Estimation of crop coefficients of dry-seeded irrigated rice–wheat rotation on raised beds by field water balance method in the Indo-Gangetic plains, India

Page 1

ErI

Ba

b

c

ARAA

KRRCCI

1

saw(2iusewIc

0h

Agricultural Water Management 123 (2013) 20– 31

Contents lists available at SciVerse ScienceDirect

Agricultural Water Management

j ourna l h o me pa ge: www.elsev ier .com/ locate /agwat

stimation of crop coefficients of dry-seeded irrigated rice–wheatotation on raised beds by field water balance method in thendo-Gangetic plains, India

.U. Choudhurya,b,∗, Anil Kumar Singhc, S. Pradhana

Division of Agricultural Physics, Indian Agricultural Research Institute, New Delhi 110 012, IndiaDivision of Soil Science, ICAR (RC) for NEH Region, Umiam, Meghalaya-793 103, IndiaRajmata Vijayaraje Scindia Krishi Vishwa Vidyalaya, Gwalior-474 002, India

a r t i c l e i n f o

rticle history:eceived 18 December 2012ccepted 10 March 2013vailable online 9 April 2013

eywords:aised bedsice–wheatrop coefficientrop evapotranspiration

ndo-Gangetic plain

a b s t r a c t

Lack of information on crop coefficient (kc) values of bed planted rice–wheat system in the Indo GangeticPlains (IGP) of India has become a constraint for irrigation planning to improve the crop water produc-tivity. In this paper, we estimated kc values from field water balance measured crop evapotranspiration(ETc) and Penman–Monteith estimated reference ET0 for dry-seeded irrigated bed planted rice–wheatrotation and also compared with conventional dry-seeded flat system of planting. The experiment wasconducted in 2001–2003 at New Delhi, India. Estimated kc values at initial stage (4–6th weeks of sow-ing) on raised beds were comparable with flat conventional planting for rice but significantly higher forwheat. However, in later stages of growth, kc values for both rice and wheat were lesser on beds than flatland. The kc values of rice during initial, crop development, mid-season and late-season stages on bedswere 0.62, 0.75, 1.16 and 0.67, respectively while in conventional flat land, corresponding kc values were0.61, 0.97, 1.42 and 0.91, respectively. The kc values for wheat at four crop growth stages (initial, crop

development, mid- and late-season) on raised beds were 0.98–1.06, 1.10–1.14, 1.25–1.26 and 0.46–0.47,respectively while on flat land with conventional row (20 cm) spacing, corresponding kc values were 0.87,1.12, 1.45 and 0.55, respectively. Flat beds similar spacing (20–47 cm) with raised beds had comparablekc values. Bed geometry led variation in plant population density influenced strongly both crop ETc lossesas well as kc values. The results provide estimates of ET0, ETc and kc for use in irrigation scheduling in bedplanted rice–wheat system in the IGP of India and elsewhere with similar environmental conditions.

© 2013 Elsevier B.V. All rights reserved.

. Introduction

Water use efficiency in lowland transplanted rice (TPR) basedystem is very low (20–30%) (Walker and Rushton, 1984; Tuongnd Bhuiyan, 1999). In this system, a minimum of 2000–2500 mmater depending on soil type is needed during rice growth periods

transplanting to harvest) (Bouman and Tuong, 2001; Tuong et al.,005; Choudhury et al., 2007; Kukal et al., 2010) and majority of

t is lost (50–80%) in the form of unproductive deep percolation,nderbund seepage and high evaporation losses from the pondedurfaces (Bouman, 2001; Bouman and Tuong, 2001; Choudhuryt al., 2007; Kukal et al., 2010). This high water intensive and low

ater productive system occupies nearly 13.5 M ha areas in the

ndo-Gangetic plains (IGP) of India (Singh et al., 2009) and alsoontributes significantly to the total food grain production of the

∗ Corresponding author. Tel.: +91 364 2570589; fax: +91 364 2570306.E-mail address: [email protected] (B.U. Choudhury).

378-3774/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.agwat.2013.03.006

country (Dhillon et al., 2010). However, with the increasing lack inwater availability day by day, sustainability of this system vis-à-visfood security in the region is at peril.

Because of the combined increasing demand for food with theincreasing scarcity of water, one of the recent innovations i.e.furrow irrigated bed planted rice–wheat (RW) system has beengaining importance to improve crop water productivity by reduc-ing the water requirements, particularly in the irrigated rice basinof semi-arid IGP of India (Hobbs and Gupta, 2003; Kukal et al., 2005,2010; Choudhury et al., 2007; Humphreys et al., 2008; Singh et al.,2009). In this system, land is prepared conventionally and raisedbed of different widths (37–107 cm) and heights (12–22.5 cm)with furrows are freshly prepared (mostly by using tractor drawnbed planters) either in puddled (wet tillage) or non-puddled (drytillage) conditions. Then on top of the puddled beds, rice seedlings

are transplanted while in non-puddled beds, rice seeds are sownby bed planter cum seed drill and irrigation water is applied inthe furrows of different widths between the beds. Number of seed-rows per bed varies depending on the crop types and bed widths.

Page 2

l Wat

AoKCtibauao2

ediTo(eeetppcee2aodwC

tiotlrdmbr(Re2(kgiacibcfp2kwKad

B.U. Choudhury et al. / Agricultura

fter the harvest of rice crops, ensuing wheat crops are drilled sownn those beds without dismantling them (Hobbs and Gupta, 2003;ukal et al., 2005, 2010; Humphreys et al., 2008; Singh et al., 2009;houdhury and Singh, 2013). Similar to zero-tillage, no additionalillage is used except to reshape the beds as required, by pass-ng a winged shovel in the furrow to maintain the shape of theed (Morrison et al., 1990). This systems is called permanent bedsnd it offers several potential advantages in improving resourcese efficiency comparable to that obtained in other most widelydopted resource conserving technologies like minimum, reducedr zero tillage for both rice and wheat crops (Hobbs and Gupta,003; Choudhury et al., 2007; Singh et al., 2009; Kukal et al., 2010).

Over a decade, adequate amount of information have been gen-rated on the comparative performance of transplanted as well asry-seeded bed planted (RB) rice and wheat crops on crop growth,

rrigation water use and yield performance over continuous floodedPR system as well as direct-seeded rice on flat land in many partsf the world including light textured alluvial soils of IGP of IndiaBalasubramanian et al., 2003; Hobbs and Gupta, 2003; Choudhuryt al., 2007; Kukal et al., 2005, 2010; Bakker et al., 2005; Beechert al., 2006; Holland et al., 2007; He et al., 2008; Humphreyst al., 2008; Singh et al., 2009; Choudhury and Singh, 2013). Allhese studies were in agreement that intermittently irrigated trans-lanted or dry-seeded RB system saved the water input at varyingroportions (9–58%) but came with significant yield reduction overontinuous flooded TPR system (Kukal et al., 2005, 2010; Bakkert al., 2005; Beecher et al., 2006; Choudhury et al., 2007; Hollandt al., 2007; He et al., 2008; Humphreys et al., 2008; Singh et al.,009). Similarly, there are reports that with the same water man-gement practices, the amount of water savings in raised bedsver flooded TPR system could also be achieved by conventionalirect-seeded rice on flat land with equal or even higher crop andater productivity (Sharma et al., 2002; Kukal et al., 2005, 2010;houdhury et al., 2007).

The knowledge of crop water requirement (CWR) is an impor-ant practical consideration to improve water use efficiency inrrigated agriculture (Tyagi et al., 2000a). CWR means the amountf water required to compensate the evapotranspiration loss fromhe cropped field (ETc) and accurate estimation of crop ETc andocal weather based atmospheric evaporative demand known aseference evapotranspiration (ET0) are very much important inetermining the exact amount of CWR. To extrapolate the measure-ent of ETc for irrigation planning at regional scale, a relationship

etween ETc vs. ET0 through crop coefficient (kc), which is theatio of ETc to reference evapotranspiration (ET0) is often usedDoorenbos and Kassam, 1979; Tyagi et al., 2000a; Kar et al., 2007).eference ET0 was estimated most widely by Penman–Monteithquation (Allen et al., 1998, 2006; Tyagi et al., 2000a; Kingra et al.,004; Kar et al., 2007) and depends mainly on weather variablesSentelhas et al., 2010). Knowing weekly/stage wise crop ETc andc values help in proper irrigation scheduling based on sensitiverowth stages and thus, optimization of water use in field cropsncluding rice–wheat system (Tyagi et al., 2000a, 2000b; Norwoodnd Dumler, 2002; Kingra et al., 2004; Kar et al., 2007). Stage wiserop coefficient (kc) values of a particular location/region also helpsn estimation of CWR from local atmospheric demand (ET0), eveny forgoing laborious and time consuming crop ETc estimation. Theoncept of CWR, ETc, ET0 and finally, kc values has been exploredairly in India for irrigation water management in lowland trans-lanted rice (TPR) followed by wheat cropping system (Tyagi et al.,000a; Tripathi, 2004; Kingra et al., 2004; Kar et al., 2007). Sincec values vary with location, local climate, topographic conditions,

ater table situations and management practices (Doorenbos andassam, 1979; Allen et al., 1998, 2006), therefore, several studiescross different agro-ecological regions of India have been con-ucted to derive daily/weekly or growth stage basis kc values of

er Management 123 (2013) 20– 31 21

rice and wheat crop using lysimeter and/or field water balanceapproach (Tyagi et al., 2000a, 2000b; Tripathi, 2004; Kingra et al.,2004; Kar et al., 2007). However, in all these studies, CWR andkc values were estimated only for lowland flooded transplantedrice followed by conventional flat planting of wheat systems (Tyagiet al., 2000a, 2000b; Tripathi, 2004; Kingra et al., 2004; Kar et al.,2007).

Till date, no such efforts were made in estimation of kc valuesand irrigation scheduling in the one decade old resource conservinginnovations i.e. bed planted rice–wheat system. Irrigation water isstill applied flatly either flooded (He et al., 2008; Singh et al., 2009)or at 2/3rd bed heights or using tensiometers at 10–40 kPa tensions(Sharma et al., 2002; Choudhury et al., 2007; Kukal et al., 2005,2010) and as a result, saving in irrigation water is coming only atthe cost of significant yield reduction (Choudhury et al., 2007; Kukalet al., 2005, 2010; Singh et al., 2009). It is still debatable whetherthe water savings in beds were because of the bed geometry orthe result of intermittent irrigation (Cabangon et al., 2002; Tabbalet al., 2002; Kukal et al., 2005, 2010) since the amount of water sav-ing in beds can also be achieved by conventional direct-seeded flatsystem (Sharma et al., 2002; Kukal et al., 2005, 2010; Choudhuryet al., 2007). Sensitive growth stage based irrigation scheduling mayimprove the low productivity in RB systems and even enhancesfurther saving of irrigation water as well as crop water productiv-ity without any penalty to the production. However, due to lack ofinformation on crop coefficient (kc) values even in the most popularbed planting region of India (IGP plains), it is difficult to implementthe irrigation scheduling. Realizing the necessity, an attempt hasbeen made to estimate the kc values of dry-seeded rice–wheat sys-tem in bed planting systems with due consideration of local climaticand management practices of IGP of India. To attribute whetherbed geometry has any effect on crop coefficients values, compar-isons were also made with conventional dry seeded flat plantedrice–wheat system with same water management practices.

2. Materials and methods

2.1. Study area description

The experiment on bed planted rice (June–December) andwheat (December–May) rotation was conducted in 2001–2003at the research farm of the Indian Agricultural Research Insti-tute, New Delhi, India, at an elevation of 228 m from mean sealevel (28◦36′N, 77◦12′E). The climate of the area is semi-aridwith an average annual temperature of 25 ◦C and average annualrainfall of 650 mm. Site description, experimental details, measure-ment of water balance components, analysis and estimation ofcrop evapotranspiration are provided in Choudhury et al. (2007),however, a brief summary of information related to crop ETc

and derivation of kc values have been provided here for bet-ter understanding. Soils were classified as Typic Haplustept, withloamy textures, medium in organic carbon (0.45–0.51%), low nitro-gen (<215 kg N ha−1 against ≥273 kg N ha−1 as medium level inIGP), medium phosphorus (15–23 kg P205 ha−1) and potassium(340 kg K20 ha−1) contents in the surface layer. The soils were neu-tral in reaction with saturated hydraulic conductivity of 0.38 to0.45 cm h−1 and having available water content of 16–18%.

2.2. Experimental design

The experiment was laid out in Randomized Complete Block

Design with four replicates and plot sizes of 10 m × 7 m. Dry-seeded rice (cultivar Pusa-44) was grown in summer (June/July toNovember/December) with dry tillage on “flat land” (DSR), “flatbeds” (FB) and raised beds irrigated at field capacity (RB00) and

Page 3

22 B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31

bed s

2m3wbibbiimrtsrfft2ldou(ptiaztwtaeop

ws4t

Fig. 1. Configuration of furrow irrigated raised

0 kPa (RB20) tensions, respectively. The raised beds (RB) wereechanically prepared by a bed planter cum seed drill and were

7 cm wide with a bed height of 22 cm, separated by furrows thatere 30 cm wide and 22 cm deep. Dry rice seeds were sown by a

ed planter cum seed drill mechanically at the rate of 60 kg ha−1

n 2 rows spaced 20 cm apart on the beds, leaving a space of 47 cmetween the rows across the furrows (20–47–20 cm) (Fig. 1). Ineds, irrigation water was applied in the furrows at field capac-

ty (RB00) and 20 kPa (RB20) tensions monitored with tensiometersnstalled at 20 cm depth in the beds. To separate the effect of water

anagement from that of the raised beds themselves, the sameow spacing (20–47–20 cm) was used on flat land in the “flat-bed”reatment (FB)(Fig. 1) and dry-seeds were sown manually. Widerpacing (47 cm “gap”) in raised and flat beds resulted in 30% lessows compared to normal spacing of 20 cm row apart. In success-ul raised bed systems, bed geometry helps the plant in making upor this loss in covered area by increased leaf growth and produc-ion at the edges of the beds (Hobbs and Gupta, 2003; Kukal et al.,005). To test the phenomenon of bed geometry effect on crop ETc

oss and more particularly, crop coefficient values, the conventionalry-seeded treatment (DSR) was included. In DSR, dry-seeded ricen flat land at the rate of 60 kg ha−1 was sown manually in rows reg-larly spaced 20 cm apart (Fig. 1) and the same water managementi.e., keeping the root zone around field capacity, 10 ± 2 kPa tension)ractices with raised (RB00) and flat (FB) beds were adopted. In allreatments, irrigation water was applied using flexible hoses. Flushrrigation (across the flat land or through the furrows in beds) waspplied in every alternate day in DSR, FB and RB00 to keep the rootone close to field capacity while in RB20, beds were irrigated whenhe soil water tension at 20 cm depth rose above 20 kPa (monitoredith installed tensiometers) (Choudhury et al., 2007). Nitrogen in

he form of urea was applied at the rate of 120 kg ha−1, of which 50%t basal, 25% at maximum tillering and the remaining 25% at flow-ring. Other basal fertilizer applications were 30 kg P in the formf single super phosphate (SSP), 60 kg K in the form of murate ofotash (MOP), 25 kg ZnSO4 and 50 kg FeSO4 ha−1.

After harvest of the rice crop, wheat cultivar HD-2687 was sown

ith a seed drill in rows spaced 20-cm apart in the formerly dry-

eeded flat land plots (DSR), and in 2 rows spaced 20 cm apart, then7 cm till the next 2 rows, in the formerly dry-seeded on FB. Inhe raised-bed treatments (RB00, and RB20) of previous rice crop,

ystem for rice–wheat establishment methods.

wheat was sown using a bed planter with two rows on the bed ata spacing of 20 cm apart similar to the rice crop (2 rows per bed)(Fig. 1); the sides of the beds were slightly reshaped manually andare considered as permanent beds. Fertilizers P and K were appliedbasally at 30 and 60 kg ha−1, respectively. Nitrogen was appliedat 120 kg ha−1 in three splits: 50% basal, 25% at crown root initi-ation (CRI) and 25% at flowering, respectively. Sources of N, P and Kwere urea, SSP and MOP, respectively. In wheat, four irrigations of6–7.5 cm each were given to all treatments at crown root initiation,tillering, flowering and the dough stage (Choudhury et al., 2007).

For measuring green leaf surface area, samples were col-lected from two row sections of 50 cm each in rice plot atmid tillering, panicle initiation (PI), flowering and physiologicalmaturity. For wheat, two row sections of 50 cm each were col-lected at crown root initiation (CRI), tillering, flowering and thedough-ripe stages. Leaf area index (LAI) was measured with anautomatic leaf area meter and the detail has been described inChoudhury et al. (2007).

2.3. Computation of reference evapotranspiration (ET0)

Reference evapotranspiration (ET0), was calculated accordingto the FAO Penman–Monteith (PMo) method recommended byFAO 56. Daily weather data were collected from the meteorolog-ical observatory of experimental farm, New Delhi which is 100 maway from the experimental plot. The FAO PMo equation for calcu-lating daily ET0 (mm day−1) using daily average data as expressedby Eq. (1) is shown below (Allen et al., 1998):

ET0 = 0.408� (Rn − G) ± �(900/T ± 273) U2(es − ea)� + �(1 + 0.34U2)

(1)

where Rn is the net radiation at the crop surface (MJ m−2 day−1); Gthe soil heat flux density (MJ m−2 day−1); T the mean daily air tem-perature at 2 m height (◦C); U2 the wind speed at 2 m height (ms−1);

es the saturation vapor pressure (kPa); ea the actual vapor pressure(kPa); (es − ea) the saturation vapor pressure deficit (kPa); � theslope of vapor pressure curve (kPa C−1); and � the psychrometricconstant (kPa C−1).

Page 4

B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31 23

Table 1Components of the seasonal water balance (mm) of rice and wheat in 2001–2002 and 2002–2003.

Treatment Rice Wheat

Ia Rb Pc DWd Ia Rb Pc DWd

2001–2002DSR 685a* 361 466b 24b 298a 25 53c −42bFB 669a 361 510a 26b 282b 25 78a −46bRB00 567b 361 433b 20b 256c 25 59b −53aRB20 497c 361 344c 48a 252c 25 44d −46b

2002–2003DSR 813a 265 493a 25a 260a 95 48b −37cFB 764ab 265 518a 23a 252a 95 65a −42bRB00 727b 265 497a 18a 228b 95 49b −46aRB20 648c 265 427b 39a 226b 95 45bc −40b

Source: Choudhury et al. (2007).* Means in the column followed by common letters (a–d) are statistically not different at 5% level of significance.a Irrigation.

nd po

2c

a

I

wtcIwSahdcrltRi(citpcm2

w(

k

wstmoc(r

b Rain.c Deep percolation beyond root zone.d Change in soil water stored in root zone (negative values mean net extraction a

.4. Computation of crop evapotranspiration (ETc) and cropoefficient values (kc)

Crop ETc was estimated with water balance of root zone of rices well as wheat crop fields as in Eq. (2):

+ R = ETc + Roff + P + S + dW (2)

here I is the irrigation, R the rainfall, P the percolation belowhe root zone, S the seepage, Roff the runoff and dW is thehange in soil water storage in the root zone. In rice and wheat,

and R were directly measured. Roff was 0 because the plotsere bunded (30 cm height) and no bund overflow occurred.

was assumed to be 0 because of the installed plastic sheetsnd because the light-textured soil favored vertical rather thanorizontal water movement. The total change in stored water,W, was calculated from the difference in measured soil waterontents just before crop establishment (but after land prepa-ation) and straight after harvest of rice and wheat. Percolationoss (P) in rice was measured using percolation cylinders fromhe bottom of the 0–30 cm root zone (Choudhury et al., 2007).ice ETc was calculated as the residual of Eq. (2), and thus

ncludes all measurement errors in the other water balance termsChoudhury et al., 2007). In wheat, ETc was calculated from thehange in measured soil water content before and after eachrrigation, taking account of the rainfall during those periods. Onhe days of irrigation, ET was estimated to be 80% of measuredan evaporation. The deep percolation beyond 90 cm depth, P, wasalculated as the residual of Eq. (2), and thus includes all measure-ent errors in the other water balance terms (Choudhury et al.,

007).Crop coefficient (kc) values (weekly as well growth stage wise)

ere determined for each crop using the following equationEq. (3)):

c = ETc

ET0(3)

Weekly and stage wise kc values for both crops (rice andheat) were estimated from the weekly crop ETC losses and atmo-

pheric evaporative demands (ET0). For rice, sowing to tillering,illering to PI, PI to flowering and post flowering to physiological

aturity/dough-ripe stages were considered as initial, crop devel-

pment, mid-season and late-season stages while for wheat, theorresponding four stages were sowing to crown root initiationCRI), CRI to tillering, tillering to flowering and flowering to dough-ipe stages of growth.

sitive values mean net addition).

3. Results and discussion

3.1. Water balance components of rice and wheat during cropgrowth periods

Irrigation water applied during rice growth period in 20 cmrow spaced flat land (DSR) and 20–47–20 cm wider spaced flatbed (FB) was not significantly different (669–813 mm) in 2001and 2002 (Table 1). Raised beds irrigated at field capacity (RB00)received 567–727 mm water and beds with irrigation at 20 kPa(RB20) received 70–79 mm less (significantly different, p < 0.05)irrigation water than RB00 in both 2001 and 2002 (Choudhury et al.,2007). Rainfall received from sowing to harvest were 361 mm in2001 and 265 mm in 2002. The percolation loss measured in flatDSR and FB were 466–517 mm and statistically (p < 0.05) similarwith beds irrigated at field capacity (RB00) (Table 1). The percola-tion of raised beds irrigated at 20 kPa was significantly (p < 0.05)70–90 mm lower than that of beds kept at field capacity. Flat (DSR,FB) and raised beds (RB00) kept at field capacity stored some extra18–26 mm water in the soil after harvest, while the raised bedsirrigated at 20 kPa stored 39–48 mm (Table 1)(Choudhury et al.,2007).

In wheat, flat treatments DSR and FB received irrigation waterof 282–298 mm in 2002 and 252–260 mm in 2003 (Table 1). Theraised beds used significantly (p < 0.05) 30–35 mm less water thanflat DSR and FB treatments (Table 1). Unlike in rice, wheat cropin flat (DSR and FB) and beds (RB00 and RB20) extracted stored soilwater of 36–46 mm and 48–78 mm, respectively from 0–90 cm pro-file depth. Percolation losses beyond 90 cm depth were 48–78 mmin flat (DSR and FB) and 44–59 mm in beds (RB00 and RB20) (Table 1)(Choudhury et al., 2007).

Crop ETc loss of both rice and wheat estimated as the residualof Eq. (2) has been presented in Tables 2 and 3 and discussed inresults and discussion Sections 3.2–3.3.

3.2. Reference evapotranspiration

During rice growing season of 2001 (5th July to 8th November),estimated daily ET0 varied from 1.43–7.82 mm day−1 with a meanvalue of 4.29 (±1.47) mm day−1 while in 2002 (29th June to 16thNovember), it varied from 1.41–8.86 mm day−1, with a mean value

of 4.24 (±1.85) mm day−1. The corresponding daily Pan Evapora-tion (US Class A: Pan E) rate varied from 1.60 to 7.60 mm day−1

(mean: 4.34 ± 1.48 mm day−1) during 2001 and in 2002, it was0.15–9.30 mm day−1 (3.23 ± 1.66 mm day−1). There was a good

Page 5

24 B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31

Table 2Crop growth stage wise leaf area index, actual evapotranspiration and crop coefficient (kc) of flat and raised bed planted dry-seeded rice.

Stage/DASa Tillering-45-49DASa bPI-74-77DAS Flowering-105–108 DAS cP. Maturity-131-135DAS Seasonal

Rice-2001 dLAI eETcfkc LAI ETc kc LAI ETc kc LAI ETc kc ETc

DSR 2.41a* 3.42a 0.69a 3.73a 5.28a 0.94a 5.23a 5.46a 1.49a 2.00a 2.42a 0.87a 556aFB 1.60c 3.33b 0.67a 3.01c 4.45b 0.79b 3.98bc 4.56b 1.25b 0.97c 2.32a 0.83a 494bRB00 1.73b 3.46a 0.70a 3.28b 4.16bc 0.74b 4.12b 4.40bc 1.20b 1.15b 1.81b 0.65b 476cRB20 1.60c 3.47a 0.70a 2.73d 4.07c 0.72b 3.90c 4.10c 1.13c 0.96c 1.77bc 0.64b 466c

Stage/DASa Tillering-44-47DAS PI-73-75DAS Flowering-106–109 DAS P. Maturity-140-144DAS Seasonal

Rice-2002 dLAI eETcfkc LAI ETc kc LAI ETc kc LAI ETc kc ETc

DSR 1.92a 2.97b 0.52a 3.53a 4.25a 1.00a 4.67a 5.08a 1.36a 1.58a 3.15a 0.95a 560aFB 1.41b 3.09b 0.53a 2.90b 3.30b 0.78b 3.57bc 4.30b 1.16b 1.27b 2.60b 0.79b 490bRB00 1.55b 3.14a 0.53a 3.00b 3.22b 0.76b 3.72b 4.17c 1.13b 1.31b 2.36c 0.69c 478bRB20 1.58b 3.17a 0.53a 2.77b 3.01c 0.71b 3.38c 3.74d 1.00b 1.20b 2.04d 0.60d 447c

* Means in the column followed by common letters (a–d) are statistically not different at 5% level of significance.a Days after sowing.b Panicle initiation.c Physiological maturity.

avsp22iuwatEt(s((0E(ms

TC

d Leaf area index.e Crop evapotranspiration.f Crop coefficient.

greement between PMo and Pan E (R2: 0.70–0.72) and PMos. net solar radiation (R2: 0.85–0.87) (Fig. 2). During both sea-ons, mean weekly ET0 loss decreased consistently as the monthrogressed from July to November (Fig. 2) and it varied from.57–6.08 mm day−1 (mean value: 4.53 ± 1.13 mm day−1) during001 and 2.60–7.67 mm day−1 (mean value: 4.54 ± 1.55 mm day−1)

n 2002. Comparable weekly average PMo estimated ET0 val-es (3.5–5.9 mm day−1) during rice growing season (July–October)ere also reported by Tyagi et al. (2000a) in the similar semi-

rid climate of Karnal (29◦43′N, 76◦58′E), which is near tohe present study area. Correlation studies affirmed that meanT0 loss was strongly influenced by weather variables, par-icularly net solar radiation (r = +0.75–0.93), mean temperaturer = +0.879–0.897) and wind speed (r = +0.75–0.87). In winter sea-on, estimated reference ET0 loss during wheat growing periodsNovember/December to April) varied widely: from 1.04–8.59 mmmean: 4.07 ± 1.77 mm) in 2002 and during 2003, it varied from.97–6.43 mm (mean: 4.86 ± 1.68 mm). Compared to reference

T0, mean Pan E was 8.5–10% higher during the winter seasonsNovember–April) of both years. However, there was good agree-

ent between estimated PMo and Pan E (R2: 0.86–0.87) and netolar radiation (R2: 0.90–0.92) (Fig. 2). Similar to rice growing

able 3rop growth stage wise leaf area index, actual evapotranspiration and crop coefficient (kc

Stage/DASa bCRI-22DAS Tillering-45DAS

Wheat-2002 cLAI dETcekc LAI ETc kc

DSR 1.23a* 1.35d 0.84c 2.21a 2.45a 1.21a

FB 0.81c 1.68a 1.05a 1.23c 2.39b 1.18a

RB00 0.96b 1.47c 0.91bc 1.58b 2.32c 1.15a

RB20 0.93bc 1.55b 0.96ab 1.47bc 2.45a 1.21a

Stage/DASa CRI-24DAS Tillering-48DAS

Wheat-2003 cLAI dETcekc LAI ETc kc

DSR 1.07a 1.63b 0.89b 1.89a 2.29b 1.02a

FB 0.73b 1.96a 1.06a 1.01c 2.52a 1.13a

RB00 0.81b 1.91a 1.04a 1.17bc 2.34ab 1.05a

RB20 0.84ab 1.95a 1.06a 1.22b 2.39ab 1.07a

* Means in the column followed by common letters (a–c) are statistically not different

a Days after sowing.b Crown root initiation.c Leaf area index.d Crop evapotranspiration.e Crop coefficient.

season, weekly mean ET0 during wheat growing season was alsostrongly influenced by net solar radiation (r = +0.97), mean temper-ature (r = +0.98), wind speed (r = +0.68–0.75) and relative humidity(r = −0.98). Estimated cumulative seasonal (summer) reference ET0losses during rice growth periods were 621 mm and 657 mm during2001 and 2002, respectively.

3.3. Seasonal crop evapotranspiration (SETc) of rice and wheatcrops

Estimated seasonal crop evapotranspiration (SETc) loss of ricecrops (presented in Choudhury et al., 2007) during 2001 and 2002in dry-seeded 20-cm row spaced flat land (DSR) was 556–560 mm(Table 2). The net solar radiation (Rn) during the rice growing sea-sons was 1598 MJ m−2 in 2001 and 1668 MJ m−2 in 2002. Wideningthe row spacing from 20 cm (DSR) to 20–47 cm in flat bed (FB)resulted in reduction in SETc by 11–12.5% and further changing theland configuration from flat (FB) to raised beds (RB00) with simi-

lar row spacing (20–47 cm) and moisture regime at field capacity,SETc loss dropped by an additional 12–18 mm over FB and regis-tered 475.9 to 477.7 mm during 2001 and 2002. On imposition ofsoil moisture stress at 20 kPa tensions in beds (RB20), SETc reduced

) of flat and raised bed planted wheat.

Flowering-90DAS Dough-112 DAS Seasonal

LAI ETc kc LAI ETc kc ETc

4.81a 3.61a 1.45a 1.05a 2.58a 0.59a 311.6a3.41c 2.98b 1.19b 0.85c 2.04c 0.47b 279.2b3.87bc 3.06b 1.23b 0.91b 2.13bc 0.49b 274.4b3.70bc 3.08b 1.24b 0.92b 2.17b 0.49b 274.5b

Flowering-95DAS Dough-116 DAS Seasonal

LAI ETc kc LAI ETc kc ETc

4.03a 3.62a 1.44a 0.97a 2.39a 0.51a 344.8a2.93c 3.14b 1.25b 0.79c 2.18b 0.46ab 324.9b3.14b 3.21b 1.28b 0.85b 2.11bc 0.45b 317.1c3.05bc 3.19b 1.27b 0.81bc 2.00c 0.43b 313.5c

at 5% level of significance.

Page 6

B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31 25

y = 0.870x + 0.502R² = 0.704

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

ET

0, m

m (

P&

M)

da

y-1

Pan Eva poration, mm day-1

Rice-2001

y = 0.894x + 1.368R² = 0.72 5

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7. 0 8.0 9.0

ET

0, m

m (

P&

M)

da

y-1

Pan Eva poration, mm day-1

Rice-2002

y = 1.027x + 0.308R² = 0.86 6

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

ET

0, m

m (

P&

M)

da

y-1

Pan Evapora tion, mm day-1

Wheat-2002

y = 0.84 4x + 0.533R² = 0.87 2

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.0 1. 0 2.0 3. 0 4. 0 5.0 6. 0 7. 0 8.0

ET

0, m

m (

P&

M)

da

y-1

Pan Eva poration, mm day-1

Wheat-2003

y = -0.188x2 + 3.116x + 0.3 51R² = 0.92 6

y = -0 .160 x2 + 2.9 54x + 1.538R² = 0.916

0

2

4

6

8

10

12

14

16

0 1 2 3 4 5 6 7 8

Net

ra

dia

tio

n, M

J m

-2d

ay

-1

ET0, mm day-1

Wheat-20 02 Wheat- 200 3

y = -0.217 x2 + 4.145x - 2.429R² = 0.86 0

y = -0 .343 x2 + 5.075x - 3.876R² = 0.90 6

0

2

4

6

8

10

12

14

16

18

20

0 1 2 3 4 5 6 7 8 9

Net

ra

dia

tio

n, M

J m

-2d

ay

-1

ET0, mm day-1

Rice-200 1 Ric e-2002

F ated(

fDiwS2piRi

ig. 2. Relationships among daily pan evaporation, net solar radiation and estimJune/July–November)–wheat (December–April) growing periods.

urther by 10–31 mm from stress free beds (RB00). In closer spacedSR, 23–34% higher plant population density (288–296 panicle m−2

n DSR) and 15–18% higher above ground biomass production thanider spaced stress free beds (RB00) resulted in 80–82 mm higher

ETc loss (Table 2). Similarly, on imposition of moisture stress at0 kPa tensions on beds (RB20), population density and biomass

roduction decreased by 4–10% over beds irrigated at field capac-

ty (RB00) (Table 2) and this has resulted in reduction of SETc inB20. Although SETc loss of flooded or alternate wetting and dry-

ng puddled rice has been estimated both by field water balance

daily reference evapotranspiration (ET0) by Penman and Monteith during rice

as well as lysimeter extensively across IGP of India (Sandhu et al.,1982; Bhardwaj, 1983; Tyagi et al., 2000a; Tripathi, 2004; Kingraet al., 2004; Kar et al., 2007; Chahal et al., 2007; Yadav et al., 2011),information on ETc loss of dry-seeded rice on raised and flat bedslimited. Among the few studies similar to our study, Sharma et al.(2002) measured SETc by field water balance of direct-seeded rice

on flat land (DSR) and raised beds (RB) irrigated at 10–20 kPa ten-sions at semi-arid IGP of Modipuram (at 29◦01′N and 77◦45′E andclose to our experimental site). They reported SETc loss of 760 mmwhen the crop was irrigated at 10 kPa tensions on raised (RB) and

Page 7

2 l Wat

watdwphrboogcilaiEe

(3ntaRmaerewspm(vb22f

3o

r0ssti2uwt7usrciot(d

6 B.U. Choudhury et al. / Agricultura

ith the increase in soil water tension to 20 kPa, SETc loss on DSRnd RB reduced to 670 mm. These values were 35–40% higher thanhe measured values of the present investigation. This was mostlyue to the application of 2.4–3.3 times higher amount of irrigationater by Sharma et al. (2002) as well 2 times higher grain biomassroduction than in our study. Similarly, in the semi-arid IGP of Lud-iana, Yadav et al. (2011) also measured SETc loss of dry-seededice on flat land (DSR) irrigated at 20 kPa tension by field wateralance method and they reported SETc loss of 655.5 mm (averagef two years) and this value was 17–18% higher than the valuesf SETc measured for DSR in the present study which was irri-ated lightly every second day to keep the root zone close to fieldapacity. Application of 20% higher total water input (rain plusrrigation: 1297 mm) in relatively drier weather condition, particu-arly higher mean temperature, atmospheric evaporative demandnd solar radiation in experimental site of Yadav et al. (2011) thann the present investigation might have resulted in higher seasonalTc loss of rice (Sharma et al., 2002; Choudhury et al., 2007; Yadavt al., 2011).

Similarly, in post rice wheat crop, SETc losses during crop growthsowing to maturity) in 20 cm row spaced flat land (DSR) was12 mm during 2002 and 345 mm in 2003 (Table 3). This was sig-ificantly higher (11.6–13.5% in 2002 and 6.1–8.9% in 2003) thanhe wider row spaced (20-47-20 cm) flat (FB) and raised beds (RB00nd RB20). In wheat crop, water management in beds (RB00 andB20) were same with flat planting (DSR and FB) (details given inethodology). Similar to rice, 1.6–1.7 times higher panicle density

nd 14–17% higher above ground biomass production (Choudhuryt al., 2007) in closer spaced flat DSR over wider spaced beds (RB’s)esulted in higher SETc loss (Table 3). Estimated seasonal refer-nce ET0 losses during wheat growth periods (December to April)as 17–35% higher compared to measured crop SETc loss for the

ame period. The net solar radiations (Rn) during the wheat growingeriods were 1021 MJ m−2 in 2002 and 1134 MJ m−2 in 2003. Totaleasured SETc losses of wheat on flat land with normal spacing

DSR) in our study was, however, fairly comparable to the reportedalues by field water balance studies of irrigated wheat measuredy several workers in the semi-arid IGP plains of India (Tyagi et al.,000b; Jalota et al., 2006; Chakraborty et al., 2008; Singh et al.,011). However, information on SETc losses of wheat on raised bedsrom the region was inadequate.

.4. Weekly crop evapotranspiration and crop coefficient valuesf rice

In the initial week after sowing (WAS), wider spaced bed plantedice (RB00 and RB20) had kc values of 0.49–0.50 in 2001 and.36–0.37 in 2002 and this was 30–86% higher than the closelypaced (20 cm row) flat lands (DSR: 0.26–0.28) (Figs. 3 and 4). Widerpaced (20–47 cm) flat beds (FB) had also 25–49% less kc valueshan the beds. This trend of higher kc values in beds (0.71–0.78n 2001 and 0.54–0.57 in 2002) over DSR and FB (0.38–0.45 in001 and 0.38–0.43 in 2002) continues till 2nd WAS. Higher kc val-es in beds in the initial WAS were mostly due to 30–86% highereekly average crop ETc losses (2.64–3.48 mm day−1) compared

o flat lands (DSR). In bed system of planting (for wheat), only0% area is covered by crops while the remaining area remainsncovered (He et al., 2008). Owing to bed configuration, widerpacing led less plant population density in the initial 1–2 WASesulted in more surface area exposed to un-shaded or sunnyonditions. Since weekly average net solar radiation was highern the initial periods (12.6–15.6 MJ m−2) compared to later parts

f crop growth (<13 MJ m−2), therefore, ETc loss was higher andhus higher kc values in beds. Fynn et al. (1993) and Tyagi et al.2000a) also reported in separate experiments that higher ETc

uring un-shaded or sunny conditions in the initial weeks were

er Management 123 (2013) 20– 31

primarily related to net solar radiation. In 3rd to 6th WAS, cropETc losses were comparable between flat (DSR and FB) and bedplanted rice (RB00 and RB20) (Fig. 3). Weekly average ETc loss dur-ing 3–6 WAS varied from 3.19–4.68 mm day−1 in flat planting whileit was 3.06–4.06 mm day−1 in bed planted rice. As a result, kc val-ues at 3–6th WAS were comparable in flat (DSR and FB: 0.81)and bed planted (RB00 and RB20: 0.75) rice. However, from 7thWAS onwards, kc value in DSR was higher (0.83–0.85) comparedto beds irrigated at field capacity (RB00: 0.68–0.69). In DSR, kc val-ues exceeded one (1.03–1.16) in the 10th to 12th WAS during2001 and 2002, which coincides with PI stages of rice growth. Inwider spaced flat (FB) and raised beds (RB00), kc values reachedunity (1.06–1.12 in 2001 and 1.00 during 2002) in the 12–13thWAS. Effect of soil moisture stress (20 kPa tension) on beds (RB20)was visible inconsistently from 7th WAS, which resulted in 7–8%less kc value compared to stress free beds (RB00). Rice on bedsat 20 kPa tensions (RB20) reached kc value of ≥1 (1.01–1.14) oneweek later (14th WAS) of RB00. On 7th WAS, weekly average ETc

loss was 4.08–4.83 mm day−1 during 2001–2002 in closely spacedDSR against 3.29–4.01 mm day−1 in stress free widely spaced beds(RB00) and 2.96–3.72 mm day−1 in 20 kPa tension beds (RB20). Max-imum ETc loss was measured at 13–14th WAS and was 19–47%higher in closely spaced flat land (DSR: 5.61–6.39 mm day−1) overbeds (4.33–4.83 mm day−1) (Figs. 3 and 4). Crop ETc increased asmost of the radiation was intercepted by crop canopy since LAI atthis stage of rice growth exceeded 4.6. Similar observation of higherETc at peak LAI of rice crop was also reported by Tyagi et al. (2000a)in the semi-arid Karnal of IGP plain. Like ETc, maximum kc valuewas estimated in flat land (DSR: 1.531–1.551) but in the 15–17thWAS which coincides with reproductive phase. Compared to DSR,maximum kc values were 8–12% less in wider spaced flat (FB) andstress free beds (RB00) (1.353–1.387 in 2001 and 1.392–1.416 in2002). Imposing moisture stress on beds (RB20), maximum kc val-ues further reduced by 8–16% (1.19–1.24) over RB00. In the 19-21stWAS, kc values decreased to 0.66–0.76 in flat planted rice (DSR)and 0.476–0.525 in bed planted rice (RB00 and RB20). Crop ETc lossduring 2001–02 on flat (DSR and FB) and raised beds (RB00 andRB20) also declined to less than 2.5 mm day−1 in the 19–21st WAS.Flat (FB) and raised beds (RB00) with similar row spacing and mois-ture regime had comparable ETc as well as kc values in most ofthe growth periods, except in the initial WAS (Figs. 3 and 4). Thisreflects that land configuration, more particularly bed geometry ledvariation in plant population density influenced crop evapotrans-piration loss as well as kc values, since closely spaced flat DSR had23–34% higher panicle density (288–296 panicle m−2) compared towider spaced flat (FB) and raised beds (RB) (Choudhury et al., 2007).

The average weekly ETc of direct-seeded rice on closely spacedflat land (DSR) of our study increased from <2 mm per day inthe early growing period to a peak value of 5.61 – 6.39 mm perday at reproductive phase (13–14th WAS) and then declined to<2.5 mm day−1 at physiological maturity (19–21st WAS). Similarly,kc values increased from <0.5 in the early stages to a maximum of1.53–1.55 at reproductive phase and then declined to <0.8 at physi-ological maturity (Figs. 3 and 4). Crop ETc in our study was relativelylow compared to reported values of <5 mm per day in the earlygrowing period to a peak ETc of 6.61 mm per day at reproductivephase (11th WAS) and then declined to 2.88 mm day−1 at matu-rity (17th WAS) from lysimetric study for transplanted rice in thesimilar semi-arid climate of Karnal (Tyagi et al., 2000a). Lysimetricinitial kc value of one for transplanted flooded rice was reportedby several workers (Midmore et al., 1984; Singh and Kumar, 1993;Tyagi et al., 2000a) and a peak value of 1.77 at 10th WAS by Kingra

et al. (2004) under comparable semi-arid IGP of Ludhiana, Indiawhich was relatively higher than the present study. This mightbe due to transplanting of rice in flooded water in those reportedstudies and thus, more water was available for surface evaporation

Page 8

B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31 27

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

ET

0, E

Tc,

mm

da

y-1

Weeks after sowing

Rice-2001

DSR FB RB00

RB20 ET0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

ET

0, E

Tc,

mm

da

y-1

Weeks after sowing

Rice-2002

DSR FB RB00

RB20 ET0

(a)

(b)

F der diP ity) of

ceEI22t

3s

(esisIr3(g

ig. 3. Variation in weekly (sum of daily/7) mean crop evapotranspiration (ETC) unenman and Monteith during crop growing periods (sowing to physiological matur

ompared to present aerobic dry-seeded rice of our study. How-ver, pattern of initial increase, reaching peak and then decline inTc of the present study resembles other lysimetric studies acrossndia (Singh and Kumar, 1993; Tyagi et al., 2000a; Kingra et al.,004) and other parts of the world (Midmore et al., 1984; Kuo et al.,006). The weekly variation in ETc was also strongly influenced byhe fluctuation of net solar radiation (r = > + 0.87).

.5. Leaf area index, crop ETc and kc values of rice at differenttages of growth

Measured leaf area index (LAI) increased from initial 7th WAStillering), reached a maximum value during 16–17th WAS (flow-ring) and thereafter decreased due to leaf senescence in both theeasons. Maximum LAI was significantly (p < 0.05) higher (26–27%)n flat 20 cm row spaced DSR (LAI: 4.67–5.23) compared to widerpaced (20–47 cm) stress free beds (RB00: 3.71–4.12) (Table 2).mposition of moisture stress (20 kPa) on beds (RB20) further

educed LAI by 5–9% over RB20. Flat beds (FB) had maximum LAI of.57–3.98, comparable to similar row spaced beds (RB00). Flat landsDSR) recorded significantly (p < 0.05) higher LAI in all the stages ofrowth compared to wider space flat (FB) and beds (RB00 and RB20).

fferent rice planting methods and estimated reference evapotranspiration (ET0) by 2001 (a) and 2002 (b).

In the initial stages of rice growth (sowing to tillering period,1–7th WAS average), in-spite of 23–34% higher planting densityand higher LAI (Table 2), mean daily ETc loss was comparable(3.33–3.47 mm day−1 in 2001 and 2.97–3.17 mm day−1 in 2002)among flat (DSR and FB) and beds (RB00 and RB20).Wider spacingand land configuration in beds attributed to lower plant populationdensity (Choudhury et al., 2007; He et al., 2008), lower LAI (Table 2)and higher specific surface area exposed for evaporation loss com-pared to flat system (DSR). As a result, initial (sowing to tillering) kc

values by PMo method were comparable among DSR, FB and RB’s:0.67–0.70 in 2001 and 0.52–0.53 in 2002. However, at crop develop-ment (tillering to PI, 8–12th WAS), 23–34% higher planting densityand LAI in flat land (DSR) increased crop ETc losses significantly (atp < 0.05; 27–31% higher) over stress free beds (RB00: 3.22–4.16 mmday−1) and an additional 3–9% over beds at 20 kPa soil water ten-sion (RB20). During this phase, kc values were 27–31% higher in DSR(0.941–1.00) than stress free beds (RB00). Imposition of stress onbeds (RB20) resulted in an additional drop of kc values by 3–10% over

RB00. At reproductive phase (PI to flowering, 12–17th WAS), cropETc loss was 21.9–24.0% higher in DSR than RB00 and attained themaximum value of 5.08–5.46 mm day−1. Moisture stress on beds(RB20) further reduced ETc significantly over stress free beds (RB00).

Page 9

28 B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Cro

p c

oef

fici

ent,

K

c

Weeks after sowing

Rice-2001

DSR FB

RB00 RB20

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Cro

p co

effi

cien

t, K

c

Weeks after sowing

Rice-2002

DSR FB

RB00 RB20

(a)

(b)

Fig. 4. Weekly average (sum of daily/7) crop coefficient values (kc) of rice crop (sowing to physiological maturity) under flat (DSR and FB) and bed planting (RB00 and RB20)methods during the experimental season 2001 (a) and 2002 (b).

Ava(mi3hepsk(stbbDiElaa

t this stage, kc values exceeded one and attained the maximumalue in both flat (DSR and FB) and bed planted (RB00 and RB20) rice,lthough closely spaced flat DSR registered 16–19% higher kc values1.35–1.48) than stress free beds (RB00: 1.09–1.23) (Table 2). This

ay be attributed to cumulative effects of 23–34% higher plant-ng density in DSR (288–296 panicle m−2 (Choudhury et al., 2007),0–31% higher leaf area index (LAI: 4.67–5.23) as well as 19–29%igher crop ETc loss in DSR (ETc: 4.80–5.80 mm day−1) (Choudhuryt al., 2007). Similarly, by irrigating beds at 20 kPa tensions (RB20),opulation density, LAI and ETc decreased further (4–10% less) overtress free beds (RB00) (Table 2) and this has resulted in reduction ofc values by 4–8% at reproductive (PI-flowering) and 7–14% at latepost flowering to physiological maturity) stages of growth overtress free beds (RB00). In the late stages of growth (post-floweringo physiological maturity; 17–21st WAS), kc values decreased inoth flat (DSR and FB: 0.83–0.86 in 2001 and 0.78–0.95 in 2002) andeds (RB00 and RB20: 0.636–0.739 in 2001and 0.69–0.75 in 2002).ecline in crop ETc in the post-flowering stages across all plant-

ng methods resulted in reduction in kc values. Leaf area influenced

Tc loss substantially as evident from a very strong positive corre-ation between LAI and kc values (r = +0.86–0.89) as well as a goodgreement (R2: 0.795–0.851) between the estimated values of kc

nd measured LAI (Fig. 5) during 2001 and 2002.

Finally, computed kc values (average of two seasons) during ini-tial, crop development, reproductive and late stages in 20 cm rowspaced flat lands (DSR) were 0.61, 0.97, 1.42 and 0.91, respectivelywhile in 20-47-20 cm row spaced raised beds (RB00 and RB20), cor-responding values were 0.62, 0.71–0.75, 1.06–1.16 and 0.62–0.67,respectively. Flat beds (FB) similar in row spacing with raised beds,kc values were 0.60, 0.79, 1.20 and 0.81, respectively (Table 2). Inwider spaced dry-seeded aerobic flat (FB) and raised beds (RB00and RB20), kc values at all stages (except reproductive phase) wereconsiderably lower than those suggested by FAO for flat floodedtransplanting rice. The FAO estimated values for the correspondinggrowth periods in semi-arid condition with light to moderate windspeeds were 1.1–1.15, 1.10–1.15, 1.1–1.3 and 0.95–1.05, respec-tively (Allen et al., 1998; Tyagi et al., 2000a). Comparatively highervalues of FAO derived kc were mostly due to the maintenance ofsubmergence with standing water depth of 10–20 mm from initial(transplanting) till crop development stages of rice growth. Stand-ing water encouraged more evaporation losses from free watersurface and thus, resulted in higher ET losses and kc values (Allen

et al., 1998). However, in the present study, instead of floodedtransplanted rice, kc values were estimated for direct-seeded ricein raised and flat beds. No standing water was maintained; rathercrops were irrigated at soil moisture tensions of 10–20 kPa tensions.

Page 10

B.U. Choudhury et al. / Agricultural Water Management 123 (2013) 20– 31 29

y = 0.17 6x + 0.41 2

R² = 0.795

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

LAI

Rice-2001

y = 0. 215x + 0.262

R² = 0. 851

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

LAI

Rice-2002

y = 0.1 11x + 0.87 6

R² = 0.718

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.0 0.5 1.0 1.5 2.0 2.5 3. 0 3.5 4.0 4.5 5.0

Kc

valu

esK

c valu

es

Kc

valu

esK

c valu

es

LAI

Wheat-2002

y = 0.109x + 0.94 3

R² = 0.72 3

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

LAI

Wheat-2003

(a)

(b)

F oeffic(

Al

3s

Cfl(shrhflcedoweeaAuae

ig. 5. Relationship between measured leaf area index (LAI) and estimated crop cRB’s) rice (a) and wheat (b) system.

s a result, less water was available for soil evaporation and thusower kc values were estimated.

.6. Leaf area index, crop ETc and kc values of wheat at differenttages of growth

In post rice wheat crop, at initial stages of 22–24 (sowing to -RI) days after sowing (DAS), kc values in wider spaced (20–47 cm)at (FB) and raised beds (RB’s: RB00 and RB20) were significantlyp < 0.05) higher (10–17%) and exceeded one compared to closelypaced (20–20 cm) flat land (DSR: <0.90) (Table 3). Despite 27–51%igher LAI (1.07–1.23) of wheat at initial stage (CRI) in DSR overaised beds (RB’s), the corresponding crop ETc loss was 8–24%igher in RB’s during 2002 and 17–20% higher in 2003 compared toat lands (DSR: 1.35–1.63 mm day−1) (Table 3). The initial higherrop ETc loss in bed planted wheat might be due to higher soilvaporation rather than transpiration losses at LAI of <2.0, sinceue to bed geometry, specific surface length exposed for evap-ration loss was 44 cm higher for every 67 cm length and withider spacing, 37–44% less plant population density (Choudhury

t al., 2007) compared to 20 cm row spaced flat lands (DSR). Het al. (2008) also reported that in bed planted wheat, 30% surfacerea remains uncovered by crop canopy and is exposed to sun.

t crop development stage (CRI to tillering, 45–48 DAS), kc val-es of wheat on flat (DSR) and raised beds (RB’s) were comparablend statistically not different at 5% level of significance. In DSR, kc

xceeded one (1.02–1.21) and was 15–44% higher than CRI stage.

ient (kc) values at different stages of growth under flat (DSR, FB) and bed planted

Similarly, in beds (RB’s), during 2002, kc values at crop develop-ment stage (1.15–1.21) were 26% higher over CRI stage. Estimatedkc values were comparable between similar row spaced flat (FB)and raised beds (RB’s) and FB recorded 12% increase in kc valuesat crop development over initial stages of growth. At reproductivestage (tillering to flowering, 90–95 DAS), kc values reached max-imum in both flat (DSR and FB) and bed planted wheat but was13–16.9% higher in closely spaced DSR (1.44-1.45) compared tobeds (RB’s: 1.23–1.28) (Table 3). However, flat beds (FB) had com-parable kc values (1.19–1.25) with similar row spaced beds (RB’s).Significant decline in mean crop ETc (>10%) as well 19–24% lesserLAI at flowering in bed planted wheat (RB’s) compared to closelyspaced flat land (DSR) might have reduced kc values in beds. Atflowering (90–95 DAS) during 2002 and 2003, estimated crop ETc inDSR was maximum (3.61–3.62 mm day−1) when LAI was 4.03–4.81.At late stage (post flowering to dough-ripe stages, 112–116 DAS),crop ETc declined by more than 30% (2.0–2.58 mm day−1) and LAIreduced to <1 over flowering stage in both flat (DSR and FB) andbeds. In the IGP of Delhi, India, Chattaraj et al. (2011) also reportedsimilar pattern of peak ETc loss of wheat at flowering stage (80–95DAS) when LAI was maximum and then declined with the declinein LAI at milking stages (90–105 DAS) of growth. Crop coefficientvalues in flat land (DSR) decreased to 0.51–0.59 during 2002–2003

and the corresponding decrease in beds (RB’s) was 0.43–0.49 whichwas comparable to wider spaced flat beds (FB: 0.46–0.47).The kc

values (average of both seasons) during initial (CRI), crop devel-opment (tillering), reproductive (flowering) and late (dough-ripe

Page 11

3 l Wat

sasrvsmad(oItd

otr(tfdva(drcr(gtspdiEahbi

3

(t2ea(cb0aLsr0

4

wt

0 B.U. Choudhury et al. / Agricultura

tages) stages in closely spaced flat land (DSR) were 0.87, 1.12, 1.45nd 0.55, respectively while in widely spaced beds (RB’s), corre-ponding kc values were 0.98–1.06, 1.10–1.14, 1.26 and 0.46–0.47,espectively. Flat bed (FB) with similar row spacing of beds had kc

alues of 1.05, 1.16, 1.22 and 0.47, respectively (Table 3). In theemi-arid climate of Karnal, Tyagi et al. (2000b) reported PMo esti-ated lysimetric kc values of 0.50, 1.36, 1.24, and 0.42, respectively

t the four crop growth stages (initial, crop development, repro-uctive, and maturity). Our kc values for conventional plantingDSR) was comparable to mini-lysimeter based reported kc valuesf Singh et al. (2011) in the similar semi-arid climate of Ludhiana,ndia. They also reported an initial increase in kc values from 0.6–0.8o 1.0–1.3 at 90 DAS, reached a peak of 1.5–1.7 at anthesis and theneclined to 0.4–0.5 at grain filling.

In both flat (DSR and FB) and raised beds (RB00 and RB20) ofur study, ETc and simultaneously kc values increased from ini-ial (CRI) to crop development (tillering), reaching a maximum ateproductive (flowering) and then declined at late phases of growthdough-ripe stages) (Table 3). Singh and Kumar (1993) stated thathe cumulative ETc values increase at a decreasing rate with timerom sowing to harvesting. A similar behavior of lower averageaily ETc (<1 mm day−1) in the early growing period to a maximumalue at flowering (2.7 mm day−1) was reported by Bandyopadhyaynd Mallick (2003) in a humid tropical climate of West Bengal23◦1′N latitude, 88◦5′E longitude), India. In the present study,uring both seasons, till flowering stages of growth, estimatedeference ET0 was lesser than crop ETc but afterwards increasedonsiderably (1.6–1.9 times) than crop ETc. A strong positive cor-elation among LAI vs. crop ETc (r = +0.875 to 0.887), crop ETc vs. kc

r = +0.587 to 0.626) and LAI vs. kc (r = +0.690 to 0.711) across therowth stages of both seasons of wheat (sowing to dough) affirmedhat land configuration (flat to beds) led variation in planting den-ity influenced considerably the crop ETc loss and kc values. In beds,lant population density was 40–44% less in 2002 and 37–40% lessuring 2003 compared to flat DSR where wheat had 369–417 pan-

cles per m2 area (Choudhury et al., 2007). As a result, LAI and cropTc were significantly high(er) in flat closely spaced DSR acrossll phenological stages of crop growth and might have resulted inigher kc values in DSR compared to widely spaced beds. Duringoth seasons, the estimated values of kc and measured LAI was also

n good agreement (R2: 0.718–0.723) (Fig. 5).

.7. Crop water productivity of rice and wheat

Crop water productivity with respect to evapotranspiration lossWPET: grain yield/ETc) of both rice and wheat also varied amongreatments but inconsistently between two years (2001–02 to002–03) (Choudhury et al., 2007). In 2001, WPET of rice was high-st in 20 cm row spaced flat land (DSR: 0.76 g grain kg−1 water)nd was less in flat (FB: 0.65 g grain kg−1 water) and raised bedsRB: 0.67–0.68 g grain kg−1 water). In 2002, however, WPET wasomparable (statistically same) in the three treatments on raiseded and flat land kept around field capacity (DSR, FB and RB00:.75–0.78 g grain kg−1 water) but declined in raised beds irrigatedt 20 kPa (RB20: 0.67 g grain kg−1 water) (Choudhury et al., 2007).ike in rice, WPET of wheat was significantly higher in 20 cm rowpaced flat land (DSR: 1.16–1.58 g grain kg−1 water) compared toaised and flat beds (1.37–1.51 g grain kg−1 water in 2001–02 and.89–1.01 g grain kg−1 water in 2002–03) (Choudhury et al., 2007).

. Conclusions

In the initial weeks (2nd WAS), crop coefficient (kc) value of riceas significantly (p < 0.05) higher on raised beds than on conven-

ional dry-seeded flat land. In the later part (7th WAS onwards), kc

er Management 123 (2013) 20– 31

value of rice on raised beds was significantly less (19–26%) com-pared to flat land. Similarly, compared to conventional flat land,kc value of bed planted wheat was significantly less (13–15%) atmid- and late-seasons. From 7th WAS onwards, moisture stress at20 kPa tensions on beds further reduced kc value of rice by 5–8%over beds irrigated at field capacity. Bed geometry led reduction incrop canopy coverage and planting density in both rice and wheatstrongly influenced crop ETc loss and kc values as well. Inadequacyof information on kc values of bed planted rice and wheat acrossIndia and other parts of the world including FAO derived kc valueslimited the comparison of the presently estimated kc values. Cropcoefficient values of rice and wheat on conventional dry-seeded flatlands, however, differ considerably from those suggested by FAO fortransplanted rice and dry-seeded wheat. This unique crop estab-lishment method i.e. bed planting system specific kc estimationunder semi-arid climate will certainly help in efficient manage-ment of water resources through precise irrigation scheduling inthe arid and semi-arid regions of world including IGP of India.

References

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration Guidelinesfor Computing Crop Water Requirements. Irrigation and Drain, Paper No. 56.FAO, Rome.

Allen, R.G., Pruitt, W.O., Wright, J.L., Howell, T.A., Ventura, F., Snyder, R., Itenfisu, D.,Steduto, P., Berengena, J., Yrisarry, J.B., Smith, M., Pereira, L.S., Raes, D., Perrier,A., Alves, I., Walther, I., Elliott, R., 2006. A recommendation on a standard-ized surface resistance for hourly calculation of reference ET0 by the FAO56Penman–Monteith method. Agricultural Water Management 81, 1–22.

Bakker, D.M., Hamilton, G.J., Houlbrooke, D.J., Spann, C., 2005. The effect of raisedbeds on soil structure, water logging, and productivity on duplex soils in WesternAustralia. Australian Journal of Soil Research 43, 575–585.

Balasubramanian, V., Ladha, J.K., Gupta, R., Naresh, R.K., Mehla, R.K., Singh, B., Singh,Y., 2003. Technology options for rice in the rice–wheat system in South Asia. In:Ladha, J.K., Hill, J.E., Duxbury, J.M., Gupta, R.K., Buresh, R.J. (Eds.), Improving theProductivity and Sustainability of Rice–Wheat Systems: Issues and Impacts. ASAInc., CSSA Inc., SSSA Inc., Madison, USA, pp. 115–147 (ASA Special Publication65).

Bandyopadhyay, P.K., Mallick, S., 2003. Actual evapotranspiration and cropcoefficients of wheat (Triticum aestivum) under varying moisture levels of humidtropical canal command area. Agricultural Water Management 59 (1), 33–37.

Beecher, H.G., Dunn, B.W., Thompson, J.A., Humphreys, E., Mathews, S.K., Timsina, J.,2006. Effect of raised beds, irrigation and nitrogen management on growth,water use and yield of rice in south-eastern Australia. Australian Journal ofExperimental Agriculture 46, 1363–1372.

Bhardwaj, S.P., Studies on Consumptive Use Rates in Weighing Type Lysimeters forIrrigation. Report, Central Soil and Water Conservation Research Institute, DeharDun, 1983, p. 100.

Bouman, B.A.M., 2001. Water-efficient management strategies in rice production.International Rice Research Notes 16 (2), 17–22.

Bouman, B.A.M., Tuong, T.P., 2001. Field water management to save water andincrease its productivity in irrigated rice. Agricultural Water Management 49,11–30.

Cabangon, R.J., Tuong, T.P., Abdullah, N.B., 2002. Comparing water input and waterproductivity of transplanted and direct-seeded rice production systems. Agri-cultural Water Management 57, 11–31.

Chahal, G.B.S., Sood, A., Jalota, S.K., Choudhury, B.U., Sharma, P.K., 2007. Yield, evapo-transpiration and water productivity of rice (Oryza sativa L.)–wheat (Triticumaestivum L.) system in Punjab-India as influenced by transplanting date of riceand weather parameters. Agricultural Water Management 88, 14–27.

Chakraborty, D., Nagarajan, S., Aggarwal, P., Gupta, V.K., Tomar, R.K., Garg, R.N.,Sahoo, R.N., Sarkar, A., Chopra, U.K., Sarma, K.S.S., Kalra, N., 2008. Effect ofmulching on soil and plant water status, and the growth and yield of wheat(Triticum aestivum L.) in a semi-arid environment. Agricultural Water Manage-ment 95, 1323–1334.

Chattaraj, S., Chakraborty, D., Garg, R.N., Singh, R., Singh, G.P., Sehgal, V.K., Sahoo, R.N.,Singh, S., Gupta, V.K., Chand, D., 2011. Evaluating the effect of irrigation on cropevapotranspiration in wheat (Triticum aestivum L.) by combining conventionaland remote sensing methods. Journal of Agricultural Physics 11, 35–52.

Choudhury, B.U., Singh, A.K., 2013. Bed planted rice–wheat rotation on soil prop-erties, root growth, nitrogen uptake and system productivity. Paddy and WaterEnvironment 11 (1–4), 265–275.

Choudhury, B.U., Bouman, B.A.M., Singh, A.K., 2007. Yield and water productivity of

rice–wheat on raised beds – results from a field experiment at New Delhi, India.Field Crops Research 100, 229–239.

Dhillon, B.S., Kataria, P., Dhillon, P.K., 2010. National food security vis-à-vis sus-tainability of agriculture in high crop productivity regions. Current Science 98(1), 1–4.

Page 12

l Wat

D

F

H

H

H

H

J

K

K

K

K

K

M

M

B.U. Choudhury et al. / Agricultura

oorenbos, J., Kassam, A.H., 1979. Yield Response to Water. Irrigation and Drain,Paper No. 33. FAO, Rome.

ynn, R.P., Al-shooshan, A., Short, T.H., Mc Mahon, R.W., 1993. Evapotranspirationmeasurement and modeling for a potted chrysanthemum crop. Transactions ofthe ASAE 3 (6), 1907–1913.

e, J., Li, H., McHugh, A.D., Ma, Z., Cao, X., Wang, Q., Zhang, X., Zhang, X., 2008. Springwheat performance and water use efficiency on permanent raised beds in aridnorthwest China. Australian Journal of Soil Research 46, 659–666.

obbs, P.R., Gupta, R.K., 2003. Rice–wheat cropping systems in the Indo-Gangeticplains: issues of water productivity in relation to new resource-conserving tech-nologies. In: Kijne, J.W., Barker, R., Molden, D. (Eds.), Water Productivity inAgriculture: Limits and Opportunities for Improvement. CABI, Wallingford, UK,pp. 239–253.

olland, J.E., White, R.E., Edis, R., 2007. The relation between soil structure andsolute transport under raised bed cropping and conventional cultivation insouth-western Victoria. Australian Journal of Soil Research 45, 577–585.

umphreys, E., Kukal, S.S., Kaur, A., Thaman, S., Yadav, S., Singh, Y., Singh, B., Timsina,J., Dhillon, S.S., Prashar, A., Smith, D.J., 2008. Permanent beds for rice–wheat inPunjab, India. Part 2. Water balance and soil water dynamics. In: Humphreys, E.,Roth, C.H. (Eds.), Permanent Beds and Rice-residue Management for Rice–wheatSystems in Indo-Gangetic Plains. Proceedings of International Workshop, PAULudhiana, India, 7–9 September 2006. ACIAR Proceedings No. 127.

alota, S.K., Sood, A., Chahal, G.B.S., Choudhury, B.U., 2006. Crop water productivityof cotton (Gossypium hirsutum L.)–wheat (Triticum aestivum L.) system as influ-enced by deficit irrigation, soil texture and precipitation. Agricultural WaterManagement 84, 137–146.

ar, G., Kumar, A., Martha, M., 2007. Water use efficiency and crop coefficients ofdry season oilseed crops. Agricultural Water Management 87, 73–82.

ingra, P.K., Hundal, S.S., Sharma, P.K., 2004. Characterization of crop coefficientsfor wheat and rice crops in Punjab. Journal of Agrometeorology 6, 58–60.

ukal, S.S., Humphreys, E., Singh, Y., Timsina, J., Thaman, S., 2005. Performance ofraised beds in rice–wheat systems of northwestern India. In: Roth, C.H., Fischer,R.A., Meisner, C.A. (Eds.), Proceedings of Evaluation and Performance of Perma-nent Raised Bed Systems in Asia, Australia and Mexico. Workshop on PermanentRaised Beds Held in Griffith. Australia, 1–3 March 2005, ACIAR Proceedings No.121.

ukal, S.S., Yadav, S., Humphreys, E., Kaur, A., Singh, Y., Thaman, S., Singh, B., Timsina,J., 2010. Factors affecting irrigation water savings in raised beds in rice andwheat. Field Crops Research 118, 43–50.

uo, S.F., Ho, S.S., Liu, C.W., 2006. Estimation irrigation water requirements withderived crop coefficients for upland and paddy crops in ChiaNan Irrigation Asso-ciation, Taiwan. Agricultural Water Management 82, 433–451.

idmore, D.J., Cartwright, J.B., Fischer, R.A., 1984. Wheat in tropical environmental.II. Crop growth and grain yield. Field Crops Research 8, 207–227.

orrison, J.E., Gerik, T.J., Chichester, F.W., Martin, J.R., Chandler, J.M., 1990. Ano-tillage farming system for clay soils. Journal of Production Agriculture 3,219–227.

er Management 123 (2013) 20– 31 31

Norwood, C.A., Dumler, T.J., 2002. Transition to dry land agriculture. Limitedirrigation vs. dry land corn. Agronomy Journal 94, 310–320.

Sandhu, B.S., Khera, K.L., Singh, B., 1982. A note on the use of irrigation water andyield of transplanted rice in relation to time of last irrigation. The Indian Journalof Agricultural Sciences 52, 870–872.

Sentelhas, P.C., Gillespie, T.J., Santos, E.A., 2010. Evaluation of FAOPenman–Monteith and alternative methods for estimating reference evapo-transpiration with missing data in Southern Ontario, Canada. AgriculturalWater Management 97, 635–644.

Sharma, P.K., Bhushan, L., Ladha, J.K., Naresh, R.K., Gupta, R.K., Balasubramanian,B.V., Bouman, B.A.M., 2002. Crop-water relations in rice wheat cropping underdifferent tillage systems and water-management practices in a marginally sodic,medium-textured soil. In: Bouman, B.A.M., Hengsdijk, H., Hardy, B., Bindraban,B., Toung, T.P., Ladha, J.K. (Eds.), Water-Wise Rice Production. International RiceResearch Institute, Los Banos, Philippines, pp. 223–235.

Singh, P., Kumar, R., 1993. Evapotranspiration from wheat under a semiarid climateand a shallow-water table. Agricultural Water Management 23, 91–108.

Singh, Y., Humphreys, E., Kukal, S.S., Singh, B., Kaur, A., Thaman, S., Prashar, A., Yadav,S., Timsina, J., Dhillon, S.S., Kaur, N., Smith, D.J., Gajri, P.R., 2009. Crop perfor-mance in permanent raised bed rice–wheat cropping system in Punjab, India.Field Crops Research 110, 1–20.

Singh, B., Gaydon, D.S., Humphreys, E., Eberbach, P.L., 2011. The effects of mulchand irrigation management on wheat in Punjab, India – evaluation of the APSIMmodel. Field Crops Research 124 (1), 1–13.

Tabbal, D.F., Bouman, B.A.M., Bhuiyan, S.I., Sibayan, E.B., Sattar, M.A., 2002. On-farm strategies for reducing water input in irrigated rice; case studies in thePhilippines. Agricultural Water Management 56, 93–112.

Tripathi, R.P., 2004. Evapotranspiration and crop coefficient for rice, wheat andpulses under shallow water table conditions of Tarai region of Uttaranchal.Journal of Agrometeorology 6 (1), 17–29.

Tuong, T.P., Bhuiyan, S.I., 1999. Increasing water-use efficiency in rice production:farm levels perspectives. Agricultural Water Management 40, 117–122.

Tuong, T.P., Bouman, B.A.M., Mortimer, M., 2005. More rice, less water – integratedapproaches for increasing water productivity in irrigated rice based systems inAsia. Plant Production Science 8, 231–241.

Tyagi, N.K., Sharma, D.K., Luthra, S.K., 2000a. Determination of evapotranspirationand crop coefficients of rice and sunflower with Lysimeters. Agricultural WaterManagement 45, 41–54.

Tyagi, N.K., Sharma, D.K., Luthra, S.K., 2000b. Evapotranspiration and cropcoefficients of wheat and sorghum. Journal of Irrigation and Drainage Engineer-ing: ASCE 126, 215–222.

Walker, S.H., Rushton, K.R., 1984. Verification of lateral percolation losses

from irrigated rice fields by a numerical model. Journal of Hydrology 71,335–351.

Yadav, S., Tao, Li., Humphreys, E., Gill, G., Kukal, S.S., 2011. Evaluation and applicationof ORYZA2000 for irrigation scheduling of puddled transplanted rice in NorthWest India. Field Crops Research 122, 104–117.