Muhammad Ishfaq ,Nadeem akbarshakeel ahmed aNjuMMuhammad aNwar-ul-haq

1 Department of Agronomy,University of Agriculture,Faisalabad 38040,Pakistan

2 Department of Plant Science,University of California,California 95616-8780,USA

3 Institute of Soil and Environmental Sciences,University of Agriculture,Faisalabad 38040,Pakistan

abstract Sustainability of traditionally cultivated rice in the rice-wheat cropping zone (RWCZ) of Pakistan is dwindling due to the high cost of production,declining water resources and escalating labour availability.Thus,farmers and researchers are compelled to find promising alternatives to traditional transplanted rice (TPR).A field study was conducted in Punjab,Pakistan,in 2017 and 2018 to explore the trade-offs between water saving and paddy yield,water productivity and economics of two aromatic rice varieties under dry direct seeded rice (DDSR) and TPR.The experiment was comprised of three irrigation regimes on the basis of soil moisture tension (SMT) viz.,continuous flooded (＞-10 kPa SMT),alternate wetting and drying (AWD) (-20 kPa SMT) and aerobic rice (-40 kPa SMT),maintained under TPR and DDSR systems.Two aromatic rice verities:Basmati-515 and Chenab Basmati-2016 were used during both years of study.In both years,DDSR produced higher yields (13-18%)and reduced the total water inputs (8-12%) in comparison to TPR.In comparison to traditional continuous flooded (CF),AWD under DDSR reduced total water input by 27-29% and improved the leaf area index (LAI),tillering,yield (7-9%),and water productivity (44-50%).The performance of AWD with regard to water savings and increased productivity was much higher in DDSR system as compared to AWD in TPR system.Cultivation of DDSR with aerobic irrigation improved water savings (49-55%) and water productivity (22-30%) at the expense of paddy yield reduction (36-39%) and spikelet sterility.With regard to variety,the highest paddy yield (6.6 and 6.7 t ha-1) was recorded in DDSR using Chenab Basmati-2016 under AWD irrigation threshold that attributed to high tiller density and LAI.The economic analysis showed DDSR as more beneficial rice establishment method than TPR with a high benefit-cost ratio (BCR) when the crop was irrigated with AWD irrigation threshold.Our results highlighted that with the use of short duration varieties,DDSR cultivation in conjunction with AWD irrigation can be more beneficial for higher productivity and crop yield.

keywords:rice establishment,aerobic,alternate wetting and drying,water input,tillering,yield

1.Introduction

Rice,is a staple crop cultivated on more than 167 million hectares annually worldwide (FAO 2018).It feeds more than half of the world’s population and provides 20% of daily calories (Carrijoet al.2017),13% per capita protein and 19% per capita energy requirements globally (GRSP 2013).Per capita,rice consumption is growing throughout the world (Ricepedia 2009).Asia contributes 87% of global rice cultivation area and consumes 90% of total rice production (FAO 2018).However,rice production also utilizes a large proportion of irrigated freshwater resources(Bouman and Tuong 2001).This is particularly the case in the rice-wheat cropping zone (RWCZ) of Pakistan,where the productivity of wheat and rice is critical.In the RWCZ,the most common production system is traditional transplanted rice (TPR) practiced in puddled fields which requires an enormous amount of irrigation (Rao and Nagamani 2007).Traditional rice cultivation in the RWCZ is dwindling as the water table declines and water supplies for rice production which become increasingly limited (Akramet al.2013) due to uneven distribution in precipitation (Aminet al.2016) and increasing temperature (Aminet al.2018) under changing climate (Aminet al.2017).Moreover,creating inundation condition for TPR for puddling the soil requires higher water and is a tedious practice.Traditional rice cultivation also requires labour for nursery raising,uprooting,transporting,and transplanting which is becoming unavailable during the peak season (Aslamet al.2008).In addition,unreliable and expensive electricity supply is also affecting the vitality of traditional TPR.Thus,water,energy and labour scarcity are the major limiting factors compelling growers to shift towards water saving dry direct seeded rice (DDSR).

DDSR involves growing rice in dry and un-puddled fields(Farooqet al.2011;Ishfaqet al.2018).Moreover,DDSR provides many benefits over TPR such as:efficient water use,conduciveness to mechanization (Farooqet al.2011)and higher economic returns (Bhushanet al.2007).It may also maintain higher yield and quality (Ishfaqet al.2018),improve stand establishment (Ishfaqet al.2018) and ensure earlier maturity (Farooqet al.2011).Additionally,it facilitates the timely sowing of the succeeding wheat crop and avoids edaphic conflict,thus improving the rice-wheat cropping system sustainability (Anjumet al.2019).Research studies in Pakistan and India revealed that direct seeded rice (DSR) requires less water input than conventional TPR and increased water productivity while achieving comparable yield to flooded TPR (Jehangiret al.2005;Liuet al.2015).Aliet al.(2014) elaborated that yield potential of DSR is higher as compared to TPR when managed properly.Contrary to these findings,Qureshiet al.(2004)reported a yield penalty in DSR as compared to TPR.This yield penalty might be due to water stress because rice production declines when the soil dries below saturation(Sudhir-Yadavet al.2011).

Different irrigation management approaches tested to reduce water consumption are alternate wetting and drying(AWD) (Carrijoet al.2017),aerobic rice (AR) (Maheswariet al.2007),and saturated soil culture (Tabbalet al.2002).Additionally,system of ground cover rice production (Qinet al.2006),combining shallow water depth with wetting and drying (Mao 2001),transplanting in non-puddled soil(Maliket al.2011),and intermittent dry spells (Fenget al.2007) are adopted to maintain or increase water productivity.Traditional TPR with AWD is another option to mitigate the water scarcity.In AWD,the field is re-irrigated when the water level in water pipes/water tubes reaches 15-20 cm below the soil surface (Boumanet al.2007).Globally,AWD is gaining importance in rice production because it increases water use efficiency (WUE) by reducing 23-33% water consumption (Carrijoet al.2017).In the case of yields,AWD has been shown to decrease (Jabranet al.2015),maintain (Lianget al.2016) or increase the yields (Nortonet al.2017) as compared to CF depending upon the severity and duration of drying period.Similarly,AR is another less water and labour demanding technique in comparison to continuous flooded transplanted rice (CF-TPR) (Jabranet al.2015).AR cultivation reduced the total water input by 35-40% and increased the yield in comparison to CF-TPR(Bhushanet al.2007).Role of AWD and AR in DDSR in comparison to AWD in TPR is rarely explored in RWCZ of Pakistan.

Basmati cultivars are a type ofindicarice with extra grain length and cooking quality (Bashiret al.2007).Excellent performance of Basmati cultivars under TPR culture has been reported in several studies where the crop never undergoes water stress.However,in water-saving techniques such as AR and AWD in comparison to CF under TPR as well as under DDSR production system can pose soil moisture stress which retards the growth and performance of rice.On the other hand,some studies indicated that consumptive water use is similar when either cultivated under puddled TPR or water-saving systems (Jabranet al.2017).If we consider that rice cultivation under water-saving techniques does not impose serious water stress,then the growth and yield of low-land rice varieties should perform similarly in either water-saving DDSR or CF-TPR.In this scenario,the selection of a promising cultivar is one of the key factors behind the success of rice cultivation under a resource-efficient production system (DDSR) and water saving techniques (AWD or AR).

The large-scale adoption of DDSR and water-saving practices like AWD or AR is progressing in many countries.However,how the crop performs in terms of yield,yield components and water productivity as influenced by DDSR and water-saving techniques has not been explored in the RWCZ of Pakistan.Therefore,this study tested the hypothesis that the adoption of DDSR and water-saving techniques in rice can increase the productivity and profitability of rice systems by increasing yields and cutting the cost of production by reducing water use.Particularly,this study compares yield,water productivity and economic returns of two aromatic rice varieties under DDSR and TPR production systems in combination with three types of irrigation management:CF,AWD and AR.

2.Materials and methods

2.1.Experimental site

A two-year field experiment was conducted on sandy clay loam soil (Lyallpur series) at the Agronomic Research Area,University of Agriculture,Faisalabad (31°N,73°E,184.4 m a.s.l.) during summer season of 2017 and 2018.Physicochemical analysis of experimental soil is represented in Table 1.

The climate of the experimental site is semiarid to subtropical having mean minimum temperatures (6 and 17°C) to mean maximum (47.5 and 38.8°C) in 2017 and 2018,respectively.Total water input (rainfall) during the study period at Faisalabad in both years is given in Fig.1.

2.2.Experimental design and treatments

The experiment consisted of three factorsviz.,production system (PS),irrigation management (IM) and varieties(V).Three-time replicated sets of treatments were laid out in randomized complete block design (RCBD) with split-split-plot arrangement for both years.Production systems (TPR,DDSR) were randomised in main plots,irrigation managements (AR,AWD and CF) were kept and randomized in sub-plots while varieties were randomized in sub-sub plots.Two varieties:Basmati-515 and Chenab Basmati-2016 used in this experiment are fine-grain,medium duration (135-140 d) and commonly recommended for RWCZ of Punjab,Pakistan.

2.3.Experimental materials

Rice seed was acquired from the Rice Research Institute,Kala Shah Kaku,Lahore.Fungicide Topsin M?70% wettable powder was obtained from a registered dealer of the local market.Weedicides:Oxadiargyl (pre-emergence) and Ethoxysulfuron (post-emergence herbicide) were provided by Bayer Crop Science?(Pakistan).Fertilizers (Urea,Diammonium Phosphate,Murate of Potash,and Zinc Sulphate) were provided by the Agronomy Farm,University of Agriculture,Faisalabad.

2.4.Crop husbandry

All details pertaining to crop husbandry operations during each season is presented in Table 2.

field preparationIrrigation at a 10-cm depth was applied after rotavating the stubbles of berseem and wheat in 2017 and 2018,respectively.Following irrigation,sesbania(Sesbania rostrataL.) seed was broadcasted in muddy conditions to promote weed germination.After 25 d,weeds and sesbania seedlings were incorporated by using rotavator.Afterwards,the field was divided into two halves:one for DDSR and the other for conventional TPR.The two sections were separated by making a water channel in the centre.Between sub-sub plots (having net plot size 8 m×3 m),1.5 m wide and 30 cm high earthen double bunds were made and compacted to serve as a buffer and control the lateral movement of water.

Crop establishment/PlantingDirect seeded rice wassown in the well pulverized dry field using a hand drill with 22.5 cm spaced rows.Seed rate for DDSR was 30 kg ha-1in both years while for transplanted rice 10 kg ha-1.The nursery was sown on the same day in well-prepared and levelled small plots of 5 m×5 m dimensions.Part of the field was allocated for transplanting culture and was soaked 1 d before transplanting,after making the layout as for direct seeded rice to puddle the soil and to facilitate the transplanting of 30-d-old seedlings.Two seedlings per hill were transplanted on 11 July in 2017 and on 27 July in 2018 manually,maintaining plant-to-plant and row-to-row distance of 22.5 cm.

Table 1 Physico-chemical soil properties of the experimental site1)

fertilizers and weeds managementThe entire application of phosphorus+potassium and a 40% dose of nitrogen was broadcasted onto the soil surface and incorporated before sowing at the time of seedbed preparation,and the remaining nitrogen was top-dressed at 30 days after sowing (DAS) in DDSR and 20 days after transplanting (DAT) in TPR and panicle initiation (PI) stage.

Weeds were controlled by stale seedbed technique and pre-and post-emergence herbicides.DDSR and the nursery raising field were irrigated after sowing and in standing water condition weedicides:oxadiargyl?as pre-emergence was applied using a shaker bottle.While post-emergence(ethoxysulfuron?) herbicide was applied at 20 DAS in DSR using flood irrigation to control sedges.Similarly,TPR preemergence was applied in standing water conditions after transplanting the seedlings.While post-emergence was used at 15 DAT.

Irrigation managementBoth canal and tube well water were available to meet the irrigation requirements of the crop.Cut-throat flume was used to measure the total volume of applied water content.While field water tubes(pani pipes/water pipes) were used to inspect field water depth.Irrigation requirement of each treatment was fulfilled after reaching the respective pre decided threshold level.There were three threshold levels of soil moisture tension,viz.,＞-10 kPa (CF),-20 kPa (AWD) and -40 kPa (AR) for both DDSR and TPR.Tensiometers were installed in each plot to measure soil water potential.All three replicates of respective treatments were irrigated when the pre-defined threshold level was attained.In each treatment,initial fully soil saturated condition was maintained to achieve better weed control and stand establishment as according to Tuonget al.(2005).

fig.1 Relative humidity (RH),rainfall,minimum temperature (temp.min.),average temperature (temp.avg.),and the maximum temperature (temp.max.) of experimental site during the course of study in 2017 and 2018.

In the CF treatment of DDSR,muddy conditions (＞-10 kPa soil water tension) were maintained soon after seeding,to 2 wk after sowing.The rest of the crop season flood conditions was maintained,while the CF treatment of TPR flood conditions were maintained from the day of transplanting.In AWD treatments:flood conditions (5 cm water depth) were maintained from 10 to 30 DAS,and from the beginning of panicle initiation (PI) to the flowering stage.For the remaining growth period,fields were re-irrigated(3 cm water depth) when water level in water pipes dropped to 20 cm below the soil surface or when soil water tension reached-20 kPa.Similarly,for AWD in TPR,flooding condition was maintained from day of transplantation to 20 DAT,and from PI to the flowering stage while rest of crop duration field was re-irrigated when soil moisture tension of -20 kPa reached.In AR treatments,plots were re-irrigated when soil moisture tension reached to -40 kPa.However,to establish crop stand,AR plots were kept fully saturated or flooded from day of sowing to 30 DAS (DDSR) and from transplanting to 20 DAT (TPR).Last irrigation was applied to DSR treatments on 29 September in 2017 and 15 October in 2018.While in transplanting treatments,last irrigation was applied on 7 and 20 October in 2017 and 2018,respectively.

Disease and insect managementRice stem borer and other chewing insects were controlled by using Regent?GR at 75 g ha-1.In both years applications took place at 25 and 40 DAT in TPR treatments and 40 and 55 DAS in DSR treatments.Nativo was sprayed in both years at 165 g ha-1after 60 and 75 DAS in DSR and 45 and 60 DAT in TPR using a Knapsack sprayer to safeguard the crop from neck blast,brown leaf spot,leaf blast,and glume discoloration.To safeguard rice crop from bacterial leaf blight,Aliette?was used at 65 and 75 DAS in 2017 and 2018,respectively in DSR while 55 and 65 DAT of TPR in 2017 and 2018,respectively.

harvesting and threshingRice plants were harvested from 7.5 m2manually and left for 4-5 days to dry.At 14% moisture content,grains were threshed.A portable microprocessor grain moisture meter (MC-7825G with a range of 5-31%,Swastik Scientific Co.) was used to measure paddy moisture and final paddy yield was presented in t ha-1.

2.5.Data recording procedure

Total applied water and water productivityTotal applied water was measured using a cut-throat flume.It was installed in the main water channel that irrigating the rice field.In both years of study,total depth of water to each treatment was calculated by using the following formula:

where D is irrigation water depth in mm,Q represents the amount of discharged water in ft (cubic feet)3s-1,A indicates the area to be irrigated in m2,and t denotes the time of water application.Rainfall data for both rice-growing seasons were obtained from observatory equipped with rain gauge and located at the University of Agriculture,Faisalabad.Water productivity (WP) in kg m-3was calculated as grain yield (kg ha-1) produced per unit water (m3ha-1) used (Ali and Talukder 2008).Total water used included both rainfall and applied irrigation:

Crop growth and phenologyIn 2017,total numbers of tillers and productive tillers were counted at physiological maturity for both production systems (PS).In 2018,tillering was recorded periodically at effective tillering stage,PI stage,heading stage,and physiological maturity.From each treatment,two parallel rows of 50 cm length of DDSR and nine hills (0.5 m2area) of TPR were tagged to count periodic tillering as described by Karet al.(2018).Periodic tillering for both varieties was taken at 45,65,85,and 110 DAS of DDSR and at 22,42,62,and 80 DAT in TPR treatment.In 2018,leaf area was measured at 65,80,90,110,and 125 DAS of DDSR and at 25,40,55,70,and 85 DAT in TPR treatment as suggested by Yoshida (1981).Leaf area index(LAI) was measured as the ratio of leaf area and land area.Days to heading were considered as the number of days until the appearance of the first panicle.Ten hills of each TPR treatment and 10 mother tillers of DDSR treatment were tagged to record 50% heading as per method of Buenoet al.(2010).Harvest maturity was considered when 95%spikelets of any treatment had changed their colour from green to yellow.

Yield and yield componentsEffective and ineffective tillers were counted and averaged from two locations (1 m2) in the DDSR plots,whereas in TPR system,tillers of 20 adjacent hills (1 m2) of each sub-sub plot were counted.1 000 grains were counted using a Coutador electric seed counter(automatic,0.13-15 mm seed size,Pfeuffer,Germany).Two samples (1 000 grains) of each sub-sub plot were oven dried for 2 d at 60°C and then dry weight was taken for the test weight (1 000-grain weight).

The average length of 20 panicles (primary tillers) at harvest maturity was recorded by using a wooden meter rod.Furthermore,panicles of 10 randomly selected primary tillers were clipped off with a scissor and the number of branches,filled and sterile kernels were counted.Spikelets were pressed between the thumb and forefingers to identify filledvs.sterile spikelets.Spikelet sterility was calculated as:

Sun-dried threshed crop straws from an area of 7.5 m2were tied into bundles,weighed and expressed as t ha-1.Harvest index (HI) was computed as:

Economic analysisTotal expenditure was calculated by adding fixed cost (land rent,transport charges and expenditure on crop protection measures) and variable cost(seed,land preparation,sowing,irrigation,and fertilizers).Economic analysis was performed following the guidelines of CIMMYT (1998).Total income in USD ha-1was calculated on the basis of per unit price of grain and straw in the market.Net returns of each treatment were calculated by subtracting total expenditure from the total income.Dividing total income(USD ha-1) by total expenditure (USD ha-1) of respective treatments gave benefit-to-cost ratio (BCR).

statistical analysesData-collected were analysed using statistics 10,a computer program.Fisher’s analysis of variance (ANOVA) techniques and Tukey’s (HSD) test were applied at aα=0.05 probability level to compare the treatment means (Steelet al.1997).Line graph data were presented using Sigma Plot Software and differences among treatments means were indicated by standard error bars.Higher order interaction (PS×IM×V) and lower order interaction (IM×V) were non-significant so only main factors and lower order interactions (PS×IM and PS×V) are presented in tables.

3.results

3.1.Experimental weather condition

Mean minimum temperature in the summer season of 2017 was lower (16°C) as compared to the summer of 2018(17.8°C),while the mean maximum temperature was higher in 2017 (39.5°C) than that in 2018 (38.8°C).Total water input in the form of rainfall during the crop season in 2017 was 260 mm,whereas in 2018 total rainfall totalled 286 mm.More rainfall was observed in 2018 before the experiment began which delayed sowing,but total number of rainy days during crop growth was more in 2017 (25) than in 2018 (20)(Fig.1).Rainfall pattern was similar for both years till the first week of August,but September 2017 received more rainfall as compared to September 2018.Similarly,October 2018 received more rainfall which coincides the grain filling periods as compared to 2017.

3.2.Total water inputs and water saving

In each PS,before commencement of irrigation application according to different thresholds of SMT,irrigation events were similar for both years.After commencement of irrigation according to defined SMT thresholds,irrigation events varied significantly.Therefore,after initiation of irrigation according to respective SMT thresholds,number of irrigation events varied between PS as well as within the PS.A brief sketch of irrigation application in both systems was described in Fig.2.In 2017,CF,AWD and AR of DDSR received 6,8 and 3 more irrigations than those of TPR,respectively.However,in 2018,the difference of irrigation events between DDSR and TPR was less as compared to 2017.CF,AWD and AR of DDSR received 5,3 and 1 more irrigations than those of TPR,respectively.With respect to variety,in each IM technique of DDSR,Basmati-515 received two more irrigations in 2017 and one more irrigation event in 2018 as compared to that in IM of TPR.It was observed that frequency of irrigation events was higher in each IM technique of DDSR but total water input was significantly higher in TPR due to heavy water inputs at crop establishment stage (data not shown).

fig.2 Water management in dry direct seeded rice (DDSR) and transplanted rice (TPR) during the whole season.A,traditional continuous flooded (CF) under DDSR.B,alternate wetting and drying (AWD) under DDSR.C,aerobic rice (AR) under DDSR.D,CF-TPR.E,AWD-TPR.F,AR-TPR.Grey shaded region indicates flooding.+T,transplanting;+N,nitrogen fertilization;+S,sowing.

In comparison to TPR,average water inputs in DDSR was 12 and 8% lower in 2017 and 2018,respectively.When irrigation was applied according to threshold level of AR-DDSR and AWD-DDSR,water input was reduced by 50-55% and 26-29%,respectively,relative to CF-DDSR.Similarly,AE-TPR and AWD-TPR reduced water input by 49-52% and 25-29% respectively in comparison to CF-TPR(Table 3).A significant interaction of PS×IM techniques and IM techniques×variety on total water input was observed in both production years.Water inputs decreased significantly in AWD and AR in both PS,as SMT was allowed to decrease to -20 and -40 kPa,respectively.In both years,influence of different IM techniques under different PS on total water inputs was observed in the following trend:CF-TPR (2 895-3 087 mm)＞CF-DDSR (2 705.0-2 790.0 mm)＞AWD-TPR(2 105.0-2 297.5 mm)＞AWD-DDSR (1 955.0-2 005.0)＞ARTPR (1 420.0-1 527.5 mm)＞AR-DDSR (1 245.0-1 335.0 mm)(Table 3).Under different PS,total water requirement of both varieties varied significantly.Cultivation of Basmati-515 in DDSR saved 4-6% while Chenab Basmati-2016 saved 4-8% water inputs in comparison to TPR (Table 3).

3.3.Total water productivity

Total water productivity (WPt) was substantially influenced by PS,IM and selection of variety and lower order interaction of PS×variety.Total water productivity varied from 0.7 to 1.6 kg m-3in 2017 and from 0.8-1.7 kg m-3in 2018.Higher WPt(30% and 27%) was observed in DDSR as compared to TPR in 2017 and 2018,respectively.Total water productivity was substantially higher in all AWD treatments as compared to CF and AR treatments,regardless of PS,due to maintenance of high grain yield despite reduced total water inputs.When irrigation was applied according to CF or AR threshold,WPtwas decreased due to either higher water inputs (CF) or reduction in grain yield (AR).In both years,WPtof AWD was higher (44-50%) in comparison to WPtof CF treatment.Similarly,WPtof AWD was 15.3-18%higher than WPtof AR treatment.In case of varieties,water productivity of Chenab Basmati-2016 was 9% higher in 2017 and 8% higher in 2018 than water productivity of Basmati-515.Among interactive effect of PS×variety,the highest water productivity was recorded in Chenab Basmati-2016 under DDSR in 2017.Total water productivity was similar for both varieties under DDSR in 2018 and under TPR in both 2017 and 2018 (Table 4).

3.4.Tillering and leaf area index (laI)

In 2017,the average plant population at 15 DAS of DDSR was 205 plants m-2for Basmati-515 and 160 plants m-2for Chenab Basmati-2016.In 2018,226 plants m-2for Basmati-515 and 186 plants m-2for Chenab Basmati-2016 were recorded.In TPR,plant population was similar (19hills m-2each having two seedlings) for both varieties in both production years.Tillers density in 2018 was significantly higher in DDSR as compared to TPR.In DDSR,28 and 19% higher tiller densities were recorded at PI and heading stages respectively,but there was no significant difference at physiological maturity (Fig.3-A).Across both PS,AWD improved tiller density at PI and heading by 7 to 9%,but there was not a significant difference at effective tillering stage and physiological maturity stage (Fig.3-B).Among varieties,Chenab Basmati-2016 performed better and gave higher tiller density at PI and heading stages (Fig.3-C).The interaction of PS and IM indicated that tiller density started to decline after 70-80 DAS in DDSR but in TPR the decline was a bit late (except in AR treatment).In AR treatment of either DDSR or TPR,tillering density started to decline earlier as compared to PI (Fig.3-D).

Table 3 Influence of traditional continuous flooded (CF) and water-saving irrigation management (AWD (alternate wetting and drying) and AR (aerobic rice)) on total water input and water saving of two aromatic rice varieties under DDSR and TPR systems

Table 4 Influence of traditional continuous flooded (CF) and water-saving irrigation management (AWD and AR) on yield and water productivity of two aromatic rice varieties under DDSR and TPR systems

LAI was significantly influenced by PS and IM but most lower and higher order interactions did not influence the LAI significantly until 95 DAS or 55 DAT.LAI of both varieties increased up to 90-95 DAS or 50-55 DAT and then declined(Fig.4).Additionally,LAI of DDSR (6.5) was significantly higher than the LAI (3.93) attained in TPR (Fig.4-A).In both PS,the highest LAI was attained by AWD treatments followed by CF and AR,respectively (Fig.4-B).In case of varieties,the highest LAI (3.48) until 110 DAS/70 DAT was shown by Chenab Basmati-2016,but then for the remainder of the growing season,Basmati-515 had a higher LAI(Fig.4-C).Lower order interaction of PS and IM significantly influenced LAI at 125 DAS or at 85 DAT.In both production systems,the highest LAI (3.65 and 1.75) was recorded in AWD-DDSR and AED-TPR,respectively,followed by CF and AR treatments (Fig.4-D).

3.5.Phenology and crop duration

fig.3 Influence of traditional continuous flooded (CF) and water-saving irrigation management (alternate wetting and drying(AWD) and aerobic rice (AR)) on periodic tillering of two aromatic rice varieties in 2018 under dry direct seeded rice (DDSR) and transplanted rice (TPR) systems.P.Maturity,physiological maturity.Bars mean SE (n=3).

fig.4 Influence of traditional continuous flooded (CF) and water-saving irrigation management (alternate wetting and drying(AWD) and aerobic rice (AR)) on the leaf area index of two aromatic rice varieties in 2018 under dry direct seeded rice (DDSR) and transplanted rice (TPR) systems.Bars mean SE (n=3).

Rice PS,IM,varieties,and years indicated a substantial influence on the days taken to heading,50% heading and harvest maturity.In 2017,the growth duration of both varieties in all combination of PS and IM was about 5-12 d more as compared to that in 2018.DDSR crop matured a bit earlier and attained harvest maturity 3-10 days earlier in comparison to TPR.Chenab Basmati-2016 attained harvest maturity earlier (9-10 days) than Basmati-515.In IM,crop grown under AR treatment matured at the end as compared to AWD and CF treatments (Table 5).

3.6.spikelet sterility

Although production system did not influence spikelet sterility,comparatively more spikelet sterility was observed in DDSR as compared to TPR.Under both PS,spikelet sterility was higher for AWD and AR relative to CF treatments.There was not a significant difference between AWD and CF,but under AR treatment,spikelet sterility increased significantly in both years and this increase was larger in DDSR as compared to TPR.Substantially,higher sterile spikelets were observed in Basmati-515 in both years as compared to Chenab Basmati-2016 (Fig.5).

3.7.Yield causative attributes and yield

A significant influence on total tillers m-2and productive tillers m-2was observed due to:PS,IM,varieties,the interaction of PS×IM and PS×varieties in both years of study (Table 6).More tillers m-2were observed in DDSR as compared to TPR in both years.In case of IM,in both PS,total tillers m-2as well as productive tillers m-2were the highest when irrigation was applied according to AWD,while AR gave the lowest total tillers m-2and productive tillers m-2under both systems.Interactive effect of PS and variety indicated that the highest total tillers m-2(450.3 and 484.6) and productive tillers m-2(422.6 and 463.2) were recorded for Chenab Basmati-2016 under DDSR in 2017 and 2018,respectively.Similarly,Chenab Basmati-2016 performed better in the TPR system than Basmati-515 as indicated in Table 6.

All main factors significantly influenced the panicle length and branches per panicle in both years.However,the influence of variety on panicle length and branches per panicle in 2018 was non-significant (Table 6).Longer panicle length (29.0 and 27.5 cm) and branches per panicle (12.0 and 12.5) were recorded under TPR as compared to DDSR in 2017 and 2018,respectively.In both PS,AWD led to significant improvement in panicle length and branches per panicle,but this improvement was greater in TPR relative to DDSR.Panicle length and branches per panicle were reduced under AR,but this reduction was more pronounced in DDSR as compared to TPR.In terms of variety,Chenab Basmati-2016 produced longer panicles and more branches per panicle in DDSR as compared to TPR (Table 6).

A significant influence of PS,IM,variety,and interaction of PS and variety was observed on kernels per panicle and 1 000-grain weight in both years,but the effect of variety on 1 000-grain weight was non-significant in 2017.The highest kernels per panicle (125.3 and 126.6) were recorded in Chenab Basmati-2016 grown under DDSR,while the highest grain weight (25.2 and 22.1 g) was obtained by Basmati-515 when it was grown under TPR (2017 and 2018,respectively).In both PS,AWD improved both kernels per panicle and 1 000-grain weight.Alternatively,AR reduced both kernels per panicle and 1 000-grain weight (Table 6).

Comparatively higher straw yield was recorded in DDSR in both years but in 2018 significantly higher (25% more)straw yield was observed.Chenab Basmati-2016 produced more straw in comparison to Basmati-515.Furthermore,in both PS,Chenab Basmati-2016 had higher straw yield under AWD than under CF and AR (Table 4).

Table 5 Influence of traditional continuous flooded (CF) and water-saving irrigation management (AWD and AR) on days taken to heading,50% heading and harvest maturity of two aromatic rice varieties under DDSR and TPR systems

fig.5 Influence of traditional continuous flooded (CF) and water-saving irrigation management (alternate wetting and drying (AWD)and aerobic rice (AR)) on spikelet sterility of two aromatic rice varieties in 2017 (A and B) and 2018 (C and D) under dry direct seeded rice (DDSR) and transplanted rice (TPR) system.Bars mean SE (n=3).Same case lettering for a parameter do not differ significantly (P≤0.05) by the tukey’s HSD (honestly significant difference) test.

In both years,paddy yield and biological yield were significantly higher in DDSR as compared to TPR.In each year,paddy yield and biological yield in CF and AWD were statistically similar,but when SMT decreased to-40 kPa (in AR treatment),a significant yield decline was observed.Additionally,the increase in paddy yield due to AWD was greater in DDSR as compared to TPR.Yield reduction under AR was greater in TPR than that in DDSR.Additionally,performance of Chenab Basmati-2016 in DDSR was significantly better than that in TPR (Table 6).In both years,harvest index (HI) of CF and AWD was similar but was affected by PS.HI decreased under AR in both DDSR and TPR,but this decrease was more pronounced in TPR.Moreover,HI of Basmati-515 was lower (34.6) than the HI of Chenab Basmati-2016 (36.1) (Table 6).

3.8.Economic analysis

Economics analysis indicated that use of AWD in DDSR significantly increased the net field’s benefit (USD),net returns (USD) and benefit cost ratio (BCR) in both years.Highest net returns (USD 1 482.3 in 2017) and BCR (3.4 in 2018) were obtained when Chenab Basmati-2016 was cultivated under DDSR with AWD.Net returns and BCR of DDSR were comparatively higher than those in TPR in both years.Lowest net returns,net benefits and BCR were observed when Chenab Basmati-2016 was cultivated using TPR under AR.Lowest net returns were observed in ARTPR then increased in the following order:AR-DDSR＜CFTPR＜CF-DDSR＜AWD-TPR and the highest net returns were observed in AWD under DDSR system (Table 7).

4.Discussion

Overall,AWD was optimal in terms of water saving,grain yield and WPtof both TPR and DDSR.In terms of WPt,DDSR-AWD outperformed TPR-AWD due to increased yields and simultaneously decreased water inputs.

4.1.Total water input and crop duration

Total water input in the form of rainfall and irrigation(canal or tube well) was higher in 2017 than that in 2018,irrespective of irrigation treatments.This reflects that there was more evaporative demand due to higher temperature and evapotranspiration in 2017.This variation can beattributed as the temperature is increasing due to climate change and anomalies in precipitation and temperature are uncertain and affecting the crop production (Aminet al.2016).Average water input of DDSR was lower than that of TPR during both rice seasons despite having more irrigation events.The reason being that DSR is devoid of large amounts of early season water inputs to puddle the field (Kaur and Singh 2017).Water savings in DDSR as compared to TPR were consistent with previous studies which found 50-60% water savings from DDSR (Bouman and Tuong 2001).In our study,extent of water savings in DDSR was low (only 8-12%) in comparison to TPR,which might be attributed to the sandy loam composition of the soil or the absence of a plough pan or a hard pan (Sudhir-Yadavet al.2011).However,within each system,water input to AWD or AR treatment was significantly lower than that to CF.DDSR with AWD (-20 kPa threshold level) reduced the water input by 26-29%,while maintaining higher yield,as compared to CF-DDSR,and TPR with AWD by 25-29% in comparison with CF-TPR.On the other hand,AR reduced water inputs by 50-55%,but with yield penalties of 32 to 37% (in DDSR) and 38-40% (in TPR).

Table 6 Influence of traditional continuous flooded (CF) and water-saving irrigation management (AWD and AR) on yield causative attributes of two aromatic rice varieties under DDSR and TPR systems

Previous studies have reported 15-40% reduction in water inputs in TPR with AWD irrigation management(Humphreyset al.2010).Carrijoet al.(2017) concluded that rice cultivation under AWD has the potential to reduce water input by 23.4% in comparison to CF.Furthermore,Boumanet al.(2007) indicated that deep drainage water losses could be significantly reduced when TPR was cultivated with AWD irrigation.Similarly,under AR conditions,unsaturated field conditions offer great opportunities to save water (Boumanet al.2002).Conversely,CF is much less efficient at saving water relative to AWD and AR as there are greater water losses from CF system due to increased seepage and percolation,puddling and evaporative losses (Boumanet al.2007).

Additionally,water saving in DDSR can be attributed to early maturity of the crop (Farooqet al.2011).In our study,DDSR treatments also matured earlier as compared to TPR.On average,TPR took 3-9,4-15 and 3-10 days longer to achieve heading,50%heading and harvest maturity,respectively.Chenab Basmati-2016 consumed less water (5-6%) as compared to Basmati-515 because it matured 6 d earlier in 2017 and 12 d earlier in 2018 under DDSR.Under TPR,Chenab Basmati-2016 matured 6 days earlier than Basmati-515 in both 2017 and 2018.Hence,it can be suggested that a significant amount of water input can be reduced by growing early maturing varieties under DDSR with AWD.

4.2.Growth attributes

Irrigation regimes did not influence the initial crop stand because at germination stage of DDSR or transplanting stage of TPR,all treatments were irrigated equally.Varietal differences were observed in terms of tiller density and LAI at different growth stages under varying PS.Chenab Basmati-2016 produced higher tiller density as compared to Basmati-515.DDSR was better in terms of crop stand,tiller density and leaf area development than TPR.The reasons might be attributed to higher seedling vigour in DDSR (Farooqet al.2011) which enable DDSR to attain early canopy cover,more light interception and increased WUE (Anwaret al.2011).Conversely,in TPR root damage due to nursery uprooting and transplanting shock slows down the early growth and vigour.Sudhir-Yadavet al.(2011)also reported that DSR achieved relative higher growth than TPR,as was observed in our study.Tiller density at different growth stages in 2018 was influenced due to water regimes.AWD improved the number of productive tillers,while AR decreased the tillering density and number of productive tillers.Our results are consistent with findings of previous studies which show that AWD has an ability to increase productive tillers and reduces redundant growth by altering leaf angles,improving root health,shoot growth and LAI (Nortonet al.2017;Pascual and Wang 2017).Whereas in AR rice cultivation,reduced water input leads to hampered shoot growth and LAI due to water stress (Nguyenet al.2015).

4.3.Grain yield

More paddy yield was observed in 2018 as compared to 2017.Yields from DDSR were comparatively higher than those from TPR due to more productive tillers per unit area.However,the increase in number of tillers under DDSR resulted in reduced panicle length,branches per panicle,number of kernels per panicle,and 1 000-grain weight relative to TPR.The reduction in yield causative parameters of DDSR might be due to intraspecific competition among higher tiller density (Hasanuzzamanet al.2009).These results are contrary with results represented by Singhet al.(2001) that more grain yield in TPR might be due to better weed management under the TPR system.On the other side,when DSR constraints are properly managed,it gave higher or comparable yield to TPR (Bhushanet al.2007).

In the case of IM,in both years,paddy yield with CF and AWD was statistically similar,however,reduction in paddy yield was observed under AR.These results are consistent with the findings of Sudhir-Yadavet al.(2011) who found that CF yielded higher than severe-AWD (-60 kPa).In comparison to CF,paddy yield of AWD was higher in both years of our study,but this difference was not significant.Our results are in consistent with the findings of Nortonet al.(2017) and Carrijoet al.(2018) who reported that AWD has the potential to increase yield because safe/mild AWD implications maintain soil water potential ≥?20 kPa which ensure the soil moisture for optimal growth.Other studies postulated that more yield under AWD might be attributed to better grain filling and increased root proliferation to uptake more water and nutrient (Ndiiriet al.2012;Pascual and Wang 2017).Moreover,AWD improves effective tiller percentage and translocation of carbohydrates to grain,reduces spikelet sterility,and increases grain weight (Yang and Zhang 2010).Post-anthesis AWD enhances the grain-filling rate due to enhanced activity of enzymes involved in grain filling and ultimately increases the grain yield (Zhanget al.2012).Chenab Basmati-2016 outperformed Basmati-515 in both years,which might be attributed to more LAI,increased effective tillering,panicle length,grains per panicle,and test weight due to better grain filling (Table 6).In contrary to AWD,AR reduced the paddy yield significantly.This can be attributed as the water stressed conditions of AR,which can hamper yields due to stunting,reduced chlorophyll content and increased electrolyte leakage (Jahanet al.2014).AR has been shown to result in many anatomical changes in the rice plant due to water stress,such as hampered cell division and decreased intercellular space which ultimately leads to decreased paddy yield (Maheswariet al.2007).

4.4.Total water productivity

Total water productivity across all treatments ranged from 1.0-1.6 kg m-3in our study.Our results align with the findings of Sudhir-Yadavet al.(2011) who reported a range of 0.26 to 1.58 g kg?1across all treatments in their study.In our study,WPtwas higher under DDSR as compared to TPR.This is contrary to the findings of Bhushanet al.(2007) and Sudhir-Yadavet al.(2011) who reported higher WPtfrom TPR.We are not sure why this difference between studies,but the higher WPtfrom DDSR in our study can be directly attributed the higher grain yield and reduced water input.Among IM,WPtof AWD was the highest due to a reduced number of irrigation events,yet simultaneously high yield.As one might expect,the fewest irrigation events were recorded from AR,but this did not result in high WPtas yields were also low.In CF,WPtwas the lowest because of high water input needed to maintain flooded conditions throughout the growing season (Table 4).Our results agreed with Joshiet al.(2009) who concluded that at the farm level,AWD increased WPtbecause of water input reduction.These results were also in agreement with the findings of Yaoet al.(2012) and Lianget al.(2016) who reported AWD to increase WPtover CF.The high WPtof AWD under DDSR presents an attractive option for farmers in the RWCZ of Indo-Gangetic Plains.DDSR in combination with AWD not only produces high yields while saving water,but also has a reduced demand for labor,cuts the energy costs of pumping groundwater,and improves the overall sustainability of the system by allowing for timely planting of succeeding crop.

4.5.Economic analysis

Commercial viability and reduction in production costs are motivating factors behind the adoption of new and innovative practices by farmers.Given the rise of industrialization,it is becoming increasingly difficult to find labour in a timely manner to transplant a rice nursery.Moreover,dwindling water resources threatens the sustainability of the transplanting culture of rice.This study found that DDSR is an economical option for farmers due to water saving,lower costs of production and comparable or higher yield.By comparison,CF-TPR is more expensive than DDSR due to the costs associated with nursery raising,uprooting,transporting,transplanting,higher water inputs,and puddling.As compared to TPR,economics of DDSR can be further improved by using AWD as it improved the paddy yield and saved a significant amount of water(Table 7).Overall,we found that each water-saving rice PS (DDSR with AWD or AR irrigation,TPR with AWD or AR irrigation) has the potential to reduce production costs relative to CF rice.

4.6.The trade-off between rice establishment methods,water management practices and varieties

In this study,the performance of DDSR was higher in both years as compared to TPR.Within each combination of PS and IM,the performance of Chenab Basmati-2016 was better than that of Basmati-515 in terms of paddy yield,water saving and economics.However,an immaculate trade-off among water inputs,water productivity,yield,and BCR was observed within each establishment method.In AR,water was saved at the expense of yield penalty in TPR as well as in DDSR crop.While in CF,higher yield was achieved at the expense of more water input.

5.Conclusion

Among the various water management practices,AR cultivation managed to increase water savings and water productivity,but at the cost of yield reduction.Specifically,AR cultivation under DDSR enhanced spikelet sterility,an important factor in yield reduction.In comparison to CF and AR,AWD not only increased water savings and yield,but also improved growth,grain-filling rate and water productivity in both TPR and DDSR.In regard to varietal performance,Chenab Basmati-2016 outperformed Basmati-515 in both production systems and among the three irrigation management practices.Chenab Basmati-2016 achieved higher water productivity due to early maturity,increased yield and reduced water input.Furthermore,less spikelet sterility,better tillering capacity and higher LAI were observed in Chenab Basmati-2016.

The supremacy of water saving irrigation techniques was consequently evaluated by an economic analysis.DDSR in combination with a water-saving irrigation management technique (AWD) reduced the cost of production due to less water,energy and labour.These reductions led to higher income due to increased yield.Traditional TPR and CF irrigation techniques not only jeopardised the rice production but also increased the cost of production due to additional water input and labour requirement.

This study shows that DDSR in combination with AWD has the potential to improve yields while reducing water inputs.However,further assessments of these management practices are needed under different environmental conditions (e.g.,climate and soil types) to assess the benefits under varying growing conditions and to determine the feasibility of these practices.

Acknowledgements

We are especially obliged to the Directorate of farms,Agronomic Research Farm and Department of Agronomy,for administrative,manpower and technical assistance during the course of study.