<strong>in</strong>g that time was some 940 mm, of which 110mm was used for soak<strong>in</strong>g, 225 mm disappearedas surface runoff, 445 mm was lost by seepageand percolation, and 160 mm was lost by evaporation.In UPRIIS, farmers raise seedl<strong>in</strong>gs <strong>in</strong> part oftheir ma<strong>in</strong> field. Because of a lack of tertiary fieldchannels, the whole ma<strong>in</strong> field is soaked when theseedbed is prepared and rema<strong>in</strong>s flooded dur<strong>in</strong>gthe entire duration of the seedbed. In systems suchas UPRIIS, the turnaround time can be m<strong>in</strong>imizedby the <strong>in</strong>stallation of field channels, the adoptionof common seedbeds, or the adoption of direct wetor dry seed<strong>in</strong>g. With field channels, water can bedelivered to the <strong>in</strong>dividual seedbeds separately andthe ma<strong>in</strong> field does not need to be flooded. Commonseedbeds, either communal or privately managed,can be located strategically close to irrigation canalsand be <strong>irrigated</strong> as one block.With direct seed<strong>in</strong>g, the crop starts grow<strong>in</strong>gand us<strong>in</strong>g water from the moment of establishmentonward. Direct dry seed<strong>in</strong>g can also <strong>in</strong>creasethe effective use of ra<strong>in</strong>fall and reduce irrigationneeds as shown for the MUDA irrigation scheme<strong>in</strong> Malaysia (Cabangon et al 2002). However, dryseed<strong>in</strong>g with subsequent flood<strong>in</strong>g is possible only<strong>in</strong> heavy (clayey) soils with low permeability andpoor <strong>in</strong>ternal dra<strong>in</strong>age. A major driv<strong>in</strong>g force forthe adoption of direct seed<strong>in</strong>g <strong>in</strong> Asia is scarcityof labor s<strong>in</strong>ce direct seed<strong>in</strong>g does not use labor fortransplant<strong>in</strong>g and can be a mechanized operation.3.3 Crop growth period3.3.1 Saturated soil cultureIn saturated soil culture (SSC), the soil is kept asclose to saturation as possible, thereby reduc<strong>in</strong>gthe hydraulic head of the ponded water, which decreasesthe seepage and percolation flows. SSC <strong>in</strong>practice means that a shallow irrigation is given toobta<strong>in</strong> about 1 cm of ponded water depth a day orso after the disappearance of ponded water. Tabbalet al (2002) reported water sav<strong>in</strong>gs under SSC <strong>in</strong>transplanted and direct wet-seeded <strong>rice</strong> <strong>in</strong> puddledsoil, and <strong>in</strong> direct dry-seeded <strong>rice</strong> <strong>in</strong> nonpuddled soil(Table 3.1). Analyz<strong>in</strong>g a data set of 31 publishedfield experiments with an SSC treatment, Boumanand Tuong (2001) found that water <strong>in</strong>put decreasedon average by 23% (range: 5% to 50%) from thecont<strong>in</strong>uously flooded check, with a nonsignificantyield reduction of 6% on average. Thompson (1999)found that SSC <strong>in</strong> southern New South Wales,Australia, reduced both irrigation water <strong>in</strong>put andyield by a bit more than 10%.Raised beds can be an effective way to keepthe soil around saturation. <strong>Rice</strong> plants are grown onbeds and the water <strong>in</strong> the furrows is kept close to thesurface of the beds. In Australia, Borell et al (1997)experimented with raised beds that were 120 cmwide and separated by furrows of 30-cm width and15-cm depth to facilitate SSC practices. Comparedto flooded <strong>rice</strong>, water sav<strong>in</strong>gs were 34% and yieldlosses 16−34%. More <strong>in</strong>formation on raised bedsis found <strong>in</strong> Chapter 3.4.1.Practical implementationAlthough conceptually sound, SSC will be difficult toimplement practically s<strong>in</strong>ce it requires frequent (dailyor once every two days) applications of small amountsof irrigation water to just keep a stand<strong>in</strong>g water depthof 1 cm on flat land, or to keep furrows filled just to thetop <strong>in</strong> raised beds.3.3.2 Alternate wett<strong>in</strong>g and dry<strong>in</strong>gIn alternate wett<strong>in</strong>g and dry<strong>in</strong>g (AWD), irrigationwater is applied to obta<strong>in</strong> flooded conditions after acerta<strong>in</strong> number of days have passed after the disappearanceof ponded water. The number of days ofnonflooded soil <strong>in</strong> AWD before irrigation is appliedcan vary from 1 day to more than 10 days. ThoughTable 3.1a. Yield, water <strong>in</strong>put, and water productivity with respect to total water <strong>in</strong>put (WP IR) <strong>in</strong> transplanted and wet-seeded<strong>rice</strong> under cont<strong>in</strong>uous flood<strong>in</strong>g and SSC, Muñoz, 1991 dry season. Data from Tabbal et al (2002).TreatmentTransplantedWet-seededYield <strong>Water</strong> <strong>in</strong>put WP IRYield <strong>Water</strong> WP IR(t ha −1 ) (mm) (g gra<strong>in</strong> kg −1 water) (t ha −1 ) <strong>in</strong>put (mm) (g gra<strong>in</strong> kg −1 water)Flooded 7.4 694 1.06 7.6 631 1.20SSC 6.7 373 1.81 7.3 324 2.2719
Table 3.1b. Yield, water <strong>in</strong>put, and water productivity with respect to total water <strong>in</strong>put (WP IR) and irrigation (WP I) <strong>in</strong> dry-seeded<strong>rice</strong> under cont<strong>in</strong>uous flood<strong>in</strong>g and SSC, San Jose City, Philipp<strong>in</strong>es, 1996-97. Data from Tabbal et al (2002).Treatment<strong>Water</strong> <strong>in</strong>put (mm)<strong>Water</strong> productivity (g gra<strong>in</strong> kg −1 water)Yield (t ha −1 ) Irrigation + ra<strong>in</strong>fall Irrigation WP IRWP I1996Flooded 4.3 1,417 531 0.31 8.16SSC 4.2 1,330 432 0.32 9.651997Flooded 4.7 1,920 941 0.25 4.99SSC 4.5 1,269 355 0.36 12.81Table 3.2. Yield, water use, and water productivity with respect to irrigation and ra<strong>in</strong>fall of <strong>rice</strong> under alternate wett<strong>in</strong>g anddry<strong>in</strong>g (AWD) and cont<strong>in</strong>uously flooded conditions. Data from Bouman et al (2006a).Location Year TreatmentYield Total water <strong>Water</strong> productivity(t ha −1 ) <strong>in</strong>put (mm) (g gra<strong>in</strong> kg −1 water)Guimba, Philipp<strong>in</strong>es 1988 Flooded 5.0 2,197 0.23(Tabbal et al 2002) AWD 4.0 880 0.461989 Flooded 5.8 1,679 0.35AWD 4.3 700 0.611990 Flooded 5.3 2,028 0.26AWD 4.2 912 0.461991 Flooded 4.9 3,504 0.14AWD 3.3 1,126 0.29Tuanl<strong>in</strong>, Huibei, Ch<strong>in</strong>a 1999 Flooded 8.4 965 0.90(Belder et al 2004) AWD 8.0 878 0.952000 Flooded 8.1 878 0.92AWD 8.4 802 1.07Muñoz, Philipp<strong>in</strong>es, 2001 Flooded 7.2 602 1.20(Belder et al 2004) AWD 7.7 518 1.34some researchers have reported a yield <strong>in</strong>creaseus<strong>in</strong>g AWD (Wei Zhang and Song 1989, Stoopet al 2002), recent work <strong>in</strong>dicates that this is theexception rather than the rule (Belder et al 2004,Cabangon et al 2004, Tabbal et al 2002; Table 3.2).In 31 field experiments analyzed by Bouman andTuong (2001), 92% of the AWD treatments resulted<strong>in</strong> yield reductions vary<strong>in</strong>g from just more than 0%to 70% compared with those of the flooded checks.In all these cases, however, AWD <strong>in</strong>creased waterproductivity (WP IR) with respect to total water <strong>in</strong>putbecause the reductions <strong>in</strong> water <strong>in</strong>puts were largerthan the reductions <strong>in</strong> yield. The large variability <strong>in</strong>results of AWD <strong>in</strong> the analyzed data set was causedby differences <strong>in</strong> the number of days between irrigationsand <strong>in</strong> soil and hydrological conditions.Experiment<strong>in</strong>g with AWD <strong>in</strong> lowland <strong>rice</strong>areas with heavy soils and shallow groundwatertables <strong>in</strong> Ch<strong>in</strong>a and the Philipp<strong>in</strong>es, Cabangonet al (2004), Belder et al (2004), Lampayan et al(2005), and Tabbal et al (2002) reported that total(irrigation and ra<strong>in</strong>fall) water <strong>in</strong>puts decreased byaround 15–30% without a significant impact onyield. In all these cases, groundwater depths werevery shallow (between 10 and 40 cm), and pondedwater depths almost never dropped below the rootzone dur<strong>in</strong>g the dry<strong>in</strong>g periods (Fig. 3.3), thus turn<strong>in</strong>gAWD effectively <strong>in</strong>to a k<strong>in</strong>d of near-saturatedsoil culture. Even without ponded water, plant rootsstill had access to “hidden” water <strong>in</strong> the root zone(Chapter 1.4). More water can be saved and waterproductivity further <strong>in</strong>creased by prolong<strong>in</strong>g the20