Research programs Rainfed lowland rice ecosystem

Research programs Rainfed lowland rice ecosystem

MANAGING CROP, SOIL, AND WATER RESOURCES FOR ENHANCED PRODUCTIVITY AND SUSTAINABILITY OF LOWLAND AREAS 26 Growth and variability of rice production in eastern India 26 Improving timeliness and reducing cost of crop establishment 28 Weed communities of gogorancah rice in Indonesia 31 CROP AND RESOURCE MANAGEMENT FOR DEEPLY FLOODED AND COASTAL AREAS Adoption of rice-rice and rice-aquaculture farming systems in coastal West Bengal: determinants and impact 32 GERMPLASM IMPROVEMENT FOR RAINFED LOWLAND AND FLOOD-PRONE AREAS Germplasm for the rainfed lowland ecosystem 36 Breeding for the flood-prone ecosystem 37 Genotype by environment interactions in rainfed lowland environments 37 The role of active O2 scavenging systems in submergence tolerance 39 Inhibitory effect of ethylene on recovery of rice seedlings after submergence 40 Variety dependence of microbe-mediated nitrate supply: potential for microbial interventions to improve N use efficiency 41 Building partnership between scientists and farmers 42

32

36

VALIDATION AND DELIVERY OF NEW TECHNOLOGY FOR INCREASING PRODUCTIVITY OF FLOODPRONE RICELANDS OF SOUTH AND SOUTHEAST ASIA 44 Developments in boro rice farming 44 Identification of suitable boro varieties 45 Crop establishment 45 Nursery management 45 Crop management 45 PROGRESS OF UNREPORTED PROJECTS 46 Rainfed Lowland Rice Research Consortium 46 Facilitating technology transfer among NARES for the flood-prone rice ecosystem PROGRAM OUTLOOK

47

47

Rainfed lowland rice ecosystem

Rainfed lowland rice covers about 40 million ha and flood-prone rice covers about 12 million ha and supplies about 25% of the world’s rice production. Farmers in the ecosystem are confronted with drought, submergence, and problem soils that constrain the adoption of high-yielding modern varieties. Improvements required for crop management are increased tolerance for drought and submergence, better tolerance for poor soils, and better tolerance for biotic stresses. The Rainfed Lowland Rice Ecosystem program’s research activities are implemented in five projects— crop and resource management, germplasm improvement for rainfed and flood-prone areas, validation and adaptation of technologies through farmer participatory experiments, and facilitation of research partnership with national agricultural research and extension systems (NARES).

Managing crop, soil, and water resources for enhanced productivity and sustainability of lowland areas This project develops information and technological options for improved crop, soil, and water management for sustainable increases in productivity and farmers’ income. It also identifies and analyzes the processes governing productivity and sustainability of crop production systems. Research covers nutrient and water management, crop establishment and weed management, and risk management. Growth and variability of rice production in eastern India S. Pandey, S. Pal, and R. Villano Eastern Indian states account for more than twothirds of rice area and more than half of the rice output in India. Growth rates were an impresive 3.07% y–1 for yield and 3.60% y–1 for production during 1982-97. That growth was not evenly distributed among different states of eastern India, however, and the effect of production growth on output variability at the aggregate level has not been investigated. This research examines the extent to which growth in output has resulted in an increase in production variability. The analysis is based on district-level, time-series data on rice area, yield, and production for 71 districts covering eastern Uttar Pradesh, eastern Madhya Pradesh, Bihar, West Bengal, and Orissa. The trends in rice yield and production are seen in Figure 1. Growth in yield during 1961-81 was negligible. During 1982-97, yield increased dramatically with the spread of modern varieties sup-

26

IRRI program report for 2000

Area (million ha), production (million t) 60 50

Yield (t ha–1) 2.6

Production Yield Area

2.4 2.2

40

2.0

30

1.8 1.6

20

1.4 1.2

0

1.0

19 69 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97

10

Year

1. Trends in area, yield, and production of rough rice in eastern India, 1969-97.

ported by expansion of irrigation by shallow tubewells and pumps. Since the early 1980s, Uttar Pradesh and West Bengal have shown the strongest growth in production of rice, although all states registered a significant increase in this period (Table 1). The effect of production growth on variability in rice area yield and production was examined by using a 10-year moving coefficient of variation (CV), which adjusted the data for trends in mean values through time (Fig. 2). The results indicated that the CV of yield started to decline for most states after 1975-84 (Fig. 2A). The stabilizing effect was strongest in West Bengal, eastern Uttar Pradesh, and eastern Madhya Pradesh. On the other hand, the decline in yield variability has been modest in Orissa, while the yield variability showed little trend, or tended to increase, in Bihar. The moving CV of area indicates a clear pattern of increase in area instability in eastern Uttar Pradesh (Fig. 2B). The area instability in Bihar remained high throughout the period. The trend in production instability

generally mirrors that of the yield instability (Fig. 2C). Districts were classified into two groups for district-level analysis based on whether the average rice yield was below or above 2.28 t ha–1. The below-average group covered 55%, and the above-average group covered 45%, of total rice area. A large number of districts had statistically significant change in growth or variability, or both (defined as average % deviation from the trend) in rice yield and production between 1969-81 and 1982-94 (Table 2). Most of those districts had either a decrease in variability with no changes in growth rates of yield and production, or an increase in growth rate with no change in variability. The districts in the nochange category are in eastern Uttar Pradesh and those with an increase in growth rate with no change in variability are mostly in West Bengal. In contrast, growth in yield and production was accompanied by an increase in variability in Puri district of Orissa. Districts with average rice yield below 2.28 t ha–1 and districts with increased variability but no improvement in growth were mostly in Bihar. Thus the analysis of district-level data corroborated the finding at the state level as shown in Figure 2. A regression analysis examined the variation in district-level variability of yield (defined as average percentage deviation from the trend) across districts. Application of NPK fertilizer at the district level was the explanatory variable. It was used as a proxy for the adoption of modern varieties and irrigated areas because farmers tend to use more fertilizer when they have adopted modern varieties and when the area is irrigated. Dummy variables were used to account for agroclimatic differences across districts. The results show that the variability in rice yield during 1982-94 was negatively correlated with

Table 1. Growth rates (% y–1) in rice area, yield, and production in various parts of eastern India, 1961-97. Area

Yield

Production

State 1961-81 Assam Bihar Madhya Pradesh Orissa Uttar Pradesh West Bengal Eastern India

1.08** 0.26 0.79** 0.03 1.02** 0.81** 0.63**

1982-97 0.78* –0.18 0.61** 0.52** 0.43 1.11** 0.53**

1961-81 0.59** 0.79 0.40 –0.06 2.02* 0.91** 0.79

1982-97 1.81** 2.83* 1.31 2.52* 3.93** 3.81** 3.07**

1961-81 1.67** 1.06 1.19 –0.04 3.04** 1.73** 1.42**

1982-97 2.59** 2.65* 1.91* 3.04* 4.37** 4.92** 3.60**

Rainfed lowland rice ecosystem

27

CV (%) 8

A

6 4 2 0 8 B

6 4 2 0 8

C

6 4 2

Bihar

Eastern Madhya Pradesh

84 -9 3 85 -9 4 86 -9 5 87 -9 6 88 -9 7

76 -8 5 77 -8 6 78 -8 7 79 -8 8 80 -8 9 81 -9 0 82 -9 1 83 -9 2

19 69 -7 8 70 -7 9 71 -8 0 72 -8 1 73 -8 2 74 -8 3 75 -8 4

0

Year Orissa

Eastern Uttar Pradesh

West Bengal

2. Moving coefficient of variation (CV) of rice (A) yield, (B) area, and (C) production in eastern India, 1969-97.

the rate of NPK application. This indicates that where modern technology is adopted, the variability in rice yield is lower than where the traditional technology is used. Where adoption of modern varieties was supported by an assured source of irrigation, such as in eastern Uttar Pradesh and West Bengal, there was an overall decline in yield variability as well as an increase in the growth rate of yield.

28

IRRI program report for 2000

Improving timeliness and reducing cost of crop establishment R. Bakker, T. Alihamsyah, and P. Borlagdan A major constraint in rainfed lowland farming systems is the provision of sufficient labor and draft power for adequate, timely farming activities, especially land preparation, seeding, and weeding for direct-seeded rice (DSR). A survey of farmers in

Table 2. Change in growth and variability in the production and yield of rice in eastern India by districts. Figures in parentheses are rice area of the group as percentage of total rice area in eastern India during 1992-94. Changes in the growth and variability are statistically significant up to 10% level. None of the districts in the low-productivity group showed significant decrease in the growth rate. Change in growth rate

Change in variability (annual percentage deviation from trend) Production Decrease

No change

Yield Increase

Decrease

No change

Increase

Districts with yield 2.28 t ha–1 or higher Increase

Sultanpur (0.71)

Raipur, Bankura, Burdwan, Darjeeling, Midnapore, Murshidabad, Purulia, West Dinajpur (18.41)

Puri (1.82)

Gonda, Gorakhpur, Sultanpur (3.0)

Raipur, Bankura, Darjeeling, Midnapore, Murshidabad, Purulia, West Dinajpur (15.77)

No change

Allahabad, Azamgarh, Faizabad, Ghazipur, Jaunpur, Mirzapur, Pratapgarh, Nadia, Varanasi, Howrah (7.70)

Ballia, Gonda, Gorakhpur, Sahabad, Sambalpur, 24-Paragana, Birbhum, Hooghly, Malda (13.78)

Ganjam (1.47)

Allahabad, Azamgarh, Deoria, Faizabad, Ghazipur, Jaunpur, Mirzapur, Pratapgarh, Varanasi (6.88)

Ballia, Sahabad, Ganjam, Sambalpur, 24-Paragana, Birbhum, Burdwan, Hooghly, Howrah, Malda, Nadia (17.48)

Decrease

Deoria (1.06)

Puri (1.82)

Districts with yield less than 2.28 ha–1 Increase

No change

Basti, Bilaspur, Balasore, Bolangir, Dhenkanal, Cooch-Behar, Jalpaiguri (13.04) Balaghat, Jabalpur, Raigarh, Shahdol, Surguja, Kalahandi (7.71)

Durg, Mandla, Seoni, Bhagalpur, Dhanbad, Gaya, Hazaribagh, Palamau, Purnea, Ranchi, Saharsa, Santhal Paragana, Saran, Singhbhum, Cuttack, Keonjhar, Koraput, Mayurbhanj, Phulbani, Sundergarh (25.85)

Bahraich, Bastar, Champaran, Darbhanga, Monghyr, Muzaffarpur, Patna (9.56)

Basti (1.76)

Bahraich, Bolangir, Keonjhar, Cooch-Behar, Jalpaiguri (5.95)

Balaghat, Jabalpur, Raigarh, Shahdol, Surguja, Dhanbad, Kalahandi (7.92)

Bilaspur, Durg, Mandla, Seoni, Bhagalpur, Champaran, Gaya, Hazaribagh, Palamau, Patna, Purnea, Ranchi, Saharsa, Santhal Paragana, Saran, Sighbhum, Balasore, Cuttack, Dhenkanal, Koraput, Mayurbhanj, Phulbani, Sundergarh (33.34)

Bastar, Darbhanga, Monghyr, Muzaffarpur (6.05)

Rainfed lowland rice ecosystem

29

rainfed lowland areas of Central Java, Indonesia, indicated that farmers are experiencing changes in the availability and cost of rural labor and are adopting labor- and cost-saving methods. Central Java farmers prepare soil following first monsoon rains and rice is sown before fields are flooded. This rainfed lowland cultivation system is locally known as gogorancah (dry-seeded rice). Rice may germinate at the onset of the monsoon rains, but rainfall is unpredictable, planting is often delayed, and drought stress occurs at different stages of crop growth. Farmers use water buffaloes for tillage with traditional implements and planting is done by hand (dibbling) or a combination of manual labor and animal traction (furrow seeding). The traditional seeding methods are labor-intensive, require high seeding rates (up to 200 kg ha–1), and often lead to late establishment and irregular crop stand. Surface broadcasting of seed is not popular among farmers because of limited access to herbicides and sprayers to control weeds. Yield of the first DSR crop can be as high as 6 t ha–1 but yields are highly variable because of fluctuations in rainfall. Farmers may follow the first DSR crop with a second crop by soil puddling and transplanting of seedlings in an attempt to use the remaining soil moisture. Farmers in Central Java have started using simple two-wheel tractors for cultivation. The tractors are manufactured in Indonesia and often owned by farmers who offer contractual tillage services to other farmers in their village. The tractors were initially used for puddling to quickly establish the second rice crop, but farmers have gradually started using them for land preparation for the DSR crop. Mechanizing tillage at the start of the season enables farmers to start operations earlier in the season, with additional improvements in timeliness of the second crop. A case study in one village shows that the cost for mechanized tillage is lower (US$18.25–25 –1 ha) than the cost for traditional tillage (US$26.50–37.50 ha–1). The availability of tractors through contract systems provides opportunities to improve the traditional methods used for direct seeding. During 1994-99, IRRI and the Central Research Institute for Food Crops (CRIFC) developed improved mechanical seeders through a farmer partici-

30

IRRI program report for 2000

patory approach. The seeders, manufactured in small village workshops, mounted on the locally available two-wheel tractor and seed placement simulates the conventional DSR methods used by farmers. Animal-drawn versions of the seeders were also developed. On-station and farm-level research found that time requirements for planting were reduced from as much as 40 d ha–1 for traditional hand dibbling (five persons) and 30 d ha–1 for furrow seeding behind a country plow (four persons + one animal), down to 1.2 d ha–1 for a tractor mounted 4-row seeder (one operator + tractor). The mechanical seeding gave more uniform seeding depth and seed distribution and reduced seed rates. The best performing seeder was a simple seed drill (Fig. 3) that has tines that open slits in the soil.

3. CRIFC-IRRI seeder for rice in farmer's field, Meteseh and Rembang districts, Central Java.

Seed is deposited through a seed tube located behind the tine. Seed metering is by a simple fluted roll inside the seed hopper, which is driven by a groundwheel. Row spacing can be adjusted by changing tine spacing. The simple drill was selected by farmers on the basis of weight (implements often have to be carried to adjacent fields), ease of operation, and capacity to use it in different soil types. Estimated costs of using the tractor-mounted seeder in Central Java are $15.30 ha–1, assuming a 6-year economic life for the seeder, use for 20 d season–1, and 100 d y–1 tractor use (20 d for seeding and 80 d for other operations). This represents about half the costs of furrow seeding behind a country plow ($31.25 ha–1). A sensitivity analysis showed

Table 3. Assessment of a mechanical seeder in 20 farmers’ fields in Tawangrejo village, Blora, Central Java, 1999-2000. Farmers’ (%) rating of tractor-mounted seeder

Parameter

Ease of operation Ease of handling Ease of transporting Ease of field-to-field transfer Straightness of the row Neatness of seed drop in the row Uniformity of seed emergence Overall performance

Farmers’ (%) rating of traditional furrow seeding

Good

Fair

Poor

Good

Fair

Poor

70 80 70 70 90 90 90 90

20 10 20 20 0 0 0 10

0 0 0 0 0 0 0 0

30 20 20 30 10 0 10 10

60 70 60 50 90 70 70 70

10 10 10 10 0 10 20 20

Zone C

that the tractor-mounted seeder is competitive with the established seeding methods at utilization rates of 10 ha season–1 or more. To further evaluate the seeder’s acceptability to farmers, it was made available to farmers in three villages, and seeding in farmers’ fields was done in collaboration with local tractor owners. Responses of farmers (Table 3) confirmed that the mechanical seeder provided more even seed placement and improved plant stand. The study showed that relatively simple interventions can greatly improve the productivity and profitability of DSR. The seeder design was made available to the Indonesian Assessment Institute of Agricultural Technology to test its performance and acceptability in other provinces. Weed communities of gogorancah rice in Indonesia H. Pane9, E. Sutisna Noor9, M. Dizon, and M. Mortimer Dry-seeded rice is grown in the Jakenan region, Central Java, Indonesia, from the onset of the wet season (WS)(Oct-Nov) through Jan-Feb, followed by minimum-tillage transplanted rice (TPR) until May. Weed control in both crops is manual, with labor at 80 d ha–1 in DSR and 48 d ha–1 in TPR. Yield reductions due to weeds can be 76% in DSR and 45% in TPR. Weed communities vary in composition, are higher in DSR than in TPR, and severest in low-lying areas. Documentation of the efficacy of farmer weed control practices and of the abundance and diversity

of weeds throughout crop growth provides a baseline for an analysis of the impact of existing weed control systems and for an ex ante assessment of proposed changes in management. A study was designed to obtain a baseline as a precursor to the design of on-farm yield gap trials. The objectives were to quantitatively describe the on-farm weed flora in DSR, to analyze variation in that flora in relation to toposequence and soil characteristics, and to provide implications for improved weed management. Twenty-five farm sites were chosen in subdistricts within 50 km of CRIFC’s Jakenan station in the districts of Pati and Rembang. Three positions (upper, middle, and lower) on the sloping lands (5–30°) of the toposequence were chosen for each farm. The weed flora was assessed during Feb 1998 when rice was at the booting stage and the density of individual weed species (plants m–2) enumerated by destructive removal of all plants beyond the small seedling stage. Farmers were interviewed prior to data collection to confirm that rice weeding had been finished and to record nutrient management practices. Bulk soil samples were taken at each toposequence position, and pH, % total organic carbon, % total N, soluble phosphorus (mg kg–1), exchangeable K (meq 100 g– 1), and cation exchange capacity (CEC) (meq 100 g– 1 ) were measured. Farmyard manure was the principal source of fertilizer. Univariate analysis of variance was used to explore variation in weed densities and in soil nutrient status among sites. Weed-site associations were assessed using correspondence analysis and unimodal

Rainfed lowland rice ecosystem

31

species responses to environmental variables examined with canonical correspondence analysis. Considerable intersite variation was found in mean levels of K, P, organic C, and CEC, with the latter significantly correlated with all variables except phosphate. No differences in N status were found. Soils from upper positions were always more acid, with the average difference within a site being 0.65 of a pH unit from upper to lower position. Contrastingly, K, P, organic C, and CEC varied among sites at every farm and differences were not correlated with toposequence position. Levels of nutrients fell within reported ranges for other rainfed lowland environments. The sites were characteristically nutrient-deficient. Species commonly observed in the flowering stage were Eclipta alba, Echinochloa crus-galli, E. colona, Fimbristylis miliacea, Cyperus difformis, and C. rotundus. The overall mean total weed density was 175 plants m–2, which did not differ significantly in relation to toposequence when averaged across sites. Figure 4 shows the predicted distribution of six major weed species in relation to toposequence. The study showed that surprisingly dense, diverse weed communities persisted in DSR during the later stages of crop development. Their presence may reflect the fact that farmers perceive little economic benefit in weed removal at late stages of crop development, particularly because manual weeding may damage the crop. However on-farm yields have been reported as typically less than 3 t ha–1 in contrast to yields of weed-free research station trials of 3 and 5 t ha–1 depending on water availability and nutrient management. The 30-d duration between last weeding and crop harvest is sufficient for populations of E. colona, F. miliacea, and Lindernia spp. to develop, and for species with strong developmental plasticity (e.g., Ammania baccifera) to complete their life cycle and to pose a competitive threat (for light and nutrients) to yield during grain filling. The weed diversity in the ecosystem also implies the strong potential for transient, longer term shifts in relative abundance of weed species in response to changes in agronomy and water management and herbicide use. The existence of grass weeds such as E. crus-galli, Leptochloa chinensis, and Ischaemum rugosum, and of C. difformis clearly poses a signifi-

32

IRRI program report for 2000

cant threat to rice intensification and underlies the importance of effective early weed control.

Crop and resource management for deeply flooded and coastal areas Productivity of deeply flooded and coastal areas depends on the combination of cultural methods and selection of suitable crops. Research focus for 2000 was on the productivity, sustainability, and the environmental impacts of the emerging farming systems. Adoption of rice-rice and rice-aquaculture farming systems in coastal West Bengal: determinants and impact M. Hossain, S.K. Bardhan Roy, N.K. Saha, and F. B. Gascon One-fifth of the net cultivable area of India’s West Bengal state is in the coastal region. Salinity and flooding are main constraints to agricultural production. The coastal area has traditionally grown rainfed rice monocropped with photoperiod-sensitive, long-duration, low-yielding varieties. Substantial changes in land use have occurred in the last decade due to the availability of modern rice varieties and small-scale irrigation, a growing knowledge of semi-intensive shrimp and prawn culture, and a strong market for fish and shrimp at home and abroad. There has been an increase in adoption of ricerice and rice-aquaculture farming systems but there is inadequate information on these emerging farming systems. Furthermore, there has been concern that the rice-aquaculture farming system, which uses brackish water in the dry season (DS), may lead to increased soil salinity and toxicity. That could make land unsuitable for both rice cultivation and aquaculture in the long run. A study of the farming systems sought to ● document the farming systems practices at different levels of soil salinity and land forms, ● analyze the determinants of adoption of the emerging farming systems, and ● assess the overall impact of rice-aquaculture and other farming systems on farmers’ income and sustainable management of natural resources.

The study included 18 sample villages in four coastal districts representing different levels of soil salinity and natural resource management practices in West Bengal. We used focus group discussions with village leaders to collect information on the land use pattern, ecosystem characteristics, infrastructure facilities, farming practices, prices of agricultural inputs and outputs, terms and conditions of tenancy, and labor market. One hundred and seventy-nine households (10% of the total households) in the villages were selected for an in-depth household survey. The survey data consisted of ● socioeconomic background of the household, farmers’ perception of the biophysical characteristics (salinity, flooding depths, elevation) of different parcels of land owned and operated by the household; ● utilization of the parcels in different seasons; ● inputs used and production for different enterprises; and ● farmers’ perceptions regarding the sustainability of the emerging farming systems. Secondary information on changes in the farming system was also collected from various local organizations and from unpublished documents. Multivariate regression models were used to analyze factors affecting the adoption of rice-rice and rice-aquaculture farming systems, the effect of the new farming systems on the yield of aman (WS) rice, and the impact on sustainability of the system. The survey identified three major land use patterns (Table 4). The rice-based system was found to be significantly higher across the salinity level. A nonrice based cropping system was practiced in nonsaline to moderate-saline areas. Although a Table 4. Distribution of area (%) by land use system and salinity level in 18 villages in four coastal districts. West Bengal, India, 2000. Salinity level Land use system

Rice-based Nonrice-based Noncrop-based Homestead and fallow Total a b

Non-saline

Moderate

Severe

63.4 7.4 12.9a 14.3 100.0

69.6 7.8 14.9b 7.7 100.0

56.0 0.0 40.0c 4.0 100.0

Mostly vegetables and spices and orchards in the homestead. Mostly pulses, oilseeds, and spices. cMostly fish.

noncrop (aquaculture)-based system was practiced across different salinity levels, it was significantly higher in severe saline areas than in other salinity levels. Single-cropped rice occupied a major part of moderate-saline areas. Rice-rice systems were adopted in 27% of the area, mostly on land of medium flooding depth and moderate soil salinity. Among the noncrop-based system, rice-aquaculture accounted for 7% of the surveyed area. Nearly 15% of the parcels with severe salinity were used for salt making, and another 14% for raising fish. Focus group discussions revealed that the shift from rice-fallow to rice-rice and rice-aquaculture, although started in the 1970s, occurred between the mid–1980s and early 1990s. The shift had taken place on 41% the total land area. Table 5 indicates that there is a higher probability of adopting the rice-rice system if the parcel has access to irrigation, is situated in the medium elevation with lower depth of flooding, and has clay soil instead of loamy or sandy soils, which require more irrigation and increase the cost of cultivation of DSR. The probability of lower adoption at higher levels of moderate soil salinity affects the adoption of the system. This suggests that availability of moderate salinity-resistant rice varieties and the development of irrigation management practices that preclude intrusion of saline water during the critical stages of crop growth have been effective in inducing farmers to grow another rice crop during the DS. It is interesting to note that variables representing farm size and the availability of credit and educational status of the farmers have negative signs, and all coefficients are statistically significant. That shows the smaller and less educated farmers are adopting the intensive rice system faster than larger, better educated ones, and the availability of finance is not a constraint to the adoption of the rice-rice system. It is presumably subsistence pressure that is inducing farmers to adopt rice-rice, and the amount of investment needed for growing modern varieties is not high enough to preclude cash-starved farmers from doing so. The most important variable affecting the adoption of the rice-aquaculture systems is the soil salinity of the parcel and the depth of flooding. There is a higher probability of adopting the system if the soil is highly saline and of loamy or sandy type, and

Rainfed lowland rice ecosystem

33

Table 5. Factors affecting adoption of rice-rice and rice-aquaculture cropping patterns. Regression analysis of data from 179 households. West Bengal India, 2000.

Factor

Unit

Farm size

ha

Education

Illiterate = 0 Primary = 1 Secondary = 2 Yes = 1 No = 0 Irrigated = 1 Rainfed = 0 Yes = 1 No = 0 Yes = No = 0 Yes = 1 No = 0 Yes = 1 No = 0 Yes = 1 No = 0

Access to credit Irrigation Moderate salinity Severe salinity High elevation Medium elevation Loamy soil

Mean values of the variables

Rice-rice system (n = 112)

Rice-aquaculture system (n = 78)

1.02 0.87 1.21

–.46** (–3.80) –0.20* (–2.18)

0.26** (2.56) –0.005 (–0.006)

0.28

–0.42* (–2.18) 1.61** (8.55) 0.46* (2.44) –0.18 (–0.63) 0.59 (1.81) 0.74* (2.68) –0.72** (–3.68) –1.07** (–3.49) 137 –233

0.30 (1.65) –0.80** (–3.69) –0.096 (–0.47) 1.383** (5.44) –0.376 (–1.25) –0.592* (–2.23) 0.40* (2.08) –0.78** (–2.68) 93 –147

0.37 0.59 10.14

Constant

0.16 0.72 0.29 -

Chi-square Log-likelihood ratio

the parcel is at a low elevation with high depth of flooding. Such parcels are not suitable for growing rice during the DS (rice-rice system). Among the socioeconomic variables, only farm size has a statistically significant association with the adoption of the rice-fish system; the positive value of the coefficient indicates that it is the larger farmers who adopt the system more than the smaller ones. The findings on the costs and returns of the major cropping systems are reported in Table 6. Farmers in the rice-rice system got about 85% higher value of production than farmers in the traditional rice-fallow system. But, the cost of inputs was nearly four times higher for the rice-rice system compared with the rice-fallow systems. The cost of hired labor was, however, proportionately lower than the increase in outputs because of the greater use of family labor during the DS, which is a slack time for agricultural activities. The total paid-out cost was 64% of the gross value of production for the rice-rice system compared with 55% for the traditional rice-fallow system. The family income ha–1 of land was US$185 for the rice-rice system and US$125 for the rice-fallow system.

The yield of rice in the rice-aquaculture system was lower (19%) than the yield in the rice-fallow system, but the farmer got an additional 477 kg of fish and shrimp ha–1, of which 75% was shrimp, which fetched a premium price in domestic and inTable 6. Cost and returns a for different cropping patterns incoastal areas of West Bengal, 1995-96.

Yield (kg ha–1) Rice Shrimp Fish Costs and returns (US$ ha–1) Gross value of output Current inputs Seeds or fingerlings Fertilizer and lime Pesticides Irrigation cost Fish meal Taxesc Hired power Hired labor Total cost Net income a

Ricefallow

Ricerice

Riceaquaculture

2,229 – –

4,125 – –

1,815 360 117

280 36 21 13 2 1 – 2 9 108 155 125

524 481 23 57 18 50 – 6 28 153 336 188

2,279 1,175 1,162 10 b b

19 27 3 42 1,263 999

Weighted av of WS and DS. bLess than US$1.00. cIncludes rent paid for land.

34

IRRI program report for 2000

ternational markets. The gross value of production in the rice-aquaculture system was about eight times higher than the rice-fallow system, and 4.4 times higher than the rice-rice system. Shrimp cultivation, however, requires high investment. The cost of fingerlings was US$1,140 ha–1 compared with the seed cost of only US$21 ha –1 for rice cultivation. Because the farmers practiced traditional (semi-intensive) shrimp cultivation, the cost of fish meal was much lower than the cost of irrigation and fertilizer for the cultivation of the DS rice. The use of hired labor in the riceaquaculture system was also lower than for the ricefallow or rice-rice system. This finding suggests that the landless and the marginal landowners who supply labor in the market do not gain much from this emerging system. The net family income for the farmer from the rice-aquaculture system was estimated at US$1,000 ha–1, nearly four times of the income from the ricerice system and 10 times the income from the traditional rice-fallow system. The land productivity and family incomes from the adoption of riceaquaculture were positively associated with soil salinity.

Due to the high cost of inputs in the riceaquaculture system, small landowners often lease their land (sometimes collectively) after the rice harvest to landowners or outsiders interested in aquaculture during the DS. The rent received for this seasonal tenancy varied from US$143 to US$214 ha–1,which was higher than the net family income from the cultivation of the DS rice, or alternative economic activities that land owner could have pursued. Thus the tenancy system for aquaculture farming also provided benefits to the small landowners who are constrained from taking up aquaculture due to lack of capital. But the benefit accruing to the aquaculture farmers (who belong to the high-income group) was much larger, suggesting that the emergence of the riceaquaculture system has accentuated the inequality in the distribution of rural income. Table 7 shows the results of analysis of the effects of various parameters of the rice-rice and riceaquaculture on the yield of the aman rice crop. The major factor affecting the rice yield was the use of modern varieties (MV BORO). The yield level was higher in the low-lying parcels (LOW LAND), which are regularly silted by flooding and hence

Table 7. The effect of rice-aquaculture and rice-rice farming system on WS (aman) rice yield. Regression estimates with parcel-level data from 179 households, West Bengal, India, 1995-96. Explanatory variablea LABOR FERTILIZER MV BORO CLAY SALINITY 1 SALINITY 2 HIGH LAND LOW LAND RICE-RICE RICE-AQUACULTURE 1 RICE-AQUACULTURE 2 RICE-AQUACULTURE Constant R2

Unit

Regression coefficient

days ha–1 kg ha–1 MV=1; TV= 0 Clay soil = 1; Others = 0 Moderately saline =1 Others = 0 Severely saline = 1 Others = 0 High elevation = 1 Others = 0 Low elevation = 1 Others = 0 Rice-rice cropping = 1 Others = 0 Rice-aquaculture cropping = 1 Others = 0 Rice-aquaculture sequence = 1 Others = 0 Years rice-aquaculture system in production

Estimated T value

Significance of T

–0.005 0.53 373 59

–0.05 1.61 6.26 1.50

0.963 0.108 0.000 0.135

–15

–0.72

0.737

–51

–0.72

0.475

53

0.99

0.324

253

3.69

0.000

–219

–3.42

0.000

–35

–0.36

0.729

–50 8

–0.73 1.24

0.468 0.216

12.8

0.000

0.78 0.26

a

The dependent variable is rice yield in the aman-season (traditional crop) measured in kg ha–1.

Rainfed lowland rice ecosystem

35

more fertile. The yield was positively related with fertilizer use (FERT), and was relatively higher on clay soils (CLAY), presumably because of higher moisture-holding capacity. The rice-rice system has a negative effect on yield as shown by the negative and the statistically significant coefficient of the variable representing parcels that were used for growing DS rice (BORO) before the cultivation of aman rice. However, no significant relationship was found between the rice yield and the number of years the farmer had been practicing riceaquaculture system in the parcel (YEAR) or whether the farmer had been growing fish in the parcel as a mixed crop with rice (MIXED). The indirect analysis with cross-section parcel-level data thus fails to validate the hypothesis that the riceaquaculture system has had adverse impact on soil quality and rice yields. Because salinity status will determine the sustainability of the changing farming system, further study with actual measurement of soil characteristics is suggested. Development of salinity-tolerant varieties, improved management practices for rice-aquaculture system, early maturing rice varieties for the rice-rice system, and management options for irrigation and drainage were identified as the major technological needs of the farmers.

Germplasm improvement for rainfed lowland and flood-prone areas Germplasm improvement for the rainfed lowlands is complicated due to the heterogeneity and complexity of the environment. A breeding program based on physiological understanding of plant adaptation and environmental characterization operates through a shuttle breeding partnership in collaboration with NARES at key representative sites. A methodology for field evaluation of genotype by environment (G × E) interaction was developed, and preliminary analysis identified principal factors influencing G × E interaction. An important constraint to adoption of improved germplasm that tolerates drought and submergence is the preservation of traits found in traditional cultivars that are highly valued by farmers and consumers. This constraint is approached through farmer participatory breeding and selection of advanced lines based on indigenous knowledge of farmers.

36

IRRI program report for 2000

Germplasm for the rainfed lowland ecosystem S. Sarkarung and R.K. Singh Thailand. The elite line, IR62558-SRN-17-2-1-B, was released as Surin 1 for drought-prone conditions in northeastern Thailand in 2000. This variety combines resistance to blast and bacterial blight and tolerance for salinity. Because it is insensitive to daylength, farmers can grow it in the DS. Its grain quality is suitable for industrial use. This is the first nonglutinous variety ever released from the Thailand-IRRI shuttle breeding program. A total of 12 advanced breeding lines have been evaluated in farmers’ fields in northeastern Thailand (e.g., Ubon, Surin, Nakorn Ratchsima, Sakon Nakorn, and Khonkaen). The most promising lines were IR68796-27-3-B-6-1-B and IR69515-27KKN-1-UBN-1-1-1-B. In drought-screening tests, these lines showed high drought resistance at the vegetative stage. India. A release proposal for IR66363 (NDR96005) was submitted to the government of Uttar Pradesh, India. The variety did well in on-station and on-farm trials with yields ranging from 3 to 5 t ha–1. It has adequate submergence tolerance and tolerates delayed planting, making it suited to the rainfed lowland ecosystem. Farmers in submergence-prone areas in eastern India are frequently forced to delay transplanting due to late onset of monsoon rains. The breeding lines IR66366-M-7-1-1-1-1, IR66876-11-M-1-1-1, and IR67471-8-M-1-1-1-1 exhibit low yield losses due to delayed transplanting. The promising line, IR67493-M-2 (NDR 8002), designated as IET15848 in the slender grain trials of the All India Coordinated Rice Improvement Program (AICRIP) since 1997, was recommended for release for semi-deepwater ecosystems (40-70 cm) in Orissa, West Bengal, eastern Uttar Pradesh, and Madhya Pradesh. It has grain quality similar to PR106 and is resistant to whitebacked planthopper and moderately resistant to blast. The line IR54112B2-1-6-2-2-2, which has high submergence tolerance has been nominated for the same ecosystem in West Bengal and Orissa. The lines SBIR66366-7-M-1-1-1-1, SBIR66876-11M-1-1-1, SBIR69051-M-1-1-1-1-1, SBIR67471-M-8-11-1-1, SBIR67051-15-M-1-1-1-1, and SBIR67471-M-9-

1-1-1-1 were nominated to AICRIP from 1997 to 2000 and are at different stages of testing. Five advanced lines were nominated for the AICRIP 1999 Initial Variety Trial for Shallow Water: TTB522-SBIR70237-3-1 (IET16679), TTB532-SBIR70197-23-2 (IET16680), TTB528SBIR70251-15-2 (IET16681), TTB517-44SBIR70149-33-1 (IET16682), and TTB517-44SBIR70149-33-2 (IET16683). A few of them were promoted to the next stage of testing. Philippines. The elite line IR54068-B-60-1-3-3, designated as PSBRc 102, was recommended for release in drought-prone environments of Central Luzon, Philippines. It is resistant to bacterial blight, brown planthopper (BPH), BPH biotype 1 (BPH1), and moderately resistant to BPH2, BPH3, and green leafhopper (GLH). It has long slender grain with 23% amylose content. Breeding for the flood-prone ecosystem G.B. Gregorio, R. Mendoza, A.N. Monroy, and P. Bonilla Three hundred and twenty-eight elite breeding lines were generated from crosses between high-Fe rice parents and modern varieties. These Fe- and Zndense rices will be distributed for NARES field tests in the Philippines, Vietnam, Bangladesh, and Indonesia. Three important quantitative trait loci (QTLs) were detected for the high grain-Fe trait. These are located on chromosomes 7, 8, and 9. A double haploid population of IR64/Azucena was used to map those QTLs. Simple-sequence repeat (SSR) markers RM09 and RM24, flanking the salinity tolerance genes on chromosome 1, were identified. These markers will be tested in breeding populations to confirm their applicability for marker-aided selection for salinity tolerance. Genotype by environment interactions in rainfed lowland environments L.J. Wade and C.G. McLaren The nature of G × E interactions in rainfed lowland rice was examined from 1994 to 1997 using data for 37 genotypes across 36 environments in India, Bangladesh, Thailand, Indonesia, and the Philippines.

More than 47% of the G × E sum of squares was captured by nine genotype groups and nine environment groups. Sites with similar characteristics were tightly grouped, as were related genotypes. Environment groups included some with favorable water supply, and others with early drought, late drought, rapid-onset late drought, and submergence. Groupings of genotypes could be explained by their performance in relation to those conditions (Fig. 5). Genotype groups 74 and 85 comprising varieties and hybrids with high yield potential, semidwarf stature, and short duration (90-95 d) yielded well in favorable environment groups 60-61, but yielded poorly when subjected to drought (groups 62 and 46) or submergence (group 63). Genotype group 76, including the reference line IR62266-42-6-l, showed tolerance for late-season drought, but not for flooding. Groups 81 and 79, typified by CT9993-5-10-1-M, which flowered at 100 d, and Mahsuri, which flowered at 110 d, were stable in yield across most environments. Group 52 (NSG19) was preferentially adapted to environments with rapid-onset late drought, and group 82 (Sabita and KDML 105) to environments favoring late maturity or recovery after drought. A reference set of nine lines was identified— Sabita, NSG19, IR58821-23-1-3-1, Mahsuri, IR52561-UBN-1-1-2, CT9993-5-10-1-M, IR62266-426-l, PSBRc 14, and CT9897-55-2-M-3M. They represent broad and specific adaptations to the major target subecosystems in the rainfed lowlands. G × E interactions were also examined for phenology, biomass, yield components, and drought measures. Flowering date, as a measure of phenology, and plant height, as a measure of relative biomass, showed little G × E interaction, but were important in explaining grain yield. Flowering date and plant height were both negatively correlated with yield interaction scores, indicating that the taller, long-duration cultivars (the more traditional, photoperiod-sensitive lines) showed positive yield interaction in stress sites and were disadvantaged in favorable sites in comparison with the modern cultivars. The number of days from onset of late-season drought until physiological maturity at sites with lethal late-season drought was used as one measure of drought tolerance. It showed a strong, nonlinear relationship with G × E interaction in grain yield, leading to a partitioning of genotypes into three

Rainfed lowland rice ecosystem

37

Response patterns Sabita KDML 105 IR57546-PMI-1-B-2-2

GGP-82

NSG 19 GGP-52

IR58821-23-1-3-3 IR66469-17-5-B IR66516

GGP-78

MAHSURI IR66883

GGP-79

GGP-84

IR52561-UBN-1-1-2 IR54071-UBN-1-1-3-1 IR57515-PMI-8-1-1-S IR66506-5-1-B IR66879 IR66883

GGP-81

IR20 CT9993-5-10-1-M IR66893-5-2-B IR58307-210-1-2-3-3

GGP-76

IR54977-UBN-6-1-3-3 IR57514-PMI-5-B-1-2 IR62266-42-6-1 IR63429-23-1-3-3 IR66882-4-4-B

GGP-74

PSBRc14 IR64 IR36 CT9897-55-2-M-3-M IR64615H IR68877H

GGP-85

Environment group 63

46

62

57

60

61

General hydrology

10

16

21

Stagnation Rapid onset late drought Severe rapid onset late drought Favorable (gogorancah) Favorable (long season) Mild late drought Early drought and severe late drought Prolonged late drought Submergence

5. Interaction response patterns for grain yield for nine genotype groups over nine environment groups. Response values are averages, over group members, of mean polish interaction values with a range of –3 to +3 for each genotype group. IRRI, 2000.

38

IRRI program report for 2000

Drought interval (d) 30 IR665’6x3

Sabita Mahsuri

25

IR66883x4

KDML 105 IR66879x4

22

IR20 NSG19

20

15 –1.0

IR64

IR36

–0.5

0.0

0.5

1.0

Yield interaction score on PCA axis 2 6. Number of days from onset of lethal, late-season drought to physiological maturity (drought interval) against grain yield interaction scores for the second PCA interaction axis. IRRI, 2000.

groups (Fig. 6). One group, which included the irrigated lines IR64, IR36, and IR20, as well as NSG19, had short tolerance intervals and positive interaction in sites with no late-season drought. A second group with long tolerance intervals, including the photoperiod-sensitive cultivars, had positive interaction with the light-textured sites in northeast Thailand. A third group with long tolerance intervals comprised test lines with positive interaction in stressed sites but negative interaction in sites with no late-season drought. Grain size and percent filled spikelets showed strong G × E interactions that were related to interaction in grain yield. Mahsuri, which had the smallest grains of all lines tested, showed positive interaction for spikelet fertility at the favorable Philippine sites, but negative interaction at stressed sites. On the other hand, Mahsuri showed negative interaction for 1,000 grain weight at the Philippine sites. High-yield-potential cultivars IR36, IR64, and PSBRc 14 showed the opposite relationships, indicating alternative adaptation strategies. The examination of G × E interaction for yield and yield components indicated there were different strategies of cultivar adaptation across rainfed lowland environments. The use of G × E analysis on measures of phenology and drought tolerance aided the interpretation of these cultivar responses and

improved the understanding of mechanisms and traits likely to confer an adaptive advantage in specific subecosystems. It is important to clearly identify target environments and to know how well actual test environments represent those targets. A methodology for using measurements on a set of reference lines to classify sites was developed and tested. Strategies for choosing reference lines, classifying new sites, and deducing their environmental characteristics were examined. The results showed that the reference set was able to capture repeatable G × E patterns, provided it contained representatives of all discriminatory genotype groups. The methodology for characterizing new environments on the basis of reference line responses relied heavily on an ability to impute missing values. Although no optimal solution was available, a heuristic solution with the pattern analysis algorithm was satisfactory. Reference lines should be chosen according to how well they match the discriminatory pattern of their genotype group, their agronomic features, knowledge of their physiological responses, and practical issues such as the availability of seed. Based on this analysis, we conclude that a series of small field trials at many sites could obtain a useful characterization of new environments and allow breeders to appropriately weight responses of test lines. If detailed physical and climatic measurements are also made in these environments, the responses can be integrated with geographical information, physiological understanding, and crop modeling to quantify environment frequency, predictability, repeatability, and risk. The role of active O2 scavenging system in submergence tolerance E. Ella, O. Ito, and A. Ismail Active O2 species like hydrogen peroxide (H2O2) can cause dysfunction of enzymes and oxidative damage to lipids producing toxic malondialdehyde (MDA) and damaging cell membranes. Fourteen-day-old seedlings of submergence-tolerant (FR13A) and -susceptible (IR42) rice cultivars were submerged for 6 d and allowed to recover under low (300 µE m–2 s–1) or high (1,000 µE m–2 s–1) light intensities from 0600 to 1800 daily. Leaf samples for chemical analyses were collected during the

Rainfed lowland rice ecosystem

39

Value of SPAD meter 40

NADPH (µM oxidized mg–1 protein min–1) 100

A

FR13A

30

75

Y=19.395X + 1.454 r = 0.86

20

50 IR42

10

25

0 Low

High

Low

High

7. Activity of glutathione reductase (GR) enzyme of 12 nonsubmerged ( ) and submerged (12) rice cultivars. IRRI, 2000.

recovery period. Under high light intensity during recovery, FR13A and IR42 had comparable H2O2 content. However, FR13A had less MDA than IR42 under both light intensities. Among the four active O2-scavenging enzymes studied (ascorbate peroxidase, catalase, glutathione reductase [GR]), and superoxide dismutase), only the level of GR activity is significantly different between the two cultivars with higher activity (Fig. 7). The level of ascorbate antioxidant was also higher in submergence-tolerant FR13A (Fig. 8). The high levels of ascorbate antioxidant and GR activity ensure a better functioning of the ascorbic acid-glutathione cycle in FR13A. H2O2 is detoxified more efficiently in this cycle, thereby lowering the extent of exposure of membrane lipids to high levels of H2O2 and alleviating oxidative damage. This explains the difference in MDA content between these two cultivars that had comparable H2O2 production. Our results suggest the involvement of an active O2-scavenging system during recovery of submerged rice seedlings. AsA (% of AsA + DHAsA) 100 FR13

0 0

0.5 1.0 1.5 Total ascorbate content (mg g–1FW)

2.0

Total ascorbate content (mg g–1 FW) 2.0 B

1.6 1.2 Y=4.858X + 1.805 r = 0.93

0.6 0.4 0

0

0.05

0.1

0.15

0.2

0.25

Malondialdehyde content (nmol g–1FW) –1 Malondialdehyde content (nmol g FW) 0.25

C

0.20 0.15 0.10

Y=0.0023X + 1.2536 r = 0.98

0.05 0 0

10 20

30 40 50 60 70 80 90 100 110 Percentage survival

9. Association of (A) chlorophyll content after 8 d of submergence and total ascorbate content after 3 d of recovery, (B) malondialdehyde (MDA) and ascorbate content after 3 d of recovery, and (C) MDA and percentage survival after 3 d of recovery. IRRI, 2000.

IR42

75 50 25 0 Low

High

Low

High

8. Ascorbate content (% of the total) of nonsubmerged (filled bars) and submerged (unfilled bars) rice cultivars after 4-d light treatment during recovery.

40

IRRI program report for 2000

Inhibitory effect of ethylene on recovery of rice seedlings after submergence N. Kawano, E. Ella, O. Ito, Y. Yamauchi, K. Tanaka, and A. Ismail Earlier studies have documented the accumulation of ethylene during submergence, causing chlorosis

and decreasing the photo protection of the photosynthetic system. Ascorbic acid is synthesized in plants from glucose produced during photosynthesis and serves as an antioxidant or a cofactor in the xanthophyll cycle. This may play an important role in enhancing photo protection. We attempted to clarify the role of ethylene in recovery of submerged rice seedlings using AgNO3, an ethylene antagonist. Submergence-tolerant BKNFR76106-16-0-1-0 and submergence-intolerant Mahsuri and IR42 were used. Fourteen-day-old rice seedlings were sprayed with AgNO3 1 d before complete submergence for 8 d. They were then allowed to recover for 7 d after submergence. A significant decrease in ascorbic acid was observed in all cultivars during submergence, with greater reduction for the submergence-intolerant cultivars. Reduction in ascorbic acid during submergence may have decreased photo protection and slowed recovery of rice seedlings after submergence. Treatment with AgNO3 decreased stem elongation and improved survival of the susceptible cultivars but had no effect on the tolerant cultivar. Ethylene-induced stem elongation probably depleted the carbohydrates needed to provide energy for survival and maintenance processes. Chlorophyll content was positively correlated with total ascorbate content (r = 0.86; Fig. 9A). A strong negative correlation was observed between lipid peroxidation and total ascorbate content (r = –0.93; Fig. 9B) and with percentage survival (r = –0.98; Fig. 9C). Treatment with AgNO3 was, therefore, associated with suppressed chlorophyll degradation during submergence. It increased ascorbate content in the submerged seedlings during recovery of susceptible cultivars but had no effects on the tolerant cultivar. Our results suggest strong inhibitory effects of ethylene on the recovery of rice seedlings after submergence. Variety dependence of microbe-mediated nitrate supply: potential for microbial interventions to improve N use efficiency A.M. Briones and W. Reichardt

tolerant rice varieties can be viewed as a function of nitrate supply by root-associated AOB. The oxidized microenvironment of rice roots ensures the activity of root-associated AOB. Hence, the impact of AOB on fine tuning of the nitrate supply under changing redox conditions would ultimately depend on the diversity of substrate kinetic types among the ammonium-oxidizing populations. Sequencing of cloned fragments of the functional gene amoA revealed fundamental varietyspecific differences of AOB types associated with drought-tolerant rice. Ninety-six percent of polymerase chain reaction (PCR)-amplified, cloned amoA from the root environment of Mahsuri represented Nitrosospira-like sequences, which are frequent in N-limited soils. Only 85% of amoA sequences from the ammonium-rich root environment of IR63087-1-17 represented Nitrosomonas-like populations. This latter AOB population went almost undetected when using viable counts or PCR for AOB-specific 16S rRNA. 15NH Cl oxidation rates in the rhizosphere of 4 Mahsuri and IR63087-1-17 suggested an association of these rice varieties with different kinetic types of ammonium oxidizers (Fig. 10). Predominance of AOB showing adaptation to high N and low N levels seemed to match the presumed ecological niches of the varieties IR63087-1-17 and Mahsuri. This first evidence of variety-specific, root-associated microflora governing nitrate supply in the Cumulative N uptake (mg N plant–1) 25

A

15NH + 4 15 NO3–

20 15 10 5 0 25

B

20 15 10 5

Fluctuating redox potentials influence soil nitrate supply in rice cropping systems exposed to drought and flooding. This depends on the activity of aerobic, ammonium-oxidizing (nitrifying) bacteria (AOB). Therefore N uptake efficiencies of drought-

0

0

3 6 Days after 15N application

9

10. Cumulative uptake of 15N by drought-tolerant rice varieties Mahsuri (A) and IR63087-1-17 (B). IRRI, 2000.

Rainfed lowland rice ecosystem

41

rhizosphere of rice provides clues to the regulation of nitrate supply and hence of N uptake efficiencies in drought-adapted rice varieties. This may lead to the identification of options to manipulate root-associated, AOB, or genes in order to increase N use efficiencies. Building partnership between scientists and farmers T. Paris, R.K. Sahu10, V.N. Sahu10, R.K. Singh, S. Sarkarung, K. McAllister11, and M.L. Sharma10 A farmer participatory breeding project was established in eastern India in 1997 as a collaborative project among plant breeders and social scientists at IRRI and in six national agricultural research institutions. The objectives were to develop and test a methodology for effectively involving farmers in the breeding program, to improve understanding of male and female farmers’ criteria for selecting spe-

cific rice varieties, and to develop rice varieties that meet farmers’ preferences. A survey of area under rice varieties at the farm level in three villages in Madhya Pradesh shows substantial diversity (Table 8). The diversity was due to large proportion of rainfed land, heavy soils, and variation of land forms within a village. Traditional variety Safri 17 (late duration) is preferred by farmers due to stable yield, insect pest and disease resistance, drought tolerance, and weed competitiveness. However, its yield is perceived to be lower than Swarna and Kranti, and it is susceptible to lodging. Swarna is a high-yielding, late-maturing semidwarf variety. Farmers perceive it to be drought-tolerant and responsive to fertilizer. It has a dark green color, which allows weedy rices to be identified. It has good eating quality, keeps well when cooked, and has a high milling recovery. Mahamaya has medium duration, high yield poten-

Table 8. Area (ha) planted to modern and traditional rice varieties by farming households in three villages. Raipur, Madhya Pradesh, 1997. Tarpongi (n = 25)

Saguni (n = 50)

Khairkut (n = 50)

Varietya

Duration (d) Upland

Modern Swarna Mahamaya Kranti 262 Purnima IR36 Culture Others Total modern Traditional Safri-BD Safri-17 Chepti Gurmatia Ranikajar Bhata Safri Anjan Safri Ganga Safri Nankershar Dubraj Chepti Total traditional) Total all varieties % modern % traditional

Lowland

Upland

Lowland

Upland

Lowland

7.8 2.6 6.9 2.1 0.4

27.6 2.2 8.8

9.9 1.4 1.8 0.1

38.7 6.6 4.9 0.8

5.0 1.0

0.8

1.9

1.2

17.5

20.6

40.5

0.7 15.1

0.4 51.4

2.9 1.2 10.8 1.8 4.4 0.5 0.3 0.2

28.4 10.7 7.0 1.4 7.8 0.1

7.7 12.3 3.2 6.3

4.1 3.7 3.8 1.9 0.4

40.6 0.4 0.6 5.7 2.1

5.2

4.7 18.6 33.7 44.81 55.19

49.4 100.86 50.92 49.08

12.2 18.8 35.11 64.89

0.8 6.8 7.5 2.4

0.6

6.6

5.0 0.4 1.6

Late (150) Late (155) Medium (130) Medium (130) Medium (130) Late (145) Late (145 ) Late (135)

1.6 22.1 39.6 44.19 55.81

57.0 77.6 26.55 73.45

a

29.5 70.02 57.87 42.13

Medium (130)

Modern: semidwarf high-yielding varieties. Traditional: tall-statured, improved or not improved.

42

Late (150) Medium (130) Medium (130) Medium (125) Late (145) Early (120) Medium (130)

IRRI program report for 2000

Table 9. Comparison of rankings attributed by farmers and breeders at different growth stages in trials in Raipur villages, eastern India, 1997-99. Agreement among farmers Wb

Agreement among breeders Wb

Correlation between farmers' and breeders’ rankings (Rc)

Trial location

Year

Stagea

Varieties (no.)

Farmers (no.)

Breeders (no.)

Station

1997 1997 1997 1997 1997 1997 1997 1997 1997 1997 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 1999 1999 1999 1999 1999 1999 1999 1999 1999 1999 1999 1999 1999 1999

F M F M F M F M F M F M F M F M Crop failure F M F M M M M M M M M M V F M V F M

16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16

8 8 5 4 5 7 7 6 5 5 8 6 8 6 5 4

1 1 – 2 – 2 – 2 – 2 2 2 2 2 1 1

0.34** 0.51** 0.51** 0.55** 0.50** 0.34** 0.30** 0.44** 0.79** 0.54** 0.32** 0.26 0.31** 0.67** 0.55** 0.30***

– – – 0.47 – 0.53 – 0.30 – 0.56 0.77 0.60 0.54 0.70 – –

–0.20 0.11 – 0.13 – –0.03 – -0.18 – –0.06 0.16 0.50* –0.04 0.28 0.46 0.20

16 16 16 16 16 16 16 16 16 16 16 16 20 20 20 20 20 20

4 4 6 4 7 7 6 5 7 7 7 6 5 5 5 5 5 5

1 1 1 1 3 3 3 3 3 3 3 3 3 3 3 3 3 3

0.56** 0.59** 0.38** 0.44* 0.49** 0.65** 0.65** 0.62** 0.53** 0.34** 0.50** 0.66** 0.98** 0.98** 0.96** 0.98** 0.94** 0.90**

– – – – 0.91** 0.89** 0.94** 0.84** 0.81** 0.76** 0.93** 0.91** 0.94** 0.98** 0.97** 0.95** 0.99** 0.97**

0.07 0.02 0.51* –0.01 0.33 0.62* 0.61* 0.46 0.15 0.11 0.66** 0.64** 0.90** 0.91** 0.89** 0.87** 0.92** 0.41**

Tarpongi

Saguni

Station

Tarpongi

Saguni Khairkhut Station Station Tarpongi 1 Tarpongi 2 Station Station Saguni 1 Saguni 2 Station Station Station Khairkhut Khairkhut Khairkhut a

F = flowering, M = maturity, V = vegetative. bW = Kendall’s coefficient of concordance. cR = Spearman’s coefficient of correlation.

tial, and some disease resistance. It also has the dark green color plus good straw and grain qualities. Its bold, heavy grains remain soft after cooking and poor consumers prefer it because even a small quantity makes them feel full for a long time. Two farmers in each village volunteered to grow a set of diverse materials using their labor and level of management. They experimented with two sets of medium-duration rice genotypes in Tarpongi village and one set each of late-duration varieties in Saguni and Khairkut villages, which have heavytextured soils. The set of genotypes included pre-

released F7-F8 and a local check. The same sets (16 genotypes) were tested on an experiment station. During different stages of crop growth (vegetative, flowering, and maturity), farmers and plant breeders ranked the varieties on the station and in farmers’ fields. The assessment of these varieties by breeders and farmers is shown in Table 9. Correlation between breeders’ and farmers’ evaluation at all sites and in all the years was consistently low. However, there was high agreement in varietal ranking among farmers and among breeders. This may indicate that farmers and breed-

Rainfed lowland rice ecosystem

43

ers consider different criteria. Farmers’ ranking is not correlated with yield, indicating that farmers are considering other criteria in their rankings. The four late-duration lines preferred by breeders in the 1999 trials were BKP232, R650-1817, R304-34, and R738-1-64-2-2. These are all modern varieties and also gave the highest yields. Farmers preferred Swarna, Safri 17, R738-1-64-2-2, Mahsuri, and R650-1817. These were not always highest yielding varieties and Safri 17 is a traditional variety. At Tarpongi, the top-ranking medium-duration varieties for breeders were R574-11, IR42342, Chepti Gurmatia, BG380-2, R703-1-52-1, and OR1158-261. All of those are also the top-yielding varieties. All are modern varieties except for Chepti Gurmatia. For farmers, the top-ranking varieties included BG380-2, OR1158-261, R714-2-9-3-3, IR63429, and R574-11. These are all modern varieties, but not always top-yielding. Farmers and breeders agreed only on R574-11, BG380-2, and OR1158-261. A weighted participatory ranking method was used in assessing the trade-off between traits preferred by male and female farmers. The most important traits that both men and women value in rice varietal selection are grain yield, eating quality (taste), market price, duration or maturity, drought tolerance, and resistance to insect pests and diseases. Women placed higher weights on multiple uses of straw across all land types and size of landholding. This may be because women are responsible for caring for livestock and gathering fuel, and rice straw is important in both activities. Men gave more importance to grain size and shape for varieties grown on the uplands. Men having small farms considered adaptation of varieties to specific soil conditions to be extremely important (second to yield), but were the only group to rank that factor highly. The challenge facing the plant breeders is to develop new cultivars better than Swarna and Mahamaya, which meet the requirements that farmers have for their rice environments. Giving farmers an opportunity to test the performance of different rice genotypes on their own fields and to evaluate their cooking and eating qualities can lead to development and fast dissemination of varieties appropriate to farmers’ needs.

44

IRRI program report for 2000

Validation and delivery of new technology for increasing productivity of flood-prone ricelands of South and Southeast Asia The main project objective is to evaluate and disseminate new technologies for sustainable increase in rice yields and rice land productivity in the floodprone areas. There are two major areas of emphasis: ● farmer-participatory testing of adaptation of new rice production and resource management technologies including improved cultivars of rice in flood-prone ecologies, and ● socioeconomic analysis to assess the viability and social acceptability of the alternative technologies and thus support dissemination of the most appropriate ones. Bangladesh, India, Vietnam, Sri Lanka, and Thailand participate. Activities are identified for different subecosystems (boro, deepwater, tidal nonsaline, and tidal saline) of the flood-prone rice lands. Developments in boro rice farming V.P. Singh and M. Dhanapala The boro rice crop in South Asia is essentially an irrigated DS crop that takes advantage of abundant sunlight and residual soil moisture and is practically free from climatic adversities, except in some areas of Bangladesh where early floods coincide with crop maturity. Boro is known as spring-summer rice in Southeast Asia. The cultivation of boro rice with traditional varieties used to be limited to the river basins and deltas because of nonavailability of water during the DS. With the availability of high-yielding, DS rice varieties, and with rapid expansion of irrigation coverage, boro rice farmers started growing modern rices in the DS and leaving the deepwater land fallow in the WS. Boro rice is now spread across Bangladesh and is grown in Assam, West Bengal, Orissa, and Bihar states in India. Major concerns for researchers are development of shorter duration and cold-tolerant boro varieties to escape early floods and increase cropping intensity in the system, and improvement of crop management for increased efficiency of crop inputs.

IDENTIFICATION OF SUITABLE BORO VARIETIES

Advanced boro rice lines found superior in varietal screening were evaluated for single boro cropping pattern in seven farmers’ fields at Kuliarchar Thana and Karimgonj Thana, Bangladesh, with the objective of identifying short-duration, cold-tolerant, and high-yielding varieties. BR4828-54-4-1-4-9 yielded highest (5.9 t ha–1) and matured in 162 d. A recommended rice variety, BRRI dhan 29 (5.6 t ha–1), was the next best yielder, with maturity at 168 d. Farmers in Kuliarchar Thana preferred BR4828-54-4-1-4-9 to the recommended variety (BRRI dhan 29). In low-lying areas that flood prior to boro harvest, BR5877-21-2-3, which matures in 145 d was preferred by the farmers in despite lower grain yield (5.2 t ha–1). The advantage was its growth duration, which helps in escaping early flooding. It would also fit a double-cropped boro-deepwater aman system at medium flooding depth. Six rice lines (PSRM2-1-4B-15, RAU1344-3-2, IR59471-28-20-2-1, Panjasali, TRB7, and Banglami) were identified in Assam, India, as superior to the local checks. These promising lines were subjected to farmer participatory validation of their performance in Assam and Bihar. In Bihar, RAU1345-3-2 and RAU1344-3-2 performed well. Both had higher cold tolerance and shorter duration than the local check Gattu. RAU1345-3-2 was preferred by most farmers. It was recommended and released as variety Richharia and genotype RAU1344-3-2 was released as Dhan Laxmi for boro cultivation. Panjasali (192 d maturity) recorded the highest grain yield (9.37 t ha–1) in Assam. IR50 had a yield of 7.85 t ha–1 with maturity at 183 d. Both were superior to Mahsuri, the farmers’ check variety, which recorded a yield of 6.6 t ha–1 in 209 d. CROP ESTABLISHMENT

Timely crop establishment is of essence in boro rice farming because of temperature and growth duration, and the risk of flooding at maturity stages. The boro crop is traditionally a transplanted crop, wherein the requirements for a nursery delays transplanting.

A comparison of wet-direct seeding with conventional transplanting was done in farmers’ fields in Basail, Bangladesh, using the commonly grown varieties BRRI dhan 28, BRRI dhan 29, and BRRI dhan 36. The wet-direct seeded rice had a 10–14 d earliness in maturity and a yield advantage of 0.7– 1.1 t ha–1 over the transplanted system, mainly because of timely crop establishment. NURSERY MANAGEMENT

The nonavailability of early maturing, high-yielding boro rice varieties with cold tolerance at the seedling stage is a major constraint to boro rice cultivation because seedlings have to be raised during the cold winter season to be ready for timely crop establishment. Nursery management techniques were compared for three varieties of rice (Pusa 221, Jaya, and IR56383-7-1-1-1): ● covering the seedlings with polythene during the night to conserve the heat through restricted air circulation, ● early -morning removal of dew drops from the seedling leaves to prevent leaf scotching caused by evaporating dew drops, and ● dusting cowdung ash over seedling canopy to insulate against heat loss from the seedlings. All the three management techniques significantly contributed to the survival of seedlings, including the susceptible genotype Pusa 2-21. Covering seedlings with a polythene sheet provided the best results. CROP MANAGEMENT

Boro rice requires several irrigations during the growing season. Farmers in Assam, India, for example, provide as many as 32 irrigations per crop, each with about 7 cm water depth, mostly from shallow tubewells. A field demonstration of applying 7 cm water, 3d after the disappearance of water from previous irrigation (i.e., about 7 cm irrigation every after 8 d) in 10 farmers’ fields at Katimari village, Shillongani, Nagaon, Assam, achieved a water economy of more than 50%, without sacrificing any crop yield, when compared with farmers’ practice using same rice variety, crop management, and crop

Rainfed lowland rice ecosystem

45

growing periods. Average yields were 7.7 t ha–1, exactly the same as obtained with the farmers’ practice of applying 15 cm water every 8 d.

Progress of unreported projects Rainfed Lowland Rice Research Consortium ●

●

●

46

At CRRI-Cuttack in India, the farm mechanization group and IRRI collaborators worked to identify opportunities and priorities for adaptive research on agricultural mechanization. Initiated a farmer participatory experiment on validation and evaluation of drought-tolerant breeding lines. At NDUAT-Faizabad in India, five genotypes from the participatory varietal selection (PVS) program from 1997 to 1999 were tested in shallow, drought-prone rainfed lowland farmers’ fields in Mungeshpur and Sariyawan, Faizabad District. In Mungeshpur, PVS7 (NDRSB9830102) and PVS10 (NDRSB9730020) were the preferred varieties grown by nine farmers. Yields of PVS7 ranged from 2.5 to 3.8 t ha-1. Farmers in Sariyawan grew PVS1 (GayaPrasad Maurya) and PVS7. Farmers preferred PVS7 because of its good yield, suitability to land type, medium bold and cylindrical grains, good milling recovery, good eating quality, good quantity and quality of rice straw for animals, and ease in threshing due to their long stalks. Because of its short maturity (less than 120 days), rabi crops can be grown after PVS7, whose seeds are now spreading in nearby villages. At MMSU-Batac in the Philippines, recent economic analyses have shown that positive trends in total factor productivity in 1992-95 had turned negative in the subsequent period from 1996 to 1997. At an N rate of 120 kg ha1 , as recommended to improve input-use efficiency and sustainability, crop residue reincorporation further increased rice yields by 0.4 t ha-1. Responses of the five dry-season rotation crops to this new fertilizer recommendation varied, but sweet pepper crops continued to accumulate nitrate in the groundwater. This highlights the need for green manure or catch crops to reduce

IRRI program report for 2000

●

●

●

leaching of nitrate. To improve farmer awareness, at least 8,000 rice farmers in the region were given updated information on the potential of indigo as a green crop to enhance productivity in these intensifying rice-based cropping systems. AT RRIT-Ubon Ratchathani in Thailand, 12 promising lines were selected from observation nurseries under drought-prone rainfed lowlands and evaluated on-farm at five locations in the northeast and one in the north. The promising lines IR68796-27-3-B-5-2-B, IR69515-27-KKN-1-UBN-1-1-1-B, and IR57549-PMI-13-1-2-1, considered well suited to drought-prone poor soils and resistant to leaf and neck blast and iron toxicity conditions, will be nominated for release. Ten other promising lines, IR70182-18-PMI-7-2B, IR70215-70-CPA-3-4-1-B, IR71506-27-21-B, IR69515-27-KKN-1-UBN-1-1-1-B, IR68796-27-3-B-5-2-B, IR70172-1-SRN-4-1B-1-B, IR68835-133-3-B-B-4-2-B, IR6883544-5-B-B-9-3-B, IR68835-130-B-B-2-B-B1B, and IR68835-130-B-B-2-B-B2-B, had high yield, tolerance for some drought at early and flowering stage and resistance to blast. These lines will be evaluated on-farm to test adaptability at farm level. Of 1,668 breeding lines from Thai-IRRI shuttle research and 1,201 from the Thai breeding program, 70 and 36%, respectively, showed resistance to leaf blast; of these, more than 50% were resistant to neck blast. AT IGAU-Raipur in India, a regional analysis of rainfed ecosystems is being conducted at the micro scale, leading to a climatological assessment of drought incidence in Madhya Pradesh. The results are being mapped as an agroclimatic atlas for distribution to farmers. Nutrient management studies for improving the residual and applied P and nitrogen-use efficiency in rice-based systems are leading to an improved management package for the entire cropping system, which is being documented as a resource book for scientists, extensionists, and farmers. At CRIFC-Jakenan in Indonesia, recent findings in tillage, crop establishment, and weed management were published in the

●

international workshop on direct seeding. The results demonstrated the clear associations of weed species with toposequence position and soil fertility levels. Further experiments were set up at selected sites in central Java to quantify nutrient and water variability and weed community dynamics across toposequences in rainfed lowland rice. At BRRI-Rajshahi in Bangladesh, a long-term study of productivity and sustainability in rainfed rice-rabi cropping systems has begun. During 2000, an additional component on weed succession in the different rotations was added to the long-term experiment. ON-farm yield gap studies have indicated that any change from transplanting of rice to direct seeding increased the requirement for handweeding from one to three times. Although farmers are interested in direct seeding to bring the rice harvest forward to allow timely planting and reduced drought risk in the rabi crop, they are unlikely to change from transplanting unless the weeding issue is resolved. Accordingly, the third treatment in the long-term experiment now includes the use of Ronstar as a preemergent herbicide in the direct-seeded rice crop.

Facilitating technology transfer among NARES for the flood-prone rice ecosystem ●

●

●

●

Completed benchmark socioeconomic survey in flood-prone environments and its data analysis in Bangladesh. Developed a new plant type of deepwater rice in which elongation is triggered at a certain water depth and field-tested 52 lines in Bangladesh in the 2000 wet season. Conducted participatory testing and evaluation of new tidal saline and tidal nonsaline rice types for two seasons in Vietnam, Sri Lanka, Bangladesh, and India. Demonstrated the viability of wet seeding and the “dapog method” of raising seedlings and nursery management options for lowtemperature survival of boro rice in Bangladesh and India.

●

Promoted need-based nitrogen management in boro rice, including the field testing of the leaf color chart in Bangladesh.

Program outlook Research activities in the rainfed lowlands and upland rice ecosystems were combined into one program, Improving Productivity and Livelihood for Fragile Environments, in the new medium-term plan (MTP), implemented in 2001. The new program will focus on developing stress-tolerant, high-yielding varieties and efficient crop management practices that will reduce risk in rainfed and upland rice cultivation. Research partnerships with NARES to facilitate collaborative research aimed at understanding and developing technology and expertise to improve the productivity of rainfed lowland and upland systems will continue. Problems and issues will be addressed in three projects: ● Genetic improvement for improving productivity and human nutrition in fragile environments, ● Natural resource management for rainfed and upland systems, and ● Rice research consortia for fragile environments. Work on improved crop and resource management for the deeply flooded and coastal areas will continue to focus on understanding the nature and implications of emerging changes in these areas and strategic research needed for improved productivity and sustainability of these fragile systems. Genetic enhancement of rice for unfavorable environments is seen as an activity with high potential for gains in food security, human nutrition, poverty reduction, and environmental protection. Under the new MTP, work in this area, involving many disciplines and utilizing new and traditional methodologies will be in a project, Genetic Enhancement for Improving Productivity and Human Health in Fragile Environments.

Rainfed lowland rice ecosystem

47