Water Management in Irrigated Rice:Coping with Water Scarcity

B.A.M. Bouman, R.M. Lampayan, and T.P. Tuong

Los Baños, Philippines

2007

The International Rice Research Institute (IRRI) was established in 1960 by the Ford and Rockefeller Foundations with the help and approval of the Government of the Philippines. Today, IRRI is one of the 15 nonprofit international research centers supported by the Consultative Group on In-ternational Agricultural Research (CGIAR – www.cgiar.org).

IRRI receives support from several CGIAR members, including the World Bank, European Union, Asian Development Bank, International Fund for Agricultural Development, Rockefeller Foundation, Food and Agriculture Organization of the United Nations, and agencies of the fol-lowing countries: Australia, Brazil, Canada, Denmark, France, Germany, India, Iran, Japan, Malaysia, Norway, People’s Republic of China, Repub-lic of Korea, Republic of the Philippines, Sweden, Switzerland, Thailand, United Kingdom, United States, and Vietnam.

The responsibility for this publication rests with the International Rice Research Institute.

This publication is copyrighted by the International Rice Research Institute (IRRI) and is licensed for use under a Creative Commons Attribution-Non-

Commercial-ShareAlike 3.0 License (Unported). Unless otherwise noted, users are free to copy, duplicate, or reproduce, and distribute, display, or transmit any of the articles or portions of the articles, and to make translations, adaptations, or other derivative works under the following conditions:

Attribution: The work must be attributed, but not in any way that suggests endorsement by IRRI or the author(s).

NonCommercial: This work may not be used for commercial purposes. ShareAlike: If this work is altered, transformed, or built upon, the resulting

work must be distributed only under the same or similar license to this one.

To view the full text of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/.

Mailing address: DAPO Box 7777, Metro Manila, PhilippinesPhone: +63 (2) 580-5600Fax: +63 (2) 580-5699Email: irri@cgiar.orgWeb: www.irri.org.Rice Knowledge Bank: www.knowledgebank.irri.orgCourier address: Suite 1009, Security Bank Center 6776 Ayala Avenue, Makati City, Philippines Tel. +63 (2) 891-1236, 891-1174, 891-1258, 891-1303

Suggested Citation:Bouman BAM, Lampayan RM, Tuong TP. 2007. Water management in irrigated rice: coping with water scarcity. Los Baños (Philippines): Inter-national Rice Research Institute. 54 p.

ISBN 978-971-22-0219-3

Editing: Bill HardyCover design: Juan Lazaro IVLayout and design: George ReyesFigures and illustrations: Emmanuel Panisales

iii

Preface iv

1 Riceandwater 11.1 Riceenvironments 11.2 Irrigatedlowlands 11.3 Thericefieldanditswaterbalance 11.4 Groundwaterunderricefields 51.5 Ricewaterproductivity 51.6 Globalricewateruse 71.7 Waterscarcityinrice-growingareas 8

2 Theplant-soil-watersystem 112.1 Watermovementinthesoil-plant- 11

atmospherecontinuum2.2 Thericeplantanddrought 14

3 Copingwithwaterscarcity 173.1 Landpreparation 17 3.1.1 Fieldchannels 17 3.1.2 Landleveling 17 3.1.3 Tillage:reducingsoil 17

permeability 3.1.4 Bundpreparationand 18 maintenance

3.2 Cropestablishment 183.3 Cropgrowthperiod. 19 3.3.1 Saturatedsoilculture 19 3.3.2 Alternatewettingand 19

drying

3.3.3 TheSystemofRice 22 Intensification

3.4 Aerobicrice 23 3.4.1 Raisedbeds 27 3.4.2 Conservationagriculture 283.5 Whatoptionwhere? 293.6 Sustainability 32

4 Ecosystemservicesand 35theenvironment

4.1 Ecosystemservices 354.2 Environmentalimpacts 36 4.2.1 Ammoniavolatilization 36 4.2.2 Greenhousegases 36 4.2.3 Waterpollution 364.3Effectsofwaterscarcity 37

5 Irrigationsystems 395.1 Waterflowsinirrigationsystems 395.2 Fieldversusirrigationsystemlevel 415.3 Integratedapproaches 415.4 Irrigationsystemmanagement 425.5 Irrigationsystemdesign 43

Appendix:Instrumentation 45A.1Fieldwatertube 45A.2Groundwatertube 46References 49

Contents

iv

Worldwide,about79millionhaofirrigatedlow-lands provide 75% of the total rice production.Lowland rice is traditionally grown in bundedfields that are continuously flooded from cropestablishment to close to harvest. It is estimatedthatirrigatedlowlandricereceivessome34–43%of the total world’s irrigation water, or 24–30%ofthetotalworld’sfreshwaterwithdrawals.Withincreasingwaterscarcity,thesustainability,foodproduction,andecosystemservicesof ricefieldsarethreatened.Therefore,thereisaneedtodevelopanddisseminatewatermanagementpracticesthatcanhelpfarmerstocopewithwaterscarcityinir-rigatedenvironments.

Thismanualprovidesanoverviewoftechnicalresponseoptions towater scarcity. It focusesonwhatindividualfarmerscandoatthefieldlevel,withabriefdiscussiononresponseoptionsattheirrigationsystemlevel.Themanualismeantasasupportdocument for trainingonwatermanage-ment in rice production. It combines scientificbackgroundinformation(withmanyliteraturerefer-encesforfurtherreading)withpracticalsuggestionsforimplementation.Thetargetaudienceispeopleinvolvedinagriculturalextensionortrainingwithan advanced education in agriculture or watermanagement,whowishtointroducesoundwatermanagementpracticestoricefarmers(suchasstaffofagriculturalcollegesanduniversities,scientists,irrigationoperators,andextensionofficers).

Introductorychaptersanalyzethewateruseandwaterbalanceofricefields,andwatermovementintheplant-soilsystem,anddiscusstheconceptsofwaterscarcityandwatersavings.Consequencesofwaterscarcityforsustainability,environmentalimpacts,andecosystemservicesof irrigatedricefieldsarediscussedattheend.Anappendixintro-duces twosimple instruments tocharacterize thewaterstatusofricefieldsthatcanhelpfarmersinapplyingwater-savingtechnologies.

ThismanualwasdevelopedthroughtheWaterWorkGroupoftheIrrigatedRiceResearchCon-sortium(whichisco-fundedbytheSwissAgencyforDevelopmentandCooperation).Thesectionsonaerobicricewereco-developedbytheCGIARChallenge Program on Water and Food throughthe project “Developing a System of TemperateandTropicalAerobicRiceinAsia(STAR).”Manypartners from national agricultural research andextensionsystemsinAsiahavecontributedtotheworkdescribed in thismanual.The manual was reviewed by Dr. Ian Willet (Australian Centre for International Agricultural Research) and Dr. Moh-sin Hafeez (CSIRO Land and Water).

TheauthorsLosBaños,2007

Preface

�

1

1.1 Rice environments

Worldwide,thereareabout150millionhectaresofriceland,whichprovidearound550–600milliontonsofroughriceannually(Macleanetal2002).Riceisuniqueamongthemajorfoodcropsinitsability to grow in awide range of hydrologicalsituations,soiltypes,andclimates.Riceistheonlycerealthatcangrowinwetlandconditions.

Dependingonthehydrologyofwherericeisgrown,thericeenvironmentcanbeclassifiedintoirrigatedlowlandrice(79millionha),rainfedlow-landrice(54millionha),flood-pronerice(11mil-lionha),anduplandrice(14millionha).Lowlandriceisalsocalled“paddyrice.”Lowlandricefieldshavesaturated(anaerobic)soilconditionswithpon-dedwaterforatleast20%ofthecrop’sduration.Inirrigatedlowlands,theavailabilityofirrigationassuresthatpondedwaterismaintainedforatleast80%of thecrop’sduration. In rainfed lowlands,rainfallistheonlysourceofwatertothefieldandnocertaindurationofpondedwatercanbeassured(dependingonvagariesofrainfall).Inflood-proneenvironments, thefields suffer periodically fromexcesswateranduncontrolled,deepflooding(morethan25cmfor10daysormore).Deepwaterriceandfloatingricearefoundintheseenvironments.Uplandricefieldshavewell-drained,nonsaturated(aerobic)soilconditionswithoutpondedwaterformorethan80%ofthecrop’sduration.

1.2 Irrigated lowlandsThe79millionhaofirrigatedlowlandsprovide75%oftheworld’sriceproduction(Macleanetal2002;Fig.1.1).AttheturnoftheMillennium,country-averageirrigatedriceyieldsinAsiarangedfrom3

to9tha–1,withanoverallaverageofabout5tha–1(Macleanetal2002).Irrigatedriceismostlygrownwithsupplementaryirrigationinthewetseason,andisentirelyreliantonirrigationin thedryseason.Significantareasofricearegrowninrotationwitharangeofothercrops,suchasthe15–20millionhaofrice-wheatsystems(TimsinaandConnor2001,Daweetal2004).Irrigationsystemsvarywidely,andinclude

Individual pump irrigation from shallowtubewells(downtoabout15-mdepth).

Small- tomedium-scale community-basedpump irrigation fromdeepwells (down to200–300-mdepth).

Small- tomedium-scale community-basedsurface irrigation where water is divertedfrom ponds or reservoirs (for example,the tank system in southern India andSriLanka).

Small- tomedium-scale community-basedsurface irrigation where water is directlydiverted from a river (run-off-the-river ir-rigation).

Large-scalesurfaceirrigationwherewaterisdivertedfromreservoirsorlakes.

Conjunctive groundwater-surface-waterirrigation schemes (can be small to largescale).

In each type of system, the ownership andcontrolofwatermayvarywidely.

1.3 The rice field and its water balanceIrrigated lowland rice is grown under floodedconditions.Mostly,riceisfirstraisedinaseparateseedbed and subsequently transplanted into thericefieldwhen theseedlingsare2–3weeksold.

Rice and water

�

Fig.

�.�

. Dis

trib

utio

n of

irrig

ated

low

land

rice

in A

sia.

Sou

rce:

IRR

I GIS

uni

t, �0

06.

Ricecanalsobeestablishedbydirectwetseeding(broadcastingpregerminatedseedsontowetsoil)ordirectdryseeding(broadcastingdryseedsontodryormoistsoil)inthemainfield.Asofthelate1990s,itwasestimatedthatone-fifthoftheareainAsiawasdirectseeded(PandeyandVelasco2002).Aftercropestablishment,themainfieldisusuallykept continuously flooded as this helps control

weeds andpests.Before crop establishment, themainfieldispreparedunderwetconditions.Thiswetlandpreparationconsistsofsoaking,plowing,andpuddling(i.e.,harrowingorrotavatingundershallowsubmergedconditions).Puddlingisdonetocontrolweeds,toreducesoilpermeability,andtoeasetransplanting.Puddlingleadstoacompleteorpartialdestructionofsoilaggregatesandmacro-

120°

90°

60°

50°

30°

10°

-10°

010

00

km

Each

dot

rep

rese

nts

1000

0 ha

of

irri

gate

d ri

ce

�

porevolume,andtoalargeincreaseinmicropores(Moorman and vanBreemen 1978).A typicalverticalcross-sectionthroughapuddledricefieldshowsalayerof0–10cmofpondedwater,apud-dled,muddytopsoilof10–20cm,aplowpanthatisformedbydecadesorcenturiesofpuddling,andanundisturbed subsoil (Fig. 1.2).Rice roots areusuallycontainedwithinthepuddledlayerandarethereforequiteshallow.Theplowpanreducesthehydraulicconductivityandpercolationrateofricefieldsdramatically.

Becauseofitsfloodednature,thericefieldhasawaterbalancethatisdifferentfromthatofdrylandcropssuchaswheatormaize.Thewaterbalanceofaricefieldconsistsoftheinflowsbyirrigation,rainfall, and capillary rise, and the outflows bytranspiration, evaporation, overbundflow, seep-age,andpercolation(Fig.1.2).Capillaryriseistheupwardmovementofwaterfromthegroundwatertable. Innonflooded (aerobic) soil, this capillaryrisemaymoveintotherootzoneandprovideacropwithextrawater.However,infloodedricefields,thereisacontinuousdownwardflowofwaterfromthepuddled layer tobelow theplowpan (called“percolation”; seebelow) thatbasicallypreventscapillaryriseintotherootzone.Therefore,capil-laryriseisusuallyneglectedinthewaterbalanceofricefields.

Beforethecropactuallystartsgrowing,waterinputisalreadyneededforwetlandpreparation.Afterpuddling,thefieldisusuallyleftfallowand

floodedforafewdays(or1to4weeks)beforetheseedlings are transplanted. The amount of waterused forwet land preparation can be as low as100–150mmwhenthe turnaroundtimebetweensoaking and transplanting is a fewdays only orwhenthecrop isdirectwetseeded.However, inlarge-scaleirrigationsystemsthathavepoorwatercontrol,theturnaroundtimebetweensoakingandtransplantingcangoupto2monthsandwaterin-putsduringthisperiodcanreach940mm(Tabbalet al 2002).After crop establishment, the soil isusuallykeptpondedwitha5–10-cmlayerofwateruntil 1–2weeksbefore harvest.Duringboth theturnaroundtimeandthecropgrowthperiod,wateroutflowsarebyoverbundrunoff,evaporation,seep-age, andpercolation. During crop growth, wateralso leaves the ricefieldby transpiration.Of allwateroutflows,runoff,evaporation,seepage,andpercolationarenonproductivewaterflowsandareconsideredlossesfromthefield.Onlytranspirationisaproductivewaterflowasitcontributestocropgrowthanddevelopment.

Whenrainfallraisesthelevelofpondedwaterabovetheheightofbunds,excessrainleavesthericefieldassurfacerunofforoverbund flow.Thissurface runoff canflow into aneighboringfield,but,inasequenceoffields,neighboringfieldswillpassontherunoffuntilitislostinadrain,creek,orditch.

Evaporationleavesthericefielddirectlyfromthepondedwaterlayer.Transpirationbyriceplants

Fig. 1.2. Water balance of a lowland rice field. C = capillary rise, E = evaporation, I = irrigation, O = overbund flow, P = percolation, R = rainfall, S = seepage, T = transpiration.

�

withdrawswaterfromthepuddledlayer.Sincetherootsof riceplantsgenerallydon’tpenetrate thecompactedlayer,thecontributiontotranspirationfromthesubsoilismostlyabsent.Sinceevaporationandtranspirationaredifficulttomeasureseparatelyinthefield,theyareusuallytakentogetheras“eva-potranspiration.”Typicalevapotranspirationratesofricefieldsare4–5mmd–1inthewetseasonand6–7mmd–1inthedryseason,butcanbeashighas 10–11 mm d–1 in subtropical regions (Tabbaletal2002).Duringthecropgrowthperiod,about30–40%ofevapotranspirationisevaporation(Bou-manetal2005,Simpsonetal1992).

Seepage is the subsurface flow of waterunderneath the bunds of a ricefield.Withwell-maintainedbunds,seepageisgenerallysmall.Inatoposequenceofricefields,seepagelossfromonefieldmaybeoffsetbyincomingseepagefroman-otherfieldlocatedhigherup.Considerableseepagecanoccurfromtop-endfieldsandfrombottom-endfieldsthatborderdrains,ditches,orcreeks.Seepageratesareaffectedbythesoilphysicalcharacteristicsofthefieldandbunds,bythestateofmaintenanceandlengthof thebunds,andbythedepthof thewatertableinthefieldandinthesurroundingdrains,ditches, or creeks (Wickham and Singh 1978).

Percolationistheverticalflowofwatertobelowthe root zone.Thepercolation rateof ricefieldsisaffectedbyavarietyofsoilfactors(WickhamandSingh1978):structure,texture,bulkdensity,mineralogy,organicmattercontent,andsalttypeandconcentration.Soilstructureischangedbythephysicalactionofpuddling.Inaheavy-textured,montmorilloniticclay,sodiumcationsandahighbulkdensityare favorable foreffectivepuddlingtoreducepercolationrates.Thepercolationrateisfurtherinfluencedbythewaterregimeinandaroundthefield.Largedepthsofpondedwaterfavorhighpercolation rates (Sanchez 1973,Wickham andSingh1978).InafieldsurveyinthePhilippines,Kampen(1970)foundthatpercolationrateswerehigherforfieldswithdeepgroundwatertables(>2mdepth)thanforfieldswithshallowgroundwatertables (0.5–2mdepth). In practice, seepage andpercolationflowsarenoteasilyseparatedbecauseoftransitionflowsthatcannotbeclassifiedaseitherpercolationorseepage(WickhamandSingh1978).Typicalcombinedvaluesforseepageandpercola-tionvaryfrom1–5mmd–1inheavyclaysoilsto25–30mmd–1insandyandsandyloamsoils(Bou-manandTuong2001).Someexamplesofseepageandpercolationratesmeasuredatdifferentsitesare

Table 1.1. Total seasonal water input and daily seepage and percolation rates from lowland rice fields with continuously ponded water conditions. Data collected from field experiments and from farmers’ fields in China and the Philippines.

Total seasonal water Seepage andSite input by rain plus percolation References irrigation (mm) rate (mm d−1)

Zanghe Irrigation System, Hubei, China • Field experiment 750 to 1,110 4.0 to 6.0 Cabangon et al (2001, 2004) • Farmers’ fields 650 to 940 1.6 to 2.8 Dong et al (2004), Loeve et al (2004a,b) • Irrigation system level 750 to 1,525 4.0 to 8.0 Dong et al (2004), Loeve et al (2004a,b)Shimen, Zhejiang, China Cabangon et al (2001, 2004) • Early rice 850 to 950 1.0 to 6.0 • Late rice 575 to 700 1.0 to 6.0 Guimba, Philippines Tabbal et al (2002) • Experiment 1988 2,197 18.3 • Experiment 1989 1,679 12.5 • Experiment 1990 2,028 16.4 • Experiment 1991 3,504 32.8 Muñoz, Philippines, 1991 1,019 to 1,238 5.2 to 7.0 Tabbal et al (2002)Muñoz, Philippines, 2001 600 1.1 to 4.4 Belder et al (2004)Talavera, Philippines 577 to 728 0.3 to 2.0 Tabbal et al (2002)San Jose, Philippines Tabbal et al (2002) • Experiment 1996 1,417 9.6 • Experiment 1 1997 1,920 15.2 • Experiment 2 1997 2,874 25.8

�

given inTable1.1.Water lossesbyseepageandpercolationaccountforabout25–50%ofallwaterinputsinheavysoilswithshallowgroundwaterta-blesof20–50-cmdepth(Cabangonetal2004,Dongetal2004),and50–85%incoarse-texturedsoilswithdeepgroundwater tables of 1.5-mdepth ormore(Sharmaetal2002,Singhetal2002).Thoughseepageandpercolationarelossesatthefieldlevel,theyareoftencapturedandreuseddownstream(justlikeoverbundflow;Chapter5.1).Theactualamountofwaterreuseinrice-basedirrigationsystemsisnotknownandisexpectedtovarywidelyamongirrigationsystems.

Dailyseepageandpercolationlossesfromthepondedwaterdonotoccurindrylandcropssuchaswheatandmaize.Percolationofwaterbelowtherootzonecanalsooccurindrylandcropswhentheamountofwaterinfiltratingintothesoil(eitherafterheavyrainfallorafterirrigation)islargerthanthestoragecapacityoftherootzone,butthisisnotadailywaterflowasinlowlandrice.Also,evapora-tionfrompondedwatersurfacesishigherthanfromsoilsurfaces(asindrylandcrops).Therefore,itisthe relatively largewaterflowsby seepage,per-colation,andevaporationthatmakelowlandricefields heavy “water users.”Total seasonalwaterinputtoricefields(rainfallplusirrigation)canbeupto2–3timesmorethanforothercerealssuchaswheatormaize(Tuongetal2005).Itvariesfromaslittleas400mminheavyclaysoilswithshal-lowgroundwatertables(thatdirectlysupplywaterforcroptranspiration)tomorethan2,000mmincoarse-textured(sandyorloamy)soilswithdeepgroundwater tables (Bouman andTuong 2001,Cabangonetal2004).Around1,300–1,500mmisatypicalvalueforirrigatedriceinAsia.Table1.1listssomevaluesforwaterinputsanddailyseep-ageandpercolationratesforlowlandricefieldsinChinaandthePhilippines.

It isuseful todistinguishbetween thewateroutflowsfromricefieldsthatcan,inprinciple,bereusedandthosethatcannotbereused.Nonreus-ableoutflowsarecalled“depletedwater”(Molden1997)andareevaporationandtranspiration.Over-bundflow,seepage,andpercolationaregenerallyreusableflows.Onlywhentheseflowsenterverydeeporsalinegroundwater(orsalinesurfacewa-ter)fromwheretheycannotberecoveredaretheynotreusableanymoreandtheybecomedepletionflowsaswell.

1.4 Groundwater under rice fields

Theroleofgroundwaterinprovidingwatertoriceplantsmaybelarge,butithasbeenneglectedinmoststudiesofthericewaterbalance.Recentdatacollection suggests that through the (decade- toage-old)practiceofcontinuousflooding,thelargeamountsofpercolatingwaterhaveraisedground-water tables toveryclose to the surface.This isespeciallytrueinsoilswithaheavytexturethatarepoorlydrainedinthesubsoil,asisthecaseinmanytraditionalirrigatedriceenvironments.

Figure1.3givesthegroundwatertablemeas-uredunderfloodedricefieldsatsomesitesinChinaandthePhilippines.Whenthegroundwaterislessthan20cmdeep,itprovidesa“hidden”sourceofwatertothericecropastherootsoftheplantscandirectlytakeupwaterfromthegroundwater.Whenfor some reasonfields are not flooded (Chapter3),capillaryrisemayreachintotherootzoneandagain provide extra water to the crop. In mostwater balance studies, the effect of groundwateronwatersupplyisnottakenintoaccount,andthebeneficialeffectofwater-savingtechnologiescanbeoverestimated.Withshallowgroundwater,cropgrowthwithasmallirrigationwatersupplycanstillbegoodbecauseofthe“hidden”watersupplyofgroundwater.

1.5 Rice water productivity Water productivity (WP) is a concept ofpartial productivity anddenotes the amountor valueofproduct(inourcase,ricegrains)overvolumeorvalueofwaterused.Discrepanciesarelargeinre-portedvaluesofWPofrice(Tuong1999).Thesearepartiallycausedbylargevariationsinriceyields,withcommonlyreportedvaluesrangingfrom3to8tonsperhectare.Butthediscrepanciesarealsocausedbydifferentunderstandingsofthedenomi-nator(waterused)inthecomputationofWP.ToavoidconfusioncreatedbydifferentinterpretationsandcomputationsofWP,itisimportanttoclearlyspecifywhatkindofWPwearereferringtoandhowitisderived.CommondefinitionsofWPare

WPT : weightofgrainsovercumulativeweight ofwatertranspired.

WPET:weightofgrainsovercumulativeweight ofwaterevapotranspired.

WPI :weightofgrainsovercumulativeweight ofwaterinputsbyirrigation.

6

WPIR:weightofgrainsovercumulativeweight ofwaterinputsbyirrigationandrain.

WPTOT:weightofgrainsovercumulativeweight of allwater inputs by irrigation, rain, andcapillaryrise.

Breedersareinterestedintheproductivityofthe amount of transpiredwater (WPT), whereasfarmersandirrigationengineers/managersarein-terestedinoptimizingtheproductivityofirrigationwater(WPI).Toregionalwaterresourceplanners,whoareinterestedintheamountoffoodthatcanbeproducedbytotalwaterresources(rainfallandirrigationwater)intheregion,waterproductivitywithrespecttothetotalwaterinputbyirrigationandrainfall(WPIR)ortothetotalamountofwaterthatcannolongerbereused(WPET)maybemorerelevant.

Modern rice varieties, when grown underfloodedconditions,havesimilarwaterproductiv-itywithrespecttotranspiration(WPT)asotherC3cerealssuchaswheat,atabout2ggrainkg–1water

transpired(BoumanandTuong2001,Tuongetal2005).Thefewavailabledataindicatethatwaterproductivitywithrespecttoevapotranspirationisalsosimilartothatofwheat,rangingfrom0.6to1.6ggrainkg–1ofevapotranspiredwater,withameanof1.1ggrainkg–1(Tuongetal2005,Zwartand Bastiaanssen 2004, Fig. 1.4A). Comparedwithwheat,thehigherevaporationratesfromthewaterlayerinricethanfromtheunderlyingsoilinwheatareapparentlycompensatedforbythehigheryieldsofrice.Formaize,beingaC4crop,thewaterproductivitywithrespecttoevapotranspirationishigher,rangingfrom1.1to2.7ggrainkg–1water,withameanof1.8ggrainkg–1.Waterproductivityofricewithrespecttototalwaterinput(irrigationplusrainfall) rangesfrom0.2 to1.2ggrainkg–1water,with0.4astheaveragevalue,whichisabouthalfthatofwheat(Tuongetal2005,Fig.1.4B).

ComparingWPamongseasonsandlocationscanbemisleadingbecause of differences in cli-matic yieldpotential, evaporativedemands from

18 Jan 7 Feb 27 Feb 18 Mar 7 Apr 27 Apr

20

0

–20

–40

–60

–80

–100

Groundwater depth (cm)

16 Jun 11 Jul 5 Aug 30 Aug 7 Sep 19 Oct

20

0

–20

–40

–60

–80

–100

Groundwater depth (cm)

Panicleinitiation

Flowering Harvest

A B

Groundwater depth (cm)

14 Jan 13 Feb 15 Mar 14 Apr 14 Jan 13 Feb 15 Mar 14 Apr

–20

20

–60

–100

–140

–180

–220

–260

–300

0

20

–20–30–40

–70–80–90

–100

–60–50

–10

10

Groundwater depth (cm)C D

Fig. 1.3. Groundwater depth under flooded rice at Tuanlin, Hubei Province (A), and Changle (B), Beijing, China, in 2002. Adapted from Belder et al (2004) and unpublished data from China Agricultural University/IRRI. Groundwater depth under flooded rice in Dolores (C) and Gabaldon (D) in 2002, Central Luzon, Philippines. Adapted from Lampayan et al (2005).

Day Day

Day Day

�

the atmosphere, or crop management practicessuchasfertilizerapplication.ItismorerelevanttostudywhatthepotentialandactualWPvaluesareinaparticularenvironment,andtoidentifymeas-ures toclose thegapsbetween them, rather thantocompareWPvaluesacrossenvironments(andsometimesyears).Forexample,inrainyseasons,asmallamountofsupplementaryirrigationcanleadtoveryhighWPIlevels(e.g.,>10ggrainkg

–1ir-rigationwaterinTuanLin,China,Cabangonetal[2003])becauserainfallsuppliesmostofthewaterneededforcropgrowth.Thisdoesnotmeanthatirrigationwaterisbetterusedintherainyseasonthaninthedryseason.Theimpactofthesupple-mentary irrigation canbe better assessedby the“incrementalirrigationwaterproductivity,”definedastheincreaseintheamountorvalueoftheproduct(comparedwithnoirrigation)overthevolumeof

supplementaryirrigationwater.Unfortunately,dataonthiskindofwaterproductivityarescarce.

The concept ofwater productivity becomesimportantwhenwaterisscarce.Examplesoftheuseofwaterproductivityinthedesignormanage-ment of irrigation systems are given in Chapter5.4and5.5.

1.6 Global rice water useTherearenodataavailableon theamountof ir-rigationwater used by all the rice fields in theworld.However,estimatescanbemadebasedontotalworldwidewaterwithdrawalsfor irrigation,therelativeareaof irrigatedrice land(comparedwith other crops), and the relative water use ofricefields.Totalworldwidewithdrawalsoffreshwaterareestimatedat3,600km3annually,ofwhich

Frequency (%)35

30

25

20

15

10

5

0

0.0–0.1

0.1–0.2

0.2–0.3

0.3–0.4

0.4–0.5

0.5–0.6

0.6–0.7

0.7–0.8

0.8–0.9

0.9–1.0

1.0–1.1

1.1–1.2

1.2–1.3

1.3–1.4

1.4–1.5

1.5–1.6

1.6–1.7

1.7–1.8

1.8–1.9

1.9–2.0

2.0–2.1

2.1–2.2

2.2–2.3

2.3–2.4

2.4–2.5

WPET (g grain kg–1 water)

Frequency (%)35

30

25

20

15

10

5

0

0.1–0.2

0.2–0.4

WPIR (g grain kg–1 water)0.4–0.6

0.6–0.8

0.8–1.01.0–1.2

1.2–1.4

1.4–1.61.6–1.8

1.8–2.0

2.0–2.2

2.2–2.4

A

B

Fig. 1.4 (A) Frequency distribution of water productivity of rice with respect to evapotranspiration (WPET). Data from Zwart and Bastiaanssen (2004). (B) Frequency distribution of water productivity of rice with respect to water inputs by irrigation and rainfall (WPIR). Data from Tuong et al (2005).

�

2,500km3isusedtoirrigatecrops(FalkenmarkandRockström2004).Therestisusedinindustryandfordomesticpurposes.Approximately56%oftheworld’s271millionhaofirrigatedareaofallcropsisinAsia,wherericeaccountsfor40–46%ofthenetirrigatedareaofallcrops(Dawe2005).Atthefieldlevel,ricereceivesupto2–3timesmorewaterthanotherirrigatedcrops(Chapter1.3),butanunknownproportionofthewaterlossesfromindividualfieldsis reused by other fields downstream (Chapter5.1).Assumingareusefractionof25%,itcanbeestimatedthatirrigatedricereceivessome34–43%ofthetotalworld’sirrigationwater,or24–30%ofthe total world’s freshwater withdrawals. Figure1.5givesirrigatedareasandvolumesofirrigationwaterusedinagricultureandinrice.

1.7 Water scarcity in rice-growing areasWorldwide,waterforagricultureisbecomingin-creasinglyscarce(Rijsberman2006).Thecausesare diverse and location-specific, but includedecreasingresources(e.g.,fallinggroundwaterta-bles,siltingofreservoirs),decreasingquality(e.g.,

chemical pollution, salinization), malfunctioningof irrigation systems, and increased competitionfromother sectors such as urban and industrialusers.Thereisnosystematicinventory,definition,orquantificationofwaterscarcityinrice-growingareas.TuongandBouman(2003)estimatedthat,by 2025, 15–20million ha of irrigated ricewillsufferfromsomedegreeofwaterscarcity.Thereare no indications yet of water scarcity in someofAsia’s largest irrigated riceecosystems in theriver deltas of theYangtze, theMekong, or theIrrawady. However, in South Asia, the Gangesand Indus rivershave littleoutflow to thesea inthe dry season, thus affecting downstream rice-growing areas (Postel 1997).OverexploitationofgroundwaterduringrecentdecadeshascausedseriousproblemsinnorthernChinaandSouthAsia(Postel1997,ShuGengetal2001,Singh2000),affectingrice-wheat-growingareas.Groundwatertableshavedroppedonaverageby1–3my–1intheNorthChinaPlain;by0.5–0.7my–1intheIndianstatesofPunjab,Haryana,Rajasthan,Maharashtra,Karnataka,andnorthernGujarat;andbyabout1m y–1 inTamilNadu and southern India,wherefloodedrice is thedominantcroppingsystem.InBangladesh,theheavyuseofgroundwaterhasledtoshallowwellsfallingdrybytheendofthedryseasonandtosevereproblemsofarsenicpollutioninrice-growingareas(Ahmedetal2004).Heavycompetition for riverwater between states anddifferent sectors (city, industry) iscausingwaterscarcityinsouthernIndia’sCauverydeltaandinThailand’sChaoPhrayadelta(Postel1997),whicharemajorregionalricebowls.Severalcasestudiessuggestlocalhot-spotsofwaterscarcitybecauseofincreasedcompetitionbetweendifferentusersofwater,eveninareasgenerallynotconsideredwaterscarce,forexample,theZangheIrrigationSysteminthemiddlereachesoftheYangtze(Dongetal2004)andAngatreservoirnearManila,Philippines(BhuiyanandTabbal,asreferencedinPingalietal1997).Inprinciple,waterisalwaysscarceinthedryseasonwhenthelackofrainfallmakescroppingimpossiblewithoutirrigation.

Usually,interventionstorespondtowaterscar-cityarecalled“watersavings”andimplyareduceduseofwater.Formanyofus,theterm“watersav-ings”suggestsanactiveactionordecisiontosavewater.This suggests thatwater is available, butfarmersmayoptnottouseitforirrigationbuttosaveitforalatertimeorforadifferentpurpose,or

250

200

150

100

50

0

300Million ha

World irrigated area Asia irrigated area World irrigated rice area

km3

World water withdrawal

World water use in irrigation

Rice water use in irrigation

4,0003,5003,0002,5002,0001,5001,000

500

0

Fig. 1.5. Irrigated areas (A) and volumes of irrigation water used (B) in the world, in Asia, and in rice production.

B

A

�

tocutdownoncostsifirrigationwaterisexpensive.However,formostricefarmers,thereisnodeliber-atechoicetosavewaterastheyarejustconfrontedbyalackofwater.Undersuchconditions,savingwater simply means coping with scarcity. Thus,theobjectivesofsavingwaterdependonthenatureofthewaterscarcityandthecontrolfarmershaveover the water. Drawing parallels with generaldefinitionsof“savings,”wediscussthreereasonsforsavingwater:

Definition 1:toreducecurrentexpendituresononecommoditytoallowforredirectedexpendituresonothercommodities.Inwaterterms,thistranslatesinto“reducingwaterusedforirrigationsothatitcanbeusedforanotherpurpose.”Inagriculture,themotivationforthistypeofwatersavingsisusually

notanabsoluteshortageofwaterbutadesiretousetheavailablewaternotforirrigationbutforotherpurposessuchasdomesticorindustrial.Increasingcompetitionforwaterbetweensectorsofsocietyisthedrivingforcebehindsuchsavingsinagriculturalwateruse.TheexamplesofwhatishappeningintheZangheIrrigationSysteminChinaandattheAngat reservoir in the Philippines are a case inpoint:themanagersofthesereservoirsarereduc-ing the amount of water released for agricultureandredirectingthiswatertocities(Manila,inthecaseofAngat)andtoindustryandhydropower(inthecaseofZanghe)(Loeveetal2004a,b;Fig.1.6).Thesewatersavingsinagriculturearenotanactiveanddeliberate choice by farmers.The choice towithdrawwaterfromagricultureismadeatahigher

Fig. 1.6. (A) “Imposed water scarcity”: change in percentage water allocation from Angat reservoir, Philippines, to through-flow in the river for potential agricultural use downstream and to the city of Manila. Data from Pingali et al (1997) and unpublished data from the National Irrigation Administration, Philippines. (B) “Imposed water scarcity”: change in water allocation to agricultural and nonagricultural use in Zanghe Irrigation System, Hubei, China, between 1965 and 2002. Data points are 5-year moving averages. Data from Hong et al (2001) and unpublished data.

100Distribution (%)

90

80

70

60

5040

30

20

10

01968 1973 1978 1983 1988 1993 1998

Through-flowManila

Year

A

1965 1971 1977 1980 1989 1992 19981968 1974 1983 1986 1995

Water allocation (106 m3)

Year

800

700

600

500

400

300

200

100

0

IrrigationOther uses

B

�0

level:theirrigationsystem,provincial,ornationallevel.Farmersmustfacetheconsequencesofthesedecisions:theyreceivelesswaterandhavetocopewith“imposedwaterscarcity.”Thenotionthatweshouldaskfarmerstoactivelysavewatersoitcanbeusedelsewhere,suchasbyindustryandcities,isafallacy.Farmerscanbeencouragedtovoluntar-ilyreducewateruse,forexample,byintroducingvolumetricwaterpricing,butthisistheexceptionratherthantheruleand“enforcedwaterscarcity”isprevalent.Voluntarywatersavingsbyfarmersforredirecteduseworkwellonlywithproperlyfunc-tioningwatermarkets.Forexample,inAustralia,rice farmers in irrigation schemes in theMurraybasincanselltheirwaterrightsinawatermarkettootherusers,suchasotherfarmerswhogrowhigh-valuecropssuchasfruits(Thompson2002).

Definition 2: toreduceexpendituresbecauseofreducedincome.Inwaterterms,thistranslatesinto“reducingtheuseofirrigationwaterbecausethere is lessof it.”This typeofwatersavingsinagricultureisinducedbyactualandphysicalwaterscarcity.Anexampleofthissituationforfarmersisthe“enforcedwatershortage”discussedabove.However,anabsolutewatershortagecanalsobeinducedbynaturalcauses.Forexample,whensea-sonalrainshavefailedtofillupreservoirsorponds,theamountofwatermaynotbesufficienttokeepallricefieldsfloodedthroughouttheyear.Whenthereservoirformspartofalarge-scaleirrigationsystem, reservoirmanagers usually respond byreducingthe“programarea”forirrigation:fewerfarmerswillreceiveirrigationwater.However,withsmallerreservoirssuchasindividualponds,farmersthemselvescandecidehowtocopewiththewaterscarcity.Theymaydecidetoreducetheirlandunderirrigationortheycoulddecideto“reducecurrentexpenditurestoallowforfutureexpenditures”:todeliberatelysavewaterearlyintheseasontohaveitavailablelaterintheseason.Farmerscansavewater

byreducingtheamountofirrigationappliedtotheirfieldsearlyintheseason.Thebestwaytodothisisbyreducingthenonproductiveoutflowsseepage,percolation,andevaporation(Chapter3).

Definition 3:toreducecoststoincreaseprofit.Inwater terms, this translates into“reducing theuse of irrigationwater to lower the costs.”Thisscenarioisapplicablewhenfarmerspayahighcostforwaterandhavethemeanstoreducetheirwaterusetoincreasetheirprofits.Theremaybeplentyofwater,butitisrelativelyexpensive(“economicwaterscarcity”).InmostsurfaceirrigationsystemsinAsia,farmerseitherpaynocostfortheirwaterorpayaflatrate(afixedsumperunitlandarea),andwatercostscannotbereducedbyreducingwateruse.Whenfarmerspumptheirownwater,eitherindividuallyorcollectively, theypayarelativelyhighpricefortheirwaterwhenpumpingisfromdeepaquifersand/orwhen theprice forelectric-ityorfuel ishigh.Inthiscase,watersavingsbyfarmers are a voluntary anddeliberate choiceoftheirown.Themeanstosavewaterarethesameasinthescenarioabove:toreduceirrigationwatertotheirfields.

Theseexamplesillustratethatwaterscarcityisusuallyimposeduponfarmers(eitherbynatureorbydecisionmakersathigherlevels)andthatsavingwaterishardlyavoluntarychoice(exceptindefi-nition3).Farmersjusthavetocopewithphysicalwaterscarcity,andtheterm“water-scarcitycopingtechnology”maybemoreappropriatethantheterm“water-savingtechnology.”

��

The plant-soil-water system

2.1 Water movement in the soil-plant- atmosphere continuum

Riceplantstakeupwaterfromthesoilandtransportitupwardthroughtherootsandstemsandreleaseitthroughtheleavesandstemsasvaporintheat-mosphere(calledtranspiration).Themovementofwaterthroughtheplantisdrivenbydifferencesinwaterpotential:waterflowsfromahighpotentialtoalowpotential(imaginefreewaterflowoveraslopingsurface:waterflowsfromthetop,withahighpotential,tothebottom,withalowpotential).Differentunitsexpresswaterpotential(Table2.1)and,unfortunately,differentauthorsreportdiffer-entunits. In this report,weusuallyuse the termPascal(P).

In the soil-plant-atmosphere continuum, thewaterflows from the soil,with a relativelyhighpotential,throughtheplanttotheatmospherejustoutsidetheleaves,whichhasarelativelylowpo-tential.Potentialsinthesoil-plant-atmosphereareusuallynegativeandwealsousetheterm“tension,”whichhastheoppositevalue.Forexample,aten-sionof+10kPaistheequivalentofapotentialof

–10kPa.Thetermtensionisintuitivelyeasiertounderstand:ahightensionsuggestsahigh“pullingforce.”Thus,waterflows froma low tension inthesoil(lowpullingforce)toahightensionintheatmosphere(highpullingforce).Thewatertensionintheatmosphereoutsidetheleavesisdeterminedbyclimaticfactors:relativehumidity,windspeed,temperature,andsolarradiation.Thisatmospherictension translates into the “evaporativedemand”of the atmosphere, which determines potentialtranspiration rates. The water tension in the soilisdeterminedby theamountofwater in thesoilandbysoilphysicalpropertiessuchastextureandbulkdensity.Thespeedwithwhichwatermovesthroughtheplantisdeterminedbythedifferenceinwatertensionbetweenthesoilandtheatmosphere(thehigherthedifferences,thefasterthewaterwillflow)andbytheresistancetowaterflowintheplant(EhlersandGoss2003).First,soilwaterneedstoovercome a physical resistance (the epidermis)toenter the roots.Then itflows throughcellsorthroughspacesbetweencellsintoso-calledxylemorvascularbundlesthatwilltransportitupward.Thewaterflowseasierandfasterinwidebundles

Table 2.1. Units to express water potential and some corresponding values. pF is calculated as �0log(–H), where H is water height in cm.

Unit name Corresponding value

Water height (cm) 0 −10 −100 −1,000 −15,850pF (–) –∞ 1 2 3 4.2Bar (bar) 0 0.01 0.1 1 15.85Pascal (Pa) 0 1,000 10,000 100,000 1,585,000Kilo Pascal (kPa) 0 1 10 100 1,585Mega Pascal (MPa) 0 0.001 0.01 0.1 1.585

2

��

(withlessresistance)thaninnarrowbundles(withlargeresistance).Fromthebundles,thewaterflowsthrough the cells or spaces between the cells ofthe leaves to the“stomata”:smallcavities in theleavesthatconnecttotheoutsideworld(Fig.2.1).Thestomataarethelastbarrier(resistance)towa-terflowingoutoftheplant.Theprocessofwaterrelease through the stomata is called (stomatal)transpiration(thereisalsoacuticulartranspirationdirectlythroughcellsoftheleaves,butthisismuchlowerthanstomataltranspiration).Figure2.2givesanexampleofwaterpotentialsinthesoil-plant-at-mospheresystemaswatermovesgraduallyfromthesoilthroughtheplantintotheatmosphere.

Theflowofwater through the plant servesseveral purposes: it transports nutrients (in itsstream)fromthesoiltotheplantorganswheretheyareneeded,itprovidestheplantwithwaterinitscellssoitwillstayerect(thisiscalled“turgor”),andtranspirationcoolstheplantsoitdoesn’tgetoverheated.Plantscanactivelyregulatetherateofwaterflow(transpiration)byregulatingthesizeoftheopeningofthestomata.Ifthereisnotenoughwater in the soil to satisfy the demand from theatmosphere(thatis,togiveintothepullingforce

oftheatmosphere),theplantcancloseitsstomataandreduceorevencompletelystoptranspiration.Besides the reduction in transpiration, severalgrowth processes of the plant become affectedwhenthere isnotenoughwater.Weusuallycallthese“droughteffects,”andtheyaresummarizedinChapter2.2.SometypicaltensionlevelsofsoilwaterthatareimportantforuplandcropsaregiveninTable2.2.

Nonricesoilsusuallyhaveamixtureofwater,air,andsolidsoilparticles,andthewaterpotentialisnegative(positivetension).However,underfloodedconditions,asinthemuddylayerabovetheplowpanofricefields,thesoilissaturatedwithwaterand thepotential is positive.Negativepotentials(positivetensions)occurinthislayeronlywhenitdriesout.Generally,riceplantsexperiencetheshiftfromfloodedconditions(negativetensions)tonon-floodedconditions(positivetensions)as“droughtstress.”Afloodedsoilthatissaturatedwithwaterisalsocalledan“anaerobicsoil,”whereasasoilthatisnotsaturatedbuthasamixtureofairandwaterintheporesiscalledan“aerobicsoil.”Thewaterstatus of a soil hasmajor implications for watersupply to the crop and forpH (acidity), nutrient

Stomatal cavity

[CO2] air

Intercellular space

Mesophyll cell

Chloroplast

Boundary layer

Epidermis cell

CO2 H2O

[CO2] Chloroplast

[CO2] Stomatal cavity

Boundary layerresistance

Stomatalresistance

Mesophyllresistance

BA

Fig. 2.1. Schematic cross section of a leaf stomata (A) indicating components of resistance (B). Source: Lövenstein et al (1992).

��

availability,soilmicrobialpopulation,soilorganicmatterbuildup/decomposition,andoccurrenceofsoilpestsanddiseases(Chapter3.6). Figure2.3illustratessoilwatertensionsmeasuredinanaero-bicsoilwhere ricewasgrownundernonfloodedconditionslikeanuplandcrop.

Eachsoilhasaspecificrelationshipbetweenthetensionofthewaterandtheamountofwater:thelowertheamount,thehigherthetension.Thus,a small amount of water in the soil and a hightensionbothreflectaconditionofrelative“waterscarcity” anddrought.The relationship betweensoilwatertensionandsoilwatercontentiscalledthe“soilwaterretentioncurve”orthe“pFcurve.”

Theshapeof thepFcurvedependsonsoil type,especiallyonitstexture(mixtureofclay,silt,andsandparticles),bulkdensity(theweightofasoiloveritsvolume),mineralogy,andorganicmattercontent. Examples are given in Figure 2.4 for atypicalclaysoilandatypicalsandsoil.TherearefourimportantpointsonthepFcurveforuplandcrops:saturation=pF–∞(H=0cm),fieldcapac-ity=pF2,permanentwiltingpoint=pF4.2,andairdryness=pF7(seeTable2.2).Theamountofwaterateachofthesefourtensionsisdifferentforeach soil type.When ricefields areflooded, thepFcurveisnotneededassoilwatercontentsarealways at saturation. However, when a rice soil

Xylem Stomatal cavity

Endodermis

Water flow

Soil particle

Root

Stem

Leaf

Demand

Supply

Ψroot – Ψsoil Ψleaf – Ψroot Ψair – Ψleaf= =гroot гstem гleaf

Ψ Root

Ψ Soil

Ψ Leaf

Ψ Air

г Root

г Stem

г Leaf

–100 MPa

–1 MPa

–0.4 MPa

–0.02–0.1 MPa

Wat

er p

hase

Vapo

r pha

se

Fig. 2.2 Schematic overview of water flow and water potentials in the soil-plant-atmosphere continuum. Adapted from Lövenstein et al (1992).

Table 2.2. Some typical soil water tension levels in relation to the growth of upland crops (e.g., wheat, maize, cotton).

Name pF value Explanation

Saturation –∞ All soil pores are filled with waterField capacity 2 Soil water content that is considered optimum for upland cropsPermanent wilting point 4.2 Soil water content at which most upland crops cannot extract water from the soil any more and show permanent wiltingAir dryness 7 No free water in soil pores any more

��

100

Day number175 200 225 250 275 300

HarvestFloweringPanicleinitiation

Soil moisture tension (kPa)

90

80

70

60

50

4030

20

100

Fig. 2.3. Time course of soil water tension in the root zone of rice grown in a nonpuddled, nonflooded soil (this system is called “aerobic rice,” Chapter 3.4). The dips in tension are associated with rainfall or an irrigation event. Adapted from Yang Xiaoguang et al (2005).

7

6

5

4

3

2

1

00 0.1 0.2 0.3 0.4 0.5 0.6

Sand

Clay

Soil water tension (pF = log(–H))

Soil water content (cm3 cm–3)

Air dryness (pF 7)

Permanent wilting point (pF 4.2)

Field capacity (pF 2)

Saturation (pF –∞)

Fig. 2.4. Typical soil water retention curve (or pF curve) for a clay soil and a sand soil. Notice the differences in water content at the four critical tension levels (Table 2.2) for the two soil types. Adapted from Lövenstein et al (1992).

becomesdry,thepFcurvebecomesmeaningfulasitindicatestheamountofwateravailableforplantuptakeatdifferenttensions.

2.2 The rice plant and droughtCultivated rice evolved from a semiaquatic per-ennial ancestor (Lafitte andBennett 2002).Thewetlandancestryofriceisreflectedinanumberofmorphological and physiological characteristicsthatareuniqueamongcropspecies.Lowlandrice

isextremelysensitivetowatershortageanddroughteffectsoccurwhensoilwatercontentsdropbelowsaturation.Rice has a variety ofmechanismsbywhichitreactstosuchconditions.ThefollowingisalistcompiledbyBoumanandTuong(2001):

1. Inhibitionofleafproductionanddeclineinleaf area, leading to retarded leaf growthandlightinterception,andhencetoreducedcanopyphotosynthesis.Droughtstressaf-fects both cell division and enlargement,thoughcelldivisionappearstobelesssensi-

��

tivetowaterdeficitthancellenlargement.Leafareaexpansionisreducedassoonasthe soil dries below saturation (tensionshigher than1kPa) inmostcultivars, andwhenonlyabout30%oftheavailablesoilwaterhasbeenextractedincultivarswithaerobicadaptation(LilleyandFukai1994,Wopereis etal1996).

2. Closureofstomata,leadingtoreducedtran-spirationrateandreducedphotosynthesis.Leafstomatadonotcloseimmediatelywithdroughtstress,however,andthecropkeepsonphotosynthesizing for a certainperiodbefore stomata close.The assimilates arenotusedforleafgrowthorexpansion(seepoint1),butarestoredintheexistingleaves,stems, and roots. When drought stress isrelieved, these assimilatesmay becomeavailableandleadtoaflushinleafgrowth.Inthemodernhigh-yieldingvarietyIR72,stomatalclosurestartsatsoilwatertensionsof75kPa(Wopereisetal1996).

3. Leafrolling, leadingtoareduction inef-fectiveleafareaforlightinterception.LeafrollinginIR72startsatsoilwatertensionsof75kPa(Wopereisetal1996).Leavesun-rollagainwhendroughtstressisrelieved.

4. Enhanced leaf senescence, leading to re-duced canopy photosynthesis. EnhancedsenescenceinIR72startsatsoilwaterten-sionsof630kPa(Wopereisetal1996).

5. Changes in assimilate partitioning. Rootsgrow more, at the expense of the shoot,during vegetative development, whereaspartitioningofassimilatesamongvariousshootcomponentsisnotaffected.Deeperrootsareeffectiveforexploringwaterstoredindeepersoillayers.

6. Reducedplantheight(thoughitisnotlikelythatreducedplantheightinitselfwillresultinyieldreduction).

7. Delayedflowering.Droughtinthevegeta-tivedevelopmentstagecandelayfloweringupto3to4weeksinphotoperiod-insensi-tive varieties. The delay in flowering islargestwithdroughtearlyinthevegetativestageandissmallerwhendroughtoccurslater.

8. Reducedtilleringandtillerdeath.Droughtbeforeorduringtilleringreducesthenumberoftillersandpaniclesperhill.Ifthedroughtis relieved on time, and the source size(i.e.,photosynthesizingleavesandstems)issufficientlylarge,thereducednumberoftillers/paniclesmaybecompensatedforbyanincreasednumberofgrainsperpanicleand/orbyanincreasedgrainweight.

9. Reducednumberofspikeletswithdroughtbetweenpanicle initiation andflowering,resultingindecreasednumberofgrainsperpanicle.

10. Rice is very sensitive to reduced wateravailabilityintheperiodaroundfloweringasthisgreatlyaffectsspikeletsterility(CruzandO’Toole1984,Ekanayakeetal1989).Increasedspikeletsterilitywithdroughtatflowering results in decreasedpercentageoffilledspikeletsand,therefore,decreasednumberofgrainsperpanicle.Especiallyatanthesis, there is a short time span whenspikelet fertility is especially sensitive todrought.

11. Decreasedgrainweightwithdroughtafterflowering.

Theaboveprocessesappearroughlyinorderof crop development and/or severity of drought,thoughnumbers2–4alsooccurinthereproductivestage.Someeffectsleadtoirreversibleprocessesofyieldreduction,suchasnumbers4,9,10,and11,whereasothersmayberestoredwhendroughtis relieved,suchasnumbers2and3,andothersmaybecompensatedforbyothereffectslaterinthegrowingseason,suchasnumbers1,2,and8.Droughtmayalsoaffectnutrient-useefficiencybythecropsincewaterflowistheessentialmeansofnutrienttransport.Howyieldisfinallyaffectedbydroughtdependsonitstiming,severity,duration,andfrequencyofoccurrence.Themostsensitivestageofricetodroughtisaroundflowering.

��

Coping with water scarcity

Inthischapter,wepresenttechnologyoptionstohelpfarmerstocopewithwaterscarcityatthefieldlevel.Thewaytodealwithreduced(irrigationorrain)water inflows to ricefields is to reduce thenonproductiveoutflowsbyseepage,percolation,orevaporation,whilemaintainingtranspirationflows(asthesecontributetocropgrowth).Thiscanbedoneatlandpreparation,atcropestablishment,andduringtheactualcropgrowthperiod.

3.1 Land preparationLandpreparationlaysthefoundationforthewholecroppingseasonanditisimportantinanysituationto“getthebasicsright.”Especiallyimportantforgoodwatermanagement arefieldchannels, landleveling,andtillageoperations(puddling,andbundpreparationandmaintenance).

3.1.1 Field channels Many irrigation systems inAsia have no fieldchannels(or“tertiary”irrigationordrainagechan-nels)andwaterflowsfromonefieldintotheotherthroughbreachesinthebunds.Thisiscalled“plot-to-plot” irrigation.The amount ofwaterflowinginandoutofaricefieldcannotbecontrolledandfield-specificwatermanagement is not possible.Thismeansthatfarmersmaynotbeabletodraintheir fields before harvest becausewater keepsflowinginfromotherfields.Also,theymaynotbeabletohavewaterflowinginifupstreamfarmersretainwater in their fields or let their fields dryouttoprepareforharvest.Moreover,anumberoftechnologies to cope with water scarcity requiregoodwatercontrolfor individualfields(Chapter3.5). Finally, thewater that continuously flowsthroughricefieldsmayremovevaluable(fertilizer)

nutrients.Constructingseparatechannelstoconveywatertoandfromeachfield(ortoasmallgroupoffields)greatlyimprovestheindividualcontrolofwaterandistherecommendedpracticeinanytypeofirrigationsystem.

3.1.2 Land levelingAwell-leveledfieldisaprerequisiteforgoodcrophusbandry.Whenfieldsarenot level,watermaystagnateindepressions,whereashigherpartsmaybecomedry.Thisresultsinunevencropemergenceandunevenearlygrowth,unevenfertilizerdistribu-tion,andmaybeextraweedproblems.Informationontechnologiesforlandlevelingcanbefoundatwww.knowledgebank.irri.org.

3.1.3 Tillage: reducing soil permeabilitySeepageandpercolationflowsfromricefieldsaregovernedbythepermeability(hydraulicconductiv-ity)oftheirsoils:theircapacitytoconductwaterdownwardandsideward(Chapter1.3).Aricefieldcanbecomparedtoabathtub:thematerialofabath-tubisimpregnableanditholdswaterwell—how-ever,ifyouhaveonlyonehole(byremovingtheplug),thewaterrunsoutimmediately.Ricefieldsjustneedafewratholesorleakyspotsandtheywillrapidlylosewaterbyseepageandpercolation.

Large amounts ofwater can be lost duringsoaking prior to puddlingwhen large and deepcracksarepresentthatfavorrapid“by-passflow”tobelowtherootzone.CabangonandTuong(2000)showedthebeneficialeffectsofadditionalshallowsoiltillagebeforelandsoakingtoclosethecracks:theamountofwaterusedinwetlandpreparationwasreducedfromabout350mmtoabout250mm(Fig.3.1).

3

��

Thorough puddling results in a good com-pactedplowsolethatreducespermeabilityandper-colationratesthroughoutthecropgrowingperiod(Chapter1.3;DeDatta1981,Tuongetal1994).Theefficacyofpuddling in reducingpercolationdependsgreatlyonsoilproperties.Puddlingmaynotbeeffectiveincoarsesoils,whichdonothaveenoughfine clay particles tomigrate downwardandfillupthecracksandporesintheplowsole.Ontheotherhand,puddlingisveryefficientinclaysoilsthatformcracksduringthefallowperiodthatpenetratetheplowpan(Tuongetal1994).Althoughpuddlingreducespercolationratesofthesoil,theactionofpuddlingitselfconsumeswater,andthereisatrade-offbetweentheamountofwaterusedforpuddlingandtheamountofwater“saved”duringthecropgrowthperiodbyreducedpercolationrates.Puddlingmaynotbenecessaryinheavyclaysoils

withlowverticalpermeabilityorlimitedinternaldrainage.Insuchsoils,directdryseedingonlandthatisnotpuddledbuttilledinadrystateisverywellpossiblewithminimalpercolationlosses(Tab-baletal2002;Chapter3.2).

Soil compaction using heavy machineryhasbeen shown to decrease soil permeability insandy soilwith loamy subsoilswith at least 5%clay(Harnpichitvitayaetal2000).Althoughmostfarmerscannotaffordtocompacttheirsoils,thistechnologymaybefeasibleonalargescalewithgovernmentsupport.

3.1.4 Bund preparation and maintenanceGood bunds are a prerequisite to limit seepageandunderbundflows(Tuongetal1994).Tolimitseepage losses,bundsshouldbewellcompactedand any cracks or rat holes should be plasteredwithmud at the beginning of the crop season.Makebundshighenough(atleast20cm)toavoidoverbundflowduringheavyrainfall.Smallleveesof5–10-cmheightinthebundscanbeusedtokeeppondedwaterdepthat thatheight.Ifmorewaterneedstobestored,itisrelativelysimpletoclosetheselevees.Researchershaveusedplasticsheetsinbundsinfieldexperimentstoreduceseepagelosses.Forexample,Boumanetal (2005)demonstratedareductionof450mmintotalwateruseinaricefield by lining the bundswith plastic (Fig. 3.2).Althoughsuchmeasuresareprobablyfinanciallynotattractivetofarmers,theauthorcameuponafarmerintheMekongDeltainVietnamwhousedoldplastic sheets to block seepage throughveryleakypartsofhisbunds.

3.2 Crop establishmentMinimizing the turnaround time between landsoaking forwet landpreparation and transplant-ing reduces the period when no crop is presentandwhenoutflowsofwaterfromthefielddonotcontributetoproduction.Especiallyinlarge-scaleirrigationsystemswithplot-to-plotirrigation,waterlossesduringtheturnaroundtimecanbeveryhigh.Forinstance,inthelargestsurfaceirrigationschemeinCentralLuzon,calledUPRIIS(UpperPampangaRiverIntegratedIrrigationSystem),ittookupto63daysinacontiguous145-hablockfromthefirstdayofwaterdeliveryforlandpreparationuntilthewhole areawas completely transplanted (Tabbaletal2002).Thetotalamountofwaterinputdur-

Control Cracks plowed

Water input (mm)

200

250

300

350400

150

100

500

Fig. 3.1. Effect of shallow tillage to fill cracks before soaking on water input during land preparation, Bulacan, Philippines. Data from Cabangon and Tuong (2000).

No lining Lining

Water input (mm)

800

1,0001,200

1,4001,600

600400

2000

1,8002,000

RainfallIrrigation

Fig. 3.2. Effect on total water input of lining bunds with plastic in a field experiment at IRRI, Los Baños, Philippines. Data from Bouman et al (2005).

��

ing that timewas some940mm, ofwhich 110mmwasused for soaking, 225mmdisappearedas surface runoff, 445mmwas lost by seepageandpercolation,and160mmwaslostbyevapora-tion.InUPRIIS,farmersraiseseedlingsinpartoftheirmainfield.Becauseofalackoftertiaryfieldchannels,thewholemainfieldissoakedwhentheseedbed is prepared and remainsfloodedduringtheentiredurationoftheseedbed.InsystemssuchasUPRIIS,theturnaroundtimecanbeminimizedbytheinstallationoffieldchannels,theadoptionofcommonseedbeds,ortheadoptionofdirectwetordryseeding.Withfieldchannels,watercanbedeliveredtotheindividualseedbedsseparatelyandthemainfielddoesnotneedtobeflooded.Commonseedbeds,eithercommunalorprivatelymanaged,canbelocatedstrategicallyclosetoirrigationcanalsandbeirrigatedasoneblock.

Withdirect seeding, thecrop startsgrowingand usingwater from themoment of establish-mentonward.DirectdryseedingcanalsoincreasetheeffectiveuseofrainfallandreduceirrigationneedsasshownfortheMUDAirrigationschemeinMalaysia(Cabangonetal2002).However,dryseedingwithsubsequentfloodingispossibleonlyinheavy(clayey)soilswithlowpermeabilityandpoorinternaldrainage.AmajordrivingforcefortheadoptionofdirectseedinginAsia isscarcityoflaborsincedirectseedingdoesnotuselaborfortransplantingandcanbeamechanizedoperation.

3.3 Crop growth period�.�.� Saturated soil cultureInsaturatedsoilculture(SSC),thesoiliskeptasclose to saturation as possible, thereby reducingthehydraulicheadofthepondedwater,whichde-creasestheseepageandpercolationflows.SSCinpracticemeansthatashallowirrigationisgiventoobtainabout1cmofpondedwaterdepthadayor

soafterthedisappearanceofpondedwater.Tabbaletal(2002)reportedwatersavingsunderSSCintransplantedanddirectwet-seededriceinpuddledsoil,andindirectdry-seededriceinnonpuddledsoil(Table3.1).Analyzingadatasetof31publishedfieldexperimentswithanSSCtreatment,BoumanandTuong(2001)foundthatwaterinputdecreasedonaverageby23%(range:5%to50%)fromthecontinuouslyfloodedcheck,withanonsignificantyieldreductionof6%onaverage.Thompson(1999)found that SSC in southernNewSouthWales,Australia,reducedbothirrigationwaterinputandyieldbyabitmorethan10%.

Raisedbedscanbeaneffectivewaytokeepthesoilaroundsaturation.Riceplantsaregrownonbedsandthewaterinthefurrowsiskeptclosetothesurfaceofthebeds.InAustralia,Borelletal(1997)experimentedwithraisedbedsthatwere120cmwideandseparatedbyfurrowsof30-cmwidthand15-cmdepthtofacilitateSSCpractices.Comparedtofloodedrice,watersavingswere34%andyieldlosses16−34%.MoreinformationonraisedbedsisfoundinChapter3.4.1.

Practical implementationAlthough conceptually sound, SSC will be difficult to implement practically since it requires frequent (daily or once every two days) applications of small amounts of irrigation water to just keep a standing water depth of 1 cm on flat land, or to keep furrows filled just to the top in raised beds.

3.3.2 Alternate wetting and dryingInalternatewettinganddrying(AWD),irrigationwaterisappliedtoobtainfloodedconditionsafteracertainnumberofdayshavepassedafterthedisap-pearanceofpondedwater.ThenumberofdaysofnonfloodedsoilinAWDbeforeirrigationisappliedcanvaryfrom1daytomorethan10days.Though

Table 3.1a. Yield, water input, and water productivity with respect to total water input (WPIR) in transplanted and wet-seeded rice under continuous flooding and SSC, Muñoz, 1991 dry season. Data from Tabbal et al (2002).

Transplanted Wet-seeded Treatment Yield Water input WPIR Yield Water WPIR

(t ha−1) (mm) (g grain kg−1 water) (t ha−1) input (mm) (g grain kg−1 water)

Flooded 7.4 694 1.06 7.6 631 1.20SSC 6.7 373 1.81 7.3 324 2.27

�0

some researchers have reported a yield increaseusingAWD (WeiZhang andSong1989, Stoopetal2002),recentworkindicates that this is theexceptionratherthantherule(Belderetal2004,Cabangonetal2004,Tabbaletal2002;Table3.2).In31fieldexperimentsanalyzedbyBoumanandTuong(2001),92%oftheAWDtreatmentsresultedinyieldreductionsvaryingfromjustmorethan0%to70%comparedwiththoseofthefloodedchecks.Inallthesecases,however,AWDincreasedwaterproductivity(WPIR)withrespecttototalwaterinputbecausethereductionsinwaterinputswerelargerthanthereductionsinyield.ThelargevariabilityinresultsofAWDintheanalyzeddatasetwascausedbydifferencesinthenumberofdaysbetweenirriga-tionsandinsoilandhydrologicalconditions.

Experimenting with AWD in lowland riceareaswithheavy soils and shallow groundwatertables inChina and the Philippines,Cabangonetal(2004),Belderetal(2004),Lampayanetal(2005),andTabbaletal(2002)reportedthattotal(irrigationandrainfall)waterinputsdecreasedbyaround 15–30%without a significant impact onyield.Inallthesecases,groundwaterdepthswereveryshallow(between10and40cm),andpondedwaterdepthsalmostneverdroppedbelowtherootzoneduringthedryingperiods(Fig.3.3),thusturn-ingAWDeffectivelyintoakindofnear-saturatedsoilculture.Evenwithoutpondedwater,plantrootsstillhadaccessto“hidden”waterintherootzone(Chapter1.4).Morewatercanbesavedandwaterproductivity further increasedbyprolonging the

Table 3.1b. Yield, water input, and water productivity with respect to total water input (WPIR) and irrigation (WPI) in dry-seeded rice under continuous flooding and SSC, San Jose City, Philippines, 1996-97. Data from Tabbal et al (2002).

Water input (mm) Water productivity (g grain kg−1 water)Treatment Yield (t ha−1) Irrigation + rainfall Irrigation WPIR WPI

1996 Flooded 4.3 1,417 531 0.31 8.16 SSC 4.2 1,330 432 0.32 9.651997 Flooded 4.7 1,920 941 0.25 4.99 SSC 4.5 1,269 355 0.36 12.81

Table 3.2. Yield, water use, and water productivity with respect to irrigation and rainfall of rice under alternate wetting and drying (AWD) and continuously flooded conditions. Data from Bouman et al (2006a).

Location Year Treatment Yield Total water Water productivity (t ha−1) input (mm) (g grain kg−1 water)

Guimba, Philippines 1988 Flooded 5.0 2,197 0.23(Tabbal et al 2002) AWD 4.0 880 0.46 1989 Flooded 5.8 1,679 0.35 AWD 4.3 700 0.61 1990 Flooded 5.3 2,028 0.26 AWD 4.2 912 0.46 1991 Flooded 4.9 3,504 0.14 AWD 3.3 1,126 0.29

Tuanlin, Huibei, China 1999 Flooded 8.4 965 0.90(Belder et al 2004) AWD 8.0 878 0.95 2000 Flooded 8.1 878 0.92 AWD 8.4 802 1.07

Muñoz, Philippines, 2001 Flooded 7.2 602 1.20(Belder et al 2004) AWD 7.7 518 1.34

��

periodsofdrysoilandimposingaslightdroughtstressontheplants,butthisusuallycomesattheexpenseofyieldloss(BoumanandTuong2001).Researchinmoreloamy and sandy soils with deeper groundwater tables in India and thePhilippinesshowedreductionsinwaterinputsofmorethan50%coupledwithyieldlossofmorethan20%comparedwiththefloodedcheck(Sharmaetal2002,Singhetal2002,Tabbaletal2002).

AWD is amature technology that has beenwidelyadoptedinChina(LiandBarker2004).ItisalsoarecommendedpracticeinnorthwestIndia,andisbeingtestedbyfarmersinthePhilippines.Figures3.4and3.5giveanexampleofyieldsandwaterinputsobtainedbyfarmersinCentralLuzon,Philippines,whopracticedAWDirrigation(Lam-payan et al 2005). The farmers used communaldeep-well pumps or their own shallow tubewellpumps to irrigate theirfields.Theydivided theirfieldsintotwo,onewithAWDmanagementandtheotherwithcontinuousflooding.Table3.3givesaneconomiccomparisonamongAWDandcontinu-ouslyfloodedfields.TheAWDfieldshadthesameyieldascontinuousflooding,butsaved16–24%inwatercostsand20–25%inproductioncosts.

VerylittleresearchhasbeendonetoquantifytheimpactofAWDonthedifferentwateroutflowsofricefields:evaporation,seepage,andpercolation.The littlework done so far suggests thatAWD

150

50

–50

–150

–250

–350

140 160 180

Ponded water depth (mmm)

200 220 240

Day number

Flowering HarvestTransplanting

Fig. 3.3. Depth of ponded water on the soil surface (above the horizontal line) and below the soil surface (below the horizontal line) in an AWD experiment, Tuanlin, Hubei, China. Adapted from Belder et al (2004).

1,400Water input (mm)

2002 2003

2002 20031,200

1,000

800

600

400

200

0

1,400Water input (mm)

1,2001,000

800

600

400200

0

Upp

er

Mid

dle

Low

er

Aver

age

Upp

er

Mid

dle

Low

er

Aver

age

Dol

ores

Gab

aldo

n

Aver

age

Dol

ores

Gab

aldo

n

Aver

age

Fig. 3.4. Total water input (mm) under AWD (shaded columns) and continuous flooding (black columns) in farmers’ fields in Tarlac (A) and Nueva Ecija (B), Philippines, 2002 and 2003 dry seasons. Data from Lampayan et al (2005).

A

B

��

mostlyreducesseepageandpercolationflowsandhasonlyasmalleffectonevaporationflows.Belderetal(2007)andCabangonetal(2004)calculatedthatevaporationlossesdecreasedby2–33%com-paredwithfullyfloodedconditions.

ThefollowingpotentialbenefitsofAWDhavebeensuggested:improvedrootingsystem,reducedlodging(becauseofabetterrootsystem),periodicsoilaeration,andbettercontrolofsomediseasessuchasgoldensnail.Ontheotherhand,ratsfinditeasiertoattackthecropduringdrysoilperiods.

3.3.3 The System of Rice IntensificationAWD is the water management practice of theSystemofRiceIntensification(SRI),anintegratedcropmanagement technology developed by theJesuitpriestFatherHenrideLaulanieinMadagas-car(Stoopetal2002).SRIischaracterizedbythefollowingpractices(Uphoff2007):

Transplanting8-to12-day-oldseedlingsverycarefully(roottipdown)

Transplantingsingleseedlings Spacingtheplantswidelyapartinasquare

pattern(25×25cmorwider) Controllingweedsbyweedingwitharotating

hoe,whichaeratesthesoil Applyingcomposttoincreasethesoil’sor-

ganicmattercontent(optional)

Grain yield (t ha–1)

Upp

er

Mid

dle

Low

er

Aver

age

Upp

er

Mid

dle

Low

er

Aver

age

2002 2003

9876543210

9876543210

Yield (t ha–1)

Dol

ores

Gab

aldo

n

Aver

age

Dol

ores

Gab

aldo

n

Aver

age

2002 2003

A

B

Fig. 3.5 Average yield under AWD (shaded columns) and continuous flooding (black columns) in farmers’ fields in Tarlac (A) and Nueva Ecija (B), Philippines, 2002 and 2003 dry seasons. Data from Lampayan et al (2005).

Table 3.3. Average cost and returns of rice grown under AWD and FP (farmers’ practice = continuously flooded) in farmers’ fields at three sites in Central Luzon, Philippines, in 2002 and 2003 dry seasons. Data from Lampayan et al (2005).

Tarlac Nueva Ecija (Gab) Nueva Ecija (Dol)Item

FP AWD FP AWD FP AWD

2002 Gross returns ($ ha−1) 933 933 1,183 1,292 1,073 1,043 Production cost ($ ha−1) 441 397 1,030 870 597 574 Net profit ($ ha−1) 491 535 152 422 477 469 Net profit-cost ratio 1.1 1.3 0.1 0.5 0.8 0.8

2003 Gross returns ($ ha−1) 1,134 1,105 Production cost ($ ha−1) 519 491 Net profit ($ ha−1) 615 614 Net profit-cost ratio 1.2 1.2

��

No continuous flooding during the cropgrowth period, applying small amounts ofwater regularly or alternating wet and dry(AWD)field conditions tomaintain amixof aerobic and anaerobic soil conditions.Afterflowering,athinlayerofwatershouldbekeptonthefield,althoughsomefarmersfind alternatewetting and drying of fieldsthroughoutthecropcycletobefeasibleandevenbeneficial.

The benefits of SRI are the topic of scien-tificdebateandcontroversy(SinclairandCassman2004, Surridge 2004). Proponents of SRI claimthatSRIfarmersareusuallyable toattainyields

of7–8tha–1,while,withproperandsustaineduseoftheSRImethod,yieldscangobeyond15tha–1(Uphoff2007).Alargenumberofsuccessstoriesand “testimonies” are reported on theSRIWebsite(http://ciifad.cornell.edu/sri/).Independentre-searchers,however,havenotbeenabletoconfirmthesehighyieldsandsuccessstories(Dobermann2003, Sheehy et al 2004). When compared withlocalbest-managementpractices,SRIusuallygivessimilarorevenloweryields.InBangladesh,inanon-stationcomparison,riceyieldwithmanagementpractices recommendedby theBangladeshRiceResearchInstitute(BRRI)wassignificantlyhigherthanunderSRImanagement(Latifetal2005).Inon-farm trials, the BRRI-recommended manage-mentperformedsignificantlybetterthanSRIandresulted in higher yield, lower cost, and higherprofit.SRIisespeciallylabor-intensive.Onaver-age,SRIrequired13%morelaborthantheBRRIpracticeand19%morethanthefarmers’practice.McDonald et al (2006) analyzed 40 site-yearsof SRI versus local best-management practices(BMP)thatincludeddatafromMadagascar,Nepal,China,Thailand,Laos,India,SriLanka,Indonesia,Bangladesh,andthePhilippines.AsidefromonesetofexperimentsinMadagascar,theyfoundnoevidenceofasystematicyieldadvantageofSRI.Noneof the35experimental recordsbesides theMadagascar cases demonstrated yield increasesthatexceededBMPbymorethan22%.ExcludingtheMadagascarcases,thetypicalSRIoutcomewasevennegative,with24of35site-yearshavingonaverage11%loweryieldsthanBMP.McDonaldetal(2006)foundnoevidencethatSRIfundamentallychanges thephysiologicalyieldpotentialof rice,and advocated caution in extendingSRIbeyondits origin of development. Investigating the useof SRI in the original country of development,Madagascar,MoserandBarett(2003)reportedadisadoption of the system, partly because of itsrelatively high labor requirements and the highneedsforextensionsupport.

For the purpose of water savings, we rec-ommend the combination ofAWDwithwell-researched and site-specific bestmanagementpractices(integratedcropmanagement).

3.4 Aerobic riceAfundamentallydifferentapproachtoreducewateroutflowsfromricefieldsistogrowthecroplikean

Practical implementation of AWDA practical way to implement AWD is to monitor the water depth in the field using the “field water tube” described in the Appendix (A.1). After an irrigation application, the field water depth will gradually decrease over time. When the water level (as measured in the tube) is 15 cm below the surface of the soil, it is time to irrigate and flood the soil with a depth of around 5 cm. Around flowering, from 1 week before to 1 week after the peak of flowering, ponded water should be kept at 5-cm depth to avoid any water stress that would result in potentially severe yield loss (Chapter 2.2). The threshold of 15 cm is called “Safe AWD” as this will not cause any yield decline since the roots of the rice plants will still be able to take up water from the saturated soil and the perched water in the root zone. The field water tube helps farmers see this “hidden” source of water. In Safe AWD, water savings may be relatively small, on the order of 15%, but there is no yield penalty. After creating confidence that Safe AWD does not reduce yield, farmers may experiment by lowering the threshold level for irrigation to 20 cm, 25 cm, 30 cm, or even deeper. Some yield penalty may be acceptable when the price of water is high or when water is very scarce (Chapters 1.7 and 3.5). Remember that, in many irrigated areas, the groundwater is very shallow and may reach into the field water tube! In Safe AWD, the following rules should be observed. AWD irrigation can be used from a few days after transplanting (or a 10-cm-tall crop after direct seeding) till first heading. In the period of first heading to 1 week after flowering, keep the field flooded with 5-cm depth. After that, during grain filling and ripening, apply AWD again. When many weeds are present in the early stages of crop growth, the implementation of AWD can be postponed for 2–3 weeks until weeds have been suppressed by the ponded water. Under Safe AWD, no special N management regime is needed and local recommendations as for flooded rice can be used (Belder et al 2004). Apply fertilizer N preferably on the dry soil just before irrigation is applied.

��

uplandcrop,suchaswheatormaize.Unlikelow-landrice,uplandcropsaregrowninnonpuddled,nonsaturated(i.e.,“aerobic”)soilwithoutpondedwater.Whenrainfallisinsufficient,irrigationisap-pliedtobringthesoilwatercontentintherootzoneup tofieldcapacityafter ithasreachedacertainlowerthresholdlevel,suchashalfwaybetweenfieldcapacity andwiltingpoint (Doorenbos andPruit1984).Theamountofirrigationwatershouldmatchevaporationfromthesoilandtranspirationbythecrop(plusanyapplicationinefficiencylosses).Thepotentialwaterreductionsatthefieldlevelwhenricecanbegrownasanuplandcroparelarge,es-peciallyonsoilswithhighseepageandpercolationrates(Bouman2001;Chapter3.5).Besidesseepageandpercolationlossesdeclining,evaporationde-creasessincethereisnopondedwaterlayer,andthelargeamountofwaterusedforwetlandpreparationiseliminatedaltogether(Chapter3.1.3).

InAsia,“uplandrice”isalreadygrownaero-bicallywithminimalinputsintheuplandenviron-ment,butmostlyasalow-yieldingsubsistencecroptogivestableyieldsundertheadverseenvironmen-talconditionsoftheuplands(Lafitteetal2002).Uplandricevarietiesaredroughttolerant,buthavealowyieldpotentialandtendtolodgeunderhighlevelsofexternalinputssuchasfertilizerandsup-plemental irrigation. Alternatively, high-yieldinglowland rice varieties grownunder aerobic soilconditions,butwithsupplementalirrigation,havebeen shown to savewater, but at a severe yieldpenalty(Blackwelletal1985,WestcottandVines

1986,McCauley 1990).Achieving high yieldsunderirrigatedbutaerobicsoilconditionsrequiresnewvarieties of “aerobic rice” that combine thedrought-tolerantcharacteristicsofuplandvarietieswith thehigh-yieldingcharacteristicsof lowlandvarieties(Lafitteetal2002,Atlinetal2006).

The development of temperate aerobic ricestartedinthemid-eightiesinnorthernChinaandBrazil.InChina,breedershaveproducedaerobicricevarietieswithanestimatedyieldpotentialof6-7tha–1(WangHuaqietal2002).InexperimentswithChineseaerobicricevarietiesclosetoBeijingin2001and2002,YangXiaoguangetal(2005)andBoumanetal(2006b)obtainedaerobicriceyieldsof2.5−5.7tha−1withonly500−900mmoftotal(irrigationplusrainfall)waterinput(Table3.4).Forcomparison,theaerobicvarietiesyielded5.4−6.8tha−1underfloodedlowlandconditions,receivingabout1,300mmoftotalwaterinput.Atthesamesite,Xue et al (2007) reported yieldmaximaof3.6−4.5tha−1with688mmoftotalwaterinputin2003,and6.0tha−1with705mmofwaterinputin2004(Table3.5).TherelativelyhighyieldsofaerobicriceatBeijingwereobtainedunder“harsh”conditions for growing rice. The soil containedmore than 80% sand,was permeable, and heldwater abovefield capacity for a fewhours afterirrigationonly.Thegroundwatertablewasdeeperthan20m,thesoilmoisturecontentintherootzonewasmostlybetween50%and80%ofsaturation,andsoilmoisturetensionswentupto90kPa(seeFig.2.3inChapter2.1).Infieldexperimentsnear

Table 3.4. Water input (I = irrigation R = rainfall) and yield of two aerobic rice varieties (HD502, HD297) under flooded and aerobic conditions in 2001 and 2002, Beijing, China. Data from Yang Xiaoguang et al (2005).

Water Water input (mm) Yield (t ha−1)Year management I I + R HD502 HD297

2001 Flooded 1,057 1,351 6.8 5.4 aerobic 350 644 5.3 4.7 283 577 4.6 4.3 292 586 4.3 4.2 225 519 3.5 3.4 175 469 3.0 2.52002 Flooded 900 1,255 4.6 5.3 aerobic 522 917 5.7 5.3 374 769 4.8 4.7 225 620 4.0 3.9 300 695 4.3 4.6 152 547 3.6 2.9

��

Table 3.6. Performance of aerobic rice grown by farmers around Kaifeng, northern China, in terms of yield and water use, 2002-03. Data from Bouman et al (2007).

Farmer (code letter) A B C D E F G

Yield (t ha−1) 3.8 4.4 3.8 5.1 5.5 4.7 3.4Irrigation (mm) 225 225 80 231 230 300 225Rainfall (mm) 337 337 337 337 337 337 337Total water input (mm) 562 562 417 568 567 637 562

Farmer (code letter) U V W X Y Z

Yield (t ha−1) 1.2 3.8 2.4 3.8 3.6 3.0 Irrigation (mm) 156 159 145 169 146 162 Rainfall (mm) 674 674 674 674 674 674 Total water input (mm) 830 833 819 843 820 836

KaifengintheNorthChinaPlain,FengLipingetal(2007)obtainedrelativelylowyieldsof2.4−3.6tha−1with750−1,000mmoftotalwaterinput.Itisestimatedthataerobicricesystemsarecurrentlybeing pioneered by farmers on some80,000 hainnorthernChinausingsupplementaryirrigation(WangHuaqi et al 2002).Bouman et al (2007)reportedyieldsofaerobicriceobtainedbyfarmersaroundKaifengofupto5.5tha−1withsometimesaslittleas566mmoftotalwaterinput,withonlyoneor twosupplementary irrigationapplications(Table3.6).Table3.7comparestheperformanceindicatorsofaerobicrice,lowlandrice,andmaizeobtainedbyfarmersinthesamearea.Simulationmodelpredictionsevensuggestedthatnoirrigationwouldbeneededforhighyieldswithsome400−600mmofrainfallandgroundwatertablesof2mdeepandless.InBrazil,abreedingprogramtoimproveuplandricehasresultedinaerobicvarietieswithayieldpotentialofupto6tha–1(Piñheiroetal2006).

Farmersgrowthesevarietiesinrotationwithcropssuchassoybeanandfodderonlargecommercialfarmswithsupplementalsprinklerirrigationonanestimated250,000haofflatlandsintheCerradoregion,realizingyieldsof3−4tha−1.

Thedevelopmentoftropical aerobic riceisofrelativelyrecentorigin.DeDattaetal(1973)grewlowlandvarietyIR20inaerobicsoilunderfurrowirrigationatIRRIinthePhilippines.Watersavingswere55%comparedwithfloodedconditions,butyieldfellfromabout8tha−1underfloodedcondi-tionsto3.4tha−1underaerobicconditions.Usingimproveduplandricevarieties,Georgeetal(2002)reported aerobic rice yields of 1.5−7.4 t ha−1 inuplandswith2,500to4,500mmofannualrainfallin the Philippines. Yields of 6 t ha−1 and more,however,wererealizedonlyincidentallyinthefirstyearsofcultivation,andmostyieldswereinthe2−3tha−1range.Atlinetal(2006)reportedaerobicriceyieldsof3−4tha−1usingrecentlydevelopedaerobic

Table 3.5. Water input (I = irrigation, R = rainfall) and yield of aerobic rice variety HD297 under aerobic conditions in 2003 and 2004, Beijing, China. Data from Xue Changying et al (2007).

2003 2004

Water input (mm) Yield (t ha−1) Water input (mm) Yield (t ha−1)

I I + R Exp 1 Exp 2 I I + R

408 688 4.25 3.11 535 705 5.58408 618 3.70 2.54 535 675 5.35408 648 2.11 1.26 535 645 5.35408 578 − 0.46 535 605 4.99

�6

ricevarietiesinfarmers’fieldsinrainfeduplandsinthePhilippines.Thoughtheamountofrainfallwasnotreported,theconditionsofthetrialsweredescribedas“wellwatered.”Boumanetal(2005)andPengetal (2006)quantifiedyieldandwateruseof the recently released tropical aerobic ricevarietyApounder irrigated aerobic andfloodedconditions.Inthedryseason,yieldsunderaerobicconditionswere4−5.7tha−1andinthewetseasontheywere3.5−4.2tha−1.Theseyieldswereobtainedin relatively wet soil with seasonal-average soilmoisture tensions in the rootzoneof10−12kPaandwithmaximumvaluesofaround40kPa.Onaverage,themeanyieldofallvarietieswas32%lowerunderaerobicconditionsthanunderflooded

conditionsinthedryseasonand22%lowerinthewet season.Totalwater inputwas 1,240−1,880mminfloodedfieldsand790−1,430mminaerobicfields.Onaverage,aerobicfieldsused190mmlesswaterinlandpreparationandhad250−300mmlessseepageandpercolation,80mmlessevaporation,and25mmlesstranspirationthanfloodedfields.Successful examples of the adoption of aerobicricebyfarmersinthetropicsareinsomerainfeduplandsinBatangasProvince,Philippines(Atlinetal2006).InthehillyregionsofYunnanProvince,southernChina,farmersgrowrainfedaerobicriceunderintensifiedmanagement,realizingyieldsof3−4tha−1(Atlinetal2006).

Table 3.7. Average performance of aerobic rice, lowland rice, and maize near Kaifeng, northern China, 2002-03. Unpublished data from China Agricultural University and IRRI.

Item Lowland rice Aerobic rice Maize

Number of farmers, 2002 5 7 3 Field size (ha) 0.12 0.12 0.15 Yield (t ha−1) 7.3 4.4 7.5 Irrigation (mm) 1,407 217 77 Rainfall (mm) 337 337 337 Total water (mm) 1,744 553 414 WPIR (g grain kg−1 total water) 0.42 0.79 1.81 Input costs ($ ha−1) 379 230 140 Production value ($ ha−1) 1,097 706 1,071 Net income ($ ha−1) 718 487 906 Own labor (d ha−1) 116 93 109 Net income (including labor) ($ ha−1) 500 312 703

Number of farmers, 2003 2 6 3 Field size (ha) 0.11 0.11 0.56 Yield (t ha–1) 3,7 3.0 5.7 Irrigation (mm) 476 156 0 Rainfall (mm) 674 674 674 Total water (mm) 1,149 830 674 WPIR (g grain kg−1 total water) 0.32 0.36 0.85 Input costs ($ ha−1) 378 261 129 Production value ($ ha−1) 643 520 856 Net income ($ ha−1) 265 259 727 Own labor (d ha−1) 162 75 41 Net income (including labor) ($ ha−1) −34 120 651

��

Practical implementationTemperate environment/China. Promising aerobic rice varieties in northern China are HD277, HD297, and HD502. Before sowing, the land should be dry prepared by plowing and harrowing to obtain a smooth seedbed. Seeds should be dry seeded at 1−2-cm depth in heavy (clayey) soils and at 2−3-cm depth in light-textured (loamy) soils. Optimum seeding rates still need to be established but are probably in the 60−80 kg ha−1 range. In experiments so far, row spacings between 25 and 35 cm gave similar yields. Sowing of the seeds can be done manually (e.g., dibbling the seeds in slits opened by a stick or a tooth harrow) or using direct-seeding machinery. The total amount of fertilizer N application could probably follow local recommendations for lowland rice aiming at a 4−6 t ha−1 yield level. The total amount of N to be applied depends on indigenous soil N supply and other sources of N (such as atmospheric deposition). If no knowledge on local recommendations is available, an amount of 90 kg N ha−1 could be a useful starting point (to be subsequently optimized). Instead of basal application of the first N split, the first application can best be applied 10−12 days after emergence to minimize N losses by leaching (the emerging seedlings can’t take up N fast, so it will easily leach out). Second and third split applications can be given around maximum tillering and panicle initiation, respectively. With future research, principles of site-specific nutrient management (SSNM) for aerobic rice should be developed. If the crop is grown in a dry season, a light irrigation application (say 30 mm) should be given after sowing to promote emergence. Subsequent irrigation applications should aim to frequently restore the soil water content to field capacity, and depend on the rainfall pattern, the depth of groundwater, and on availability and/or cost of irrigation water. Irrigation can be applied by the same means as used for upland crops: flash flood, furrow, or sprinkler. Tropical aerobic rice systems for water-short irrigated environments are still in the research and development phase. More research is especially needed to develop high-yielding aerobic rice varieties and sustainable management systems. In the tropical Philippines, the most promising variety so far is Apo, but the breeding of improved varieties is in full swing. In general, the same management practices as for northern China can be followed. However, sustainability seems so far more of a problem in tropical areas than in temperate areas such as northern China. Aerobic rice should not be grown consecutively on the same piece of land, and—depending on the cropping history and soil type—even complete yield failures can occur on fields cropped to aerobic rice the very first time in their history! The main problems to overcome in the development of tropical aerobic rice are listed in Chapter 3.6.

�.�.� Raised beds

Oneoftherecentlyproposedinnovationstodealwith water scarcity in the rice-wheat system intheIndo-GangeticPlainistheuseofraisedbeds,inspiredbythesuccessofthesysteminhigh-yield-ing,irrigatedwheat-maizeareasinMexico(SayreandHobbs 2004). In the systemof raised beds,riceisgrownonbedsthatareseparatedbyfurrowsthroughwhichirrigationwateriscoursed.Inirriga-tionengineeringterms,thesystemofraisedbedsiscomparablewith“furrowirrigation.”Irrigationisintermittentandthesoilofthebedsisdominantlyin aerobic conditions; hence, the system can beconsideredanaerobicricesystem(thisisdifferentfrom the use of beds in heavy soils tomaintainsaturatedsoilconditions,Chapter3.3.1).Ingeneral,furrowirrigationismorewaterefficientthanflash-flooding(dependingonsoiltype,fielddimensions,andslopeoftheland),andfurrowirrigationshouldholdpromiseforaerobicrice.Thoughdimensionsmayvary, beds are usually around35 cmwide,separatedbyfurrowsthatare30cmwideand25cm

deep.Ricecanbetransplantedordirect-seededonthebeds.Sofar,theraised-bedsystemhasmostlybeentestedwithcurrentlowlandricevarieties,andyieldgainscanbeexpectedwhensuitableaerobicvarietiesaredeveloped/used.Tractor-pulledequip-menthasbeendevelopedthatshapesthebedsanddrillsseed(sometimestogetherwithfertilizers)inoneoperation.

Amongthesuggestedbenefitsofraisedbedsareimprovedwater-useandnutrient-useefficiency,improved water management, higher yields,and—when the operations aremechanized—re-duced labor requirements and improved seedingandweedingpractices(Connoretal2003,HobbsandGupta2003).Balasubramanianetal(2003)andHobbsandGupta(2003)reportedinitialresultsofon-stationtrialsandfarmerparticipatoryevaluationofriceonbedsintherice-wheatbeltinIndia.Yieldofricetransplantedordirect-seededonbedswasplus/minus5−6%ofthatofpuddledtransplantedrice,whileirrigationwatersavingsaveragedabout37−40%.Inarecentreview,however,Kukaletal

��

(2006)reportedthat“the performance of rice on beds in NW India has been variable, but generally disappointing to date. Even with similar irrigation scheduling, yields on permanent beds are generally 20−40% lower than puddled transplanted rice, with serious problems of iron deficiency, weeds, accurate sowing depth, and sometimes nematodes. Strategies for overcoming these problems are ur-gently needed, including breeding and selection for rice grown in aerobic soil and for the wide row spacing between adjacent beds. There are many reports of substantial irrigation water savings with rice on beds compared with continuously flooded puddled transplanted rice. However, some studies suggest that where similar irrigation scheduling is used, irrigation water use of transplanted rice on beds and puddled flats is similar, or even higher on the beds due to higher percolation rates in the nonpuddled furrows and longer duration of direct-seeded rice.”

Choudhuryetal(2007)comparedtheyield,wa-terinput(rainfall,irrigation),andwaterproductivityofdry-seededriceonraisedbedsandflatlandwiththatofflooded transplantedandwet-seededrice,andanalyzedtheeffectsofbedsonthesubsequentwheatcrop.Theirexperimentswereconductedin2001-03 atNewDelhi, India.The yields variedfrom3.2tha−1(flatlandandraisedbeds)to5.5tha−1(floodedtransplanted).Yieldsonraisedbedsthatwerekeptaroundfieldcapacitywere32−42%lowerthanunderfloodedtransplantedconditions,and21%lowerthanunderfloodedwet-seededcon-ditions.Totalwaterinputvariedfrom930mmonraisedbedsto1,600mminthefloodedtransplantedfields.Totalwaterinputinriceonraisedbedswas38−42% lower than inflooded transplanted rice,and32−37%lowerthaninfloodedwet-seededrice.However,thereducedwaterinputsinraisedbedswere also realizedwith dry seedingonflat landwiththesamewatermanagement.Reducedwaterinputs and yield reductions balanced each othersothatwaterproductivitywascomparableamongmosttreatments.Itshouldbenotedthatthisstudywasdoneinsmallplots(comparedwithfarmers’fields),whereedgeeffects (seepage lossesunderandadjacenttothebunds)candominatethewaterbalance.

Adistinctionneedstobemadebetween“per-manent beds” and “freshbeds.”Permanent bedsareconstructedonceandreshapedonlyafterwardwithsubsequentcropping.Theyareusedincrop

rotations(rice-nonricecrops)andhaveadvantagesin terms of cost savings, timeliness of planting,andopportunitiesforrapidcropdiversificationinresponse tomarketoptions.The raisedbeds canespeciallybenefitthenonricecropinheavywater-loggedsoilsbecauseofimproveddrainage(removalofwaterthroughthefurrows;bedsremainrelativelydry).However, just likewithaerobicriceonflatland(Chapter3.6),yieldofbothtransplantedanddirect-seededricehasbeennotedtodeclineundercontinuouscroppingonpermanentraisedbeds(E.Humphreys,personalcommunication).

Practical implementation Growing rice on raised beds shows promise but is still in its infancy of development (Humphreys et al 2005, Kukal et al 2006). In the Indo-Gangetic Plain, farmers are experimenting with raised beds for rice and other crops with different degrees of success. More information on raised beds can be obtained from the Rice-Wheat Consortium (www.rwc.cgiar.org/index.asp). Problems to overcome are listed in Chapter 3.6.

3.4.2 Conservation agricultureWith aerobic rice, technologies of conservationagriculture, suchasmulchingandzero-ormini-mumtillageaspracticedinuplandcrops,becomeavailabletoricefarmersaswell(HobbsandGupta2003).Variousmethodsofmulching(e.g.,usingdrysoil,straw,andplasticsheets)arebeingexperi-mentedwithinnonfloodedricesystemsinChinaandhavebeenshowntoreduceevaporationaswellaspercolationlosseswhilemaintaininghighyields(Dittertetal2002).InhillyareasinShiyan,HubeiProvince,inChina,farmersareadoptingtheuseofplasticsheetstocoverricefieldsinwhichthesoiliskeptjustbelowsaturation.Thelocalgovernmentsubsidizesandactivelypromotesthisuseofplasticsheets,and,in2006,therewereanestimated6,000haoffarmeradopters.Theproclaimedadvantagesareearliercropestablishmentby3weeks(riceisestablishedinearlyspringwhentemperaturesarestill low, and the plastic sheets increase the soiltemperature), higher yields, less weed growth,and lesswateruse (importantduringdryspells).However, little researchhasbeendone toverifythesebenefits.Theleftoverplasticafterharvestmaycause environmental degradation if not properlytakencareof.

��

Practical implementationSpecific information on minimum tillage and conservation agriculture technologies for rice can be obtained from the Rice-Wheat Consortium (www.rwc.cgiar.org/index.asp).

3.5 What option where?The relative “attractiveness” of the above tech-nologiesforfarmerstorespondtowaterscarcitydepends on the type and level of water scarcity(Chapter1.7),on the irrigation infrastructure (orthelevelofcontrolthatafarmerhasovertheir-rigationwater),andonthesocioeconomicsoftheirproductionenvironment.

With absolute, or physical,water scarcity,farmershavelittlechoicebuttoadapttoreceivinglesswaterthantheywouldneedtokeeptheirfieldscontinuouslyflooded.Figure3.6presentsagradientinrelativewateravailabilityandsomeappropriateresponseoptions.Onthefarright-handsideofthe(horizontal)water axis,water is amplyavailableand farmers canpractice continuousfloodingoflowlandriceandobtainthehighestyields.Onthefar left-handside,water isextremelyshort, suchasinrainfeduplands,andyieldsareverylow.Go-ingfromrighttoleftalongthewater-availability

axis,watergetsincreasinglyscarceandyieldswilldecline.

Evenwith sufficientwater available, goodlandleveling,bundmaintenance,constructionoffieldchannels,andthoroughpuddling(inthecaseofpuddledsystems)willcontributetogoodcropgrowthandhighyields.“Gettingthebasicsright”is something that all farmers can do, no matterwhethertheyoperateinlarge-orsmall-scaleirriga-tionsystemsorwhethertheyusetheirownsourcesofirrigation(suchastubewells)orsharedsources.After crop establishment, continuouspondingofwatergenerallyprovidesthebestgrowthenviron-mentforriceandwillresultinthehighestyields.Aftertransplanting,waterlevelsshouldbearound3cminitially,andgraduallyincreaseto5−10cmwithincreasingplantheight.Withdirectwetseeding,thesoilshouldbekeptjustatsaturationfromsowingtosome10daysafteremergence,andthenthedepthof ponded water should gradually increase withincreasingplantheight.Withdirect dry seeding,thesoilshouldbemoistbutnotsaturatedfromsow-ingtillemergence,orelsetheseedsmayrotinthesoil.Aftersowing,applyaflushirrigationifthereisnorainfalltowetthesoil.Saturatethesoilwhenplantshavedevelopedthreeleaves,andgraduallyincreasethedepthofpondedwaterwithincreasingplantheight.Inspecialproblemsoils,introducingsomeformofalternatewettinganddrying(AWD)

Yield (t ha–1)

10

5

0

Soil condition

Aerobic Flooded

FC S

Water availability HighLow

Traditionalupland rice

system

Aerobic ricesystem

Lowland ricesystem

Technologies to reduce water input

Furrow, flush,sprinkler irrigation

Soil management,reduced water depth

AWD SSC

∆Y = f (variety, management)

Fig. 3.6. Schematic presentation of yield responses to water availability and soil condition in different rice production systems and their respective technologies to reduce water inputs. AWD = alternate wetting and drying, SSC = saturated soil culture, FC = field capacity, S = saturation point, ΔY = change in yield. Adapted from Tuong et al (2005).

�0

orincreasingtheinternaldrainageratemayimprovecropgrowthandyield(Ramasamyetal1997).Theunderlyingreasonmaybeimprovedsoilaerationortheremovaloftoxicsubstances.

Thefirstresponseoptiontodecreasingwateravailabilitywouldbe to check“thebasics.”Theamountofwaterlossthatcanbereduceddependsontheinitialconditionofthericefield;ifthebasicsarerightfromthestart,thereisnotmuchthatcanbedoneanymore.Withprogressingwaterscarcity,establishmentoptionsalternativetotransplantingcanbeconsiderediftheturnaroundtimebetweensoakingandtransplantingisrelativelylarge,suchasinsomelarge-scaleirrigationsystems.Ifcom-munity seedbeds or a commercial provider ofseedlings could be organized, thiswould be theleast water-consuming method of getting a cropestablished.Seedlingscouldgettransplantedafewdaysafterlandsoakingandpuddlingonly,whilethe large-scaleraisingofseedlingswouldensureanefficientuseofwaterduringthatperiod(mainfieldsdonotyethavetobesoaked).Bothdirectwetanddirectdryseedingarealternativeoptions.Dryseedingwillbeeffectiveonlyinrelativelyclayeyandimpermeablesoilsthatdon’tneedpuddlingtoreducethepermeabilityanymore.

Withfurtherincreasingwaterscarcity,watermanagementpracticesduring thewholegrowingseasonneedtobeconsidered.Insteadofkeepinga5−10-cmdepthofpondedwaterduringthegrowingseason,thedepthcanbereducedtoaround3cm.Thiswillreducethehydrostaticpressureandmini-mizeseepageandpercolationlosses.Insaturatedsoil culture (SSC), thedepthofpondedwater isreducedto0−1cm.Aroundflowering,from1weekbeforeto1weekafterthepeakofflowering,pondedwatershouldbestbekeptat5-cmdepthtoavoidanypossiblewaterstressthatcouldresultinsevereyieldloss.ThepracticeofSSCwouldrequirefrequent(oncein2days)applicationsofsmallamountsofirrigationwater,andhencerequireahighlevelofcontroloverirrigationwater.ThepracticeofsafeAWDcanreducewaterlossesbyasmalltocon-siderableamountwithoutayieldpenalty.TowhatextentwaterlossescanbereducedunderSSCorAWDdependsmainlyon soil typeanddepthofthegroundwatertable:withaheavyclaysoilandshallowgroundwater(10−40cmdeep),waterlossesaresmalltostartwithandreductionsinwaterlossesareequallysmall.Withmoreloamyorsandysoilsand/or deeper groundwater tables, reductions in

waterlossescanbehigher,buttheriskofareduc-tioninyieldalsobecomeshigher.Ifwaterisgettingsoscarcethat“safeAWD”isnolongerpossible,theperiodsbetweenirrigationwillhavetobecomelonger(lettingthewaterinthefieldwatertubesgodeeperthan15cm)andyieldlossbecomesinevita-ble.AllformsofAWDrequirewatercontrolbythefarmer.Withownwatersources,suchastubewells,thisisnotaproblem.Incommunity-basedorlarge-scaleirrigationsystems,acommunalapproachtoAWD is required in which delivery of water togroupsoffarmersisscheduledtorealizeacertainpattern of AWD. Irrigation system upgrading ormodernizationmayberequiredtodothis,orsmallstoragefacilities(suchason-farmreservoirs)mayprovidetherequiredwatercontrol(Chapter5).

With still further increasing water scarcity,yieldoflowlandriceunderAWDwillcontinuetogodown.Atacertainpoint,aerobicricesystemsbecomeaviablealternative.Howmuchlesswaterisusedunderaerobicconditionsthanunderfloodedconditionsdependsmostlyontheseepageandper-colation(SP)lossesunderfloodedconditionsandonthedeeppercolationlossesofirrigationwaterunderaerobicconditions.TypicalSPratesoffloodedricefieldsaregiveninTable1.1inChapter1.3.Underaerobicconditions,theamountofdeeppercolationdependsonthecombinationofsoilwater-holdingcapacityandmethodofirrigation,andisreflectedintheirrigationapplicationefficiency(EA).Withaprecisedosageandtimingofirrigationinrelationtocroptranspirationandsoilwater-holdingcapacity,theEAinflash-floodirrigationcanbeupto60%(DoorenbosandPruit1984).Iffurrowirrigation(orraisedbeds)isused,theEAcangoupto70%,andwithsprinklerirrigationupto80%ormore.Assum-inganaveragegrowthdurationof100days,andmeanETvaluesforrice,wecanroughlycalculatethe“break-even”pointforSPratesinfloodedfieldsthatwouldresultinsimilarwaterrequirementsinaerobic fieldswith different irrigationmethods(Table3.8).WhentheSPrateinfloodedriceis3.5mmd−1orhigher,aerobicsystemswithflash-floodirrigation will require less water, and, if the SPrateis0.5mmd−1orlower,onlyaerobicsystemswithsprinklerirrigationrequirelesswater.Whenaerobicricesystemsaredirect(dry)seeded,asisthetypicaltargettechnology,anadditionalamountofwaterinputcanbesavedbyforgoingthewetlandpreparation.Anexampleofthecross-overpointintermsofwateravailabilitywhereaerobicricegives

��

higheryieldsthanfloodedlowlandriceisgiveninFigure3.7forfieldexperimentsatBeijinginChina(YangXiaoguang et al 2005).Two aerobic ricevarieties(HD297andHD502)andonelowlandricevariety(JD305)weregrownunderfloodedcondi-tionsandunderaerobic soil conditionswithdif-ferentamountsoftotalwaterinput.Underfloodedconditionswith1,300−1,400mmofwaterinputtotheright-handsideofthehorizontal(water)axis,lowlandvarietyJD305gavethehighestyieldsof8−9tha−1.TheyieldofJD305,however,quicklydeclinedwithincreasingwatershortageandaerobicsoilconditions.Withlessthan1,100mmofwaterinput, andunder aerobic soil conditions, aerobicricevarietiesHD297andHD502outperformedthelowlandvariety.

Whenwaterisphysicallyavailable,buthasahighcost,thechoiceofadoptinganyofthewater-savingtechnologiesbecomesmoreofaneconomicissue.Adoptingcertainwater-savingtechnologiesmayreducewaterbutattheexpenseofyieldloss.Ifthefinancialsavingsincurredbyusinglessirriga-tionwaterunderacertaintechnologyoutweighthefinanciallossofreducedyield,thentheadoptionofthattechnologybecomesattractive.Figure3.8givesso-called“water-responsecurves”obtainedfromtwodifferentfieldexperimentsinIndiawheredifferentformsofAWDwereimplemented(differ-entintervalsbetweenirrigations).Theexperimentof the lower curvewas done inCuttack,Orissa(Jhaetal1981).Theclimaticyieldpotentialwasrelativelylowsincetheexperimentwasperformed

Table 3.8. Comparison of water use in a hypothetical aerobic rice crop with that of lowland rice on different soil types characterized by their seepage (S) and percolation (P) rates.

Water flow process Aerobic rice (mm) Lowland rice (mm)

Lowland soil SP rate − − 1 mm d−1 5 mm d−1 15 mm d−1

Irrigation efficiency 85% 60% − − −

Evaporation 100 100 200 200 200Transpiration 400 400 400 400 400Seepage and percolation − − 100 500 1,500Irrigation inefficiency loss 90 335 − − −Total 590 835 700 1,100 2,100

10

9

8

7

6

5

4

3

2

1

0200 400 600 800 1,000 1,200 1,400

Water input (mm)

Yield (t ha–1)

Aerobic

Flooded

Fig. 3.7. Yield of aerobic rice varieties HD297 ( ) and HD502 ( ) and lowland rice variety JD305 ( ) at different levels of water input. Data from Yang Xiaoguang et al (2005).

��

inthewinterseason(lowradiationlevels).Ferti-lizerapplicationwasonly80kgNha−1andzeroPandK.ThesoilSPratewasabout21mmd−1.TheexperimentofthetopcurvewasdoneinPantnagar,UttarPradesh(Tripathietal1986).Theclimaticyieldpotentialwashigherbecausetheexperimentwasdoneinthesummer(highradiationlevels).Torealizethehigheryieldpotential,fertilizerapplica-tionswerealsohigher:120kgNha−1,60kgP2O5ha−1,and40kgK2Oha

−1.ThesoilSPratewas9−14mmd−1.Goingfromrighttoleftonthehorizontalaxis,waterusedecreasedwith adopting increas-ingintervalsbetweenirrigationsinAWD.Yieldswereinitiallynotaffected,but,afteracertainpoint,yieldsdeclinedwithlesswateruse.Thishappenedsomewhere below 1,750mm for the top curveandsomewherebelow1,000mmfor thebottomcurve.Thisexampleillustratesthesite-specificityof results of AWD. Farmers will decide on thetype/severityofAWDtoadoptbasedonthesite-specificfinancial trade-offbetweenyielddeclineandwatersavings.

3.6 SustainabilityWhile relativelymuchwork has been done onthedevelopmentoftechnologiestomaintaincropproductivity under water scarcity, little researchhasbeendoneontheirlong-termsustainabilityand

environmentalimpacts.Givenassuredwatersup-ply,lowlandricefieldsareextremelysustainableandabletoproducecontinuouslyhighyields,evenunder continuousdouble or triple cropping eachyear(Daweetal2000).Floodingofricefieldshasbeneficialeffectsonsoilacidity(pH);soilorganicmatterbuildup;phosphorus,iron,andzincavail-ability;andbiologicalNfixationthatsuppliesthecropwithadditionalN(Kirk2004).Whenfieldscannotbecontinuouslyfloodedanymorebecauseofwaterscarcity,thesebeneficialeffectsgraduallydisappear.Achangetomoreaerobicsoilconditions(suchasinAWDandaerobicrice)willnegativelyaffectthesoilpHinsomesituationsanddecreasetheavailabilityofphosphorus,iron,andzinc.Underthe “safeAWD”practice (Chapter 3.3.2), theseproblemsdonotoccur,but,whenmoresevereformsofAWDareimplemented,theymaystarttooccur.Underfullyaerobicconditions,whetheronflatlandor on raised beds, problemswithmicronutrientdeficiencieshavebeenreportedbyChoudhuryetal(2007),Sharmaetal(2002),Singhetal(2002),andTaoHongbinetal. (2006).The introductionofaerobicphasesinricefieldsmayalsodecreasethesoilorganiccarboncontent.Inalong-termex-perimentatIRRI,whereacontinuousricesystemis compared with a maize-rice system, 12 yearsof maize-rice cropping caused a 15% decline insoilorganicCandindigenousNsupplyrelativeto

8

6

4

2

0

Water input (mm)

Yield (t ha–1)

5000 1,000 1,500 2,000 2,500 3,000

Fig. 3.8. Yield versus water input in two experiments in India. Top curve data ( ) are from Pantnagar, Uttar Pradesh (Tripathi et al 1986), and bottom curve data ( ) are from Cuttack, Orissa (Jha et al 1981). The curved lines are fitted production functions of the shape [yield = a*(1 – e(b*(water input – c)))]. Adapted from Bouman and Tuong (2001).

��

floodedrice-ricecropping(RolandBuresh,personalcommunication).

There are indications that soil-borne pestsanddiseasessuchasnematodes,rootaphids,andfungioccurmoreinnonfloodedricesystemsthaninfloodedricesystems(Sharmaetal2002,Singhetal2002,VenturaandWatanabe1978,Venturaetal1981).Currentexperienceisthatunderfullyaerobicsoilconditions,ricecannotbegrowncon-tinuouslyonthesamepieceoflandeachyear(ascanbesuccessfullydonewithfloodedrice)withouta yield decline (George et al 2002). Figure 3.9presentsrecentdatafromacontinuousaerobicricecroppingexperimentatIRRI(Boumanetal2005,Pengetal2006).Since2001,aerobicricevarietyApohasbeencontinuouslygrownunderfloodedandaerobicconditionsinthesamefield.Floodedyields in thedryseasonareusually6.5−7tha−1,exceptin2001,whendiseasesdepressedyields.In

6

4

23

1

5

78

02001 2002 2003 2004 2005

Flooded yield (t ha–1)

2001 2002 2003 2004 2005

% Aerobic yield100

908070605040302010

0

ContinuousNewRestored

Year

A

B

Fig. 3.9. Yield of aerobic rice variety Apo under flooded conditions (A) and relative yield of Apo under aerobic conditions (as % of the Apo yield under flooded conditions) (B). In B, the black columns indicate continuous aerobic conditions, the darkly shaded column indicates yield under new aerobic conditions after conversion of flooded fields, and the lightly shaded column indicates yield under restored aerobic conditions after 1 year of fallow or flooded conditions (average is given). Data from Bouman et al (2006a).

2001,yieldunderaerobicconditionswas86%ofthat underflooded conditions, but this graduallydeclineduntilitwasonly45%in2005.In2003,halfofthefloodedfieldswereconvertedtoaerobicconditions,andaerobicyieldsreturnedto85%ofthefloodedyields,asinthefirstyear,2001.In2004,halfofthecontinuousaerobicfieldswereleftfalloworwerefloodedforthewholeyear,andreturnedtoaerobicconditions in2005.This“restoration”attempt brought aerobic yields back to 65%offloodedyields,andwasonlypartiallysuccessful.Themechanismsbehindthegradualyielddeclineandtherestorationeffectarenotyetunderstood,althoughhighlevelsofthenematodeMelodoigyne graminicola are found in the aerobic ricefields(upto3,000countsg−1freshroot)comparedwiththe floodedfields (6−400 counts g−1 fresh root;unpublisheddata).

Insomefieldexperimentsandfarmers’fields,nearlycompleteyieldfailurehasbeenobservedinfieldswhereaerobicricewasestablishedtheveryfirstyear(Fig.3.10).Usually,suchfieldshadalightsoiltextureandahistoryofbeingpartiallycroppedtouplandcropsortoriceundernonfloodedcondi-tionsbeforeaerobicricegotestablished.Always,nematodes were found when yield failures wereobserved,sometimesaggravatedbythepresenceofrootaphids,fungi,and/ornutrientdisorders.Croprotation is necessary under such conditions, andbreedersaretryingtodevelopaerobicricevarietieswithtoleranceofthesesoilsicknesses.

2,500

2,000

1,500

1,000

500

0200160120600

Yield (kg ha–1)

N (t ha–1)

Fig. 3.10. Yield of aerobic rice variety Apo under aerobic soil conditions at IRRI in 2004, at four levels of fertilizer N supply, and under three levels of water input (black column = 1,147 mm, darkly shaded column = 772 mm, and lightly shaded column = 632 mm). Unpublished data, IRRI.

��

Ecosystem services and the environment

4.1 Ecosystem services

Although themain function of rice fields is toproduce rice, they also provide a range of other“ecosystemservices” (also referred toas“multi-functionality”)(Boumanetal2006a).Ecosystemservicescanbegroupedintothefollowingcatego-ries(MillenniumEcosystemAssessment2005):

Provisioning(e.g.,freshwaterandcommodi-tiessuchasfood,wood,timber,andfuel)

Regulating(e.g.,waterpurification,andcli-mateandfloodregulating)

Supporting(e.g.,nutrientcycling,soilforma-tion,primaryproduction)

Cultural (e.g., aesthetic, spiritual, educa-tional,recreational)

Themostimportantprovisioning functionofriceenvironmentsis,ofcourse,theproductionofrice.Examplesofotherprovisioningservicesaretheraisingoffishandducksinricefields,ponds,orcanals.Frogsandsnailsarecollectedforconsump-tioninsomecountries.

As part of regulating services, bunded ricefieldsmayincreasethewaterstoragecapacityofcatchmentsandriverbasins,lowerthepeakflowofrivers,andincreasegroundwaterflow.Forex-ample,in1999and2000,20%ofthefloodwaterinthelowerMekongRiverBasinwasestimatedtobetemporarilystoredinupstreamricefields(Mas-umoto2004).Otherpossibleregulatoryservicesofbundedricefieldsandterracesincludethepreven-tionormitigationoflandsubsidence,soilerosion,andlandslides(PAWEES2005).Percolatingwaterfrom rice fields, canals, and storage reservoirsrecharges groundwater systems (Mitsuno1982).Themoderationof air temperatureby ricefieldshasbeen recognized as an important function in

peri-urbanareaswherericefieldsandurbanlandareintermingled(Oue1994).Thisfunctionisattributedtorelativelyhighevapotranspirationratesresultingin reducedambient temperatureof thesurround-ingarea in summer,and in lateralheatemissionfromthewaterbodyinwinter.Ricecanbeusedasadesalinizationcropbecausethecontinuouslypercolatingwater(Chapter1.3)leachessaltsfromthetopsoil(BhumblaandAbrol1978).Theleachateshouldberemovedbyagooddrainagesystem,orelsethereisriskforincreasedsalinizationofthegroundwater.Ricesoils thatarefloodedforlongperiodsoftheyearcontributetothemitigationofthegreenhouseeffectbytakingCO2fromtheat-mosphereandsequesteringthecarbon(C)(Bronson1997a,Dobermann2003).

Asasupporting service,floodedricefieldsandirrigation channels form a comprehensive waternetwork,which,togetherwiththecontiguousdryland,providesacomplexmosaicoflandscapes.Ir-rigatedricelandhasbeenclassifiedashuman-madewetlandsbytheRamsarConventiononWetlands(Ramsar2004).Surveysshowthatsuchlandscapessustain a rich biodiversity, including unique aswellasthreatenedspecies,andalsoenhancebio-diversityinurbanandperi-urbanareas(Fernandoetal2005).

Thecultural servicesof ricefieldsareespe-cially valued in Asian countries where, for cen-turies,ricehasbeenthemainstaplefoodandthesinglemostimportantsourceofemploymentandincome for rural people.Manyold kingdoms aswellassmallcommunitieshavebeenfoundedontheconstructionofirrigationfacilitiestostabilizericeproduction.Riceaffectsdailylifeinmanywaysandthesocialconceptofrice culturegivesmeaningtoricebeyonditsroleasanitemofproductionand

4

�6

consumption (Hamilton 2003).Many traditionalfestivalsandreligiouspracticesareassociatedwithricecultivationandricefieldsarevaluedfortheirscenicbeauty.

4.2 Environmental impactsTheproductionof lowland rice affects the envi-ronment in negative ways, such as the emissionof greenhouse gases and water pollution. In thissection,environmentalimpactsthathavearelation-shipwithwaterandthehydrologyofricefieldsaresummarized.

4.2.1 Ammonia volatilizationAmmonia(NH3)volatilizationfromureafertilizeristhemajorpathwayofNlossintropicalfloodedricefields,oftencausinglossesof50%ormoreoftheappliedurea-N (BureshandDeDatta1990).Ammonia-Nemissionsfromlowlandricefieldsareestimatedtoberoughly3.6Tgperyear(comparedwithatotalof9Tgy−1emittedfromallagriculturalfields),whichissome5−8%oftheestimated45−75Tgofgloballyemittedammonia-Nperyear(Kirk2004).Themagnitudeofammoniavolatilizationlargelydependsonclimaticconditions,fieldwaterstatus,andthemethodofNfertilizerapplication.Volatilized ammonium can be deposited on theearthbyrain,whichcanleadtosoilacidification(Kirk2004)andunintendedNinputsintonaturalecosystems.

4.2.2 Greenhouse gasesIrrigatedricesystemsareasignificantsinkforat-mosphericCO2(Chapter4.1),asignificantsourceofmethane (CH4), anda small sourceofnitrousoxide(N2O).Intheearly1980s,itwasestimatedthatlowlandricefieldsemittedsome50−100Tgofmethaneperyear,orabout10−20%ofthethenestimatedglobalmethaneemissions(Kirk2004).Recentmeasurements,however,showthatmanyricefieldsemitsubstantiallylessthanthoseinvesti-gatedintheearly1980s,especiallyinnorthernIndiaandChina.Also,methaneemissionshaveactuallydecreasedsincetheearly1980sbecauseofchangesin crop management such as a decreased use oforganicinputs(VanderGonetal2000).Currentestimatesofannualmethaneemissionsfromricefieldsareintherangeof20to60Tg,being5–10%of total global emissions of about 600Tg (Kirk2004).Themagnitudeofmethaneemissionsfrom

ricefieldsismainlydeterminedbywaterregimeandorganicinputs,andtoalesserextentbysoiltype,weather,tillage,residuemanagement,fertilizeruse,andricecultivar(Bronsonetal1997a,b,Wassmannetal2000).Floodingofthesoil isaprerequisitefor sustained emissions ofmethane.Mid-seasondrainage,acommonirrigationpracticeadoptedinmajor rice-growing regions in China and Japan,greatly reduces methane emissions. Similarly,riceenvironmentswithanunevensupplyofwater(forexample,thosesufferingfromwaterscarcity,Chapter1.7)havealoweremissionpotentialthanfullyirrigatedrice.

Fewaccurateassessmentshavebeenmadeofemissionsofnitrousoxidefromricefields(Abaoetal2000,Bronsonetal1997a,b,Dittertetal2002),andthecontributiontoglobalemissionshasnotyetbeenassessed.Inirrigatedricesystemswithgoodwater control, nitrous oxide emissions are quitesmall exceptwhen excessively high fertilizer-Nrates are applied. In irrigated ricefields, nitrousoxideemissionsmainlyoccurduringfallowperiodsandimmediatelyafterfloodingofthesoilattheendofthefallowperiod.

4.2.3 Water pollutionChangesinwaterqualityassociatedwithricepro-ductionmaybe positive or negative, dependingmainlyonmanagementpracticessuchasfertiliza-tionandbiocide(allchemicalsusedforcropprotec-tion,suchasherbicides,pesticides,fungicides,etc.)use.Thequalityofthewaterleavingricefieldsmaybeimprovedasaresultofthecapacityofthericefieldstoremovenitrogenandphosphorus(Fengetal2004,IkedaandWatanabe2002).Ontheotherhand,nitrogentransferfromfloodedricefieldsbydirectflowof dissolvednitrogen through runoffwarrantsmoreattention.Highnitrogenpollutionofsurfacefreshwaterscanbefoundinrice-grow-ing regionswhere fertilizer ratesareexcessivelyhigh, such as in JiangsuProvince inChina (Cuietal2000).

Contamination of groundwater may arisefromtheleachingofnitrateorbiocidesandtheirresidues (Bouman et al 2002).Nitrate leachingfromfloodedricefieldsisquitenegligiblebecauseofrapiddenitrificationunderanaerobicconditions(BureshandDeDatta1990).Forexample,inthePhilippines,nitratepollutionofgroundwaterunderrice-basedcroppingsystemssurpassedthe10mgL−1limitforsafedrinkingwateronlywhenhighly

��

fertilized vegetableswere included in the crop-ping system (Bouman et al 2002). In traditionalricesystems,relativelyfewherbicidesareusedaspuddling,transplanting,andthepondingofwaterareeffectiveweedcontrolmeasures.Meanbiocideuseinirrigatedricevariesfromsome0.4kgactiveingredients(a.i.)ha−1inTamilNadu,India,to3.8kga.i.ha−1inZhejiangProvince,China(Boumanatal2002).Inthewarmandhumidconditionsofthetropics,volatilizationisamajorprocessofbiocideloss,especiallywhenbiocidesareappliedonthesurfaceofwateroronwetsoil(SethunathanandSiddaramappa1978).Therelativelyhightempera-turesfurtherfavorrapidtransformationofthere-mainingbiocidesby(photo)chemicalandmicrobialdegradation,butlittleisknownaboutthetoxicityoftheresidues.IncasestudiesinthePhilippines,meanbiocideconcentrationsingroundwaterunderneathirrigatedrice-basedcroppingsystemswereonetotwoordersofmagnitudebelowthesingle(0.1µgL−1)andmultiple(0.5µgL−1)biocidelimitsforsafedrinkingwater,althoughtemporarypeakconcentra-tionsof1.14−4.17µgL−1weremeasured(Boumanet al 2002).As for nitrogen, however, biocidesand their residuesmaybedirectly transferred toopenwaterbodiesthroughdrainagewaterflowingoverlandoutofricefields.Thepotentialforwaterpollutionbybiocides is greatly affectedbyfieldwatermanagement.Differentwaterregimesresultindifferentpestandweedpopulationsanddensities,whichfarmersmaycombatwithdifferentamountsand typesofbiocides.Residualbiocides interactdifferentlywithsoilunderdifferentwaterregimes(SethunathanandSiddaramappa1978).

4.3 Effects of water scarcityWaterscarcityaffectsnotonlytheabilityofricefields to produce food but also the environmentand the other ecosystem services of rice fields.Increasingwaterscarcityisexpectedtoshiftriceproduction tomorewater-abundant delta areas,andtoleadtolessfloodedconditionsinricefieldsandtotheintroductionofuplandcropsthatdonotrequireflooding.Thesechangeswillhaveenviron-mentalconsequencesandwillaffectthetraditionalecosystemservicesofthericelandscape.

Rice that is not permanently flooded tendsto havemoreweed growth and a broaderweedspectrum than rice that is permanently flooded(MortimerandHill1999).Itisexpectedthatwatershortageswillleadtomorefrequentuseofherbi-cides,whichmayincreasetheenvironmentalloadofherbicideresidues.Withlesswater,thenumbersandtypesofpestsandpredators(e.g.,spiders)maychangeaswellaspredator-pestrelationships.Thepossibleshift in theuseofpesticidesbyfarmersinresponsetothesechanges,andwhatthismeansfor the environment, is as yet unknown.Moreleachingofnitrateisexpectedwithincreasedsoilaeration(eitherwithgrowingriceundernonfloodedconditions,orwiththeshifttouplandcrops)thanunderfloodedconditions.Lessmethaneemissionsareexpectedunderaerobicconditionsthanunderfloodedconditions,buthighernitrousoxideemis-sionsareexpected(Bronsonetal1997a,b).However,therelativeemissionsofthesegreenhousegasesvarywithenvironmentandmanagementpractices(Dittertet al2002).Flooded rice is effective in leachingaccumulated salts from the soil profile, and thechange tomoreaerobicconditionsmay result inincreasedsalinization.

There is little information on how waterscarcitywillaffecttheecosystemservicesofricelandslistedinChapter4.1.Thereisagrowingrec-ognitionthroughouttherice-growingworldthatabetterunderstandingoftheecosystemservicesofthe rice environment is needed. Although somemethodologiesexisttomeasureandestimatedif-ferentservicesofagriculturalsystems,quantifyingandvaluingthepositiveandnegativeexternalitiesstillpresentsamajorchallenge.Inmanycountries,relevantdata at the appropriategeographic levelarenotavailable.

��

Irrigation systems

5.1 Water flows in irrigation systems

Irrigatedricefieldsarecharacterizedbylargevol-umesofoutflowsbysurfacedrainage,seepage,andpercolation(Chapter1.3).Althoughtheseoutflowsare lossesfromanindividualfield, there isgreatscopeforreuseoftheseflowswithinalandscapethat consists ofmany interconnectedfields (Fig.5.1).Surfacedrainageandseepagewaterusuallyflow into downstreamfields and the loss of onefield is the gain of another.At the bottomof atoposequence,theseflowsenterdrainsorditches.

However,farmerscanusesmallpumpstoliftwaterfromdrainstoirrigatefieldsthatareinadequately,ornot,servicedbyirrigationcanals.Inmanyirri-gationsystemsinlow-lyingdeltasorfloodplainswithimpededdrainage,thecontinuouspercolationofwater(fromfields,butalsofromcanals)hascre-atedshallowgroundwatertablesclosetothesurfacethatmaydirectlyprovidethericecropwithwater(Chapter1.4).Again, farmers caneitherdirectlypumpwaterupfromtheshallowgroundwaterorpumpgroundwaterwhenitbecomessurfacewaterasitflowsintocreeksordrains.

Groundwaterinflow

PumpPump

Groundwateroutflow

SP

S

P

S

P

S

PP

DD D

Drain

age c

hann

el

Irrig

atio

n ch

anne

l

Fig. 5.1. Surface and subsurface water flows across a toposequence of rice fields. D = drainage (overbund flow), I = irrigation, P = percolation, S = seepage.

5

�0

Recentstudiesofrice-basedirrigationsystemsin China and the Philippines indicate that manywaterperformanceindicators(suchaswaterpro-ductivity,fractionofappliedwaterusedbythecrop)improvewithincreasingspatialscalebecauseofthereuseofwater(Hafeez2003,Loeveetal2004a,b).Muchofthisreuseisdoneinformallybyfarmerswhotaketheirowninitiativetopumpwater,blockdrainagewaterways,orconstructsmallon-farmres-ervoirsforsecondarystorage.Mostofthesefarmersarefoundintail-endportionsofirrigationsystemswherewaterdoesnotreachbecausetoomuchwaterislostupstream(e.g.,byupstreamfarmerstakingtoomuchwater,bycanalseepagelosses,andbyoperational losses).Hafeez et al (2007) reportedquantitativedataonwaterreuseon18,000haofDistrictIoftherice-basedUpperPampangaRiverIntegratedIrrigationSystem(UPRIIS)inCentralLuzon,Philippines.Atotalof16checkdamswerefound for reuse of surface drainage water, and12%ofallfarmersownedapumpforgroundwaterextraction.Inthewholestudyarea,57%ofallavail-ablesurfacewaterwasreusedbythecheckdamsand17%throughpumping.Theamountofwaterpumpedfromthegroundwaterwasabout30%ofthegroundwaterrechargebypercolationfromricefields.Figure5.2showsthattheamountofwaterreusedbythecheckdamsandbypumpingincreasedwithspatialscale(because,withincreasingscale,the options for reuse increase). Because of this

increaseinwaterreusewithincreasingscale,thewaterproductivityincreasedwithspatialscaleaswell(Fig.5.3).

Althoughwatercanbeefficientlyreusedthisway,itdoes,however,comeatacost,especiallytodownstreamfarmers.Thecurrentdebateontheimprovementofirrigationsystemsfocusesontherelativebenefitsandcostsofsystemmodernizationvis-à-vis thoseof internal and (mostly informal)reuseofwater.Systemmodernizationaimstoim-provetheirrigationsystemdeliveryinfrastructureandoperationschemetosupplyeachfarmerwiththerightamountofwaterattherighttime.Gainsinwater productivity are possible by providingmore reliable irrigation supplies, for example,throughprecisiontechnologyandtheintroductionofon-demanddeliveryofirrigationsupplies(e.g.,Gleick 2000,Rosegrant 1997).The argument isthat when farmers have control over timing andamountofwatersuppliestotheirfarm,theyneednot take their turn in afixed rotational scheduleofdeliveries if the soil is stillwet from rainfall.Matchingsystemdeliveryandfield-leveldemandneeds further research, as optimal scheduling ofirrigationsisdifficultwhenalargepartofthecropwater requirement is met from rainfall. This isespecially true in large irrigation systemswithaconsiderabletimelagbetweendiversionofwateratthesource(riverorreservoir)anditsarrivalatthefarmer’sgate.InsomepartsofChina,although

Water reuse (106 m3)

5,000 10,000 15,000 20,0000

1009080

70605040302010

0

y = 0.0013x + 1.1412 R2 = 0.9174

y = 0.0046x + 0.8885 R2 = 0.9049

5,000 10,000 15,000 20,00000.000.020.040.060.080.100.120.140.16

0.180.20Water productivity (g grain kg–1 water)

y = (7 × 10–6)x + 0.0561 R2 = 0.846

Area (ha)

Fig. 5.2. Volume of reuse of surface water by check dams ( ) and by groundwater pumping ( ) versus spatial scale in District I of UPRIIS. The lines are linear regressions. Data from Hafeez (2003).

Fig. 5.3. Water productivity (WPIR; g rice grains kg–1 water supplied (irrigation plus rainfall)) versus spatial scale in District I of UPRIIS. The line is a linear regression. The absolute values of WPIR are low (compare with values in Chapter 1.5) because a lot of water still flows out of the area (drainage water) but is reused by downstream irrigators. Data from Hafeez (2003).

Area (ha)

��

the main system is supply-driven, farmers havecontroloverthetimingandamountofwateratthefarmgate becausewater is stored in small farmponds,whichcanalsoprovidewaterforotheruses(Mushtaqetal2006).

5.2 Field versus irrigation system level The relationships betweenwater use at thefieldandattheirrigationsystemlevelarecomplexandinvolvehydrological,infrastructural,andeconomicaspects.Atthefieldlevel,farmerscanreducewa-ter lossesbyadoptingwater-saving technologies(Chapter3).Iftheypayforthecostofthewatertheyuse,theycantherebyincreasetheprofitabilityofricefarming.Attheirrigationsystemlevel,theadoptionoffield-levelwater-saving technologieswillreducethetotalamountofwaterlostasevapo-ration,butbyrelativelysmallamountsonly.Mostofthewatersavedatthefieldlevelisbyreducedseepage,percolation,anddrainageflows.Ontheonehand,thisresultsinmorewaterretainedatthesurface (in the irrigationcanals),which is avail-ablefordownstreamfarmers.Ontheotherhand,itreducestheamountofwaterre-enteringthehy-drologicalcycleandthusreducestheoptionsforinformalreusedownstream.Reducingpercolationfromricefieldscanlowergroundwatertables.Thiscanadverselyaffectyieldssincericeplantsmaybelessabletoextractwaterdirectlyfromtheground-water (Chapter 1.4;Belder et al 2004).Deepergroundwater tableswillalso increase thecostofpumpingforreusedownstream.Anyadoptionofwater-saving technologies requires considerablewatercontrolbythefarmers.Thisisnotmuchofaproblemforfarmersusingtheirownpump,butitissoforfarmersinlarge-scalesurfaceirrigationsystemsthatlackflexibilityin,andreliabilityof,waterdelivery.Itisalsoaproblemforfarmersusingelectricitytopumpgroundwaterwheresuppliesareunreliable,asinnorthwestIndia.Toallowfarm-erstoprofitfromwater-savingtechnologies,suchirrigationsystemsneed tobemodernized,whichcomesataneconomiccost.

The“beneficiaries”ofwatersavingsatthefieldandirrigationsystemlevelaredifferentinmostcases.Attheirrigationsystemlevel,theirrigationsystemmanagementcansavewaterinagricultureandusethiswaterforotherpurposessuchashy-dropowergenerationorindustry.Farmerswillbeinterestedinsavingwateronlyiftheycanderive

benefitssuchasreducedirrigationcosts.AdetaileddiscussionofmotivestosavewaterispresentedinChapter1.7.

5.3 Integrated approachesApproaches that integrate agronomic measures,improvedpolicies,institutionalreforms,andinfra-structuralupgradingsmayhavethebestchanceofsuccessfullyrespondingtowaterscarcity.ArecentsuccessstoryistheZangheIrrigationSystem(ZIS)inthemiddlereachesoftheYangtzeBasininChina(Loeveetal2004a,b).ZIShasacommandareaofabout160,000haandservicesmainlyriceinthesummerseason.Sincetheearly1970s,theamountofwaterreleasedtoagriculturehasbeensteadilyreduced in favor of increased releases to cities,industry,andhydropower(Fig.5.4).Sincethemid-1990s,theamountofwaterreceivedbyagriculturehasbeenlessthan30%oftheamountreceivedintheearly1970s.Inthesameperiod,however,totalrice production has increased, with a productionpeakofaround650,000tonsinthelateeightiesthatwasnearlytwicetheamountproducedinthelatesixties.Althoughriceproductionhasleveledofftoastable500,000tonsinthelastdecade,morericehasbeenproducedwithlesswateroverthepast30years.Thisfeathasbeenaccomplishedbyavarietyofintegratedmeasures(Dongetal2004,Hongetal2001,Loeveetal2001,2004a,b,Mushtaqetal2006,Moyaetal2004):

Doublericecroppinghasbeenreplacedbymorewater-efficient single rice cropping.Thiswaspossiblebecauseoftheavailabilityofmodernshort-durationhigh-yieldingvarie-ties.

The alternate wetting-drying water-savingtechnologyhas beenpromoted andwidelyadopted.

Policies, such as volumetric water pricing,andinstitutionalreforms,suchaswater-userassociations,havebeenintroducedthatdriveandpromoteefficientuseofwaterbyfarm-ers.

Theirrigationsystemhasbeenupgraded(e.g.,canallining).

Secondary storage has been developedthroughthecreationofthousandsofsmall-tolarge-sizepondsandreservoirs.

TheZIS case study suggests thatwin-winsituationscanexistwherericeproductioncanbe

��

maintained, or even increased, while freeing upwaterforotherpurposes.

5.4 Irrigation system managementWhenwaterisscarcerthanland,itmaybebeneficialto maximize water productivity rather than landproductivity(whichisyield;kgha−1)inirrigationsystemmanagement.Take,forexample,atypicalrice-basedirrigationsystemthatfaceswatershort-ageinthemainreservoir.Ausualresponseoptionof the irrigation system manager is to programlessareaforirrigationthanthedesignedcommandarea.Inpractice,thismeansthatupstreamfarmerswouldgetsufficientwaterforfloodedriceproduc-tion,anddownstreamfarmerswouldnotgetanywater at all and their landwouldbe left fallow.However, tomaximize totalproductionfromtheirrigation system, and to improve equity amongfarmers,itwouldbemorebeneficialtospreadouttheavailableamountofwaterover thecompletecommandareaand“impose”somewaterscarcityonallfarmers.Insteadofupstreamfarmerspracticingcontinuouslyflooded irrigation, anddownstreamfarmers havingnowater at all, there couldbe amixoffarmersgrowingfloodedriceandadoptingawater-savingtechnologysuchasAWD.Table5.1quantifiesthewateruseandtotalriceproductionof

ahypotheticalirrigationsystemintwocontrastingscenarios ofwater distribution.Theproductivityandwater-useparametersoffloodedriceandAWDriceusedinourexamplearetakenfromTable3.2(Guimba,1989data),andarerepeatedinTable5.2.Tosimplifythecalculations,weassumethereisnorainfallandthatallwaterissuppliedbyirrigation.Thesystemis10,000ha,withastoragecapacityofthereservoirof168106m3.Withafullreservoir,thisamountofwaterissufficienttohave10,000haoffloodedrice,producingatotalof58103tonsofrice.Supposethereisawaterscarcityandthatthereservoirisfilledonlyto80%,storing134106m3ofwater.InscenarioI,only80%ofthecommandareareceiveswater,allowingthesefarmerstogrowflooded rice,whereas the remaining 20% is leftfallow.InscenarioII,65%ofthecommandareareceivesthe“full”amountofwater,allowingthesefarmers to growflooded rice, and the remaining35%receivesareducedamount that issufficienttogrowriceunderAWD.InscenarioI,thetotalriceproductionintheirrigationsystemis46103tons,and,inscenarioII,itis53103tons.Moreover,thereismoreequityamongthefarmersinscenarioIIthaninscenarioI.

Theaboveexampleisquitesimple(andignoresthecomplexitiesofdiversifiedcroppingandofsys-temsoperationtoallowfarmerstopracticeAWD)

1965

Water for irrigation (106 m3)1,400

1,200

1,000

800

600

400

200

01969 1973 1977 1981 1985 1989 1993 1997 2001

800

700

600

500

400

300

200

100

0

Rice production (000 t)

Year

Water

Rice

Fig. 5.4. Total rice production and irrigation water supply in Zanghe Irrigation System, Hubei, China, between 1965 and 2002. Data points are 5-year moving averages. See also Fig. 1.6B from Hong et al (2001).

��

but illustrates theeffectof includingtheconceptofwaterproductivity(besidesyield)inirrigationsystemmanagement.

5.5 Irrigation system designAWDirrigationcanbetakenasastartingpointinthedesignofanewirrigationsystem.Supposethatahypotheticalreservoircanbeconstructedwithacapacityof168106m3,andthatacommandareawillbedesignedforricefarmersgetting1haeach.If AWD irrigation is the design criterion ratherthanfloodedrice,theirrigatedriceareaandtotalriceproductionarelarger,andmorefarmerswillbenefit from irrigation development (Table 5.3).Ifreuseofpercolationanddrainagewateristaken

Table 5.1. Area, water use, and total production of flooded rice and of rice grown under AWD, in a hypothetical irrigation scheme of 10,000 ha with 134 106 m� of water available.

Land use Area (ha) Water use (106 m3) Production (tons)

Scenario I (80% farmers, flooded rice; 20% farmers, no crop) Flooded 8,000 134 46,400 AWD 0 0 0 Fallow 2,000 0 0 Sum 10,000 134 46,400

Scenario II (65% of farmers, flooded rice; 35% farmers, AWD) Flooded 6,560 110 38,048 AWD 3,440 24 14,852 Fallow 0 0 0 Sum 10,000 134 52,900

Table 5.2. Yield, water productivity with respect to total water input (WPIR), and water input of flooded rice and rice grown under AWD (from Table 3.2, Guimba 1989).

Yield (t ha−1) WPIR (kg m−3) Water use (m3 ha−1)

Flooded 5.8 0.345 1,679AWD 4.3 0.614 700

Table 5.3. Area under rice production, and total water use, rice production, and number of farmers, with flooded rice and with AWD as design criteria.

Design criterion Rice area Water use Production Farmers (ha) (106 m3) (tons) (number)

Flooded 10,000 168 58,000 10,000AWD 23,986 168 103,000 23,986

intoconsiderationinthedesign,thericeareaunderirrigation can increase further in both scenarios.With100%reuseofwater,theirrigatedriceareaswould be similar in both scenarios sinceAWDmostlysaveswaterbyreducingpercolationflows(thatcanbereused)andonlyalittlebyreducingevaporation(whichcannotbereduced).However,100%waterreuseisinpracticenotattainableandthe implementation of AWD would lead to thelargestareaunderirrigation.

This example demonstrates the impact thatwater-saving technologiescanhaveon irrigationsystemdesign.Ratherthantakingcontinuousflood-ingoffieldsasadesigncriterion,arotationofwaterdeliverycanbeusedthatisdesignedtoletthefieldsdryoutforafewdaysduringeachrotationcycle.

��

Fieldwaterrequirementsduringthefloodeddaysshouldbebasedonevapotranspiration,seepage,andpercolation,whereas,duringthenonfloodeddays,theycanbebasedonevapotranspirationalone.

Asystemcanalsobedesigned tomaximizethereuseofseepageandpercolationwaterasmuchaspossible throughcheckdamsandpumping.Ifthereisasalinizationhazard,careshouldbetakentoavoidusingwaterthathasbecometoosalinebyreuse through judicious mixing of drainage andfreshwater.

Construction of field channels (irrigation,drainage)andlandlevelingshouldbeconsideredtofacilitategoodwatermanagement.Farmersinanewirrigationsystemshouldbetrainedinsoundwatermanagementpracticessuchasthosedetailedinthismanual(bundmaintenance,crackplowing,etc.).

��

Appendix: Instrumentation

Detailed descriptions and user guides of equip-menttomeasurewaterflowsandsoilphysicalandhydrologicalpropertiesofricesoilsaregivenbyIRRI(1987)andWopereisetal(1994).CalculationandmeasurementproceduresforevapotranspirationaregivenbyFAO(1998).Here,weintroducetwosimpletoolsthatarepracticalincharacterizingthewaterstatus(hydrologicalconditions)ofricefields

andcanhelpguidetheimplementationofwater-savingtechnologies.

A.1 Field water tube Thefieldwatertubeisusedtomeasurethedepthofstandingwateronthefield,beitontopofthesur-faceorjustbelowthesurface(Fig.A.1).Perforate

20 cm

15 cm

With water belowthe soil surface:read this

Holes are 5 mm in diameter, and spaced 2 cm apart

Plow pan

Puddledtopsoil

Soil surface

With standing wateron top of the surface:read this

20-cm-diameter PVC pipeUse a ruler tomeasure water depth

Fig. A.1. Field water tube for monitoring the depth of standing water on a rice field.

�6

ahollowandbottomlessPVC-tubeofabout20cmindiameterand35cmlongwithsmallholesusingadrill.Theholesshouldbeabout0.5cmindiameterandspaced2cmapart.Installthistubeinthefieldsothatthebottomofthetubeisburiedintheplowsole(about20cmdeep)andthatsome15cmofthetubeprotrudesabovethesoilsurface.Removethesoilfrominsidethetubedowntothebottomofthetube.Waterwillflowthroughtheholesintothetube,sothatthewaterlevelinsidethetubeisthesameasoutside.Afterirrigation,thelevelofthewaterinthetubecanbeseengoingdowneveryday.Thetubecanbeplacedatthesideofthefieldclosetothebund(butatleast1mawayfromthebund),soitiseasytorecordthewaterdepth(noneedtowalkverydeepintothefield).Makesurethatthespotchosentoinstallthetubeisrepresentativeforthewholefield(don’tplaceitinadepressionorinanelevatedpatch).

Thewaterdepthismeasuredfromthetopofthetubetothelevelofthewaterinthefieldusingasimpleruler.Subtract15cmfromthereadingtoobtainthedepthofpondedwater.Anegativevaluemeansthatthewaterisstandingonthefield;aposi-tivevaluemeansthatthewaterlevelisbelowthesurface.Tomakethemeasurementmoreaccurate,theheightofthetubeprotrudingabovethesurfacecanbemeasuredafewtimesduringtheseasontocheckwhetheritis15cm.

A.2 Groundwater tube Thegroundwatertubeisusedtomeasurethedepthofthegroundwaterbelowthericefields(Fig.A.2).Cutalengthof175to250cmofhollowandbot-tomlessPVC-tubeofabout5cmindiameter.Useadrilltomakeholesalongasectionof50-cmlengthatoneendofthetube.Theholesshouldbeabout

Cover(used tin can)

5-cm-diameter PVCpipe

Bund surface

Groundwater level

Holes are 5 mmin diameter, 2 cm apart

50 cm

100 cm 200 cm

Measuringdevice

50 cm

Readhere

Fig. A.�. Groundwater tube for monitoring groundwater depth below rice fields.

��

0.5cmindiameterandspaced2cmapart.Wraptheperforatedpartofthetubeinroughclothsuchasanoldjutesacktopreventtheholesfromgettingclogged.Installthistubeinagoodandsolidbundbetweenthericefieldssothatthetubeprotrudesabout50cmabovethesoilsurface.Anaugerdrillcanbeusedtodrilltheholetoplacethetubein.Once the tube is installed, anygroundwaterwillflow through the holes into the tube, so that thewaterlevelinsidethetubeindicatesthedepthofthegroundwater.Placeacap(canbemadeofanoldtin)ontopofthetubetopreventanythingfromfallinginandblockingthetube.

Thewaterdepthismeasuredfromthetopofthetubetothelevelofthewaterinthefieldusingalongstickorapieceofbamboo.Subtract50cmfromthereadingtoobtainthedepthoftheground-water (if the height of the tube above the bundis different from50 cm, subtract the real heightinsteadof50cm).Anegativevaluemeansthatthegroundwaterhasrisenabovethebundandthatthearea isflooded.Tomake themeasurementmoreaccurate,theheightofthetubeprotrudingabovethebundcanbemeasuredafewtimesduringtheseasontocheckwhetheritis50cm.Torelatethemeasuredgroundwaterdepthtothesoilsurfaceofthefields,subtracttheheightofthebundsfromthemeasurements.Forexample,withabundheightof20cm,agroundwaterdepthof100cmbelowthebundequalsagroundwaterdepthof80cmbelowthericefield.

��

AbaoJrEB,BronsonKF,WassmannR,SinghU.2000.Simultaneousrecordsofmethaneandnitrousoxideemissions in rice-based cropping systems underrainfedconditions.Nutr.Cycl.Agroecosyst.58:131-139.

AhmedKM,BhattacharyaP,HasanMA,AkhterSH,AlamSMM,BhuyianMAH,ImamMB,KhanAA,SracekO.2004.ArsenicenrichmentingroundwaterofthealluvialaquifersinBangladesh:anoverview.Appl.Geochem.19:181-200.

AtlinGN,LafitteHR,TaoD,LazaM,AmanteM,CourtoisB.2006.Developingricecultivarsforhigh-fertilityupland systems in the Asian tropics. Field CropsRes.97:43-52.

BalasubramanianV,Ladha JK,GuptaR,NareshRK,MehlaRK,Bijay-Singh,Yadvinder-Singh. 2003.Technologyoptionsforriceintherice-wheatsysteminSouthAsia.In:LadhaJK,HillJE,DuxburyJM,GuptaRK,BureshRJ, editors. Improving thepro-ductivity and sustainability of rice-wheat systems:issues and impacts.ASASpecial Publication 65,ASA Inc.,CSSA Inc., SSSA Inc.,Madison,Wis.,USA.p115-147.

BelderP,BoumanBAM,SpiertzJHJ,CabangonR,GuoanL,QuilangEJP,LiYuanhua,Tuong,TP. 2004.Ef-fectofwaterandnitrogenmanagementonwateruseand yield of irrigated rice.Agric.WaterManage.65:193-210.

BelderP,BoumanBAM,SpiertzJHJ,LuG.2007.Compar-ingoptionsforwatersavingsinlowlandriceusingamodelingapproach.Agric.Syst.92:91-114.

BhumblaDR,AbrolIP.1978.Salineandsodicsoils.In:Soilsandrice.LosBaños(Philippines):InternationalRiceResearchInstitute.p719-738.

BlackwellJ,MeyerWS,SmithRG.1985.Growthandyieldofriceundersprinklerirrigationonafree-drainingsoil.Aust.J.Exp.Agric.25:636-641.

BorellA,GarsideA,FukaiS.1997.Improvingefficiencyofwaterforirrigatedriceinasemi-aridtropicalen-vironment.FieldCropsRes.52:231-248.

Bouman BAM. 2001.Water-efficient managementstrategies in rice production. Int.RiceRes.Notes16(2):17-22.

BoumanBAM.2007.A conceptual framework for theimprovementofcropwaterproductivityatdifferentspatialscales.Agric.Syst.93:43-60.

BoumanBAM,TuongTP.2001.Fieldwatermanagementtosavewaterandincreaseitsproductivityinirrigatedrice.Agric.WaterManage.49:11-30.

BoumanBAM,KropffMJ,TuongTP,WopereisMCS,TenBerge,HFM,VanLaarHH.,2001.ORYZA2000:modelinglowlandrice.InternationalRiceResearchInstitute,LosBaños,Philippines, andWageningenUniversityandResearchCentre,Wageningen,Neth-erlands.235p.

BoumanBAM,CastañedaA,BhuiyanSI.2002.Nitrateand pesticide contamination of groundwater underrice-basedcroppingsystems:evidencefromthePhil-ippines.Agric.Ecosyst.Environ.92:185-199.

BoumanBAM,PengS,CastanedaAR,VisperasRM.2005.Yieldandwateruseofirrigatedtropicalaerobicricesystems.Agric.WaterManage.74:87-105.

BoumanBAM,HumphreysE,TuongTP,BarkerR.2006a.Riceandwater.Adv.Agron.92:187-237.

BoumanBAM,YangXiaoguang,WangHuaqi,WangZhimin,ZhaoJunfang,ChenBin.2006b.Performanceoftemperatericevarietiesunderirrigatedaerobiccon-ditionsinNorthChina.FieldCropsRes.97:53-65.

BoumanBAM,FengL,TuongTP,LuG,WangH,FengY.2007.Exploringoptionstogrowriceunderwater-short conditions in northern China using a model-lingapproach. II.Quantifyingyield,waterbalancecomponents, and water productivity. Agric. WaterManage.88:23-33.

BronsonKF,NeueHU, SinghU,Abao EBJ. 1997a.Automated chambermeasurement ofmethane andnitrous oxideflux inflooded rice soil. I.Residue,nitrogen,andwatermanagement.SoilSci.Soc.Am.J.61:981-987.

References

�0

BronsonKF,SinghU,NeueHU,AbaoEBJ.1997b.Auto-matedchambermeasurementofmethaneandnitrousoxidefluxinfloodedricesoil.II.Fallowperiodemis-sions.SoilSci.Soc.Am.J.61:988-993.

BureshRJ,DeDatta SK. 1990.Denitrification lossesfrompuddled rice soils in the tropics.Biol.Fertil.Soils9:1-13.

CabangonRJ,TuongTP.2000.Managementofcrackedsoilsforwatersavingduringlandpreparationforricecultivation.SoilTillageRes.56:105-116.

CabangonRJ,CastilloEG,BaoLX,LuGA,WangGH,CuiYL,TuongTP,BoumanBAM,LiYH,ChenCD,WangZJ.2001.Impactofalternatewettinganddrying irrigation on rice growth and resource-useefficiency.In:BarkerR,LiYH,TuongTP,editors.Water-saving irrigation for rice.ProceedingsofanInternationalWorkshopheldinWuhan,China,23-25March2001.Colombo(SriLanka):InternationalWaterManagementInstitute.p55-80.

CabangonRJ,TuongTP,AbdullahNB.2002.Comparingwaterinputandwaterproductivityoftransplantedanddirect-seededriceproductionsystems.Agric.WaterManage.57:11-31.

CabangonRJ,TuongTP,LuG,BoumanBAM,FengY,ZhichuanZ,ChenCD,Wang JC. 2003. Irrigationmanagementeffectsonyieldandwaterproductivityofhybrid,inbredandaerobicricevarietiesinChina.In:Sustainabledevelopmentofwaterresourcesandmanagementandoperationofparticipatoryirrigationorganizations.Shiang-KueenHsu,editor.Proceedingsofthe2003InternationalCommissiononIrrigationandDrainage (ICID)AsianRegionalWorkshopon10-12Nov.2003,Taipei,Taiwan.Volume1.p184-210.

CabangonRJ,TuongTP,CastilloEG,BaoLX,LuG,WangGH,CuiL,BoumanBAM,LiY,ChongdeChen, Jianzhang Wang. 2004. Effect of irrigationmethodandN-fertilizermanagementonriceyield,water productivity and nutrient-use efficiencies intypicallowlandriceconditionsinChina.RiceFieldWaterEnviron.2:195-206.

ChoudhuryBU,BoumanBAM,SinghAK.2007.Yieldandwaterproductivityofrice-wheatonraisedbedsatNewDelhi,India.FieldCropsRes.100:229-239.

ConnorDJ,TimsinaJ,HumphreysE.2003.Prospectsforpermanentbedsintherice-wheatsystem.In:LadhaJK,HillJE,DuxburyJM,GuptaRK,BureshRJ,edi-tors.Improvingtheproductivityandsustainabilityofrice-wheatsystems:issuesandimpacts.ASASpecialPublication 65.ASA Inc.,CSSA Inc., SSSA Inc.,Madison,Wis.,USA.p197-210.

CruzRT,O’Toole JC. 1984.Dryland rice response toan irrigationgradient atflowering stage.Agron. J.76:178-183.

CuiY,ChengX,HanC,LiR.2000.TheeconomicandecologicalsatisfactoryamountofnitrogenfertilizerusingonriceinTaiLakeWatershed.ActaEcol.Sin.20:659-662.(InChinesewithEnglishabstract.)

DaweD.2005.Increasingwaterproductivityinrice-basedsystemsinAsia—pasttrends,currentproblems,andfutureprospects.PlantProd.Sci.8:221-230.

DaweD,DobermannA,MoyaP,AbdulrachmanS,BijaySingh,LalP,LiSY,LinB,PanaullahG,SariamO,SinghY,SwarupA,TanPS,ZhenQX.2000.Howwidespreadareyielddeclinesinlong-termriceexperi-mentsinAsia?FieldCropsRes.66:175-193.

DaweD,FrolkingS,ChangsengLi.2004.Trendsinrice-wheatareainChina.FieldCropsRes.87:89-95.

DeDattaSK.1981.Principlesandpracticesofricepro-duction.LosBaños(Philippines):InternationalRiceResearchInstitute.618p.

DeDattaSK,KruppHK,AlvarezEI,ModgalSC.1973.Watermanagementinfloodedrice.In:Watermanage-mentinPhilippineirrigationsystems:researchandoperations.LosBaños (Philippines): InternationalRiceResearchInstitute.p1-18.

DittertK,LinShan,KreyeC,ZhengXH,XuYC,LuXJ,Shen QR, Fan XL, Sattelmacher B. 2002. Savingwater with ground cover rice production systems(GCRPS)at thepriceof increasedgreenhousegasemissions?In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Wa-ter-wise rice production.LosBaños (Philippines):InternationalRiceResearchInstitute.p197-206.

DobermannA.2003.Acriticalassessmentofthesystemofriceintensification(SRI).Agric.Syst.79:261-281.

DobermannA,WittC,AbdulrachmanS,GinesHC,Naga-rajanR,SonTT,TanPS,WangGH,ChienNV,ThoaVTK,PhungCV,StalinP,MuthukrishnanP,RaviV,BabuM,SimbahanGC,AdvientoMAA.2003.Soilfertilityandindigenousnutrientsupply in irrigatedricedomainsofAsia.Agron.J.95:913-923.

DongB,MoldenD,LoeveR,LiYH,ChenCD,WangJZ.2004.FarmlevelpracticesandwaterproductivityinZangheIrrigationSystem.RiceFieldWaterEnviron.2:217-226.

DoorenbosJ,PruitWO.1984.Guidelinesforpredictingcropwaterrequirements.FAOIrrigationandDrain-agePaper 24.Rome (Italy):Food andAgricultureOrganization.144p.

EhlersW,GossM.2003.Waterdynamicsinplantproduc-tion.CABIPublishing,CABInternational,Walling-ford,UK.273p.

EkanayakeIJ,DeDattaSK,SteponkusPL.1989.Spikeletsterilityandfloweringresponseofricetowaterstressatanthesis.Ann.Bot.63:257-264.

FalkenmarkM,RockströmJ.2004.Balancingwaterforhumansandnature:thenewapproachinecohydrol-ogy.London(UK):Earthscan.247p.

��

FAO (Food andAgricultureOrganization). 1998.Cropevaporation: guidelines for computing crop waterrequirements.IrrigationandDrainagePaper56.FAO,Rome.300p.

FengYW,Yoshinaga I, ShirataniE,HitomiT,HasebeH. 2004.Characteristics andbehavior of nutrientsinapaddyareaequippedwitharecyclingirrigationsystem.Agric.WaterManage.68:47-60.

FengLiping,BoumanBAM,TuongTP,CabangonRJ,LiYalong,LuGuoan,FengYuehua.2007.Exploringoptionstogrowriceunderwater-shortconditionsinnorthernChinausingamodellingapproach.I:Fieldexperiments and model evaluation. Agric. WaterManage.88:1-13.

FernandoCH,GoltenbothF,Margraf J, editors. 2005.Aquaticecologyofricefields.VolumesPublishing,KitchenerOntario,Canada.472p.

GeorgeT,MagbanuaR,GarrityDP,TubañaBS,QuitonJ.2002.Rapidyieldlossofricecroppedsuccessivelyinaerobicsoil.Agron.J.94:981-989.

GleickPH.2000.Thechangingwaterparadigm:a lookattwenty-firstcenturywaterresourcesdevelopment.WaterInt.25:127-138.

HafeezMM.2003.Wateraccountingandproductivityatdifferentscalesinariceirrigationsystem:aremotesensingapproach.EcologyandDevelopmentSeries,No.8,UniversityofBonn,Germany.155p.

HafeezMM,BoumanBAM,VandeGiesenN,VlekP.2007.Scaleeffectsonwateruseandwaterproductiv-ityinarice-basedirrigationsystem(UPRIIS)inthePhilippines.Agric.WaterManage.92:81-89.

HamiltonRW,editor.2003.Theartofrice:spiritandsus-tenanceinAsia.UCLAFowlerMuseumofCulturalHistory,LosAngeles,Calif.,USA.552p.

HarnpichitvitayaD,TrebuilG,OberthürT,PantuwanG,CraigI,TuongTP,WadeLJ,Suriya-ArurojD.2000. Identifyingrecommendationsforsubsoilcompactionandsoilmanagementtoimprovewater-andnutrient-useefficiencyinrainfedlowlandrice.In:TuongTP,KamSP,WadeL,PandeyS,BoumanBAM,HardyB,editors.Characterizingandunderstandingrainfedenvironments.ProceedingsoftheInternationalWork-shoponCharacterizingandUnderstandingRainfedEnvironments,5-9Dec.1999,Bali,Indonesia.LosBaños (Philippines): InternationalRiceResearchInstitute.p97-110.

HobbsP,GuptaRK.2003.Rice-wheatcroppingsystemsintheIndo-GangeticPlains:issuesofwaterproductivity.I.Relationtonewresource-conservingtechnologies.In:KijneJW,BarkerR,MoldenD,editors.Waterproductivityinagriculture:limitsandopportunitiesfor improvement.CABIPublishing,Wallingford,UK.p239-253.

HongL,LiYH,DengL,ChenCD,DaweD,BarkerR.2001. Analysis in changes in water allocation andcropproductionintheZangheIrrigationSystemandDistrict–1966to1998.In:BarkerR,LoeveR,LiYH,TuongTP,editors.Water-savingirrigationforrice.Proceedingsofan internationalworkshop,Wuhan,China, 23-25March 2001.Colombo (SriLanka):InternationalWaterManagementInstitute.p11-24.

HumphreysE,MeisnerC,GuptaR,TimsinaJ,BeecherHG,TangYongLu,Yadvinder-Singh,GillMA,Masih I, Zheng JiaGuo, Thomposon JA. 2005.Water savings in rice-wheat systems. Plant Prod.Sci.8:242-258.

IkedaM,WatanabeT.2002.Nitrogenaccumulationinricefieldriceirrigatedwithwatercontainingammoniumand nitrate during the reproductive growth period.SoilSci.PlantNutr.48:105-109.

IRRI(InternationalRiceResearchInstitute).1987.Physicalmeasurementsinricesoils:theJapanesemethodolo-gies.LosBaños(Philippines):IRRI.

JhaKP,DineshChandra, Challaiah, 1981. Irrigationrequirement of high-yielding rice varieties grownonsoilshavingshallowwater-table.IndianJ.Agric.Sci.51:732-737.

KampenJ.1970.Waterlossesandwaterbalancestudiesinlowlandriceirrigation.PhDthesis,CornellUni-versity,USA.416p.

KirkG. 2004. The biochemistry of submerged soils.Chichester,West Sussex (UK): JohnWiley andSons.291p.

Kukal SS,HumphreysE,Yadvinder-Singh,Timsina J,Thaman S. 2006. Performance of raised beds inrice-wheat systems of north west India. In: RothC,FischerAR,MeisnerC,editors.Evaluationandperformance of permanent raised bed systems inAsiaandAustralia.ProceedingsofanInternationalWorkshop, 1-3March 2005,Griffith,NewSouthWales.Canberra (Australia):AustralianCenter forInternationalAgriculturalResearch.(Inpress.)

LafitteHR,Bennett J. 2002.Requirements for aerobicrice:physiologicalandmolecularconsiderations.In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Water-wise riceproduction.LosBaños (Philippines): InternationalRiceResearchInstitute.p259-274.

LafitteRH,CourtoisB,ArraudeauM.2002.Geneticim-provementofriceinaerobicsystems:progressfromyieldtogenes.FieldCropsRes.75:171-190.

��

LampayanRM,BoumanBAM,DeDiosJL,LactaoenAT,EspirituAJ,NorteTM,QuilangEJP,TabbalDF,Llor-caLP,SorianoJB,CorpuzAA,MalasaRB,VicmudoVR.2005.TransferofwatersavingtechnologiesinriceproductioninthePhilippines.In:ThiyagarajanTM,HengsdijkH,BindrabanP,editors.Transitionsinagricultureforenhancingwaterproductivity.Proceed-ingsoftheInternationalSymposiumonTransitionsinAgricultureforEnhancingWaterProductivity,23-25September2003,Kilikulam,TamilNaduAgriculturalUniversity,TamilNadu,India.p111-132.

LatifMA,IslamMR,AliMY,SalequeMA.2005.Vali-dationofthesystemofriceintensification(SRI)inBangladesh.FieldCropsRes.93:281-292.

LiYH,BarkerR. 2004. Increasingwater productivityfor ricefield irrigation inChina.RiceFieldWaterEnviron.2:187-193.

LilleyJM,FukaiS.1994.Effectoftimingandseverityofwaterdeficitonfourdiversericecultivars.II.Physi-ologicalresponsestosoilwaterdeficit.FieldCropsRes.37:215-223.

LoeveR,DongB,ZhaoJH,WangJZ,MoldenD.2001.Operations of theZhanghe Irrigation System. In:Water-savingirrigationforrice.BarkerR,LoeveR,LiYH,TuongTP,editors.Proceedingsofaninterna-tionalworkshop,Wuhan,China,23-25March2001.Colombo(SriLanka):InternationalWaterManage-mentInstitute.p25-54.

LoeveR,HongL,DongB,MaoG,ChenCD,DaweD,BarkerR.2004a.Long-termtrends in intersectoralwater allocation and crop water productivity inZangheandKaifeng,China.RiceFieldWaterEn-viron.2:237-245.

LoeveR,DongB,MoldenD,LiYH,ChenCD,WangJZ.2004b. Issuesof scale inwaterproductivity in theZhanghe irrigation system: implications for irriga-tioninthebasincontext.RiceFieldWaterEnviron.2:227-236.

LövensteinH,LantingaE,RabbingeR, vanKeulenH.1992.Principlesof theoreticalproductionecology.Department of Theoretical Production Ecology,WageningenUniversity,Netherlands.116p.

MacleanJL,DaweD,HardyB,HettelGP,editors.2002.Ricealmanac.LosBaños(Philippines):InternationalRiceResearchInstitute.253pp.

MasumotoT,ShimizuK,HaiPT.2004.RolesoffloodsforagriculturalproductioninandaroundTonleSapLake.ACIARProduction,No.116,Canberra,Aus-tralia,p136-146.

McCauleyGN.1990.Sprinklervs.flooded irrigation intraditionalriceproductionregionsofSoutheastTexas.Agron.J.82:677-683.

McDonaldAL,HobbsPR,RihaSJ.2006.Doesthesystemofriceintensificationoutperformconventionalbestmanagement? A synopsis of the empirical record.Agric.Syst.96:31-36.

MillenniumEcosystemAssessment. 2005.Ecosystemsandhumanwell-being:synthesis.Washington,D.C.(USA):IslandPress.137p.

MitsunoT,KobayashiS,MaruyamaT.1982.Groundwaterrechargefunctionofpaddyfields:acaseofgroundwa-terbalanceanalysisinNobiPlain.J.Soc.Irrig.Drain-ageReclamationEng.50:11-18.(InJapanese.)

MoldenD.1997.Accountingforwateruseandproductiv-ity.SWIMPaper.InternationalWaterManagementInstitute,Colombo,SriLanka.

MoormanFR,vanBreemenN.1978.Rice: soil,water,land. LosBaños (Philippines): InternationalRiceResearchInstitute.185p.

MortimerAM,HillJE.1999.Weedspeciesshiftsinre-sponsetobroadspectrumherbicidesinsub-tropicalandtropicalcrops.BrightonCropProtectionConfer-ence2:425-437.

MoserCM,BarettCB.2003.Thedisappointingadoptiondynamics of a yield-increasing, low external-inputtechnology: thecaseofSRI inMadagascar.Agric.Syst.76:1085-1100.

MoyaP,HongL,DaweD,ChongdeC.2004.Theimpactofon-farmwatersavingirrigationtechniquesonriceproductivity and profitability inZanghe IrrigationSystem,Hubei,China.RiceFieldWaterEnviron.2:207-215.

MushtaqS,DaweD,HongLin,MoyaP.2006.Anassess-mentoftheroleofpondsintheadoptionofwater-saving irrigationpractices in theZanghe IrrigationSystem,China.Agric.WaterManage.83:100-110.

OueH,FukushimaT,MaruyamaT.1994.Temperaturemitigation function of paddy fields. J. Soc. Irrig.Drainage Reclamation Eng. 62:13-18. (In Japa-nese.)

PandeyS,VelascoL.2002.EconomicsofdirectseedinginAsia:patternofadoptionandresearchpriorities.In:PandeyS,MortimerM,WadeL,TuongTP,LopezK,BHardy,editors.Directseeding:researchstrategiesandopportunities.LosBaños(Philippines):Interna-tionalRiceResearchInstitute.p3-14.

PAWEES(InternationalSocietyofRicefieldandWaterEnvironmentEngineering).2005.ProceedingsoftheInternationalConference onManagement ofRiceFieldandWaterEnvironmentforSustainableRiceProduction,7-8September2005,KyotoUniversity,Kyoto,Japan.178p.

PengS,BoumanBAM,VisperasRM,CastañedaA,NieL, ParkH-K. 2006.Comparison between aerobicand flooded rice: agronomic performance in along-term(8-season)experiment.FieldCropsRes.96:252-259.

PingaliPL,HossainM,GerpacioRV.1997.Asian ricebowls:thereturningcrisis?CABInternational,Oxon,UK (in associationwith IRRI,LosBaños, Philip-pines).341p.

��

PiñheiroBdaS,deCastroEdaM,GuimarãesCM.2006.Sustainabilityandprofitabilityofaerobicriceproduc-tioninBrazil.FieldCropsRes.97:34-42.

PostelS.1997.Lastoasis:facingwaterscarcity.NewYork(USA):NortonandCompany.239p.

RamasamyS,tenBergeHFM,PurushothamanS.1997.Yieldformationinriceinresponsetodrainageandnitrogenapplication.FieldCropsRes.51:65-82.

Ramsar. 2004.Ramsar handbooks for thewise use ofwetlands.Handbook 7.DesignatingRamsar sites.http://ramsar.org/lib/lib_handbooks_e.htm.

RijsbermanFR.2006.Waterscarcity:factorfiction?Agric.WaterManage.80:5-22.

RosegrantMW.1997.Waterresourcesinthetwenty-firstcentury:challengesandimplicationsforaction.Food,Agriculture,andtheEnvironmentDiscussionPaper20. International Food Policy Research Institute,Washington,D.C.,USA.

SanchezPA.1973.Puddlingtropicalricesoils.2.Effectsofwaterlosses.SoilSci.115(4):303-308.

SayreKD,HobbsPR.2004.Theraised-bedsystemofcul-tivationforirrigatedproductionconditions.In:LalR,HobbsP,UphoffN,HansenDO,editors.Sustainableagricultureandtheinternationalrice-wheatsystem.NewYork(USA):MarcelDekker,Inc.p337-355.

SethunathanN,SiddaramappaR.1978.Microbialdegra-dationofpesticidesinricesoils.In:Soilsandrice.LosBaños(Philippines):InternationalRiceResearchInstitute.p479-497.

SharmaPK,BhushanLav,LadhaJK,NareshRK,GuptaRK,BalasubramanianBV,BoumanBAM. 2002.Crop-water relations in rice-wheat cropping underdifferent tillage systems and water-managementpracticesinamarginallysodic,medium-texturedsoil.In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Water-wise riceproduction.InternationalRiceResearchInstitute,LosBaños,Philippines.p223-235.

Sheehy JE,PengS,DobermannA,MitchellPL,FerrerA, JianchangYang,YingbinZou,XuhuaZhong,JianliangHuang.2004.Fantasticyieldsinthesystemof rice intensification: fact or fallacy?FieldCropsRes.88:1-8.

ShuGeng,YixingZhou,MinghuaZhang,ShawnSmall-woodK.2001.Asustainableagro-ecologicalsolutiontowatershortageintheNorthChinaPlain(HuabeiPlain).J.Environ.PlanningManage.44:345-355.

SimpsonHJ,HerczegAL,MeyerWS.1992.Stableiso-toperatiosinirrigationwatercanestimatericecropevaporation.Geophys.Res.Lett.19:377-380.

SinclairTR,CassmanKG.2004.AgronomicUFOs.FieldCropsRes.88:9-10.

Singh RB. 2000. Environmental consequences of agri-cultural development: a case study from thegreenrevolution state ofHaryana, India.Agric.Ecosyst.Environ.82:97-103.

SinghAK,ChoudhuryBU,BoumanBAM.2002.Effectsofriceestablishmentmethodsoncropperformance,wateruse,andmineralnitrogen.In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Water-wisericeproduction.LosBaños (Philippines): InternationalRiceResearchInstitute.p237-246.

StoopW,UphoffN,KassamA.2002.Areviewofagri-culturalresearchissuesraisedbythesystemofriceintensification(SRI)fromMadagascar:opportunitiesfor improving farming systems for resource-poorfarmers.Agric.Syst.71:249-274.

SurridgeC.2004.Feastorfamine?Nature428:360-361.TabbalDF,BoumanBAM,BhuiyanSI,SibayanEB,Sat-

tarMA.2002.On-farmstrategiesforreducingwaterinputinirrigatedrice:casestudiesinthePhilippines.Agric.WaterManage.56:93-112.

TaoHongbin,BrueckH,DittertK,KreyeC,LinShan,SattelmacherB.2006.Growthandyieldformationofrice(Oryza sativaL.)inthewater-savinggroundcoverriceproductionsystem(GCRPS).FieldCropsRes.95:1-12.

ThompsonJ.1999.Methodsforincreasingricewateruseefficiency.In:RiceWaterUseEfficiencyWorkshopProceedings.CooperativeResearchCentreforSus-tainableRiceProduction,Leeton,NewSouthWales,Australia.p45-46.

ThompsonJA.2002.Watermanagementofriceinsouth-ernNewSouthWales,Australia.In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Water-wisericeproduction.LosBaños (Philippines): InternationalRiceResearchInstitute.p63-67.

TimsinaJ,ConnorDJ.2001.Productivityandmanagementofrice-wheatsystems:issuesandchallenges.FieldCropsRes.69:93-132.

TripathiRP,KushawaHS,MishraRK.1986. Irrigationrequirementsofriceundershallowwatertablecondi-tions.Agric.WaterManage.12:127-136.

TuongTP. 1999. Productivewater use in rice produc-tion: opportunities and limitations. J. Crop Prod.2:241-264.

TuongTP,BoumanBAM.2003.Riceproductioninwaterscarceenvironments.In:KijneJW,BarkerR,MoldenD,editors.Waterproductivityinagriculture:limitsand opportunities for improvement. Wallingford(UK):CABIPublishing.p53-67.

TuongTP,WopereisMCS,MarquezJ,KropffMJ.1994.Mechanisms and control of percolation losses inirrigatedpuddled ricefields.SoilSci. Soc.Am. J.58:1794-1803.

TuongTP,BoumanBAM,MortimerM.2005.Morerice,lesswater:integratedapproachesforincreasingwaterproductivityinirrigatedrice-basedsystemsinAsia.PlantProd.Sci.8:231-241.

��

UphoffN. 2007.Agroecological alternatives: capitalis-ing on existing genetic potentials. J. Dev. Studies43:218-236.

VanderGonHAC,VanBodegomPM,HouwelingS,VerburgP,VanBreemenN.2000.Combiningupscal-inganddownscalingofmethaneemissionsfromricefields:methodologiesandpreliminaryresults.Nutr.Cycl.Agroecosyst.58:285-301.

VenturaW,WatanabeI.1978.Growthinhibitionduetocontinuouscroppingofdrylandriceandothercrops.SoilSci.PlantNutr.24:375-389.

VenturaW,WatanabeI,CastilloMB,delaCruzA.1981.Involvementofnematodesinthesoilsicknessofadrylandrice-basedcroppingsystem.SoilSci.PlantNutr.27:305-315.

WangHuaqi,BoumanBAM,DuleZhao,WangChang-gui,MoyaPF.2002.AerobicriceinnorthernChina:opportunities and challenges. In:BoumanBAM,HengsdijkH,HardyB,BindrabanPS,TuongTP,LadhaJK,editors.Water-wisericeproduction.LosBaños (Philippines): InternationalRiceResearchInstitute.p143-154.

WassmannR,NeueHU,LantinRS,MakarimK,Chare-onsilpN,BuendiaLV,RennebergH.2000.Charac-terizationofmethaneemissions fromricefields inAsia. II. Differences among irrigated, rainfed, anddeepwaterrice.Nutr.Cycl.Agroecosyst.58:13-22.

WestcottMP,VinesKW.1986.Acomparisonofsprinklerandflooded irrigation for rice.Agron. J. 78:637-640.

WeiZhang,Si-tuSong.1989.Irrigationmodelofwatersaving-highyieldat lowlandpaddyfield.In:Inter-national Commission on Irrigation and Drainage,Seventh Afro-Asian Regional Conference, 15-25October1989.Tokyo,Japan.Vol.I-C:480-496.

WickhamTH,SinghVP.1978.Watermovementthroughwetsoils.In:Soilsandrice.LosBaños(Philippines):InternationalRiceResearchInstitute.p337-357.

WopereisMCS,KropffMJ,Bouma J, vanWijkALM,WoodheadT,editors.1994.Soilphysicalproperties:measurementanduseinrice-basedcroppingsystems.LosBaños(Philippines):InternationalRiceResearchInstitute,104p.

WopereisMCS,KropffMJ,MaligayaAR,TuongTP.1996. Drought-stress responses of two lowlandricecultivars tosoilwaterstatus.FieldCropsRes.46:21-39.

XueChangying,YangXiaoguang,BoumanBAM,DengWei,ZhangQiuping,YanWeixiong,ZhangTianyi,RouziAji,WangHuaqi,WangPu.2007.Effectsofirrigationandnitrogenontheperformanceofaerobicrice in northernChina. Submitted to FieldCropsResearch,January2007.

YangXiaoguang,BoumanBAM,WangHuaqi,WangZhimin,Zhao Junfang,ChenBin. 2005. Perform-anceoftemperateaerobicriceunderdifferentwaterregimes inNorth China. Agric.WaterManage.74:107-122.

ZwartSJ,BastiaanssenWGM.2004.Reviewofmeas-ured crop water productivity values for irrigatedwheat,rice,cottonandmaize.Agric.WaterManage.69:115-133.