1

Keywords: water-saving technologies, controlled irrigation, participatory R&D and extension, processdocumentation

ADOPTION OF WATER SAVING TECHNOLOGIES IN RICEPRODUCTION IN THE PHILIPPINES1

R.M. Lampayan1, B.A.M. Bouman1, J.L. de Dios3, A.T. Lactaoen2, A.J. Espiritu3,T.M. Norte2, E.J.P. Quilang3, D.F. Tabbal1, L.P. Llorca2, J.B. Soriano2, A.A. Corpuz3,

R.B. Malasa3 and V.R. Vicmudo2

1International Rice Research Institute, Los Baños, Laguna, Philippines2National Irrigation Administration, Groundwater Irrigation System Reactivation

Project, Tarlac, Philippines3Philippine Rice Research Institute (PhilRice), Maligaya, Muñoz, Philippines

ABSTRACT

Traditional lowland rice production in Asia requires much water: it consumes more than 50% ofall irrigation water used in the region. Water resources are, however, increasingly getting scarceand expensive. There is a need to develop alternative rice production systems that require lesswater and increase water productivity. In the last decade(s), researchers have studied anddeveloped a number of water-saving irrigation technologies. Although these technologies havebeen demonstrated to save water and increase water productivity, their adoption by farmers is lowbecause of a lack of extension. Compared with the heavy investments needed to develop newwater resources, the adoption of water-saving technologies by farmers is low-cost and has greatpotential to save water. Therefore, in 2001, a project was initiated to transfer and promote water-saving technologies among farmers in the Philippines called the “Technology Transfer for WaterSavings (TTWS)” project. The first two years of the project were designed as a participatorylearning phase with project partners. Controlled irrigation or alternate wetting and drying wasthe first matured water-saving technology included in the first phase of the project while theaerobic rice trials-cum-research were also integrated in the project. This paper documents theactivities of the TTWS project, describes the results and implications of the first two-yearimplementation, and explores a future course of action including widespread training andextension of water-saving technologies in the Philippines.

per capita availability declined by 40-60%between 1955 and 1990, and is expected todecline further by 15-54% over the next 35years (Gleick 1993). The main reasons arediverse and location specific, but includeincreasing population growth, increasing urbanand industrial demand, and decreasingavailability because of pollution (chemicals,salts, silts) and resource depletion. Inagriculture, the situation is aggravated by thedramatically increasing costs for irrigationdevelopment over the past decades. Because ofthe combined increasing demand for food withincreasing scarcity of water, rice producers facethree major challenges: (1) to save water; (2) to

INTRODUCTION

Rice is the most important food crop in Asia(IRRI 1997), however, it requires most water. Infact, the majority of the world’s rice is beingproduced under flooded, so-called lowlandconditions. Of the roughly 147 million ha riceland, 79 million ha is classified as irrigatedlowland, 36 million ha as rainfed lowland, and13 million ha as flood prone (IRRI 2002). Inthese ecosystems, rice is mostly grown inbunded, puddled fields under floodedconditions or so-called anaerobic conditions.

Fresh water for agriculture is becomingincreasingly scarce. In many Asian countries,1 This paper was also presented at the International Workshop "Transitions in Agriculture for Enhancing Water Productivity" in

Tamil Nadu, India, September 2003.

2

increase productivity; and (3) to produce morerice with less water (Bouman and Tuong 2001).In the Philippines, some 61% of the 3.4 millionha of rice land is under irrigation, with themajority of the production coming from the ricebowl in central Luzon (IRRI 1997). Irrigation isprovided by gravity systems and shallow anddeep tubewells. However, the availability ofwater for irrigation has declined in the lastdecade(s). Water from the Angat reservoir inBulacan Province is increasingly divertedtoward the Greater Manila Area (Pingali et al.1997), water in the Agno River IrrigationSystem in Pangasinan Province is polluted withsediments and chemicals from mining activitiesupstream (Castañeda and Bhuiyan 1993), andmany irrigation systems were destroyed andclogged by the earthquakes of 1990 and theMount Pinatubo eruption in 1991 (NIA 1996).Because of its dense population and closeproximity to the capital Manila, rice productionin central Luzon is of strategic importance tofood security and poverty alleviation. Thegovernment of the Philippines, through itsNational Irrigation Administration (NIA), isdedicated to maintaining and enhancingirrigation water availability by infrastructuredevelopment and maintenance and by thepropagation of water-saving irrigationtechnologies (NIA 1996).

The decreasing availability of water forirrigated rice threatens food security in Asia ingeneral and the livelihood of farmers inparticular. Also, the increasing scarcity of watermeans that the costs of its use and resourcedevelopment are increasing dramatically (Postel1997, Rosegrant 1997). Therefore, researchershave been looking for ways to decrease wateruse in rice production and increase its useefficiency. Though water use can be optimizedat scale levels from field to farm, irrigationsystem, watershed and entire river basins, afundamental approach is to save water at thefield level where water and the rice cropinteract. This is also the scale level thatconcerns rice farmers most. During the pastdecades, much research has been done at thefield level and various technologies have been

proposed that save water and increase itsproductivity while maintaining high yields(Sandhu et al. 1980, Mishra et al. 1990, Li2001). In the Philippines, pioneering researchhas been done by the International RiceResearch Institute (IRRI; Bhuiyan et al. 1995,Tabbal et al. 2002, Tuong 1999, Bouman andTuong 2001) and PhilRice (de Dios et al. 2000).Despite the good results obtained in research,however, very little attention has been paid tothe dissemination, extension and adoption ofthe developed technologies among farmers inthe Philippines. At the moment, it is not wellknown how farmers actually manage their waterand to what extent they are aware of water-saving technologies. It is generally assumedthat rice farmers in Asia have gotten used tothe idea of continuously flooding their fieldsfor much of the growing period. This practiceis tied up with weed control, ease fortransplanting and on the belief that reducingthe amount of water will be harmful to theplant. To bridge the gap between research onwater-saving technologies and adoption byfarmers, IRRI, PhilRice and NIA initiated in2001 a special project called “TechnologyTransfer for Water Savings (TTWS)” in riceproduction.

The TTWS project is part of theinternational Irrigated Rice Research Consortium(see IRRC page at the IRRI websitewww.cgiar.org/irri) through the WaterWorkgroup2 and has counterpart activities inChina and India. TTWS was conceived todevelop and implement a framework fortransfer, adaptation, and adoption of knowledgeon water-saving technologies through theinteragency collaboration of the NationalIrrigation Administration, Philippine RiceResearch Institute and IRRI. The first twoyears of the project are designed as aparticipatory learning phase with farmers whoare using irrigation water from deepwell andshallow tubewell groundwater systems in Tarlacand Nueva Ecija, respectively. Project siteselection, baseline characterization and needsand opportunities assessment were conductedin 2001. The actual implementation of the

2 The international water workgroup of the IRRC aimed to (1) obtain insights in current water-saving practices by farmers andidentify their behavior in coping with water scarcity; (2) have an inventory of water-saving technologies and identify themost promising ones; and, (3) promote the spread and exchange of information on sustainable water-saving technologiesthat increase the productivity and value of water, optimize farmers' objectives, and maintain the water resource base.

3

project started in the 2002 at 21 farmers’ fieldswith controlled irrigation (CI) as the first water-saving technology in dry season, and aerobicrice in the wet season at nine farmers’ fields.In the 2003 dry season, intensive farmer-participatory aerobic rice trials were added aftersuccessful trials in the 2002 wet season withpromising varieties. Controlled irrigation oralternate wetting and drying entails anirrigation schedule where water is applied tothe field a number of days after disappearanceof ponded water. This technology is adeparture from continuously flooding the fieldsand introduces period of dry (aerobic) soilconditions. Aerobic rice, on the other hand,refers to high-yielding rice grown in non-puddled, aerobic soil (Bouman et al. 2002,Wang Huaqi et al. 2002, Yang Xiaguang et al.2002). It entails the growing of rice in aerobicsoil, with the use of external inputs such assupplementary irrigation and fertilizers, andaiming at high yields. It has characteristics ofboth upland and lowland varieties.

This paper reports on the first two-yearresults of implementation of the CI and theaerobic rice in the TTWS pilot sites. Thispaper also attempts to sketch a possible future

direction for adaptation and adoption of water-saving technologies in rice production in thePhilippines.

METHODOLOGY

Project partners and the pilot sites

The project is truly a collaborative oneinvolving a national rice research institutionmandated to undertake rice research anddevelopment (PhilRice); the National IrrigationAdministration that administers various waterresource systems (NIA), and the InternationalRice Research Institute (IRRI). Considered aspart of the project team are farmer-cooperatorswho are themselves members of Farmer IrrigatorAssociations or Cooperatives.

The project’s study area is central Luzon(Fig. 1). An important (reservoir-backed) gravityirrigation system here is the Upper PampangaRiver Integrated Irrigation System (UPRIIS),covering some 100,000 ha but scheduled to beincreased to some 130,000 ha in the comingyears. Beside UPRIIS, shallow tubewell anddeepwell pumps owned and operated byfarmers' groups and individuals are commonly

NUEVA ECIJA

PANGASINAN

Bakay River

Tarlac River

TARLAC

NIA-PMO for TGISRP

LA PAZ

To Zaragosa

Talavera

River

TARLAC

SAN MANUEL

MONCADA

PANIQUI

TG-04

GP-125

P-38

RAMOS

PURAGERONA

VICTORIA

Tarlac pilot sites

Tarlac CABANATUAN

PALAYAN

SAN JOSE

N U E VA E C I J A

T A R L A C

P A M P A N G A

B A T A A N

B U L A C A N

Munoz

Sto DomingoVictoria

MALOLOS

BALANGA Manila Bay

SAN FERNANDO

ANGELES CITY

P A N G A S I N A N

M E T R O M A N I LA

L A G U N A

R I Z A L

N U E V A V I Z C A Y A

Z A M B A L E S

IBALuzon Sea

N

PantabanganReservoir

Luz on

M AN ILA

Visaya s

M indanao

Philippines

LEGEND

City Capital of Province

Major RoadMunicipality

Provincial BoundaryTTWS Demo Sites

Capital of Province

Tarlac sites

Fig. 1. Location of pilot sites (asterixes) in central Luzon, Philippines.

4

found in Bulacan, Pampanga, Tarlac and NuevaEcija. For the initial implementation in the 2002dry season, farmers getting irrigation waterfrom deepwells and shallow tubewells wereselected for two reasons. First, since thesefarmers directly face the costs of water thatthey use, they are considered most susceptibleto use technologies that help save water andreduce costs. Second, it is easier to manageand control water in small-scale deepwell andprivate shallow tubewell systems than in largegravity irrigation systems such as UPRIIS. Ifthe practice of water saving will be acceptedby pump users, the next territory to conquerwill be those covered by the gravity systems.

Controlled irrigation or alternatewetting and drying

Table 1 shows the number of farmer-cooperators of controlled irrigation and aerobicrice from Tarlac and Nueva Ecija pilot sites.Controlled irrigation was only carried outduring the dry seasons (DS 2002 and 2003),while aerobic rice was tested for the 2002 wetseason and the 2003 dry season. A total of 21farmers volunteered to participate in CI during2002 dry season, and about 26 farmers in thefollowing 2003 dry season. The selection offarmers was based on motivation andwillingness to participate in the field trials, andon site criteria like accessibility, spread offarmers across the site, position on thetoposequence, and nearness to pump. Aspecial effort was made to select farmers ondifferent toposequence positions (high, middle,

low elevation) to capture differences ingroundwater status and soil type since theseare expected to affect the actual number ofdays the crop can be without standing water.

Each farmer-cooperator participated withtwo fields: one managed using his standardfarmer’s practice (FP), and the other managedas controlled irrigation (CI). Each field size wasabout 500-1000 m2, with an internal farm ditch.Wetland preparation was done using thestandard hand tractor driven disc plow,followed by two harrowings and one levelingfor better water control and weed management.In Canarem and Gabaldon, crops weretransplanted, spaced at 20 cm × 20 cm. Farmer-cooperators in Dolores established their cropby wet seeding. Production inputs were thesame for both CI and FP plots. Rice cropswere transplanted between the last week ofDecember to the middle of January for both2002 and 2003 dry seasons in Canarem, while amuch later crop establishment (from first tothird week of January) was done for Doloresand Gabaldon sites because the farmer-cooperators were still busy planting onions andother upland crops during the middle to lastweek of December.

In the P-38 deepwell irrigation system inCanarem, water is distributed to the servicearea rotationally, where each farmer receivedirrigation water once a week, and usuallymaintained 6-8 cm of ponded water afterirrigation. Irrigation schedules for CI and FPplots followed the rotational irrigation scheduleof the sectors. However, to differentiate thewater management of the CI and FP plots, the

Table 1. Number of farmer-cooperators in the TTWS project

Location Number of farmer-cooperatorsControlled irrigation Aerobic rice

2002 DS 2003 DS 2002 WS 2003 DS

A. Tarlac sites (deepwell)Canarem (P-38) 11 15 3 -Dapdap (TG-04) - - 2 12Pansi (GP-125) - - 1 8

B. Nueva Ecija sites (shallow tubewell)

Dolores 5 5 2 5Gabaldon 5 6 1 4

Total 21 26 9 29

5

amount of irrigation water supplied at the CIplots was made to be about 30-40% lower thanthat supplied to the FP plots at each irrigation.In the shallow tubewell systems in NuevaEcija, FP plots were almost continuouslyflooded, while irrigation in the CI plots wasdone only after 4-5 days of no standing waterin the field. Irrigation water was measuredusing trapezoidal weirs. Staff gauges were alsoinstalled to measure daily ponded water depthin the plots.

Perched water table and groundwaterlevels in the CI and FP plots were monitoredusing PVC tubes. Divers (groundwater leveldata loggers) were also installed in the pilotsites to continuously monitor groundwater levelfluctuations. Rainfall and evaporation data inCanarem were obtained from the agro-meteorological station installed in the area,which consists of Class A evaporation panand true-check rain gage. Rainfall andevaporation for the Nueva Ecija sites weretaken from the agro-meteorological station inPhilRice, Muñoz. Calendar-type monitoringsheets were given to the farmer-cooperators inall pilot sites to record all their fieldoperations, labor used, and all inputs appliedin the CI and FP plots. Yields were taken fromcrop-cut samples collected from two 2 × 2.5 m2

sampling area in the FP and CI plots. Theactual yields from the whole plots were alsotaken for comparison.

Aerobic rice participatory R&D

Initially, nine farmer-cooperators were identifiedand selected during the 2002 wet season (WS)in Tarlac and Nueva Ecija to participate in thefirst exploratory trials of aerobic rice underfarmer field conditions. In the 2003 dry season(DS), the number of farmer-cooperators wasincreased to 29. The farmer-cooperators wereselected based on representativeness of theirfields and their willingness to participate in theR&D process. For both seasons, each farmer-cooperator was requested to test one of thethree promising aerobic varieties (APO, UPLRI-5and Magat). Farmers who volunteered orselected to test these varieties were either atthe edge of the pump area or situated on thehigher sites, with relatively large water lossesand dry soil conditions. Fields were prepareddry using either animal or tractor, and rice

seeds were dry seeded (in rows) in relativelydry soil with a seeding rate of about 80-100 kgha-1. In the 2002 WS, the establishment wasdone using a lithao, a wooden implement toopen the furrows (low tech); the seeds werehand-sown; and basal fertilizer was broadcast.However, in the 2003 DS, a mechanical seeder(high tech) was also used in seedestablishment, which is pulled by big tractorfor direct seeding and direct placement of basalfertilizer. Most of the farmer-participants triedboth the low technology (low tech) and thehigh technology (high tech) level of seedingby either contributing two plots, or splittingone big plot into two subplots. Both “lowtech” and “high tech” areas were laser-leveledbefore seeding. For uniformity in thetechnology adaptation, cultural managementpractices and inputs such as rice seeds,fertilizers and chemicals calculated for the areaof the participating fields, as well as technicalsupport were provided by the project. Thefarmers provided the day-to-day management ofthe fields, as well as the labor and power forland preparation, crop establishment, weeding,spraying, harvesting and threshing.Supplementary irrigation was given to the cropfor crop growth, although the amount of waterwas not measured.

Farmers were asked to conscientiouslyrecord all their operations such as labor andother inputs (seeds, fertilizer, pesticides, etc.) inthe forms provided. Long tubes were installedto monitor daily groundwater table at eachfarmer’s field. Emergence, flowering andharvesting dates were recorded. For theestimate of grain yields, two sources of datawere utilized: (1) crop cut samples wereobtained from two 10-m2 sampling spots, and(2) the yield of the whole field per record ofthe farmers obtained during the interview. Thecut crop samples were threshed, sun-dried, andweighed and the moisture content determined.Other observations such as pest and diseaseoccurrence, rodent infestation, lodging, weedpressures, etc. were noted.

Within season extension

Field school type activities were done todemonstrate the controlled irrigation andaerobic rice concept to other (nonparticipating)farmers and to broadly discuss the progress

6

and results at harvest. The field schoolconcept was agreed upon to emphasize thelearning objective. These field schools consistof visits to the research-extension sites andbriefings conducted by the farmer-cooperatorswith the assistance of the project teammembers. After the field visits, the differenttour groups meet in plenary for summary,integration and question and answer sessions.

Process documentation

Throughout the first year, the whole processof developing the technology transferframework was documented through reports,pictures and video footage. These materials willlater be used to develop training and extensionmaterials and to formalize the developedprocedure for transfer and adaptation ofknowledge for water savings.

RESULTS AND DISCUSSIONS

Controlled irrigation

Perched water table and groundwaterlevel fluctuation. In Canarem, perched watertable and groundwater level dynamics in bothCI and FP plots were affected by the timing ofirrigation delivery and toposequence positions.As shown in the sample graphs for 2002 DS(Fig. 2), the periodic water table rise in CI andFP plots was caused by the irrigation waterapplications. However, the degree offluctuations vary across the toposequencepositions. As summarized in Fig. 3, theseasonal average perched water table depth ofplots located in the high positions was about40 cm lower than in the low positions. Theshallow water table in the low areas wasattributed to the collected seepage from thehigh and middle portions of the service area.To compare CI and FP plots, perched watertable and groundwater table depths in CI plotswere relatively deeper (but not significantlydifferent) than in FP plots caused by higherinitial ponding depths of the latter. However, inthe low toposequence, no noticeabledifferences in depths between CI and FP plotswere observed and both had water tables thatwere already very close to the ground surface.

Groundwater depths in Gabaldon weredeeper than in Canarem and Dolores, and were

more than 3 m below the ground surface (Fig.4). Previous studies by Igbokwe (1992) showedthat during dry season in Gabaldon, thedeepest groundwater depth was recorded at 8.4m below the ground surface and the dryseason average is 7.4 m. In fact during thepeak season of crop growth (February-March),some farmers had to lower their pumps to drawgroundwater. On the other hand, perched watertable dynamics (Fig. 4) were shallower, andonly fluctuated from 0-60 cm below the groundsurface. On the average, FP plots had ashallower perched water table depth than CIplots (Fig. 3).

Perched water table depths in Dolores(Fig. 5) were shallower (but not significantlydifferent) than in Gabaldon, ranging from 0-50cm below the ground surface throughout thedry season. In both FP and CI plots, perchedwater table depths did not drop below therootzone (30-40 cm depth) throughout theseason.

Based on the above results, thedifferences of the average perched water tabledepths between CI and FP in all sites in Tarlacand Nueva Ecija were not really significantwhich showed that the farmers who weredrawing water from deepwells and shallowtubewells were already practicing controlledirrigation to a certain extent, and their currentwater management practices only require minorrefinement to optimize the benefits of controlledirrigation.

Irrigation water use. Irrigation water usein this paper is defined as the water input(rainfall and irrigation) from transplanting (ordirect seeding) until harvest. Total rainfall (fromDecember 1 to April 30) in Tarlac sites wasvery low, about 55 mm and 80 mm for 2002and 2003 dry seasons, respectively. However, amuch lower total rainfall was recorded inNueva Ecija sites with 2002 dry season total of34 mm and 2003 dry season total of 10 mm.

In Canarem, mean total water use (Fig. 6)was highest in high elevations in both CI andFP plots in both years. This was attributed toits lighter soil texture (fine silty loam) andlateral seepage towards the lower toposequencepositions. During its first season ofimplementation (dry season), the difference ofthe total water used between CI and FP plotswas also highest in high elevations (24%),compared to 20% and 5% in middle and lowtoposequences, respectively. The following

7

Fig. 2. Typical perched water table and groundwater depths at threetoposequence positions in Canarem during 2002 dry season.

A. High toposequence

B. Middle toposequence

C. Low toposequence

-100-90-80-70-60-50-40

12-J

an

19-J

an

26-J

an

2-Fe

b

9-Fe

b

16-F

eb

23-F

eb

2-M

ar

9-M

ar

16-M

ar

23-M

ar

30-M

ar

6-Ap

r

-30-20-10

01020

Dep

th (c

m)

Dep

th (c

m)

Perched water table depth Groundwater depth

-90-80-70-60-50-40-30-20-10

01020

12-J

an

19-J

an

26-J

an

2-Fe

b

9-Fe

b

16-F

eb

23-F

eb

2-M

ar

9-M

ar

16-M

ar

23-M

ar

-100-90-80-70-60-50-40-30-20-10

01020

12-J

an

19-J

an

26-J

an

2-Fe

b

9-Fe

b

16-F

eb

23-F

eb

2-M

ar

9-M

ar

16-M

ar

23-M

ar

Dep

th (c

m)

Dep

th (c

m)

Perched water table depth Groundwater depth

8

Fig. 3. Seasonal average perched water table depths in Tarlacand Nueva Ecija sites for 2002 and 2003 dry seasons

Fig. 4. Typical perched water table and groundwaterdepths in Gabaldon in dry season 2002.

Fig. 5. Typical perched water table and groundwaterdepths in Dolores in 2002 dry season.

CIFP

CIFP

-100-90-80-70-60-50-40-30-20-10

01020

03-F

eb-0

2

10-F

eb-0

2

17-F

eb-0

2

24-F

eb-0

2

03-M

ar-0

2

10-M

ar-0

2

17-M

ar-0

2

24-M

ar-0

2

31-M

ar-0

2

07-A

pr-0

2

14-A

pr-0

2

CIFP

-300

-260

-220

-180

-140

-100

-60

-20

20

13-F

eb-0

2

20-F

eb-0

2

27-F

eb-0

2

06-M

ar-0

2

13-M

ar-0

2

20-M

ar-0

2

27-M

ar-0

2

03-A

pr-0

2

10-A

pr-0

2

17-A

pr-0

2

Groundwater depth is below the 2.6 m long PVC pipe

9

year, farmers became more confident on thetechnology and as a result, a much highersavings was achieved especially in highelevation (33%). The average savings for allelevations (Fig. 7) was about 16% in 2002 and24% in 2003.

In Gabaldon, the average percentdifference in water used between CI and FPpractices was relatively low (11%) in 2002 DSas shown in Fig. 6. In fact, three out of fivefarmer-cooperators had only savings of 2-8%,while only one farmer was able to achieve asaving of 31%. The small difference can beexplained by the difficulty of water pumpingdue to the lowering of groundwater table.Because of this situation, these farmers decidedto irrigated their FP plots almost like CI plots.In Dolores, CI plots used about 727 mm whilethe FP plots used 853 mm in 2002 DS, and thedifference of 15% water used was notsignificant. A similar trend of water used wasalso observed in 2003 DS under the shallowtubewell systems as shown in Fig. 6. On theaverage, number of irrigations was slightlyhigher in Dolores than in Gabaldon for both CI(13.2 vs 12.6) and FP (14 vs 15.8) plots. Thiswas because farmers in Dolores establishedtheir crops through direct wet seeding whilethe Gabaldon farmers transplanted their crops.Thus in Dolores, extra irrigations were neededto grow the seed to the seedling stage indirect-seeded rice crop establishment.

Grain yield. In Canarem, the averagegrain yields did not vary significantly betweenCI and FP plots for the two dry seasoncropping (Fig. 8). In the 2002 DS, yields didnot also vary across toposequence and rangedfrom 5.3 to 5.5 t ha-1, with middle plots gettingslightly higher yields. In DS 2003, the lowerelevation plots got the highest average yield ofabout 7.5 t ha-1, and was the only group thathad significant increase in yield betweenseasons. This was maybe because most of thefarmer-cooperators in lower portion during the2003 dry season used APO that yielded higherthan the other variety (PSBRC-28) used bymost of the farmers in the high and middleelevations. In the 2002 DS, all farmer-cooperators used PSBRC-98 in their fields.

In Gabaldon, the yield differencesbetween CI and FP plots were also notsignificantly different, although CI averageyields are slightly higher than in FP plots inboth years. There was also no significant

difference of yields between the croppingseasons (Fig. 8). Average yields of CI and FPplots in the wet-seeded rice in Dolores wasabout one ton lower than in Gabaldon duringthe 2003 DS (4.6 vs 5.8 t ha-1), however, nosignificant yield difference was observedbetween the CI and FP plots in both years.

Water productivity. Water productivity iscomputed as the grain yield in kilogramsdivided by the mean total irrigation plus rainfallin cubic meters. In Canarem, the average waterproductivity (Fig. 9) in the CI plots was higherthan in the FP plots at all three toposequencepositions. Plots at low toposequence had thehighest productivity values of 1.7 and 1.6 kgm-3 in the 2002 dry season and 1.9 and 1.6 kgm-3 in the 2003 dry season for CI and FPplots, respectively. Since yields were the same(Fig. 8), the relatively high water productivityin low toposequence was caused by the lowerwater inputs (Fig. 6). In Gabaldon and Dolores,water productivity in the CI plots was higherthan in the FP plots, but the difference wasnot statistically significant.

Cost and returns. Table 2 shows theaverage cost and returns under the two watermanagement practices. The gross returns werecalculated as the total harvest (kg ha-1)multiplied by prevailing market price of paddy(US$ kg-1). Total production cost includesmaterial costs (seeds, fertilizers, herbicides,pesticides, fuel and oil) and labor costs (landpreparation, crop establishment, crop care andmaintenance, and post-harvest labor). Noncashand imputed costs were also added to the totalproduction cost.

On the average, there was no significantdifference of the gross returns between FP andCI plots in Canarem for the two years of dryseason cropping. During the first dry seasonimplementation of the project, the average totalproduction cost per hectare under FP washigher than CI (441 versus 397 US$) whichwas attributed to higher fuel and oilconsumption in FP plots. As a result, the 2002dry season net profit in CI was slightly higherthan in FP plots by almost US$45 ha-1.However, during the next dry season (2003),the disparity of the amount (Fig. 6) and costof irrigation had reduced that resulted toalmost the same net profits of CI and FP. Thiswas probably because the farmers gained moreconfidence on CI during the past dry seasonand they tried to copy the irrigation scheme of

10

Fig. 6. Total water used (mm) in CI and FP plots for 2002 and 2003 dry seasons.

Fig. 7. Average water savings (%) in the pilot sites for 2002 and 2003 dry seasons

Fig. 8. Average yield in CI and FP plots in the pilot sites for2002 and 2003 dry seasons

05

101520253035

Mid

dle

Mid

dle

Aver

age

Aver

age

2002 Dry Season

P

38,

Canarem

%

savi

ngs

2003 Dry Season

Nueva

Ecija

%

savi

ngs

05

101520253035

2002 Dry Season 2003 Dry Season

Dol

ores

Dol

ores

Gab

aldo

n

Gab

aldo

n

Hig

h

Hig

h

Low

Low

11

CI to their FP plots. The average two-year netprofit-cost ratio was slightly higher in CI plots(1.31) than in FP plots (1.16), which showedthe slight economic advantage of CI over FP.

In Gabaldon, only the data from the 2002DS is presented in this paper. The averagetotal gross return (in US dollars per hectare)under CI was higher than under FP by aboutUS$110 (Table 2). However, the totalproduction cost was lower in CI than in FP byabout US$46, which was attributed to lowerpumping cost (fuel and oil) and fertilizer cost.

Total number of irrigations in CI plots was lessthan in FP plots (12.6 vs 14 irrigations). Dueto higher total gross returns and lower totalproduction costs, the difference of the netprofit per ha between CI and FP was aboutUS$127.

In Dolores, total gross return was slightlyhigher in FP plots than in CI plots, althoughthe difference was not significant. Totalproduction cost was not also significantlydifferent between the two water managementpractices (Table 2), although CI plots received

Fig. 9. Average water productivity (kg m-3) at three toposequences for2002 and 2003 DS.

Table 2. Average yields, cost and returns of rice crop grown under two watermanagement practices for 2002 and 2003 DS

Canarem Gabaldon* Dolores*Item (Deepwell) (Shallow tubewells) (Shallow tubewells)

FP CI FP CI FP CI

2002 dry seasonGross return (US$/ha) 932.7 932.7 1182.6 1291.6 1073.5 1042.9Total Production cost (US$/ha) 441.2 397.3 897.6 851.4 599.3 598.6Net profit (US$/ha) 491.5 535.5 297.0 425.0 474.2 444.35Net profit-cost ratio 1.14 1.37 0.46 0.55 0.34 0.35

2003 dry seasonGross return (US$/ha) 1133.6 1105.2 N/A N/A N/A N/ATotal Production cost (US$/ha) 519.0 491.3 N/A N/A N/A N/ANet profit (US$/ha) 614.6 613.9 N/A N/A N/A N/ANet profit-cost ratio 1.18 1.25 N/A N/A N/A N/A

* N/A: data are still not available (analysis in progress for 2003 DS).

Nueva Ecija

12

a slightly higher production costs. This wasdue to the fact that some of the farmers werenot able to follow the irrigation protocol forthe CI plots. Some of them put extra irrigationwithout the prior knowledge of the researchers.

On the average, the net profits obtainedin Dolores for the CI and FP plots were higherthan in Gabaldon because of the lower cost ofcrop establishment (direct seeding) in Dolores.Moreover, the groundwater table in Doloreswas shallower and the farmer-cooperators’fields were located in a contiguous area, whereone field can receive seepage water from theneighboring fields that resulted to lowerpumping cost.

Aerobic rice

About 1500 mm of rain fell in Tarlac andNueva Ecija sites from seeding to harvestduring the 2002 WS, of which more than 50%occurred from the last week of June to middleof July. However, in 2003 DS, almost norainfall was recorded in the pilot sites.

During the 2002 WS in Tarlac, APOyielded the highest among the three varietieswith an average yield of 5.5 t ha-1, whileMagat and UPLRI-5 yielded 5.0 and 4.5 t ha-1,respectively (Fig. 10). In the Nueva Ecija sites,APO only yielded an average of 4.1 t ha-1,

while the UPLRI-5 yielded about 4.5 t ha-1. Thelow yield of APO in Nueva Ecija site wascaused by severe lodging during the floweringstage. Unfortunately, the number of samplefarmers was only 9, compared to 29 farmersduring the 2003 DS, and results may beinconclusive due to the limited samples if wetry to compare the aerobic rice results with thefarmers’ varieties. Nonetheless, the yieldadvantage during the 2002 DS of the threeaerobic rice varieties in Dapdap (GP-04) wasevaluated by comparing them with the yieldssampled from the 25 neighboring farmers’ fieldsin the area. Most of these farmers grew riceby dry seeding in hills along the rows usingdifferent lowland varieties. As shown in Figure11, yields at the neighboring fields in Dapdapranged from 2.1 to 4.8 t ha-1 or an average of4 t ha-1. This average yield was about 1.5 and1.3 t ha-1 lower than the average yield of APOand Magat, respectively. However, UPLRI-5yield was not significantly different from theaverage yield of the neighboring farmers.

With more farmer-cooperators participatingin the development of the aerobic ricetechnology during the 2003 dry season, a widerange of yield results was observed. The yieldrange for all varieties in 2002 was 4-5 t ha-1,while in 2003, the yield was 2-6.6 t ha-1. Interms of varietal performance, APO was better

Fig. 10. Average grain yields (t ha-1) of aerobic rice varieties in Tarlac andNueva Ecija sites during 2002 WS and 2003 DS.

Gra

in y

ield

(t/h

a)

Wet season 2002

APO

APO

UPL

RI-5

UPL

RI-5

Mag

at

TARLAC SITES TARLAC SITESNUEVA ECIJA SITES NUEVA ECIJA SITES

13

than UPLRI-5 in Tarlac for 2003 DS. Theaverage yields were 4.0 and 3.4 t ha-1 from thecrop cut estimate. The differences in yieldsbetween the two varieties, however, were notsignificant using the paired t-test. In terms oftechnology level of crop establishment, thelevel of technology imposed on aerobic ricedid not have a consistent and significant effecton the yield performance. The crop cutestimates at the low-tech level had relativelyhigher values compared to the high-tech levelfor the two varieties, however, the differenceswere not significant. It was difficult toestablish the relationship of yield and seedingdates because the range of the seeding datewas limited (aerobic rice was established fromDecember 5-20 only).

CONCLUSIONS

The results of the first season CI trial inCanarem, Gabaldon and Dolores have provideda good indication that controlled irrigation is aviable alternative in improving waterproductivity in deepwell irrigation system andshallow tubewell irrigation systems. Asignificant water saving of about 20% wasattained in deepwell systems and about 11-15%

in shallow tubewells. Although it could go upto 40% as demonstrated by some farmers. Interms of yield penalty, there was no significantreduction in yield between CI and FP plots. InCanarem, for two dry seasons, CI plotsobtained higher profit of US$ 575 ha-1 asagainst US$ 553 ha-1 in FP plots. In Gabaldon,a net profit per ha of US$ 425 was attained inthe CI and US$297 in the FP plots, while inDolores, the net profit per ha in CI plots wasslightly lower (US$ 444) compared to FP plots(US$ 474).

The two-season trials of the promising‘aerobic’ rice varieties in Tarlac and NuevaEcija sites showed remarkable yieldperformance. Higher yields were attainedcompared to the conventional lowland varietieswhere soils were very light, and farmers werepracticing dry seeding for decades. A yieldlevel of about 6 t ha-1 is achievable underfarmer field conditions for both wet and dryseasons. Although farmers were veryenthusiastic about adopting these varieties,there are still various issues that need to beunderstood and researched before a full-scaleadoption of this technology can take place.Many problems were experienced in aerobicrice system especially during the dry season

Fig. 11. Scatter plot of grain yields (t ha-1) of the three aerobic riceand farmers’ varieties in Dapdap, Tarlac during the 2002 WS.

0

1

2

3

4

5

6

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

ApoMagat

Neighboring farmers in Dapdap

UPLRI-5

Yiel

d (t

/ha)

14

trials, including water stress, weed pressure,possible yield reduction due to continuousmonoculture (nematode), nutrient managementand others. These problems still need to beresearched before moving to a wide scaleadoption of the technology.

FUTURE DIRECTIONS

After two seasons, the project is now ready toenter into its second phase. The second phasewill have the following components:

1. Formalizing the framework for participatorytransfer and adaptation of water-savingtechnologies. The lessons learned throughthe farmer-participatory research anddevelopment of controlled irrigation willbe translated into a kind of blueprint onhow to set up pilot sites that can have alighthouse function for a wider farmercommunity.

2. Develop training and extension programto further spread water-savingtechnologies in the Philippines, still usingcontrolled irrigation as a modeltechnology. Two activities may berequired. First, a mode of knowledgetransfer and adaptation that is suitable toreach a large audience needs to bedeveloped. One mechanism can be theestablishment of a number of lighthousepilot sites in strategic areas for watersaving, from which the knowledge candiffuse to surrounding farmers. Anothermechanism can be distant learning viainternet or radio and television media.Second, training needs to be organizedfor extension agents to disseminate thetechnology among farmers. Theseextension agents can be NIA personnelinvolved in irrigation system operation,local extension officers, local NGOs, orheads of farmer irrigator associations,farmer cooperatives or other local farmerorganizations. The training of suchextension agents involves a technologicalcomponent on water-saving technologiesand a “training-the-trainer” component inwhich the techniques of knowledgeextension and adaptation are taught.Training materials need to be preparedthat the extension agents can use in the

field (e.g., leaflets, brochures, radiobroadcasts, video etc.), and that arerequired at the training of the extensionagents themselves (course curriculum).

3. Inclusion of other water-savingtechnologies beside controlled irrigation.A number of options have beenresearched to reduce water use inirrigated rice production and each hasspecific advantages and disadvantagesand specific target domains in time andspace. As the mechanism for transfer andadaptation of technology becomes clearin the course of the project, other water-saving technologies can be tested infarmer-participatory approaches and addedto the training curriculum.The speed, scope and duration of the

second phase of the project will depend verymuch on the long-term commitments of theproject partners and their (financially)supporting governments.

REFERENCES

Bhuiyan, S.I., M.A. Sattar and D.F. Tabbal.1995. Wet seeded rice: water useefficiency, productivity, and constraints towider adoption. In: Constraints,opportunities, and innovations for wetseeded rice, K. Moody (ed.), IRRI, LosBaños, Philippines, pp. 143-155.

Bouman, B.A.M. and T.P. Tuong. 2001. Fieldwater management to save water andincrease its productivity in irrigatedlowland rice. Agric. Water Manag. 49: 11-30.

Bouman, B.A.M., D.F. Tabbal, R.M. Lampayan,R.V. Cuyno, M.B. Quiamco, V.R. Vicmudo,T.M. Norte, A.T. Lactaoen, E.J.P. Quilangand J.L. de Dios. 2002. Knowledge transferfor water saving technologies in riceproduction in the Philippines. Proceedingsof the 52nd PSAE Annual Convention heldin Puerto Princesa, Palawan, April 2002.

Castañeda, A.R. and S.I. Bhuiyan. 1993.Sediment pollution in a gravity irrigationsystem and its effects on rice production.Agriculture, Ecosystems and Environment,45: 195-202.

De Dios, J.L., A.A. Corpuz, A.J. Espiritu, J.P.Quilang and M.D. Malabayabas, 2000.Controlled irrigation on transplanted

15

lowland rice for efficient use of irrigationwater. In: Philippine Rice R&D Highlights,PhilRice, Maligaya, Muñoz, Philippines. pp.51-52.

Gleick, P.H. (Ed). 1993. Water in Crisis: AGuide to the World’s Fresh WaterResources. Oxford University Press, NewYork, USA, 473 pp.

Igbokwe, K.N. 1992. Shallow GroundwaterUtilization for Irrigation in Crop DiversifiedRice-based Areas . Unpublished PhDThesis, Central Luzon State University.

IRRI (International Rice Research Institute).1997. Rice almanac, second edition. IRRI,Los Baños, Philippines, 181 pp.

Li., Y.H. 2001. Research and practice of watersaving irrigation for rice in China. In theproceedings of the International Workshopon water saving irrigation for paddy rice,March 23-25, Wuhan China. WuhanUniversity, Wuhan, China, pp. 135-144.

Mishra, H.S., T.R. Rathore and R.C. Pant. 1990.Effect of intermittent irrigation ongroundwater table contribution, irrigationrequirements and yield of rice in mollisolsof Tarai region. Agricultural WaterManagement 18: 231-241.

NIA (National Irrigation Administration). 1996.Annual report 1996. NIA, Manila,Philippines, 46 pp.

Pingali, P.L., M. Hossain and R.V. Gerpacio.1997. Asian rice bowls; the returningcrisis? CAB International, Oxon, UK, andIRRI, Los Baños, Philippines, 341 pp.

Postel, S. 1997. Last Oasis. Facing waterscarcity. Norton and Company, New York,USA, 239 pp.

Rosegrant, M.W. 1997. Water resources in thetwenty-first century: challenges andimplications for action. Food, Agriculture,and the Environment Discussion Paper 20.IFPRI, Washington D.C., USA, 27 pp.

Sandhu, B.S., K.L. Khera, S.S. Prihar andBaldev Singh. 1980. Irrigation needs andyield of rice on a sandy-loam soil asaffected by continuous and intermittentsubmergence. Indian Journal of AgriculturalScience 50: 492-496.

Tabbal, D.F., B.A.M. Bouman, S.I. Bhuiyan,E.B. Sibayan and M.A. Sattar. 2002. On-farm strategies for reducing water input in

irrigated rice; case studies in thePhilippines. Agricultural WaterManagement.

Tuong T.P. 1999. Productive water use in riceproduction: opportunities and limitations.Journal of Crop Production 2(2): 241-264.

Wang Huaqi, B.A.M. Bouman, Dule Zhao,Wang Changgui and P.F. Moya. 2002.Aerobic rice in northern China:opportunities and challenges. Proceedingsof the International Workhop on Water-wise Rice Production, 8-11 April 20002, LosBaños, Philippines, Bouman, BAM et al.,(eds), Los Baños (Philippines): IRRI.

Yang Xiaoguang, Wan Huaqi, Wang Zhimin,Zhao Junfang, Chen Bin and B.A.M.Bouman. 2002. Yield of aerobic rice (HanDao) under different water regimes inNorth China. Proceedings of theInternational Workhop on Water-wise RiceProduction, 8-11 April 20002, Los Baños,Philippines, Bouman, BAM et al (eds). LosBaños (Philippines): IRRI.