Comparing water input and water productivity of transplanted and direct-seeded rice production systems

Abstract

Increasing water scarcity threatens food production in irrigated rice systems in Asia. It is, therefore, important to identify rice production systems that require less irrigation water than traditional transplanted (TP) rice. This study investigated the effect of crop establishment methods on irrigation input and water productivity (weight of produce per unit volume of water used) in three irrigation service units (ISU) from 1988 to 1994 in the Muda Irrigation Scheme, Malaysia. Water balance components, crop establishment method, the progress of farming activities, and rice yield of individual farmers and over the whole ISU were monitored. Yields in TP ISU were higher than in wet-seeded (WS) ISU and those in WS ISU higher than in dry-seeded (DS) ISU, but the difference was significant (P<5%) only between TP and DS ISU. Land preparation duration was significantly reduced in DS and WS ISU compared with TP ISU. This led to a significant reduction in irrigation and total water input (rainfall and irrigation) before crop establishment. However, during the crop growth period in the main field, TP ISU had a significantly shorter crop growth duration and total water input than DS and WS ISU. Over the whole crop season, the three crop establishment methods had a similar total water input, but DS ISU had significantly less irrigation water and higher water productivity with respect to irrigation water than WS ISU and TP ISU. This was attributed to the ability of DS rice to capture more rainfall after crop establishment. Site-specific management of WS rice versus TP rice has to be taken into account in assessing their relative advantage for water input and water productivity.

Introduction

More than 75% of the world’s rice supply comes from 79 million hactare of irrigated rice production in Asia. This system is a major user of fresh water, accounting for approximately 50% of the total diverted fresh water in Asia. Global rice demand in 2020 is projected to increase by 35% over the 1995 level. Much of this increase will rely on irrigated rice systems. The available amount of water for irrigation, however, is becoming increasingly scarce due to decreasing resources and quality and increased competition from non-agricultural water users. For food security, it is essential to “produce more rice with less water” (Guerra et al., 1998).

In most of Asia, rice is largely established by transplanting (De Datta, 1981, Pandey and Velasco, 1999). This involves growing seedlings in a nursery bed and later transplanting them in a main field. Land preparation activities in the main field begin with pre-saturation irrigation to saturate the topsoil layer and to create a ponded water layer (about 5 cm depth) for land soaking. This is followed by plowing and puddling (harrowing under saturated soil conditions). After land preparation, the crop growth period in the main field goes from transplanting to harvest.

Water is applied to the main field during land preparation and crop growth to compensate for outflows (seepage, S, and percolation, P) to the surroundings and depletions to the atmosphere (evaporation, E, and transpiration, T). Because S and P are difficult to separate in the field, they are often expressed together in one term, S&P. Only E takes place during land preparation, whereas both E and T occur during the crop growth period. Since it is difficult to separate E and T during crop growth, they are often expressed in one term, evapotranspiration (ET).

The water input for transplanted rice depends on the magnitude of the outflows and depletions and the duration of land preparation and crop growth. In many large irrigation systems in Asia, farmers often delay plowing and puddling when waiting for the seedlings being nurtured in seedbeds. This results in a prolonged duration and high water input during land preparation. For example, it took up to 63 days and 930 mm of water to complete land preparation of 145 ha blocks in the Upper Pampanga River Integrated Irrigation Systems in the Philippines (Valera, 1977, IRRI, 1978) and 60 days and nearly 1000 mm of water for 50 ha blocks in the Muda Irrigation Scheme (Fujii and Cho, 1996). Wickham and Sen (1978) also reported that percolation during land preparation accounted for up to 40% of the total water supplied for growing a rice crop. Thus, minimizing land preparation duration can potentially reduce the water input for rice cultivation (Tuong, 1999).

Bhuiyan et al. (1995) and Tuong (1999) suggested that direct seeding (wet seeding and dry seeding) could be an option to shorten the land preparation period. In wet-seeded (WS) rice, pre-germinated seeds are sown directly onto the puddled main fields, where rice plants will grow until harvest. In dry-seeded (DS) rice, dry seeds are broadcast onto dry or moist soil. In some temperate countries, such as in the United States and Australia, the broadcast dry seeds germinate with moisture supplied by flush irrigation (Akita, 1992). In tropical monsoon Asia, DS rice is often a rainy season crop. Seeds are broadcast at the beginning of the monsoon season and germinate when rainfall is adequate to moisten the soil (My et al., 1995, Lantican et al., 1999).

In a field experiment in the Philippines, Bhuiyan et al. (1995) reported that WS rice reduced the irrigation water input compared with TP rice. The authors showed that this was due to a shorter land preparation period, and hence, less S&P and E during this period, in WS rice than in TP rice. On the other hand, authors such as Ho Nai Kin et al. (1993) argued that wet seeding requires the standing water to be drained out before seeding, resulting in substantial wastage of water. This drainage often over-irrigates neighbors’ fields. Consequently, planting activities over the whole area are delayed. Furthermore, because TP rice already spends a part of its life cycle in the nursery, it is expected that TP rice would have a shorter crop growth period and, hence, require less water in the main field than WS rice. The trade-off between water savings during land preparation and increased water input during crop growth in the main field of WS rice compared with TP rice has not been systematically investigated.

My et al. (1995), Lantican et al. (1999), and Tuong et al. (2000) reported that dry seeding advanced crop establishment, made better use of early season rainfall than transplanting did, and facilitated an increase in cropping intensification in rainfed rice systems. The effect of dry seeding on water input in irrigated rice systems has not been investigated in details. It is expected that, since land is prepared dry, the amount of water input for land preparation of DS rice would be less than that for WS and TP rice. Dry land preparation, however, means that the effects of puddling on reducing soil permeability cannot be realized. The increased water input because of higher percolation may outweigh the water savings during land preparation.

Most previous investigations on water management of WS and DS rice were carried out at the field level. Molden (1997) pointed out that the “water loss” from a field may not be a loss at a larger scale (an irrigation system or a part of it) because the outflows from a field can be re-used in another field downstream in the system. The effect of the large-scale adoption of direct seeding on water input in a rice irrigation system has yet to be investigated.

Most water saving technologies result in some yield losses (Bouman and Tuong, 2001). Molden (1997) and Tuong (1999) suggested that water productivity (weight of produce per unit volume of water used) should be used as an indicator to compare different systems/technologies in terms of their effective use of water for food production. Bhuiyan et al. (1995) showed that wet seeding saved water and increased rice yield, but more information on the effect of crop establishment methods on water productivity of rice is needed to determine the conditions where changing from transplanting to direct seeding would contribute to “producing more rice with less water”.

The objective of this study is to compare the components of water balance and water productivity of DS rice, WS rice, and TP rice grown on a large scale within an irrigation system. We will present the results of a field study within the Muda Irrigation Scheme, Malaysia, and discuss the conditions under which direct-seeding systems will lead to water savings and increase water productivity.