the planting pattern.Compared to the FP pattern,the NO3–-N contents under the RBP pattern across both seasons were significantly improved by 10.2% in the 2018/2019 wheat-growing season and 14.9% in the 2019/2020 wheat-growing season.

Fig.6 NO3–-N content in the 0–20 cm soil layer under the different planting patterns during the 2018/2019 and 2019/2020 wheatgrowing seasons.The error bars indicate standard errors (n=3).＊ indicates significance at the 0.05 level

3.6.Correlation coefficients between the grain number per spike,spike differentiation characteristics,and leaf photosynthetic and physiological indicators

The wheat grain number correlation analysis revealed significant (P<0.01) positive correlations with the number of fertile florets per spike,survival rate of florets,floret setting rate,leaf N content,chlorophyll content,Pn,green flag leaf area,LAD,biomass after anthesis,total biomass,and the SOD,CAT and POD activities (Fig.7),which indicates that the number of grains per spike is largely determined by the number of fertile flowers and photosynthetic capacity.There were also significant positive correlations between the chlorophyll content,Pn,green flag leaf area and LAD with the leaf N content,the SOD,CAT,and POD activities and the soil NO3–-N content.

Fig.7 Correlation coefficient matrix between the grain number per spike,the maximum number of florets per spike,number of fertile florets per spike,survival rate of florets,floret setting rate,fertile floret setting rate,leaf N content,chlorophyll content,photosynthetic rate (Pn),green flag leaf area,decay rate of the leaf area (DLA),leaf area duration (LAD),biomass after anthesis,total biomass,malondialdehyde (MDA),superoxide dismutase (SOD),peroxidase (POD),catalase (CAT) and soil NO3–-N content under the two planting patterns.

4.Discussion

Throughout the entire wheat-growing season in the Yangtze River region,the optimum rainfall for high-yield productivity ranged from 245.5–439.5 mm,with 299.0 mm producing the highest yield (Duet al.2019).As observed in this study,the total rainfall during the two wheat-growing seasons was higher than 500 mm,which is known to cause obvious waterlogging during the growth period (Duet al.2021).However,the RBP pattern exhibited positive effects on soil water drainage and wheat yield,with a recorded relative yield increase ranging from 10.4–12.7%(Table 1).The increase in wheat yield could be explained by the 7.6–8.6% increase in the number of grains per spike.This finding is consistent with the results of our previous research (Duet al.2021).

4.1.Raised bed planting pattern improved the spike differentiation process

The effect of the RBP pattern on the above mentioned increase in grain number per spike can be explained by the following mechanism.First,the RBP pattern ensured a sufficient N supply during floret development,which promoted balanced floret development (Demotes-Mainardet al.2004;Ferranteet al.2010),thereby reducing floret degradation and improving the floret setting rate.The number of grains per spike is largely determined by the number of surviving florets,total floret differentiation and degree of degeneration (Serragoet al.2008;Gonzálezet al.2011;Zhanget al.2021).The increase in the number of grains per spike under the RBP pattern was mainly determined by the number of fertile florets at the flowering stage (Table 1).Plants primarily acquire N from the soil,and waterlogging due to excessive rainfall damages the wheat roots and affects nutrient absorption and utilization,which results in insufficient N nutrition (Weiet al.2016).Our previous study revealed that the RBP pattern promoted wheat root growth by improving water drainage and decreasing the bed soil water content,particularly during the rainy season (Duet al.2021),which relieved the waterlogging stress and prevented soil N leaching or loss with water migration(Liuet al.2014;Weiet al.2017).This mechanism was confirmed in this study by the consistently higher NO3–-N content in the 0–20 cm soil layer during both wheatgrowing seasons under the RBP pattern (Fig.6).These results suggest that the application of the RBP pattern increases the soil N supply and the N absorption capacity of roots,which facilitates plant N absorption.Therefore,under the RBP pattern,wheat attained a higher leaf N content (Fig.4),which provided a sufficient N nutrient supply for spike differentiation.Under the FP pattern,the insufficient N nutrient supply seriously affected the quality of floret development,resulting in a large amount of floret degeneration.

4.2.Raised bed planting pattern increased the leaf photosynthetic source capacity

Another important reason for the grains per spike increase is that the RBP pattern improved the photosynthetic source capacity of wheat,which provided a suitable material basis for the development and formation of florets,which in turn facilitated the development of young wheat ears (Yamoriet al.2010;Wuet al.2018;Zhuet al.2019;Cotrozziet al.2021).This ensured that young panicles developed and could compete for more limited resources in the growth process,thus reducing the floret mortality (Bancal 2008).Wheat yields largely depend on the amount of assimilates produced during the critical period (Dinget al.2015;Smithet al.2018;Liet al.2021).Delayed leaf senescence with a longer duration of green leaves may improve the photosynthetic capacity and facilitate the production of more assimilates for grain filling to maximize the yield (Bogardet al.2011;Wuet al.2018;Chibaneet al.2021).In this study,no significant difference was observed in the spike density between the two planting patterns,but the RBP pattern produced stronger photosynthetic sources and a higher assimilate production ability during the postanthesis period,thereby supplying more photosynthetic assimilates to the panicle.However,the grain weight did not significantly increase (Table 1).Therefore,the additional enhanced photosynthetic assimilates were converted into grains,and the number of grains per spike increased.In contrast,the restricted grain production of wheat under the FP pattern was likely attributable to the diminished supply of assimilated carbohydrates induced by early leaf senescence (Arakiet al.2012;Razaet al.2019).This hypothesis was verified in previous studies(Christopheret al.2008;Gajuet al.2011),which reported that the wheat yield was more likely sourcelimited under waterlogged conditions.

4.3.Raised bed planting pattern delayed the leaf senescence

In this study,we found that under RBP conditions,wheat leaf senescence occurred at a low rate,and the plants exhibited lower rates of decline in the chlorophyll content and net photosynthetic rate than those under FP conditions (Fig.4).The effect of the RBP pattern on lateseason leaf senescence delay can be explained by the following mechanism.First,the delay in leaf senescence was related to the high N content.Many studies have demonstrated that N deficiency is one probable cause of early leaf senescence (Agüeraet al.2010;Herzoget al.2016;Kitonyoet al.2018).In the present study,a higher leaf N content in wheat was found under RBP conditions than under FP conditions (Fig.4).Another explanation for the delay in leaf senescence is the improvement in antioxidant metabolism.Wheat leaf senescence is the physiological result of disordered processes of active oxygen metabolism (Gill and Tuteja 2010).SOD,POD and CAT are key enzymes in the plant protection enzyme system,which protects cells from injury,and the MDA content reflects the level of lipid peroxidation in cell membranes (Gill and Tuteja 2010;Milleret al.2010).Under the FP pattern,the flag leaves of the sampled plants exhibited much lower antioxidant enzyme activities (SOD,POD and CAT) but higher MDA contents in the present study,which might disrupt active oxygen metabolism and thus accelerate leaf senescence (Cotrozziet al.2021).Conversely,an improvement in antioxidation metabolism was observed under the RBP pattern (Fig.5),which alleviated the damage caused by reactive oxygen species (ROS) or other peroxide free radicals to the cell membrane system and delayed the decline in leaf physiological functions (Ahmadet al.2020;Wanget al.2021).

Overall,the results of this study confirmed our hypothesis that an improved spike differentiation process and enhanced leaf photosynthetic capacity for grain filling are the main reasons for the observed increase in the grain number per spike of wheat under the RBP pattern.These results indicate that the RBP pattern shows promise for alleviating waterlogging and improving wheat production after rice cultivation under waterlogged conditions.In addition,the grain number per spike is known to be associated with hormones and carbohydrate metabolites,so this relationship should be further investigated in future research.

5.Conclusion

The results of the present study confirmed our hypothesis that the RBP pattern could improve the spike differentiation process and leaf photosynthetic sources,which significantly increased the grain number per spike.An improved soil–plant N supply provided sufficient N nutrients for spike differentiation,and an enhanced photosynthetic capacity produced more assimilates for grain filling.The improvement in leaf sources was related to a high N content and enhanced antioxidant metabolism.Therefore,the RBP pattern may be a suitable alternative for alleviating waterlogging and improving the productivity of wheat planted after rice under waterlogged conditions.

Acknowledgements

This work was funded by the National Key Research and Development Program of China (2017YFD0301306 and 2018YFD0300906).

Declaration of competing interest

The authors declare that they have no conflict of interest.