17在干旱地区最大化水分生产率的一种非充分灌溉农业战略策略.pdf

《17在干旱地区最大化水分生产率的一种非充分灌溉农业战略策略.pdf》由会员分享，可在线阅读，更多相关《17在干旱地区最大化水分生产率的一种非充分灌溉农业战略策略.pdf（10页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、ReviewDeficit irrigation as an on-farm strategy to maximize crop water productivityin dry areasSam Geerts*,Dirk RaesK.U.Leuven(University of Leuven),Division of Soil and Water Management,Celestijnenlaan 200 E,B-3001Leuven,BelgiumContents1.Rationale.12752.Crop water productivity.12762.1.The concept.1

2、2762.2.The crop water production function.12763.Deficit irrigation:deliberately tolerating drought stress.12773.1.The concept of deficit irrigation.12773.2.Research results for different crops.12773.2.1.Seasonal crop water production functions.12773.2.2.Drought stress differentiated by phenological

3、stage.12783.3.Advantages and constraints of deficit irrigation.12783.4.Reasons for increased water productivity under deficit irrigation.12794.Modeling as a tool for assessing and developing deficit irrigation strategies.12795.Conclusion.1280Acknowledgements.1280References.12821.RationaleTo sustain

4、the rapidly growing world population,agriculturalproduction will need to increase(Howell,2001),yet the portion offresh water currently available for agriculture(72%)is decreasing(Cai and Rosegrant,2003).Hence,sustainable methods to increasecrop water productivity are gaining importance in arid and s

5、emi-aridregions(DebaekeandAboudrare,2004).Traditionally,agricultural research has focused primarily on maximizing totalproduction.In recent years,focus has shifted to the limiting factorsin production systems,notably the availability of either land orwater.Within this context,deficit irrigation(DI)h

6、as been widelyinvestigated as a valuable strategy for dry regions(English,1990;Pereira et al.,2002;Fereres and Soriano,2007)where water is thelimiting factor in crop cultivation.We review recent research onthe maximization of productivity per unit of water by DI and wediscuss crop water productivity

7、 modeling as a tool for assessingand designing DI strategies.Agricultural Water Management 96(2009)12751284A R T I C L EI N F OArticle history:Received 9 May 2008Accepted 14 April 2009Available online 14 May 2009Keywords:Water use efficiencyCrop evapotranspirationWater stressArid regionsWater produc

8、tion functionA B S T R A C TDeficit irrigation(DI)has been widely investigated as a valuable and sustainable production strategy indry regions.By limiting water applications to drought-sensitive growth stages,this practice aims tomaximize water productivity and to stabilize rather than maximize yiel

9、ds.We review selectedresearch from around the world and we summarize the advantages and disadvantages of deficitirrigation.Research results confirm that DI is successful in increasing water productivity for variouscrops without causing severe yield reductions.Nevertheless,a certain minimum amount of

10、 seasonalmoisture must be guaranteed.DI requires precise knowledge of crop response to drought stress,asdroughttolerance varies considerably bygenotype and phenological stage.Indeveloping and optimizingDI strategies,field research should therefore be combined with crop water productivity modeling.?2

11、009 Elsevier B.V.All rights reserved.*Corresponding author.Tel.:+32 16 32 97 54;fax:+32 16 32 97 60.E-mail addresses:sam.geertsbiw.kuleuven.be,(S.Geerts).Contents lists available at ScienceDirectAgricultural Water Managementjournal homepage: front matter?2009 Elsevier B.V.All rights reserved.doi:10.

12、1016/j.agwat.2009.04.0092.Crop water productivity2.1.The conceptCrop water productivity(WP)or water use efficiency(WUE),asreviewed by Molden(2003),is a key term in the evaluation of DIstrategies.Water productivity with dimensions of kg m?3isdefined as the ratio of the mass of marketable yield(Ya)to

13、thevolume of water consumed by the crop(ETa):WP YaETa(1)ETarefers to water lost either by soil evaporation or by croptranspiration during the crop cycle.Since there is no easy way ofdistinguishing between these two processes in field experiments,they are generally combined under the term of evapotra

14、nspiration(ET)(Allen et al.,1998).In water-scarce regions,crops with high WP should bepreferred,although this is not the only factor.Indeed,whilehigh-energy fruit and grain crops(e.g.crops with high proteincontent)may have a lower absolute WP value(Steduto andAlbrizio,2005),their nutritional value i

15、s higher,which should beconsidered when assessing these crops for use in drought-proneareas.WP values reported in literature vary according to whetherauthors express the denominator as the amount of water applied(the sum of rainfall and irrigation)or as the amount of watertranspired(unproductive soi

16、l evaporation is not taken intoaccount).2.2.The crop water production functionThe crop water production function(CWP function)expressesthe relation between obtained marketable yield(Ya)and the totalamount of water evapotranspired(ETa)(Stewart et al.,1977;Hexem and Heady,1978;Doorenbos and Kassam,197

17、9;Tayloret al.,1983).The highest water efficiency level in the CWP functionis determined using WP as a benchmark.As shown in Fig.1,theCWP function has a logistic shape(Hanks et al.,1969;Hanks,1974).Its axes are made dimensionless by plotting relative yield(Yrel:ratio of actual,Ya,to maximum possible

18、 yield under givenagronomic conditions,Ym)versus relative evapotranspiration(ETrel:ratio of actual evapotranspiration,ETa,to crop ET undernon-stressed,standard conditions,ETc).Within the CWP function,different sections can be distin-guished that may vary in width or that may even be absent:-Section

19、a:If insufficient water is applied during the crop cycle,the crop will not develop fully,resulting in low-quality yield(shriveled grains or fruits with low market value)or even totalloss of yield(Yazar and Sezen,2006).In this section,WP is verylow,and crop yield and WP can only be increased if acons

20、iderable amount of water is added and section b is reached(Geerts et al.,2008b).More research is needed to determine thislower limit for various crops.-Section b:Once a minimum amount of water(A)is guaranteed byresidual moisture,rainfall and/or irrigation,yields(and thereforeWP)start to increase wit

21、h increasing water levels.If this sectionis present,it has a concave shape:increasing water supply willalways result in an increased WP from A to B.-Section c:With additional water application,the productionfunction can become nearly linear,with a slope ranging frommild to sharp.Doorenbos and Kassam

22、(1979)point out that therelationship between Yreland ETrelremains linear for ETrelup to alower limit of 0.5(point B in Fig.1),although this lower limit hasnot been defined for all crops.-Section d:As observed for many crops,the slope of the CWPfunction often decreases once ETrelis close to 1.Towards

23、 theupper limit of ETrel,the proportional yield increase per unit ETgradually levels off.Possible reasons are highlighted in Section3.4 of this review.Section d can be quite large,for crops such asalfalfa,sugar beets(Doorenbos and Kassam,1979),wheat(Kanget al.,2002;Zhang et al.,2008;Sun et al.,2006)

24、or cotton(Henggeler et al.,2002;Kanber et al.,2006;DeTar,2008),while itmay be almost absent for other crops,such as maize(Kipkoriret al.,2002;Farre and Faci,2006;Payero et al.,2006).In theliterature,this section is often described using combinations oflinear functions(i.e.a broken stick model).When

25、the crop water function includes excess irrigation and/or rainfall,it has a more pronounced S shape(Fig.1),creating anadditional section:-Section e:Applying more water than required by ETcwill notincrease yield,as the water is lost through unproductive soilevaporation and/or deep percolation.If too

26、much water isapplied,yield might even decline as a result of water logging orleaching of nutrients from the root zone(Sun et al.,2006;Cabelloet al.,2009).In this section,irrigation is therefore not required,unless the root zone needs to be leached to reduce salinity.ThelevelofETaorETrelcorresponding

27、withthehighestWPcanbefound by first deriving the WP function(WP versus ETa)and thensetting the first order derivative of this function to zero.MaximumWP will be found at an ETalevel within section c or d.For the linearsection c,WP is highest at point B if the extrapolated Y-intercept ispositiveandhi

28、ghestatpointCoratahigherETaiftheextrapolatedY-interceptisnegative.IfmaximumWPislocatedinsectiond(Eq.(2),it is located at the point where Eq.(3)equals zero.WPsection-d a?ETa b c?ETa?1(2)dWPsection-ddETa a?c?ETa?2(3)The distinction between drought-tolerant and-sensitive crops isnot straightforward and

29、 depends on the range of ETawithin whichit is defined(Fig.2).In Fig.2a maximum WP is reached for ETalower than ETc,whereas in Fig.2b WP increases until full waterrequirements are met(point D).Fig.1.General shape of a crop water production(CWP)function.Sections(a),(b),(c),(d)and(e)have variable relat

30、ive widths.Relative yield is the ratio betweenactual(Ya)and potential yield(Ym)under given agronomic conditions,whilerelative evapotranspiration is the ratio between the seasonal amount of water thatis evapotranspired(ETa)and seasonal crop water requirements(ETc).S.Geerts,D.Raes/Agricultural Water M

31、anagement 96(2009)127512841276If water shortage(ETrel 1)is evenly distributed over thecropping cycle,the corresponding yield decline can be derivedfrom the seasonal ETrel.Figs.1 and 2 indeed show such seasonallydetermined CWP functions(CWPsfunction).However,for manycrops,drought tolerance varies str

32、ongly between growth stages(e.g.Martyniak,2008).Hence,the CWP functions for theseindividual growth stages will differ in shape from the CWPsfunction,as the relationship does not account for the effect of thetiming of water application.Combining various CWP functions perphenological stage is difficul

33、t,in part because there are differentcombination methods,particularly with linear CWP functions(section c)(Jensen,1968;Hiller and Clark,1971;Hanks,1974;Stewart et al.,1977;Varlev et al.,1996),and each method hasadvantages and disadvantages.If a combined CWP function(CWPcfunction)could be constructed

34、,differentiating drought stresslevels over the phenological stages,the general shape wouldremain similar to that of the CWPsfunction,but increased scatterwould make it more difficult to establish general guidelines.In thiscontext,crop water productivity modeling becomes a valuabletool.3.Deficit irri

35、gation:deliberately tolerating drought stress3.1.The concept of deficit irrigationDeficit irrigation is an optimization strategy in which irrigationis applied during drought-sensitive growth stages of a crop.Outside these periods,irrigation is limited or even unnecessary ifrainfall provides a minimu

36、m supply of water.Water restriction islimited to drought-tolerant phenological stages,often the vege-tative stages and the late ripening period.Total irrigationapplicationisthereforenot proportionaltoirrigationrequirementsthroughout the crop cycle.While this inevitably results in plantdrought stress

37、 and consequently in production loss,DI maximizeswater productivity,which is the main limiting factor(English,1990).In other words,DI aims at stabilizing yields and at obtainingmaximum WP rather than maximum yields(Zhang and Oweis,1999).In the literature,the terms supplemental irrigation anddeficiti

38、rrigationarebothused.Thefirsttermgenerallyreferstoarain-fed crop receiving additional irrigation during the wholeseason or during sensitive growth stages,whereas DI generallyrefersto fullyirrigatedcropsfromwhichwateris withheldduringcertain tolerant growth stages.Both terms are often usedinterchange

39、ably,which maycauseconfusion.Toavoidambiguity,deficit irrigation is therefore used as the only term throughoutthis review.Since drought tolerance varies considerably by genotype andby phenological stage,DI requires precise knowledge of cropresponse to drought stress for each of the growth stages(Kir

40、daet al.,1999).In addition,correct application of DI requires athorough assessment of the economic impact of the yieldreduction caused by drought stress(English,1990;English andRaja,1996;Sepaskhah and Akbari,2005;Sepaskhah et al.,2006).In areas where water is the most limiting factor,maximizing WPma

41、y be economically more profitable for the farmer thanmaximizing yields(English,1990).For instance,water savedby DI can be used to irrigate more land(on the same farm or inthe water users community),which given the high opportunitycost of water may largely compensate for the economic loss dueto yield

42、 reduction(Kipkorir et al.,2001;Ali et al.,2007).Evenwater transfers from water-rich to water-poor areas are possible,as recently demonstrated in California,where DI was used foralfalfa(Hanson et al.,2007).As these examples suggest,DIrequires a highly integrated approach to agricultural waterpolicy.

43、3.2.Research results for different crops3.2.1.Seasonal crop water production functionsThis section discusses a number of studies in which irrigationapplications or levels of tolerated drought stress did not differbetweenphenologicalstages.TheresultingseasonalCWPsfunction has the theoretical shape sh

44、own in Fig.1,with variablesection widths.If different shapes are presented in the literature,this may be due to the fact that the drought stress in theexperimental design did not cover the complete relative ET range(Fabeiro et al.,2002;Oweis et al.,2004),that certain sections mayhave been small or e

45、ven absent(Payero et al.,2006),or that certainsub-sections were approximated by(a combination of)linearfunctions.The general framework presented in Fig.1 allows us to comparethe efficiency of DI versus rain-fed cultivation and/or full irrigation(FI)for a particular crop in a particular location.In m

46、any(semi)-Fig.2.Possible examples of the shapes of crop water production(CWP)functions for a relatively drought-tolerant(a)and a drought-sensitive(b)crop.S.Geerts,D.Raes/Agricultural Water Management 96(2009)127512841277arid areas,DI is more efficient than rain-fed cultivation.Themagnitude of the ef

47、fect depends on the amount of rain available tomeet minimum ET requirements.Only in a few cases is FI requiredto reach maximum water productivity:this happens when sectiond is absent and the linear section c has a negative extrapolated Y-intercept(cf.Fig.2b).For wheat in different locations,the lite

48、rature on CWPsindicates that DI should be preferred over FI due to convex CWPsfunctions(Zhang,2003;Tavakkoli and Oweis,2004)or linear CWPsfunctions with maximum WP at sub-maximal ETa(Tolk andHowell,2008).Other crops also respond favorably to a reduction inseasonal irrigation.A convex quadratic CWPsf

49、unction wasreported for lentil(Oweis et al.,2004),cotton(Henggeler et al.,2002),green gram(Webber et al.,2006),soy bean(Sincik et al.,2008)and safflower(Lovelli et al.,2007)in varying locations,whilea linear CWPsfunction with positive extrapolated Y-intercept(cf.Fig.2a,linear approximation of upper

50、sub-section of d)or a convexquadratic CPWsfunction was found for sugarbeet(Bazza(1999)and Doorenbos and Kassam(1979),respectively).For maize grownin different locations,linear CWPsfunctions with a negativeextrapolatedY-intercept(cf.Fig.2b)aremostlyreported(Farre andFaci,2006;Payero et al.,2006;Igbad