This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

Field Crops Research 149 (2013) 105–114

Contents lists available at SciVerse ScienceDirect

Field Crops Research

jou rn al hom epage: www.elsev ier .com/ locate / fc r

Combining permanent beds and residue retention with nitrogenfertilization improves crop yields and water productivity in irrigatedarid lands under cotton, wheat and maize

M. Devkotaa,∗, C. Martiusa, J.P.A. Lamersa, K.D. Sayreb, K.P. Devkotaa, R.K. Guptac,O. Egamberdievd, P.L.G. Vleka

a Center for Development Research (ZEF), Walter Flex-Str. 3, 53113 Bonn, Germanyb International Maize and Wheat Improvement Center (CIMMYT), Apdo. Postal 6-641 06600 Mexico, D.F., Mexicoc CIMMYT, NASC Complex, CG Block, Pusa, New Delhid ZEF/UNESCO, Khorezm Project, Urgench, Uzbekistan

a r t i c l e i n f o

Article history:Received 13 January 2012Received in revised form 15 April 2013Accepted 16 April 2013

Keywords:Conventional agricultureTillageCentral AsiaKhorezmUzbekistan

a b s t r a c t

Intensive soil tillage and mismanagement of irrigation water and fertilizers are increasing productioncosts, reducing soil fertility and crop water productivity and threatening the sustainability of crop pro-duction systems in the irrigated arid lands of Uzbekistan, Central Asia. Conservation agriculture (CA)practices combined with optimum nitrogen (N) management can counterbalance some of these adverseeffects. Most work has been done in rainfed areas so there is less information available for irrigated pro-duction systems. This study compared the effects of tillage, crop residue management and N rates onyield and water productivity for irrigated cotton, winter wheat and maize grown in a two-year rotationalsequence in Uzbekistan under CA practices vs. current conventional farmer practices. Permanently raisedbeds (PB) and conventional tillage (CT) were compared under two crop residue levels (retained residue –RR and removed residue – RH), and three N levels (zero, medium and high, with actual rates dependingon the crops) on a sandy loam to loam soil. Raw cotton yield, yield components and water productivitywere not affected by tillage methods. However, the following crops, wheat and maize, produced 12 and42% higher grain yields, respectively, under PB than under CT. Under PB, water productivity increasedin wheat by 27% and in maize by 84%, while 11% less water was applied during wheat and 23% duringmaize production, compared to CT. All three crops showed a considerable increment in yield and waterproductivity when N fertilizer levels were increased from zero to medium N application, and a relativelymuch lower increment when the N rate was doubled from medium to high N, for both tillage methods.In maize, the response to applied N was more pronounced with PB than with CT. Irrespective of tillagemethod, RR increased the grain yield of wheat by 5% compared to RH. In maize, RR in PB increased grainyields by 15% compared to RH. RR had no effect for CT. The positive effect of PB and RR on yield andwater productivity of wheat and maize and the lack of negative effects on cotton yield reflect that PBwith RR and proper N application may be viable alternatives to the present, unsustainable conventionalagriculture practices in these irrigated arid lands, assuming the patterns are confirmed in the long-run.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Irrigated agriculture provides about 40% of the global agri-cultural production from just 20% of the total cultivated land.However, FAO (2002) predicted that the irrigated areas will needto expand from 202 million ha (Mha) in 1999 to 242 Mha by 2030

∗ Corresponding author. Present address: International Maize and WheatImprovement Center (CIMMYT) International, P.O. Box 5168, Singh Durbar Plaza,Marga, Kathmandu, Nepal. Tel.: +977 1 4269564; fax: +977 1 4229804.

E-mail address: m.devkota@cgiar.org (M. Devkota).

to meet the food demands of the growing population in devel-oping countries. Irrigation water is particularly important in aridclimates, where potential evapotranspiration exceeds precipita-tion. Central Asia is over 80% arid and it is most important thatagricultural production is done in a sustainable manner for foodsecurity, employment, livelihoods and environmental protection.Cotton, wheat and maize are the major commercial and food cropsgrown in the five Central Asian countries, i.e., Kazakhstan, Kyrgyz-stan, Tajikistan, Turkmenistan, Uzbekistan (FAOSTAT, 2010), oftenas mono-cultures on large areas using conventional tillage andirrigation via flood and furrows. Conventional tillage (CT) involvesintensive land preparation with up to 4–5 machinery passes, and

0378-4290/$ – see front matter © 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.fcr.2013.04.012

Author's personal copy

106 M. Devkota et al. / Field Crops Research 149 (2013) 105–114

often needs deep tillage to reduce soil compaction caused by heavymachinery. In addition, poorly managed flood irrigation and dys-functional drains combined with the excessive use of chemicalinputs are typical of the conventional land use practices. The rec-ommended N rates for cotton or wheat is 160–180 kg N ha−1 and150 kg N ha−1 for short-duration maize (MAWR, 2000). Previousresearch has shown that these rates, determined decades ago,need to be adjusted (Kienzler, 2010). Previous findings confirmeda rather low nitrogen (N) use efficiency of 33% (Kienzler, 2010)which is far less than the 50–80% reported in well managed fieldexperiments for both low and upland crops (Cassman et al., 1993;Dobermann et al., 2000; Cassman et al., 2002). The wasteful use ofresources like irrigation water and chemical fertilizers is pollutingthe environment (Scheer et al., 2008), increasing production costs,raising the groundwater level which increases secondary soil salin-ization (Forkutsa et al., 2009), deteriorating soil quality, and thusthreatening the sustainability of the overall crop production systemin the irrigated arid lands of Central Asia.

Conservation agriculture (CA) principles that imply reducedtillage, proper crop rotation, and retention of optimal levels of cropresidues have been adopted by farmers on more than 100 Mhaworld wide as of 2008 (Derpsch and Friedrich, 2009). CA is pre-dominantly practiced in North and South America, Australia, andother semi-arid areas of the world (Holland, 2004) but less widelyintroduced in the world’s irrigated areas. Among the variousCA practices, reduced tillage is predominating worldwide (Yauet al., 2010). Reduced tillage technologies effectively minimize soildisturbance, reduce evaporation from soils, enhance soil carbonsequestration, and reduce energy needs, thus lowering productioncosts (Lal et al., 2007).

The use of permanently raised beds (PBs), a reduced tillagepractice, is gaining adoption in different crop production systems.Previous studies have shown that PB with residue retention hasan advantage over zero-till flat (ZT) and CT (Limon-Ortega et al.,2000; Sayre and Hobbs, 2004; Tursunov, 2009). The benefits includehigher irrigation water use efficiency (Sayre and Hobbs, 2004;Hassan et al., 2005), better plant establishment (Khaleque et al.,2008; Gürsoy et al., 2010), lower production costs (Gupta et al.,2009; Tursunov, 2009; Hari-Ram et al., 2012), and with equal orhigher crop yield (Sayre and Hobbs, 2004; Govaerts et al., 2005;Boulal et al., 2012). Crop production on PB is 9% more energy-efficient than under ZT, 12% more efficient compared to freshlyprepared beds and even 19% more efficient than conventional prac-tices (Rautaray, 2005), while the cultivation costs can be reduced byone-third with PB in Central Asia (Gupta et al., 2009). Furthermore,PBs also increase grain yield and water use efficiency in irrigatedcotton (Boulal et al., 2012), wheat and maize (Harris and Krishna,1989; McFarland et al., 1991; Sayre and Hobbs, 2004; Wang et al.,2004; Hassan et al., 2005; Gupta et al., 2009).

A major challenge for sustainable agriculture production on irri-gated croplands in Central Asia is to increase water productivitywhich has been falling for decades due to high water applicationrate (Abdullaev and Molden, 2004). The average water productivityof cotton in the Syr Darya basin is ∼0.37 kg m−3, much below theworld average 0.60 kg m−3 (Abdullaev and Molden, 2004). Severalstudies showed the potential of PBs to increase crop water produc-tivity by 25–40% in irrigated agriculture (Sayre and Hobbs, 2004;Hassan et al., 2005; Akbar et al., 2007).

Reduce tillage is especially effective when combined with asurface mulch of crop residues (Lal et al., 2007) which helps toreduced evaporative water loss and increase the soil’s water reten-tion capacity (Gant et al., 1992). However, N response in reducedtillage with and without residue retention may differ. For example,in some cases, at the onset of the conversion to conservation tillage,N fertilizer rates need to be increased by about 25% to counteract theyield-reducing effect of short-term N immobilization (Randall and

Bandel, 1991; Wienhold et al., 1999). However, in some cases thisincrease is not necessary; as the soil mineral N distribution in thetop soil is altered (Franzluebbers et al., 1995; Angas et al., 2006) andthe availability of N improves in a few years (Rice et al., 1986). Simi-larly, Torbert and Reeves (1994) also concluded that in the long-runN applications can be reduced under reduced tillage practices dueto an increase in N uptake efficiencies caused by reduced soil traffic.

The advantages of conservation over conventional practiceshave been repeatedly shown for rainfed conditions (Lopez-Bellidoet al., 1996; DeVita et al., 2007), and also more recently under irri-gated conditions (Gupta et al., 2009; Pulatov et al., 2011; Verhulstet al., 2011a; Boulal et al., 2012; Naresh et al., 2012). In the irrigateddrylands of Central Asia, the effects of N application, under CA prac-tices on crop performance, production and water productivity ofmajor crops such as cotton, wheat and maize are still poorly under-stood (Gupta et al., 2009; Pulatov et al., 2011). Based on earlier work,we theorize that combining permanent beds with residue retentionand proper nitrogen application can improve crop yield and waterproductivity relative to the conventional agriculture practices inirrigated arid-lands. Thus, yield and water productivity of cotton,wheat and maize grown in sequence under CA practice were com-pared to conventional practices with various amounts of N fertilizerapplied.

2. Materials and methods

2.1. Site description

In October 2007, a long-term experiment on a cotton–wheat-third crop rotation was started in western Uzbekistan (41◦32′12′′

N, 60◦40′44′′ E, and 100 m a.s.l.). The experimental field of 3.1 ha hadbeen mono-cropped for 20 years with cotton under heavy mech-anization. The field had received annual fertilizer applications inthe order of 200:140:100 kg NPK ha−1. The studied rotational cyclelasted from May 2008 to October 2009. Field preparations consistedof deep ploughing, laser-guided land leveling and salt leachingin October/November 2007 and March 2008. Cotton was sown inMay 2008 and harvested in October 2008. The experiment on theimpact of tillage, N and residue level was initiated after this, in thesequence of winter wheat (October 2008–June 2009) and maize(June–October, 2009).

The soil in the experimental area is an irrigated alluvial meadow,with sandy loam to loam texture, low in organic matter (0.3–0.6%)and was saline (2–12 dS m−1). The groundwater table in the areais generally shallow (0.5–2 m). The climate is arid, with long, hotand dry summers and short, very cold winters. Average precipita-tion is less than 100 mm year−1 and potential evaporation alwaysgreatly exceeds total precipitation. During the experiment cotton,wheat and maize received 14.6 mm, 72.6 mm, and 35.4 mm of rain-fall, respectively. In the wheat season 71% of the total rain fallsduring winter (November–March). In the maize season all rainfallwas concentrated in mid September, i.e., during crop maturity.

2.2. Experimental design, treatments and crop management

The experiment was conducted as a randomized complete blockwith four replications involving two tillage treatments as mainplots (permanent bed (PB) and conventional tillage (CT)) and afactorial combination of two residue levels (previous crop residueretained: RR or residue harvested/removed: RH) and three N lev-els (0, medium and high) as subplots. The medium N level was125 kg N ha−1 for cotton and 100 kg N ha−1 for wheat and maize,and the high N level was 250 kg N ha−1 for cotton and 200 kg N ha−1

for wheat and maize. The experimental subplots were 550 m2

(11 m × 50 m).

Author's personal copy

M. Devkota et al. / Field Crops Research 149 (2013) 105–114 107

Table 1Amount of irrigation water applied (m3 ha−1) during cotton, wheat and maize growing seasons (2008–2009).

Cotton Wheat Maize

Applicationtime (DASa)

Amount applied forPB and CT (m3 ha−1)

Applicationtime (DAS)

Amount applied (m3 ha−1) Application time (DAS) Amount applied (DAS)

PBb CTc PB CT

54 537 1 697 803 1 1555 236069 556 150 750 750 18 1181 143384 661 173 699 806 33 1213 1499101 1350 190 582 670 43 1218 1427115 1303 208 704 812 70 1118 1427– – 222 689 794 – – –– – 236 649 748 – – –

Total 4454 4770 5383 6285 8146

a Days after sowing.b Permanent bed.c Conventional tillage.

In the first crop cycle tilled, fresh beds were prepared with90-cm spacing between furrows for the PB treatment using a bedmaker. The beds were 15 cm high and 60 cm wide at the top; fur-rows were 15 cm wide. From this point on, this treatment was nottilled again – the beds were only reshaped as needed in the furrowsfor all succeeding crops. Cotton (Gossypium hirsutum L., cv. Khorezm127) was seeded using the cotton planter at the recommended seedrate of 60 kg ha−1 in the center of these beds in early May 2008, pro-viding an average plant density of 45,000 plants ha−1. In CT, cottonwas sown on tilled flat land using the cotton planter with the samespacing and seed rate. A defoliant (9 kg magnesium chloride ha−1

dissolved in 200 l water) was applied to induce boll opening 10–15days before the first cotton picking. Since no previous crop residueswere available at the onset of this experiment, wheat stover wasplaced in the RR plots at a rate of 3 t ha−1 in cotton. In CT, residueswere equally distributed over the plot surface and incorporated into the soil before cotton seeding, while in PB residues were surfacemulched in both beds and furrows immediately after seeding. Win-ter wheat (Triticum aestivum L., cv. Krasnodar) was relay seeded intothe standing cotton at the recommended seed rate of 200 kg ha−1

on October 10, 2008 just after the 2nd cotton picking, and was har-vested on June, 16, 2009. In PB, wheat was seeded in rows at adistance of 22.5 cm (4 rows on each 90-cm bed) with double diskseed openers into the standing cotton. Due to standing cotton oncenter of the beds, wheat rows were mostly placed on the shoul-ders and slope of the furrow rather than the top of the beds. In CT,seeds were broadcasted manually into standing cotton after a sin-gle cultivation. A second cultivation followed to cover the seeds,as is commonly practiced by local farmers. Average plant densitywas 400 plants m−2 in both tillage methods. After the final cottonpicking in the 1st week of November 2008, cotton stalk of about 5,7 and 7.8 t ha−1 in N-0, N-100 and N-200 plots, respectively, wereretained on the surface of the RR treatments in both tillage sys-tems. The cotton stalks were chopped (about 12–15 cm length) andequally distributed over the plot surface in both tillage systems. InPB residues were applied in both beds and furrows. In the RH treat-ments all cotton stalks were cut at ground level and removed fromthe plots.

After wheat harvest, the loose stover was uniformly spread overthe RR plots but both standing and loose stover was removed fromthe RH plots and hybrid maize (Zea mays L., cv. Maldoshki) wassown with a double disk seed opener as a summer crop at a seed rateof 40 kg ha−1 with 45 cm row spacing (2 rows in each bed) on June28, 2009, in both PB and CT, and harvested as grain in September2009. The plants were thinned 10 days after emergence and main-tained average plant density of 50,000 plants ha−1. In CT, maize wassown after three cultivations followed by rough leveling, whereasunder PB, no soil tillage occurred aside from the drilling in bandsof seed and N fertilizer.

Phosphorus (P) and potash (K) at 160 and 70 kg ha−1 for wheatand maize and 140 and 100 kg ha−1 for cotton, respectively, wereapplied as basal applications during sowing. The P fertilizer wasapplied as single super phosphate (10% P2O5) in the N-0 treatments,while ammonium phosphate (11% N and 46% P2O5) was appliedin all other N treatments. A basic dressing of muriate of potash(60% K2O) was applied in all plots. The remaining N fertilizer wastop dressed through urea granules (46% N) with a band applicationin cotton (at budding (38 days after sowing, DAS) and floweringstages (52 DAS)) and in maize (at 32 and 42 DAS); while urea wasbroadcasted in two equal splits at 172 DAS (F6 stage) and 190 DAS(F8) in wheat.

2.3. Measurements

2.3.1. Irrigation waterAll three crops were irrigated when needed which was indicated

by first leaf roll. Irrigation water was applied seven times duringwheat and five times during cotton and maize growth (Table 1).During the cotton season, both PB and CT received equal amountsof irrigation water, as furrow-irrigation was applied in both tillagemethods. In case of wheat and maize, flood irrigation was appliedunder CT, while in PB furrow irrigation was applied. Under floodirrigation more time was required to fill the field compared tothe furrow irrigation. Based on the irrigation needs, wheat andmaize on PB received, respectively, 11 and 23% less water thanunder CT (Table 1). Irrigation water was measured using a standardtrapezoidal Cipolletti weir combined with a DL/N 70 diver, whichmeasured water flow through the weir in one-minute intervals.With two separate weirs (with divers installed 40 cm in front ofthe weir crest) the amount of water applied was measured in thePB and CT plots. The irrigation time during each irrigation eventwas recorded independently. The height of water above crest widthwas measured 4–5 times manually during each irrigation event. Thepressure measured by the diver was transformed to height abovecrest (m) to estimate the discharge (m3 s−1) as (Kraatz and Mahajan,1975):

Q = 1.86LH3/2

where Q is the discharge (m3 s−1), L is the crest width (m), H is theheight of water above the crest width (m).

2.3.2. Yield and yield componentsA 15 m × 3.6 m area was used to measure cotton yield in each

subplot, which covered the four central rows of each plot. Thedelineated area was harvested manually during the four pickings(142, 154, 171, and 179 DAS) which is the common practice in theregion. The raw cotton yield was adjusted to 6% moisture level;

Author's personal copy

108 M. Devkota et al. / Field Crops Research 149 (2013) 105–114

the moisture level was determined with sub-samples taken at eachharvest after oven drying at 70 ◦C for 16 h. Boll density (bolls perplant) and percentage of open bolls were determined prior to defo-liation from 40 randomly selected plants (10 plants each row) inthe harvest area. Forty randomly picked bolls were oven-dried tocalculate average boll weight at each pick. In addition, five repre-sentative plants were sampled from each plot and cut at groundlevel, and plant parts were separated and oven dried and weighedto determine the total aboveground biomass and harvest index. Theginning percentage was calculated by separating the lint and seedfrom 200 g oven-dried raw cotton and weighed separately for eachpicking.

Wheat was harvested from three quadrate randomly installedin each plot, each covering 1.8 m2 (2 beds with 1 m length). Theplants were cut at ground level. The spikes were threshed toseparate grains and straw, which were weighed. Sub-samplesof 150 g each of straw and grain were oven-dried at 70 ◦Cfor 36 h to determine moisture content. To determine num-ber of grains per spike and spike weight, 50 spikes weretaken randomly from each plot, oven-dried and weighed afterwhich the grains were separated from the spikes, weighed andcounted.

Three quadrates of 3.6 m2 (2 beds with 2 m length) eachwere used to measure maize yield in each plot. The plants fromeach quadrate were cut to the ground level. Cobs were sep-arated from stover, shelled, and the grain was weighed after2–3 days of sun-drying; then a sub-sample of 200 g was takento determine moisture content. Stover was separated into tas-sel, leaf and stem, and each fraction was weighed separatelyfresh on-site; a sub-sample of each was taken and oven driedto determine the moisture content. In addition, six cobs fromeach quadrate were randomly selected to record the numberof grains, cob weight, and 1000-grain weight. Grain yields ofwheat and maize were converted to 12% moisture content. Theharvest index (HI) for each crop was calculated as the frac-tion of seed/raw cotton yield over total aboveground dry matteryield.

2.3.3. Water productivityWater productivity was calculated as the ratio of grain yield for

wheat and maize or raw cotton yield (kg ha−1) for cotton over totalwater input (m3).

2.4. Statistical analysis

Crop parameters, i.e., yield, yield components and water pro-ductivity for cotton, wheat and maize, were statistically analyzedas a split-plot factorial design with four replications. As noresidue effect was observed for cotton growth and yield (datanot shown), only the data from the residue-harvested treatmentswere used for the statistical analysis of this crop. Treatmenteffects were compared through the analysis of variance usingGenStat Discovery Edition 3. Main and interaction effects werecompared using Fisher’s protected LSD (least significant difference;p = 0.05).

3. Results

3.1. Cotton

3.1.1. Raw cotton yield, biomass and yield componentsAcross tillage and N level, average raw cotton yield was

3.9 t ha−1. Tillage method had no significant effect (p = 0.8) on rawcotton yield and aboveground biomass (AGB) production. How-ever, N application caused a significant (p ≤ 0.001) increase in yieldand AGB in both tillage methods. With N fertilizer, up to 4.3 t ha−1

raw cotton yield could be achieved (Table 2). N-125 showed a 26%higher raw cotton yield than N-0. Raw cotton yield and AGB didnot increase significantly when increasing the N rate from N-125to N-250 (Table 2).

Tillage method did not affect HI and ginning percent ofcotton, but both slightly decreased (p < 0.05) with increasedN level (Table 2). Boll density and boll weight were notaffected by tillage methods. With N-125 boll density increasedby 20% (p < 0.05) and boll weight by 6% compared to N-0. In contrast, when doubling fertilizer rate from N-125 toN-250 boll density and boll weight were not significantlyincreased.

3.1.2. Water productivityAverage water productivity of cotton was 0.88 kg m−3 irrespec-

tive of the tillage method. A similar trend as for yield was observedfor water productivity, i.e., an increase of WP with N applicationsin both tillage methods (Table 2).

Table 2Main effect of tillage and N level on raw cotton yield, aboveground biomass (AGB), harvest index (HI), yield components, ginning percent, and water productivity (WP) ofcotton.

Treatment Cotton yield (kg ha−1) AGB (kg ha−1) HI (%) Boll density (m−2) Boll weight (g) Ginning percent (%) WP (kg m−3)

TillagePBa 3929 10,723 37.4 69 6.3 39.2 0.88CTb 3916 10,221 38.4 68.4 6.2 39.9 0.88LSD (0.05) – – – – – – –

N levels0 3299 8342 39.8 61.6 5.9 40.6 0.74125 4168 11,088 37.7 73.8 6.3 39.8 0.94250 4301 11,986 36.2 70.8 6.4 38.2 0.97LSD (0.05) 257 1198 2.8 9 0.3 1.5 0.05

ANOVATillage (T) ns ns ns ns ns ns nsNitrogen (N) *** *** * * * * ***T × N nsc ns ns ns ns ns ns

Mean 3923 10,472 37.9 68.7 6.2 39.5 0.88

a Permanent bed.b Conventional tillage.c Non significant.* Significant at p ≤ 0.05.

*** Significant at p ≤ 0.001.

Author's personal copy

M. Devkota et al. / Field Crops Research 149 (2013) 105–114 109

3.2. Wheat

3.2.1. Grain yield and biomassTillage had a significant effect (p = 0.03) on grain yield of wheat.

Across the three N and the two residue levels the PB system yielded12% more grain than CT (Table 3). The main effect of N on grain yieldwas highly significant (p ≤ 0.001) peaking at 10.8 t ha−1 grain in PB(Table 3). Wheat yield was increased (p < 0.05) by three times withN-100 compared to N-0. However, when doubling the N level fromN-100 to N-200 kg ha−1, grain yield increased (p < 0.05) by only18%. Irrespective of tillage method and N level, the RR treatmentshowed a 5% higher (p = 0.05) grain yield than for RH. No significantinteraction effect between tillage and N levels was detected.

Irrespective of N and residue levels, total AGB under PB was14% higher (p < 0.05) than for CT (Table 3). Averaged over tillageand residue level, AGB was increased (p < 0.05) by 160% with N-100compared to N-0, but was only increased by another 18% when fer-tilizer rates were doubled from N-100 to N-200 kg ha−1 (p < 0.05).HI was affected (p < 0.05) by N level but not by tillage methodand residue level (Table 3). In this regard, HI in wheat increased(p < 0.05) by 5% with N-100 compared to N-0, while no differencewas observed between N-100 and N-200 (Table 3).

3.2.2. Yield componentsSpike density (spikes m−2) was significantly (p < 0.05) affected

by tillage method and N level (Table 3). Averaged over N and residuelevels, spike density under PB was 10% higher than in CT. Spike den-sity increased (p ≤ 0.001) with increasing N level, and was higher by85% with N-100 than with N-0. With N-100 to N-200, spike densityincreased (p < 0.05) by 18%. In PB, RR showed a 6% higher (p > 0.05)spike density than RH (data not shown).

The number of grains per spike was significantly affected bymain and interaction effects of N and residue levels, but the effect oftillage was non-significant (Table 3). Averaged over tillage method,RR in the N-0 plots showed 15% more (p < 0.05) grains per spikethan under RH, but RR showed no increase in grains per spike inthe N applied treatments in both tillage systems (data not given).Thousand kernel weight (TKW) of wheat was not affected (p > 0.05)by tillage method (Table 3). Irrespective of tillage method, TKWincreased by 1.1 and 1.3 g with N-100 and N-200, respectively,compared to N-0, but there was no difference between N-100and N-200. A significant (p < 0.05) interaction effect was observedbetween tillage and residue level (Table 3). Irrespective of tillagemethods, RR significantly increased the TKW by 0.8 g compared tothe RH.

3.2.3. Water productivity (WP)Effects of tillage and N level were highly significant (p ≤ 0.001)

for WP of wheat. Averaged over N and residue levels, WP under PBwas 27% higher than in CT. A similar trend as for yield was observedfor WP, i.e., an increase of WP with N applications in both tillagemethods, and highest WP of 2.2 kg m−3 was observed with N-200(Table 3).

3.3. Maize

3.3.1. Grain yield and biomassThe main effects of tillage, N and residue levels were signifi-

cant for maize grain yield, AGB and water productivity (Table 4).Averaged over N and residue level, grain yield under PB was 41%higher than in CT. There was a significant interaction effect betweentillage and N level, such that N response was greater under PB thanin CT (Fig. 1A). Grain yield in PB was 2.5 times higher with N-100than with N-0. However, with the increase from N-100 to N-200,grain yield in PB increased by only 32%. In CT, grain yield was 2times higher (p < 0.05) with N-100 than with N-0, while it increased

(p > 0.05) by only 10% when going from N-100 to N-200. Responseto applied N was higher in PB than in CT, indicating a 34 and 61%higher (p < 0.05) grain yield with N-100 and N-200, respectively,compared to CT at respective N level (Fig. 1A). Tillage had no sig-nificant effect on grain yield at N-0. Averaged over N level, RR inPB increased (p < 0.05) grain yield by 15% compared to RH. In con-trast, RR had no significant effect (p > 0.05) on grain yield under CT(Fig. 1B).

Nitrogen level and tillage method had highly significant(p ≤ 0.001) main and interaction effects for total AGB at maturity(Table 4). Similar to the grain yield, AGB production in PB with N-100 and N-200 increased (p < 0.05) by 40% and 67%, respectively,compared to CT at the same N levels, while tillage had no signif-icant effect with N-0 application (Table 5). Irrespective of the Nlevel, RR in PB increased (p < 0.05) AGB production by 17% (10.3 vs.8.8 t ha−1) compared to RR, while RR under CT had no significanteffect on AGB production at maturity (data not shown). A similartrend as that for yield was observed for AGB production.

Irrespective of residue and N levels, HI under CT was higher(p = 0.08) by 4% than in PB. Similarly, averaged over tillage method,HI was higher (p < 0.05) by 6% with N-100 and by 5% with N-200than with N-0 (Table 4).

Nitrogen level and tillage method had a significant effect oncrop growth duration of maize (Table 4). Tasseling was delayedby 6–7 days under PB and by 2–3 days in CT with N-0 as comparedto N-100 and N-200. Similarly, maturity was delayed by 2–3 dayswith N-0 compared to N-100 and N-200 under PB, but the crop wasmatured at same time in CT for all N levels (Table 5). Crop residuedid not affect growth duration of maize in either tillage treatments(Table 4).

3.3.2. Yield componentsThere was a significant tillage by N interaction for yield compo-

nents (cob density, grains cob−1, and TKW) of maize (Table 4). InPB, cob density was higher (p < 0.05) by 21% with N-100 than N-0,and for N-100 to N-200 cob density (m−2) increased (p > 0.05) by15%. In contrast, under CT, N application had no significant effecton cob density. Number of grains per cob and TKW were higher(p < 0.05) with increased N level in both tillage systems. Grains percob increased (p < 0.05) by 76% under PB and by 33% in CT withN-100 compared to N-0, and doubling the N level from N-125 toN-250 led to an increase (p < 0.05) of 8% in PB and 7% in CT. In PB,grains per cob were higher (p < 0.05) by 14% with N-100 and 16%with N-200 than under CT at the respective N level (Table 5). TKWwas affected by tillage treatments at N-0, but not at the other Nlevels; being 13% greater under CT than PB at N-0 (p < 0.05). TKWincreased (p < 0.05) by 21% in PB and 8% under CT with N-100 com-pared to N-0, and doubling the N level from N-125 to N-250 led toan increase (p < 0.05) of 9% in PB and 8% in CT.

3.3.3. Water productivityThere was a highly significant (p = 0.001) interaction effect

between tillage and N, and a significant (p = 0.02) interaction effectbetween tillage and residue level on WP in maize (Table 4). Aver-aged over N and residue level, WP in PB was almost twice as high(85% higher) as compared to CT (Table 4). Water productivity underPB was higher (p < 0.05) for all N levels, i.e., by 44% with N-0, 73%with N-100 and 110% with N-200 than in CT at the respective Nlevel (Fig. 2A). In PB, WP increased (p < 0.05) by 148% with N-100compared to N-0. With N-100 to N-200, WP under PB increased(p < 0.05) by 32%. Similarly, under CT, WP increased by 107% withN-100 compared to N-0, while no difference was observed betweenN-100 and N-200 (Fig. 2A). Irrespective of N level, WP under PB withRR was 15% higher than with RH. In contrast, RR had no significanteffect on WP under CT (Fig. 2B).

Author's personal copy

110 M. Devkota et al. / Field Crops Research 149 (2013) 105–114

Table 3Main effect of tillage, N level, and residue on grain yield (12% moisture), aboveground biomass (AGB), harvest index (HI), water productivity (WP), spike density (spikes m−2),grains spike−1, and 1000 kernel weight (TKW) of winter wheat.

Treatment Grain yield (kg ha−1) AGB (kg ha−1) HI (%) WP (kg m−3) Spike density (m−2) Grains spike−1 TKW (g)

Tillage (T)PBa 8269 15,488 46.8 1.7 685 30.7 34.4CTb 7345 13,517 47.6 1.4 620 30.1 34.9LSD (0.05) 800.9 1540 – 0.20 44 – –

Residue (R)RHc 7615 14,128 47.3 1.5 646 29.6 34.3RRd 7998 14,877 47.1 1.6 659 31.3 35.1LSD (0.05) 384 – – 0.07 – 1.14 0.76

Nitrogen (N)0 3367 6516 45.8 0.67 389 25.7 33.9100 9204 16,948 48.0 1.83 718 33.1 34.9200 10,849 20,043 47.8 2.15 851 32.5 35.2LSD (0.05) 471 949 1.5 0.09 45.3 1.39 0.94

ANOVAT * * ns ** * ns nsR * ns nse * ns ** *N *** *** ** *** *** *** *N × T ns ns ns ns ns ns nsN × R ns ns ns ns ns * nsT × R ns ns ns ns ns ns nsN × T × R ns ns ns ns ns ns ns

Mean 7807 14,503 47 1.5 685 30.7 34.4

a Permanent bed.b Conventional tillage.c Residue harvested.d Residue retained.e Non significant.* Significant at p ≤ 0.05.

** Significant at p ≤ 0.01.*** Significant at p ≤ 0.001.

Table 4Main effect of tillage, N level, and residue on grain yield, aboveground biomass (AGB), harvest index (HI), water productivity (WP), yield components and maturity days ofhybrid maize.

Treatment Grain yield (kg ha−1) AGB (kg ha−1) HI (%) WP (kg m−3) Cob density (m−2) Grains cob−1 TKWa (g) Tasseling (days) Maturity (days)

Tillage (T)PBb 5520 9551 50.5 0.88 6.2 375 200.4 38 86CTc 3910 6510 52.4 0.48 5.0 343 206.7 40 88LSD (0.05) 922 1636 – 0.12 1.0 20.7 3.7 0.65 0.9

Residue (R)RHd 4490 7598 51.8 0.64 5.5 340 201.3 39 87RRe 4940 8464 51.2 0.72 5.7 378 205.7 38 87LSD (0.05) 317 548 – 0.04 – 12.4 – 0.7 –

N levels0 2331 4161 49.7 0.33 4.9 264 180.7 42 88100 5295 8893 52.5 0.76 5.6 392 206.3 37 86200 6519 11,038 52.3 0.95 6.2 421 223.6 37 87LSD (0.05) 388 671 1.2 0.05 0.4 15.2 8.9 0.8 0.7

ANOVAT * ** nsf ** * * * ** **R ** ** ns ** ns *** ns * nsN *** *** *** *** *** *** *** *** **N × T *** *** ns *** ** *** * *** **N × R ns ns ns ns ns ** ns ns nsT × R * * ns * ns ns ns * nsN × T × R ns ns ns ns ns ns ns ns ns

Mean 4715 8031 51.5 0.68 5.6 359 203.5 8.3 87

a Thousand kernel weight.b Permanent bed.c Conventional tillage.d Residue harvested.e Residue retained.f Non significant.* Significant at p ≤ 0.05.

** Significant at p ≤ 0.01.*** Significant at p ≤ 0.001.

Author's personal copy

M. Devkota et al. / Field Crops Research 149 (2013) 105–114 111

N level (kg ha-1 )

N0 N100 N200

Gra

in y

ield

(kg

ha

-1)

0

2000

4000

6000

8000

10000PB CT(A)

LSD(0.05)

Tillage

PB CT

0

2000

4000

6000

8000

10000

RH RR(B)

LSD(0.05)

Fig. 1. Interaction between tillage and nitrogen level (A) and tillage and residue level (B) on maize grain yield (kg ha−1). PB = permanent bed, CT = conventional tillage,RH = residue harvested and RR = residue retained. LSD is the least significant difference between tillage and N (A) and N and residue level (B). Bars represent standard error.

Table 5Interaction between N and tillage on yield components and maturity days of hybrid maize.

Tillage N levels (kg ha−1) AGBa (kg ha−1) Cob density (m−2) Grains cob−1 TKWb (g) Tasseling (days) Maturity (days)

PBc 0 4469 5.1 254 170 42 87100 10,373 6.2 418 206.4 35 85200 13,810 7.1 453 224.7 36 85

CTd 0 3853 4.7 274 191.4 42 88100 7412 4.9 366 206.1 39 88200 8267 5.2 390 222.4 39 88

LSD (0.05) 1513 0.98 22.7 10.5 1.12 1.1

a Aboveground biomass.b Thousand kernel weight.c Permanent bed.d Conventional tillage.

4. Discussion

4.1. Effect of tillage

Permanent raised bed (PB) planting has the potential to increasecrop yield and water productivity for many row-grown crops ingravity irrigated systems (Sayre and Hobbs, 2004; Hassan et al.,2005). However, in this experiment the raw cotton yield for bothtillage treatments was similar. This was to be expected as the firstcotton crop was sown following extensive tillage operations thatwere part of the initial land preparation. Once the beds in the PBtreatment were stabilized, by the time of planting the wheat crop,yield benefits for wheat and subsequent maize under PB were 12%

higher for wheat and 41% higher for maize than with CT, whenaveraged over the residue and N levels. Although the results arebased on the one season’s data, similar results have been observedin research conducted under comparable climatic conditions overmore than one season. For example, Hulugalle et al. (2004), Heet al. (2008), Naresh et al. (2012), Gürsoy et al. (2010), Singh et al.(2009) and Boulal et al. (2012) also observed equal or increasedyield of cotton, wheat and maize with PB compared with CT underirrigated arid/semi-arid regions. However, Kienzler et al. (2012)reported crop yields under conservation tillage practices in theirrigated systems were inconsistent and, in contrast to our results,Singh et al. (2009) reported difficulty in maintaining wheat yieldson beds relative to CT on a sandy loam soil. This was due to poor

N level (kg ha-1)

N0 N100 N200

Wate

r pro

ductivity (

kg m

-3)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4PB CT(A)

LSD(0.05)

Tillage

PB CT

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

RH RR(B)

LSD(0.05)

Fig. 2. Interaction between tillage and nitrogen level (A), and tillage and residue level (B) on water productivity (kg m−3) in maize. PB = permanent bed, CT = conventionaltillage, RH = residue harvested, RR = residue retained. LSD is the least significant difference between tillage and N (A); and N and residue level (B). Bars represent standarderror.

Author's personal copy

112 M. Devkota et al. / Field Crops Research 149 (2013) 105–114

establishment because of sub-optimal moisture at the time ofsowing and rapid drying on the top of beds as a result low spikedensity in beds compared to the CT, while yields were similarbetween CT and PB in loam soil.

The analysis of yield components indicated that the increasedwheat grain yield under PB was the result of an increase in thenumber of spikes m−2. As the wheat was sown in rows on the shoul-ders rather than the top of the beds, it was closer to the furrow,and this translates into better soil moisture and, in turn, nutrientuse efficiency in PB than in CT. The higher soil moisture contentdown to 90 cm soil depth under PB compared to CT throughoutthe crop growing period also suggests this (Devkota, 2011). Thismay also have increased nutrient availability under PB and withthis, the number of tiller-bearing plants (143 in PB and 130 in CT,10% higher). Similarly, as wheat was drill seeded in line under PBcompared to broadcast seeding in CT, both soil seed contact andplant uniformity were good (average plant at tillering stage was452 in PB and 410 in CT). Good contact between soil and seed inPB may have produced a more vigorous crop and greater nutrientuptake, thereby contributing to increased grain yield compared toCT (Mann et al., 2008; Gürsoy et al., 2010).

The different effects combined thus explain the yield advan-tage of PB at the different N levels compared to CT. Likewise, theeffect observed by Khaleque et al. (2008) was a higher spike den-sity in PB spring wheat compared to CT in Bangladesh. However,Singh et al. (2009) reported low spike density in beds comparedto CT in a rice–wheat system in India. In maize, no grain yield dif-ferences were found between PB and CT with N-0 (averaged overresidue level). PB maize was thus superior to CT only with N fer-tilization, although we cannot rule out that this is still an effect ofthe recent conversion of this field to CA. The increased grain yieldsof maize with N application in PB than in CT was also caused by acombination of factors including an advanced seedling emergenceand stand establishment, faster growth rate, earlier tasseling, andlonger grain filling periods (Devkota, 2011). All these factors con-tributed to an increased cob density and grains cob−1 in PB, whichare prime yield-determining components in maize (Fischer et al.,2002). The same authors also reported that maize performed bet-ter under conservation tillage with N application due to earlierseedling emergence and faster growth in the first few weeks com-pared to CT in rainfed conditions. Improved early growth resultedin increased AGB (Devkota, 2011), which was possibly due to anincreased N uptake (210 kg ha−1 in PB and 134 kg ha−1 in CT with N-200; 136 kg ha−1 in PB and 100 kg ha−1 in CT with N-100 at harvest;Devkota, 2011). Under furrow irrigation, PB-planting of maize couldhave improved microclimate, soil moisture, and nutrient availabil-ity compared to the conventionally tilled, flood-irrigated system,as previously postulated (Sayre and Hobbs, 2004; Govaerts et al.,2005; Zhang et al., 2012). However, inferior crop performance inpermanent beds due to very hard and dense soil (Kukal et al., 2005)and also rapid drying of the soil on the beds under sandy loam soil(Kukal et al., 2008) compared to the conventional tillage has beenreported under rice–wheat system in India.

Hobbs et al. (2008), Sayre and Hobbs (2004), and Govaerts et al.(2005) reported PB to be superior or comparable to rainfed maizeand wheat cultivation. The present short-term results show thesuperiority of PB also for irrigated arid lands. Similarly, highercrop yield were also observed in PB than in CT under an irrigatedmaize–cotton rotation in the Mediterranean region of Spain (Boulalet al., 2012), under maize–wheat rotation in the semi-arid zone ofPakistan (Hassan et al., 2005), and under cotton–wheat rotation inthe semi-arid zone of Australia (Hulugalle et al., 2004; Hulugalleand Scott, 2008).

The differences in cotton water productivity (WP) were insignif-icant. An equal amount of water was applied as furrow irrigationin both PB and CT (Table 1), and cotton yield differences were not

significant (Table 2). However, a higher WP was observed in PBcompared to CT in the subsequent wheat and maize crops. Theincreased WP was mainly due to higher crop yield and the shorterirrigation time required for wetting up the furrows in PB comparedwith flood irrigation in CT, where water saving in PB was 11% inwheat and 23% in maize season compared to CT (Table 1). Hassanet al. (2005) reported 36 and 32% irrigation water saving and 50 and65% higher water productivity in wheat and maize, respectively,when comparing PB to CT in a semi-arid region of Pakistan, simi-lar to our study region. Reduced amounts of irrigation water wereattributed by Sayre and Hobbs (2004) to a compaction in the fur-row bottoms from machine traffic which in turn enhanced lateralwater infiltration and forward water advance in the furrow. Thismost certainly has also occurred in our experiment. Although wedid not specifically monitored this, observations indicated greatersoil cracking in CT during the maize season. This could also explainthe need for the higher irrigation water to the CT maize.

4.2. Effect of nitrogen

Our results (Table 2) are a good indication that cotton on freshlyprepared beds can be grown under the same N level as under CTduring the initial year of CA without compromising yields. It is notknown, at this early stage of the experiment, whether the lack ofN effect on cotton yield between PB and CT would occur in thelonger term. However, this contrasts with Bronson et al. (2001)who found that 19–38 kg additional fertilizer N were needed for CAto produce lint yield similar to CT during the three first CA years.Higher cotton yields with N application, compared to N-0, is theresult of increased boll density (Wiatrak et al., 2005), and also bysignificantly increased boll weight, as found in this study (Table 2).Previous research also indicated that cotton yield components, suchas boll density and weight, impact yield (Pettigrew and Jones, 2001;Wiatrak et al., 2005).

Although N fertilization increased raw cotton yield by 22%, therewas little difference in cotton yield between N-125 and N-250. Also,in a cotton-wheat rotation system, in the study region, Kienzler(2010) reported low yield effects of N fertilization, due to high resid-ual mineral N content in the soil. However, Rochester et al. (2001)indicated much larger differences in cotton yield with increasingN applications in a similar crop rotation system. Since the exper-imental field in the present study had previously been used forcontinuous cotton production over 20 years, the initial residual soilmineral N in the top 90 cm soil profile was high (135 kg N ha−1).Hence, the decrease in ginning percentage with higher N levelswas caused by an increase in seed weight. This implies that N notonly increases lint weights but even more so seed weights, whichconfirms earlier findings (Kienzler, 2010; Reiter et al., 2008).

The increase in wheat grain yield with N application in PB and CTresulted from the increased number of spikes m−2, grains spike−1,and TKW. It has been reported that an adequate N supply promotessurvival and tiller production (Davidson and Chevalier, 1992),delays leaf senescence, sustains leaf photosynthesis during thegrain-filling period, and extends the grain filling period (Fredrickand Camberato, 1995).

The increased maize grain yield with N application in both PBand CT resulted from higher crop growth rate and biomass accu-mulation which in turn increased the number of grains cob−1, andTKW. Response of applied N in maize was higher in PB than underCT. As explained earlier, the improved crop germination and fasterinitial growth in PB with N applications resulting in higher leaf areaindex, AGB and root biomass, may have increased the N demandwhich was satisfied by N applications (Devkota, 2011).

The average yields of cotton, wheat and maize in the region,with the recommended rates of N fertilizer, are 2.7, 4.9 and5 t ha−1, respectively (OblStat, 2009). Even with the application of a

Author's personal copy

M. Devkota et al. / Field Crops Research 149 (2013) 105–114 113

lower-than-recommended N rate (i.e., with medium N applica-tions) in our experiment, the average yields of cotton and wheatwere 53 and 90% higher, respectively, and was slightly higher formaize crop.

4.3. Effect of residues

It is known that RR in PB improves crop yield in both rainfedand irrigated systems (Limon-Ortega et al., 2000; Govaerts et al.,2005; Tursunov, 2009; Verhulst et al., 2011b). In our experiment,irrespective of tillage methods, RR in wheat had small but positiveeffect on crop yield and water productivity. However, in maize RRin PB increased grain yields and water productivity compared toRH while the effect was non-significant in CT. The increase in cropyield under RR could be due to the fact that surface mulch providesmore mineralizable N in the top soil layer (Campbell et al., 1993).Similar findings were reported by Limon-Ortega et al. (2000) for amaize–wheat system in Mexico and by Naresh et al. (2012) for amaize–wheat rotations in India. The experiment was however tooshort to see large benefits of this positive effect of residue retention.

The absence of a significant difference in maize yield betweenRH and RR in CT could be due to an N-immobilization effect fromincorporated residue coming into contact with band-applied N.Similarly, Van Den Bossche et al. (2009) reported higher N immo-bilization of winter wheat residues under CT than under reducedtillage.

5. Summary and conclusions

The permanent raised bed planting system with proper N appli-cation improved the yields of wheat and maize crops compared tothat of conventional system, possibly by providing a better environ-ment for crop growth and nutrient uptake. Moreover, the PB systemimproved the crop water productivity of wheat and maize mainlydue to increased crop yield. It also reduced the amount of waterapplied compared to CT. However, the yield of cotton, the first cropin the rotation, was similar for both tillage treatments. This wasexpected as it was immediately after tillage and conversion of thefield to permanent raised bed planting.

Yields of cotton, wheat and maize were furthermore increasedwith N applications, while, possibly due to remaining residual N inthe intensively used field, cotton yield under both tillage methodsresponded only up to 125 kg N ha−1. Under N-0, the grain yields ofwheat and maize under PB and CT were similar. In maize, responseto N fertilization was greater in PB than under CT. This showshybrid maize cultivated on PB with applied N can, in the short term,produce higher grain yield and biomass than under CT. Residueretention in PB increased grain yield of maize by 15% compared toRH and had no effect on wheat yield.

As this is a study of only one rotational sequence of three cropsover two cropping seasons (May 2008–September 2009), furtherresearch is required to evaluate the system in the longer term,as other problems, such as low soil temperature, appearance ofpest and new weed flora associated with no tillage and excessiveresidues may arise in later seasons.

Acknowledgements

The research described in this paper was funded by the Ger-man Ministry for Education and Research (BMBF; project number0339970A). Field work was done in the ZEF/UNESCO project “Eco-nomic and Ecological Restructuring of Land and Water Use inthe Khorezm Region (Uzbekistan): A Pilot Project in DevelopmentResearch” funded by BMBF.

References

Abdullaev, I., Molden, D., 2004. Spatial and temporal variability of water productivityin the Syr Darya Basin, Central Asia. Water Resour. Res. 40, W08S02.

Akbar, G., Hamilton, G., Hussain, Z., Yasin, M., 2007. Problems and potentials ofpermanent raised bed cropping systems in Pakistan. Pak. J. Water Resour. 11,11–21.

Angas, P., Lampurlanes, J., Cantero-Martinez, C., 2006. Tillage and N fertilizationeffects on N dynamics and Barley yield under semiarid Mediterranean condi-tions. Soil Till. Res. 87, 59–71.

Boulal, H., Gómez-Macphersona, H., Villalobosa, F.J., 2012. Permanent bed plantingin irrigated Mediterranean conditions: short-term effects on soil quality, cropyield and water use efficiency. Field Crops Res. 130, 120–127.

Bronson, K.F., Onken, A.B., Keeling, J.W., Booker, J.D., Torbert, H.A., 2001. Nitrogenresponse in cotton as affected by tillage system and irrigation level. Soil Sci. Soc.Am. J. 65, 1153–1163.

Campbell, C.A., Zentner, R.P., Selles, F., McConkey, B.G., Dyck, F.B., 1993. Nitro-gen management for spring wheat grown annually on zero-tillage: yields andnitrogen use efficiency. Agron. J. 85, 107–114.

Cassman, K.G., Dobermann, A., Walters, D.T., 2002. Agro-ecosystems, nitrogen-useefficiency, and nitrogen management. AMBIO 31, 132–140.

Cassman, K.G., Kropff, M.J., Gaunt, J., Peng, S., 1993. Nitrogen use effi-ciency of irrigated rice: What are the key constraints? Plant Soil 155/156,359–362.

Davidson, D.J., Chevalier, P.M., 1992. Storage and remobilization water soluble car-bohydrates in stems of spring wheat. Crop Sci. 32, 186–190.

Derpsch, R., Friedrich, T., 2009. Global overview of conservation agriculture no-till adoption. In: Innovations for Efficiency, Equity and Environment. 4th WorldCongress on Conservation Agriculture New Delhi, India, 4–7 February.

DeVita, P., Di Paolo, E., Fecondo, G., Di Fonzo, N., Pisante, M., 2007. No-tillage and con-ventional tillage effects on durum wheat yield, grain quality and soil moisturecontent in Southern Italy. Soil Till. Res. 92, 69–78.

Devkota, M.K., 2011. Nitrogen management in irrigated cotton-based systems underconservation agriculture on salt-affected lands of Uzbekistan. ZEF/RheinischeFriedrich-Wilhelms-Universität Bonn, Germany, Bonn, Germany, pp. 163 (Ph.D.Dissertation).

Dobermann, A., Dawe, D., Roetter, R.P., Cassman, K.G., 2000. Reversal of rice yielddecline in a long-term continuous cropping experiment. Agron. J. 92, 633–643.

FAO, 2002. The Salt of the Earth: Hazardous for Food Production.FAOSTAT, 2010. Agriculture Data, Agriculture Production, http://www.faostat.

fao.org (last accessed 28.07.10).Fischer, R.A., Santiveri, F., Vidal, I.R., 2002. Crop rotation, tillage and crop residue

management for wheat and maize in the sub-humid tropical highlands: II. Maizeand system performance. Field Crops Res. 79, 123–137.

Forkutsa, I., Sommer, R., Shirokova, Y.I., Lamers, J.P.A., Kienzler, K., Tischbein, B., Mar-tius, C., Vlek, P.L.G., 2009. Modeling irrigated cotton with shallow groundwaterin the Aral Sea Basin of Uzbekistan: II. Soil salinity dynamics. Irrigation Sci. 27,319–330.

Franzluebbers, A.J., Hons, F.M., Zuberer, D.A., 1995. Soil organic carbon, microbialbiomass, and mineralizable carbon and nitrogen in sorghum. Soil Sci. Soc. Am. J.59, 460–466.

Fredrick, J.R., Camberato, J.J., 1995. Water and nitrogen effects on winter wheatin the southeastern central plain: I. Grain yield and kernel traits. Agron. J. 87,521–526.

Gant, C.A., Gauer, L.E., Bailey, L.D., 1992. Tillage and Soil Physical Qual-ity, http://www.mandakzerotill.org/books/proceddings/proceeding/cynt (lastaccessed 28.07.10).

Govaerts, B., Sayre, K.D., Deckers, J., 2005. Stable high yields with zero tillage andpermanent bed planting? Field Crops Res. 94, 33–42.

Gupta, R., Kienzler, K., Martius, C., Mirzabaev, A., Oweis, T., de Pauw, E., Qadir, M.,Shideed, K., Sommer, R., Thomas, R., Sayre, K., Carli, C., Saparov, A., Bekenov, M.,Sanginov, S., Nepesov, M., Ikramov, R., 2009. Research Prospectus: A Vision forSustainable Land Management Research in Central Asia. Sustainable Agriculturein Central Asia and the Caucasus Series No. 1. CGIAR-PFU. ICARDA Central Asiaand Caucasus Program, Tashkent, Uzbekistan.

Gürsoy, S., Sessiz, A., Malhi, S.S., 2010. Short-term effects of tillage and residuemanagement following cotton on grain yield and quality of wheat. Field CropsRes. 119, 260–268.

Hari-Ram, Yadvinder-Singh, Sani, K.S., Kler, D.S., Timsina, J., Humphreys, E.J., 2012.Agronomic and economic evaluation of permanent raised beds, no tillage and

straw mulchig for irrigated maize–wheat system in Northern India. Exp. Agric.48, 21–38.

Harris, B.L., Krishna, J.H., 1989. Furrow diking to conserve moisture. J. Soil WaterConserv. 44, 271–273.

Hassan, I., Hussain, Z., Akbar, G., 2005. Effect of permanent raised beds on waterproductivity for irrigated maize-wheat cropping system. In: Roth, C.H., Fischer,R.A., Meisner, C.A. (Eds.), Evaluation and Performance of Permanent Raised BedCropping Systems in Asia, Australia and Mexico. ACIAR Proceeding No. 121.

He, 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. Aust. J. Soil Res. 46, 659–666.

Hobbs, P.R., Sayre, K., Gupta, R., 2008. The role of conservation agriculture in sus-tainable agriculture. Philos. Trans. R. Soc. 363, 543–555.

Holland, J.M., 2004. The environmental consequences of adopting conserva-tion tillage in Europe: reviewing the evidence. Agric. Ecosyst. Environ. 103,1–25.

Author's personal copy

114 M. Devkota et al. / Field Crops Research 149 (2013) 105–114

Hulugalle, N.R., Nehl, D.B., Weaver, T.B., 2004. Soil properties, and cotton growth,yield and fibre quality in three cotton-based cropping systems. Soil Till. Res. 75,131–141.

Hulugalle, N.R., Scott, F., 2008. A review of the changes in soil quality and profitabilityaccomplished by sowing rotation crops after cotton in Australian Vertosols from1970 to 2006. Aust. J Soil Res. 46, 173–190.

Khaleque, M., Paul, N., Meisner, C.A., 2008. Yield and N use efficiency of wheatas influenced by bed planting and N application Bangladesh. J. Agric. Res. 33,439–448.

Kienzler, K., 2010. Improving the nitrogen use efficiency and crop quality inthe Khorezm region, Uzbekistan. Ph D Dissertation. ZEF/Rheinische Friedrich-Wilhelms-Universität Bonn, Germany, Bonn, pp. 228.

Kienzler, K.M., Lamers, J.P.A., McDonald, A., Mirzabaev, A., Ibragimov, N., Egam-berdiev, O., Ruzibaev, E., Akramkhanov, A., 2012. Conservation agriculture inCentral Asia—What do we know and where do we go from here? Field CropsRes. 132, 95–105.

Kraatz, D.F., Mahajan, J.K., 1975. Small hydraulic structures. Irrigation and drainagepaper 26/2. FAO, Rome, pp. p198.

Kukal, S.S., Humphreys, E., Yadvinder-Singh, Balwinder-Singh, Yadav, S., Amanpreet-Kaur, Thaman, S., Timsina, J., Dhillon, S.S., Brar, N.K., Prashar, A., Smith, D.J.,2008. Permanent beds for rice-wheat in Punjab, India.1: crop performance. In:Humphreys, E., Roth, C.H. (Eds.), Permanent Beds and Rice-residue Managementfor Rice–Wheat Systems in the Indo-Gangetic Plain. ACIAR Proceedings No. 127., pp. 23–36.

Kukal, S.S., Humphreys, E., Yadvinder-Singh, Timsina, J., Thaman, S., 2005. Perfor-mance of raised beds in rice–wheat systems of northwestern India. In: Roth,C.H., Fischer, R.A., Meisner, C.A. (Eds.), Evaluation and Performance of PermanentRaised Bed Cropping Systems in Asia, Australia and Mexico. ACIAR ProceedingsNo. 121. , pp. 26–40.

Lal, R., Reicosky, D.C., Hanson, J.D., 2007. Evolution of the plow over 10,000 yearsand the rationale for no-till farming. Soil Till. Res. 93, 1–12.

Limon-Ortega, A., Sayre, K.D., Francis, C.A., 2000. Wheat and maize yields in responseto straw management and nitrogen under a bed planting system. Agron. J. 92,295–302.

Lopez-Bellido, L., Fuentes, M., Castillo, J.E., Lopez-Garrido, F.J., Fernandez, E.J., 1996.Long-term tillage, crop rotation, and nitrogen fertilizer effects on wheat yield

under rainfed Mediterranean conditions. Agron. J. 88, 783–791.Mann, R.A., Ramzan, M., Munir, A., 2008. Improving the sustainability of wheat

production in irrigated areas of Punjab, Pakistan through conservation tillagetechnology. Int. J. Agric. Biol. 10, 249–254.

MAWR, 2000. Agrotechnology of Cereal Production in Irrigated and Non-irrigatedAreas. HAYOT, Andijan, Uzbekistan, pp. 46 (in Russian).

McFarland, M.L., Hons, F.M., Saladino, V.A., 1991. Effects of furrow diking and tillageon corn grain yield and nitrogen accumulation. Agron. J. 83, 382–386.

Naresh, R.K., Singh, S.P., Chauhan, P., 2012. Influence of conservation agriculture,permanent raised bed planting and residue management on soil quality andproductivity in maize-wheat system in Western Utter Pradesh. Int. J. LifeSc. Bt.Pharm. Res. 1, 27–33.

OblStat, 2009. Agricultural Indicators for Khorezm. Khorezm Regional Dematmentof Statistics. OblStat, Urgench, Uzbekistan.

Pettigrew, W.T., Jones, M.A., 2001. Cotton growth under no-till production in thelower Mississippi River Valley alluvial flood plain. Agron. J. 93, 1398–1404.

Pulatov, A., Egamberdiev, O., Karimov, A., Tursunov, M., Kienzler, S., Sayre, K.D.,Tursunov, L., Lamers, J.P.A., Martius, C., 2011. Introducing conservation agri-culture on irrigated meadow alluvial soils (Arenosols) in Khorezm, Uzbekistan.In: Martius, C., Rudenko, I., Lamers, J.P.A., Vlek, P.L.G. (Eds.), Cotton, Water, Saltsand Soums – Economic and Ecological Restructuring in Khorezm, Uzbekistan.Springer, Dordrecht/Berlin/Heidelberg/New York.

Randall, G.W., Bandel, V.A., 1991. Overview of nitrogen management for conser-vation tillage systems. In: Logan, T.J., Davidson, J.M., Baker, J.L., Overcash, M.R.(Eds.), Effects of Conservation Tillage on Groundwater Quality-Nitrates and Pes-ticides: An Overview. Lewis Publ., Chelsea, MI, pp. 39–63.

Rautaray, S.K., 2005. Machinery for conservation agriculture: progress and needs.In: Abrol, I.P., Gupta, R.K., Malik, R.K. (Eds.), Conservation Agriculture: Status andProspects. Centre for advancement of sustainable agriculture, New Delhi, p. p43.

Reiter, M.S., Reeves, D.W., Burmester, C.H., Torbert, H.A., 2008. Cotton nitrogen man-agement in a high-residue conservation system: cover crop fertilization. Soil Sci.Soc. Am. J. 72, 1321–1329.

Rice, C.W., Smith, M.S., Blevins, R.L., 1986. Soil nitrogen availability after long-termcontinuous no-tillage and conventional tillage corn production. Soil Sci. Soc. Am.J. 50, 1206–1210.

Rochester, I.J., Peoples, M.B., Hulugalle, N.R., Gault, R.R., Constable, G.A., 2001. Usinglegumes to enhance nitrogen fertility and improve soil condition in cotton crop-ping systems. Field Crops Res. 70, 33–42.

Sayre, K., Hobbs, P., 2004. The raised-bed system of cultivation for irrigated produc-tion conditions. In: Lal, R., Hobbs, P., Uphoff, N., Hansen, D. (Eds.), SustainableAgriculture and the Rice–Wheat System. Ohio State University, Columbia, OH,USA, pp. 337–355.

Scheer, C., Wassmann, R., Kinzler, K., Ibragimov, N., Eschanov, R., 2008. Nitrous oxideemissions from fertilized, irrigated cotton (Gossypium hirsutum L.) in the Aral SeaBasin, Uzbekistan: influence of nitrogen applications and irrigation practices.Soil Biol. Biochem. 40, 290–301.

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 Res. 110, 1–20.

Torbert, H.A., Reeves, D.W., 1994. Fertilizer nitrogen requirements for cotton pro-duction as affected by tillage and traffic. Soil Sci. Soc. Am. J. 58, 1416–1423.

Tursunov, M., 2009. Potential of conservation agriculture for irrigated cotton andwinter wheat production in Khorezm, Aral Sea Basin. ZEF/Rheinische Friedrich-Wilhelms-Universität Bonn, pp. 94 (Ph.D. Dissertation).

Van Den Bossche, A., De Bolle, S., De Neve, S., Hofman, G., 2009. Effect of tillageintensity on N mineralization of different crop residues in a temperate climate.Soil Till. Res. 103, 316–324.

Verhulst, N., Kienle, F., Sayre, K.D., Deckers, J., Raes, D., Limon-Ortega, A., Tijerina-Chavez, L., Govaerts, B., 2011a. Soil quality as affected by tillage-residuemanagement in a wheat–maize irrigated bed planting system. Plant Soil 340,453–466.

Verhulst, N., Sayre, K.D., Vargas, M., Crossa, J., Deckers, J., Raes, D., Govaerts, B., 2011b.Wheat yield and tillage–straw management system × year interaction explainedby climatic co-variables for an irrigated bed planting system in northwesternMexico. Field Crops Res. 124, 347–356.

Wang, F.H., Wang, X.Q., Sayre, K.D., 2004. Comparison of conventional, flood irri-gated, flat planting with furrow irrigated, raised bed planting for winter wheatin China. Field Crops Res. 87, 35–42.

Wiatrak, P.J., Wright, D.L., Marois, J.J., Koziara, W., Pudelko, J.A., 2005. Tillage andnitrogen application impact on cotton following wheat lysimeter results. Agron.J. 97, 288–293.

Wienhold, B.J., Ardell, D., Halviron, A.D., 1999. Nitrogen mineralization responsesto cropping, tillage, and nitrogen rate in the northern great plains. Soil Sci. Soc.Am. J. 83, 192–196.

Yau, S.K., Sidahmed, M., Haidar, M., 2010. Conservation versus conventional tillageon performance of three different crops. Agron. J. 102, 269–276.

Zhang, X., Ma, L., Gilliamc, F.S., Wang, Q., Li, C., 2012. Effects of raised-bed planting forenhanced summer maize yield on rhizosphere soil microbial functional groupsand enzyme activity in Henan Province, China. Field Crops Res. 130, 28–37.