Plant breeding strategies for rainfed lowland rice in Northeast Thailand and Laos

• Department of Agriculture, Thailand
• Department of Agriculture and Extension, Laos

Frequent drought restricts the production of rainfed rice in Thailand and Laos. Through the growing season, rainfall in Laos and in northern and northeastern Thailand is unpredictable in quantity and timing. In addition, relatively permeable but infertile soils undermine the yields from rice cultivars presently in use.
Rice occupies a large proportion (between 63 and 87 per cent) of the 0.75 million hectares of cropped land area in Laos, but about 98 per cent of the area is rainfed. Of 9.6 million hectares of rice crops in Thailand, 50 per cent are in northern and northeastern rainfed areas that yield only 2 tonnes per hectare on average, contributing 26 per cent of the total Thai production in 1993-94.
Laos was not even self-sufficient in this staple grain - its most important crop. But the government wanted to achieve high annual rice yields and a surplus for export. Thailand was already a rice exporter, but both countries needed to ensure stable production levels because so many of their subsistence farmers relied on it.

This project had five specific objectives: 1) to evaluate the existing and alternative breeding methods, particularly selection strategies for rainfed lowland rice in northeast Thailand; 2) to develop a screening method against late-season droughts and to identify physiological and morphological traits that confer drought resistance; 3) to examine whether or not screening under high fertiliser input is appropriate for selection of genotypes and to identify reasons for superiority of some genotypes under different soil fertility conditions; 4) to evaluate the genotypic requirement for direct seeding to minimise adverse effects of drought; 5) to quantify the effect of various environmental factors on phenological development (e.g. time of flowering and ripening) of different rice cultivars, and to examine whether different phenology groups are required at different locations.

The major scope of the project was to identify better screening methods to enhance the process of selection of rice cultivars that yield better than the existing cultivars under water-limited and adverse soil environments. The main focus for Subproject 1 was to identify efficient strategies for evaluating and selecting lines in early and advanced generations of inbreeding, and identify test locations for evaluating yields of segregating populations. This together with Subproject 5 examined strategies to improve the breeding program using existing facilities.
Subproject 2 investigated the need to introduce a new method of evaluation in which plants are screened under drought conditions to select lines for their drought tolerance. This subproject, and subproject 3 (genotype and soil fertility interaction), investigated the possibility of indirect selection for physiological, morphological or nutritional traits as indirect selection criteria.
Subproject 3 also examined whether the level of soil fertility should be considered when selecting for high yield. Subprojects 1, 2, 3 and 5 were designed to improve breeding efficiency for rice cultivated by transplanting in rainfed lowland conditions. With the recent shift to direct seeding, high-yielding materials selected for transplanting conditions may not produce high yield under direct seeding. Examination of this interaction was the scope of subproject 4.
Subproject 5 also contained part of the Australian program and this examined the problem of low temperature in irrigated rice. The aim of this subproject was to provide information for development of new cultivars and agronomic practices to reduce the effect of low temperature during panicle development and increase the fertility of spikelets.

The project successfully achieved its major goal and the five specific objectives. The project has provided improved understanding of genotype response to drought, genotype selection for different fertility environments, agronomic management and cultivar requirements for direct seeding, low temperature effects on spikelet sterility, and the importance of phenology in the selection program in rainfed lowland rice. The project also guided the plant breeders in Thailand to modify the existing plant breeding program by considering genotype-by-environment interactions along with the understanding of phenology, potential yield and timing of stress. Results from the project have led to modifications to the Thai breeding program, and further modifications are scheduled.
The rice breeders in Thailand identified the importance of genotype-by-environment interactions for grain yield as a basis for modification of plant breeding methods, and the program is in the process of being modified. They also identified a few genotypic characteristics associated with higher yield under late-season drought conditions in northeast Thailand.
The scientists determined appropriate agronomic methods for direct seeding of rice. This included appropriate time of planting, planting density and row arrangement, weed control measures and choice of cultivars with appropriate phenology and other attributes. They also gained greater understanding of mechanisms of low temperature effects during the panicle development stage and the impact on grain yield. This information is useful for development of a breeding program for low temperature tolerance in rice, and a screening method for cold tolerance has been developed.
The research led to the development of a rainfed lowland rice model that can be used to estimate drought effects. The model incorporates the drought and nutrient availability interaction in estimating the effects of drought on grain yield.