Breeding for Adaptation to Rice Hydrological Megaenvironments

There are four major hydrological environments for rice production that can be defined in terms of toposequence position, or the relative elevation of a rice field within a watershed consisting of terraced fields that drain into each other (Garrity et al., 1986), and the resulting effects on the hydrological environment. Within distances of several hundred metres, the toposequence may include:

• unbunded uplands that never retain standing water;

• bunded but drought-prone upper fields that retain standing water only briefly after a rainfall or irrigation;

• well-drained mid-toposequence fields that receive a reliable supply of water from fields higher in the watershed, but that rarely experience stagnant flooding; and

• poorly drained lower fields in which water accumulates to depths of 1 m or more during the rainy season.

All four of these hydrological environments are often found within a small area in rainfed ecosystems. The latter three may also often be found within a single irrigation command area. Water shortage is mainly observed in unbunded uplands and bunded upper-toposequence fields. Drought stress in these environments varies in severity across years due to variability in the amount and distribution of rainfall, but occurs with predictable frequency in a given field, based on its toposequence position and soil texture. Yield variability under stress can be great even within a single field because of its variability in soil texture and levelness. This micro-scale variability among and within fields results in very large estimates of genotype x environment interaction and residual error in the analysis of rainfed rice trials, complicating selection strategies based solely on METs (Cooper et al., 1999). To cope with this variability, breeders need to use managed stress screening protocols that reproduce the range of hydrologies and water-related stresses that occur within the TPE they serve.

Production strategies for these hydro-logical environments are based on preexisting adapted germplasm. Over time, rice farmers have developed germplasm and management techniques adapted to each of the hydrological environments described above (Mackill et al., 1996). In unbunded fields at the top of a toposequence, farmers grow short-duration, drought-tolerant upland rice varieties established via direct seeding. Varieties used in these systems are usually tall, unimproved, and of the aus (in South Asia) or tropical japónica (in Southeast Asia and West Africa) varietal groups. In upper bunded fields, farmers tend to grow short-duration, photoperiod-insensitive modern varieties that flower before the withdrawal of the monsoon, escaping late-season drought stress. In well-drained mid-toposequence fields, farmers usually grow semi-dwarf varieties developed for irrigated systems because of their high yield potential, and usually establish their crops via transplanting. In lower and flood-prone fields, farmers usually direct-sow tall, photoperiod-sensitive varieties that flower as the rains cease and thus stagnant water begins to decrease (Mackill et al., 1996). Individual farmers often have fields at several toposequence levels, and thus often grow several varieties, each adapted to a particular hydrological environment.

Improved germplasm has been developed for each of these hydrological MEs. For unbunded uplands, upland rice varieties combining high levels of drought tolerance with improved yield potential and input responsiveness, termed aerobic rice, have been developed and are used in both rainfed upland environments and irrigated systems where it is necessary to reduce water use (Atlin et al., 2006). These varieties are developed at IRRI using a selection protocol that combines testing for yield potential in aerobic fields where soil water content is retained near field capacity, with managed stress trials conducted in the dry season, in which severe stress is imposed at flowering. Varieties adapted to upper-toposequence bunded fields, which must withstand intermittent periods of severe drying, are a major breeding target for IRRI, and are developed using managed stress protocols wherein paddies are drained intermittently throughout the growing season and then re-flooded when soil water potentials reach -70 kPA at 20 cm depths. These protocols, conducted at IRRI's main research station in Los Baños, The Philippines, have been highly successful in identifying germplasm that is broadly adapted within similar hydrological environments in different regions.

An important example of specific adaptation to a hydrological stress is submergence tolerance in rice, a case where a single major gene is the critical element in an adaptive trait. On millions of hectares where rainfed rice is grown by poor farmers, particularly in eastern India and Bangladesh, rice fields are subject to flash flooding that completely submerges plants. Most varieties will not recover from more than a week of submergence, but several landraces tolerate up to 2 weeks of complete flooding. The key trait associated with this tolerance is growth inhibition during submergence. Tolerant varieties become dormant and conserve carbohydrate reserves, while susceptible varieties grow rapidly in an effort to exert leaf tissue above the surface; if they do not succeed, they exhaust their reserves and die. Managed stress screening for the trait is easily accomplished in tanks and deep paddies that can be drained at will; seedlings are submerged for 14 days, then the tank or field is drained and survival is scored. A highly tolerant Indian landrace, FR13A, was used as a donor for the trait in genetic analyses that identified a single major quantitative trait locus (QTL), designated subl, which controlled 60-70% of pheno-typic variation for the trait in the screening system (Xu and Mackill, 1996). The Subl gene has been cloned, and was determined to code for a defective version of an ethyl-ene-responsive transcription factor (Xu et al., 2006). Cloning of subl allowed the development of gene-based markers for more accurate genotyping in marker assisted breeding (MAB). Subl has already been introgressed through MAB into 'Swarna', a widely grown rainfed rice variety that is highly preferred by farmers in India, Bangladesh and Nepal, but is highly susceptible to submergence. From project initiation, it took only 2 years to move the allele for tolerance into 'Swarna'. The improved version of 'Swarna', 'Swarna-Subl', has a two- to threefold yield increase over the recurrent parent after 12-17 days of submergence, and is currently being disseminated in submergence-prone areas of India and Bangladesh (IRRI, unpublished data). In this case, a clear genetic solution to a climate-induced stress, based on controlled imposition of stress in the breeding and genetic analysis process, was available, greatly reducing the need for multi-location testing for adaptation to a well-defined TPE. However, varieties introgressed with subl must be evaluated in METs before release to ensure that they are adapted and productive under non-flooded conditions.

0 0

Post a comment