Annual report 1994-1995

by David Spurgeon

"We never know the worth of water till the well runs dry," went the 17^th century proverb. According to some observers, we may soon know it because the world’s fresh water well is beginning to show signs of exhaustion.

Strictly speaking, this couldn’t happen: as a result of the hydrological cycle, however much water evaporates from the earth returns to it through rainfall. In this sense, there’s as much water in the world today as there ever was. But human uses of the life-sustaining fluid have increased enormously over time: between 1900 and 1990 total worldwide water withdrawals increased at twice the rate of the population increase while, compared with three centuries ago, water use rose more than 35-fold.

Moreover, water can become scarce in particular areas or regions – and in recent years, with increasing frequency, it has.

"In several parts of the world, water demands are fast approaching the limits of resources," wrote Sandra Postel of the Worldwatch Institute in 1989. "Many areas could enter a period of chronic water shortages during the nineties, including northern China, virtually all of northern Africa, pockets of India, Mexico, much of the Middle East, and parts of the western United States."

Less water for Asia
That prediction is now coming true. In Asia, where water has always been regarded as an abundant resource, per capita availability declined by 40-60% between 1955 and 1990. Projections suggest most Asian countries will have severe water problems by 2025.

As populations increase and economic development intensifies, governments will be forced to make critical decisions in their long-term planning for regulating, allocating, and using their water resources. Disputes have already arisen between some Asian farmers and industry in their competition for water. In the future, conflicting demands will increasingly be felt between the needs for safe drinking water, sanitation, and industrial activities, particularly in fast-growing urban centers. And such conflicts may well lead to social unrest.

Agriculture is by far the biggest consumer of water worldwide. In Asia it accounts for 86% of total annual water withdrawal, compared with 49% in North and Central America and 38% in Europe. Irrigated rice, in particular, is a heavy consumer of water: it takes some 5,000 liters of water to produce 1 kilogram of rice. Compared with other crops, rice production is less efficient in the way it uses water. Wheat, for example, consumes only 4,000 m³/ha, while rice consumes 7,650.

Too much water, or too little
Asia’s water problems are caused partly by its uneven distribution. On the one hand, about half of China receives less than 400 mm of rainfall a year, and extensive areas of northwest, central, and south Asia are drought-prone. On the other hand, the Ifugao rice terraces of the northern Philippines are situated in one of the wettest rice-growing regions of the world, with an average annual rainfall of 3,530 mm.

As a further complication, when rain comes in Asia it usually arrives in torrents over a short period, during a single monsoon that lasts from four to six months. The rest of the year is almost dry. As a result, much of the runoff simply flows into the ocean as waste, at the same time eroding the uplands, sometimes catastrophically. The monsoon, furthermore, is often erratic, so that in many countries, floods and seasonal water shortages occur concurrently.

Environmental costs of increased rice production
Water was a critical input to the Green Revolution, through irrigation, flood control, and drainage, and it has contributed most to the growth in rice production for the past 30 years. But this expansion has been bought at a cost to the environment: a proportion of the chemicals applied as fertilizer and as pest and weed control pollutes rivers and lakes through runoff, or groundwater through leaching.

In some upland areas, intensive agricultural practices, coupled with deforestation, have resulted in high rates of soil erosion and degradation of both land and water resources in lowland below. The effects can reach as far as coastal waters, with consequent impact on marine life.

The Manupali Watershed, which drains into Pulangi River in Bukidnon Province in southern Mindanao in the Philippines, offers an example. The upper area of this 60,000-ha watershed still has some forest cover, while the lower stretches are cultivated. More than 40% of the watershed consists of steep hills, and migration into the area from other parts of the Philippines during the 1950s and 1960s has led to a rapid population increase. This in turn has resulted in deforestation and increased agricultural activity in cleared forest areas on the upper slopes, and siltation in the lower, rice-producing regions. Reservoir, irrigation, and drainage canal capacity there have all been seriously reduced.

Another problem involves long-standing surface water, which causes waterlogging, makes the land unproductive, and leaves soils salty as the water evaporates.

In India, about six million ha of irrigated land are known to be affected by waterlogging. Nearly 10% of Pakistan’s irrigated 13.5 million ha is estimated to be affected by salinity, while northwest India and northeast China are similarly degraded.

Once waterlogging and salinization become visible, it is difficult to reverse the process, even with major investments in drainage facilities. Thirty years after commissioning of the left bank canal of the Tungabhadra irrigation project in Karnataka State, India, some 33,000 ha were waterlogged and saline: farmers abandoned about 20,000 ha there because the land was no longer productive.

Pumps, wells, and water shortages
Overexploitation of tube and shallow wells also presents problems. This is the case in large areas of India, Pakistan, and Bangladesh. The practice causes shortages of drinking water, and also pollution when aquifers are recharged with irrigation water contaminated with chemicals.

For example, in the early years of the rice-wheat rotation system in the Indo-Gangetic Plains of India, rice acreage expanded rapidly with the availability of shallow underground water resources. Centrifugal pumps and shallow wells provided cheap assured irrigation. The irrigation also reversed the trend of rising water tables and thus contributed to reclamation of saline lands. This encouraged farmers to install additional tubewells and further extend the rice-growing area.

At first there was no problem because the water tables were recharged by above-average rainfall and floods, but more recently the situation has assumed serious proportions.

"It is estimated that the groundwater table in Punjab is receding by 20 cm per annum over two-thirds of the state," reported R.S. Narang and M.S. Gill of Punjab Agricultural University in 1993. "This is also the situation in other rice-wheat areas to varying degrees. This has reached such alarming proportions that questions are now being asked as to what extent rice cultivation should be permitted in the irrigated Indo-Gangetic Plains, and how to sustain the productivity of the region without losing the battle on the water front."

Capital costs of irrigation systems have recently soared. Today in Sri Lanka it costs almost three times as much per hectare of land to set up an irrigation system as it did in the 1960s; twice as much in India and Indonesia; and nearly 50% more in the Philippines and Thailand. At the same time, market prices for rice have plummeted by nearly 40% over the past 30 years, while political pressures are mounting against large-scale projects from environmental groups.

Reducing water consumption
For all these reasons, attitudes to irrigation and appreciation of the value of water supplies are changing. In Asia, which contains 55% of the world’s estimated 253 million ha of irrigated land, there has been a sharp decline in the rate of growth of such areas. In East Asia, growth was 2% annually through the mid-1970s, but was virtually stagnant in the 1980s.

What this will mean for future rice production is that it will depend heavily on the development of water-efficient measures – producing more rice per unit of water input. (See sidebar Slipping between the cracks.) The trend now is to develop management policies for efficient operation of irrigation systems; technologies that reduce water consumption; changes in the rice plant itself and the ways in which it is grown, so as to use water more efficiently – and also to provide economic incentives to farmers to reduce water losses.

The development of rice varieties that take less time to grow, for example, would allow farmers in regions with a short rainy season to avoid early or late droughts. (Breeders have already shortened the maturation time of irrigated rice from 150 to 110 days.) And where the monsoon season is long, farmers could grow two rice crops without supplementary irrigation. Or they could grow a crop other than rice with the water left over at the end of the monsoon.

In rainfed rice systems, which is where more than half of all Asian rice is grown, farmers have little control over their water supply, and one of their main constraints to productivity is a lack of water when needed. Drought affects 40% of rainfed lowland and all upland areas in Asia. At the same time, the potential of existing rainfall for rice growing is underutilized. Improving rainwater use efficiency to alleviate drought therefore offers tremendous potential for increased rice production. Methods of doing this include reducing field losses and collecting runoff water in reservoirs and applying it to crops during droughts.

A parsimonious attitude toward water could really pay off. In 1990 the World Resources Institute estimated that between 65 and 70% of the water used around the world was lost to evaporation, leaks and other inefficiencies – but that it was possible to reduce these losses to 15%.

New rice varieties will be needed
In rainfed systems, farmers wait for heavy rains to flood ricefields before puddling transplanted seedlings: this delay may expose the crop to drought at a productive stage. Seeding rice directly into the soil without puddling could make more efficient use of rainwater. But for this to work, new rice varieties are needed with early vigor so that they can withstand sudden submergence by early rains and can compete with weeds.

Many farmers continue to grow traditional low-yielding rice varieties in rainfed environments because of their resistance to temporary submergence and prolonged droughts. Biotechnological research is helping scientists to understand the traits in these varieties that confer such resistance, so that they may incorporate them into modern cultivars to help stabilize yields.

Dry seeding – which can avoid the waste of 400-600 mm of rainfall – is assuming an important role in rice production in rainfed areas. The early harvest of dry seeded rice can allow planting of a second crop, which makes use of rainwater that arrives later in the season. The practice also reduces risk of drought where the rainy season is short, and because dry seeded cultivars can generate more roots.

The water-weed connection
One reason farmers keep rice fields continuously flooded is to keep down weeds, which compete less well with rice under such conditions. But if flooding is reduced, other ways of controlling the weeds will be necessary. Chemical herbicides are becoming less socially acceptable as a result of public concern about health hazards, and some weeds are developing herbicide resistance, so new methods of weed control will have to be sought.

IRRI scientists consider that, at present, knowledge of how to manage weeds is as primitive as was the knowledge of how to manage insects 20 years ago. That is because farmers today feel compelled to rely almost exclusively on one source of control – herbicides – just as, two decades ago, they relied principally on insecticides to control insect pests.

Worldwide, farmers spend about US$900 million each year to control weeds in rice. As labor costs for hand weeding increase, and water itself becomes more scarce and expensive, herbicide use can be expected to increase even further.

To break this herbicide dependence, IRRI scientists are turning to integrated weed management, just as they earlier turned to integrated pest management. They are examining the place of weeds in the entire rice-growing system, with the aim not simply of attacking and killing them by one means or another, but of managing them in a way that will not interfere unduly with rice production.

This involves studying the interaction of weeds with such components as water and tillage practices; and incorporating in the rice plant characteristics that increase competitiveness with weeds, such as early vigor and allelopathy. It also involves investigations of naturally occurring pathogens of weeds, and how they might be used to suppress weed growth. And it involves increasing farmers’ understanding of weeds.

By such means IRRI scientists are developing, today, environmentally sound weed management systems that they hope will prevent, a decade hence, the kind of excessive reliance on chemical herbicides that happened with insecticides a decade or so ago.

Biotechnology offers promise
In some 22 million ha of lowland and deepwater rice areas of the world, flash flooding leads to submergence of the plants. To cope with such hazards, some rice plants have developed two protective mechanisms: submergence tolerance and elongation. Submergence tolerance is appropriate to rainfed lowland areas where flash flooding lasts only for about 14 days, during which plants may become either partly or completely submerged. Elongation, on the other hand, is appropriate to areas where flooding is deep and lasts for several months.

A multidisciplinary team, involving eight institutes in India, Thailand, USA, Australia, and IRRI, is researching these protective mechanisms to see if they can be introduced genetically into new plant types. This could become the first instance in which a nonbiological environmental stress on a plant is ameliorated through biotechnology. Success would be particularly exciting because, while tolerance for stresses such as insects or disease can subsequently be overcome by mutations in the pest, the inbred tolerance of a plant for an environmental stress would be permanent.

Another major research initiative is aimed at introducing drought tolerance into plants to be grown in rainfed lowlands. The stress on the plant of varying soil conditions, from drought to ponding or even submergence, is considered the most severe limitation to rice productivity in rainfed locations.

IRRI’s early research focused on irrigated and favorable rainfed environments, and the high-yielding, pest- and disease-resistant varieties produced were generally successful. As IRRI began to consider less favored environments, however, its scientists realized that the rainfed system was more complex, and that more basic research was required to develop plants and cultural practices for these.

The research aims at characterizing the expected timing, duration, and intensity of drought stress in rainfed lowlands, and at understanding the plant’s genetic response to this stress. IRRI is collaborating with the Rockefeller Foundation, Texas Tech University, Australia’s Commonwealth Scientific and Industrial Research Organization (CSIRO), and the University of Queensland in efforts to develop molecular markers – a powerful new tool – to help find genes for root traits likely to confer drought tolerance on rainfed rice. Marker research is also being applied to upland rice.

As the CGIAR sees it
So important is the water problem worldwide in the view of the CGIAR that it designated water management as a systemwide initiative beginning in 1995. The objective is to increase the effectiveness of water management research through improved coordination and enhanced collaboration among many of the 16 international agricultural research centers and the NARS in developing countries.

The International Irrigation Management Institute in Sri Lanka has been designated as the convening center for this initiative. IRRI will provide expertise and leadership in management at the farm level.

While the full import of the water supply problem to rice production has been recognized only relatively recently by the research community, it is now fully acknowledged there. Coping with its implications will be central to IRRI’s concerns in the years ahead.

Traditionally, farmers in Southeast Asia have established their ricefields by transplanting seedlings but, during the past 10-15 years, many have changed this practice to sow seeds directly onto the field.

Direct seeding can be done by broadcasting pregerminated seed onto wet, puddled soil (wet seeding). Or it can involve sowing ungerminated seed onto dry or moist but unpuddled soil (dry seeding).

Wet seeding requires less water than does transplanting for both land preparation and crop irrigation. IRRI studies in the Philippiines showed the amount of water used by farmers for land preparation was 27% less with wet seeding than with transplanting. That was because wet seeding required less time to complete the preparation: 6 days compared with 24. Yet the yields from the two systems were almost the same: 7 t/ha for wet seeding, 6.5 t/ha for transplanting.

During irrigation, the farmers who wet seeded again used less water than those who transplanted because they maintained shallower depths of water, especially during the post-vegetative stage of the crop. This, in turn, was because the wet seeded fields were leveled better and herbicides were used rather than deep water to control weeds. The shallow water also helped the rice plants to remain upright in strong winds.

Where water was scarce, research showed that wet seeding provided a higher yield than did transplanting. Farmers using wet seeding are aware of this advantage.

In dry seeding, seeds are broadcast before or at the onset of the rainy season and germinate when rain provides enough moisture. Almost all the early rainfall contributes to crop growth. In the Mekong River Delta of Vietnam, the whole dry seeding cycle, from land preparation to harvest, used only 700-900 mm of rainfall – nearly as much water as it took to puddle the soil for transplanting.

Dry seeding, therefore, can be used to advantage in rainfed rice areas, where for transplanting farmers have to wait to prepare the land until rainfall is adequate to saturate the soil. In these areas, farmers also often use traditional rice varieties, which take longer to become established and yield less. With these delays, there is not enough time and late-season rain to grow a second crop.

Dry seeding of modern, early-maturing varieties has already led to crop intensification in large tracts of Vietnam, Indonesia, and some parts of the Philippines. And their productivity was much higher than that of a single transplanted rice crop.

Dry seeding can also be used in irrigated systems where there is not enough water in the reservoirs for land preparation and crop establishment early in the season. There, rainfall is used to establish the crop, and irrigation supplements it later on.

How important can dry seeding be in irrigated ecosystems? In 1978, an entire transplanted crop was cancelled in Malaysia’s Muda irrigation system when reservoir levels were low. Yet in 1991, when low water levels in reservoirs resulted in no water at all being used for irrigation, farmers’ dry seeded crops yielded an average of 3.9 t/ha.

More than half the water consumed in rice production is often used to prepare the land – and most of this is lost in the process through percolation and seepage. A major thrust of IRRI’s research is toward reducing these losses.

Most Asian farmers till a wet field rather than a dry one, because it facilitates transplanting of rice plants, helps level the land, plows under weeds and stubble, and improves the soil conditions for plant growth.

They first soak the land until the topsoil is saturated, shallow plow once or twice, and then harrow once or twice. Plowing and harrowing are carried out with water standing in the field.

Rice is usually grown in clay soil, and alternate soaking and drying produces deep and wide cracks in it. In fields with permeable subsoil, up to 60% of the water applied for soaking flows through these cracks. About 30% of the flow recharges the water table below, while 70% is lost through lateral drainage.

Experiments in the Philippines have shown that shallow surface tillage after harvesting of the previous rice crop can save about 200 mm water during land soaking and preparation. The tilled layer minimizes deep crack formation and surface soil particles block water flowing into the cracks.

Water savings have also been made possible with the increasing accessibility of high-powered tractors, which makes dry tillage possible in many Asian rice areas.

The main determinant of water use efficiency in a ricefield is the reduction of percolation loss. Puddling causes the formation of a semi-impermeable layer just beneath the puddled topsoil, thus reducing loss. IRRI studies show that even a small area of nonpuddled soil increases percolation losses by a factor of five. A further 2- to 5-fold increase in percolation water loss results from movement of water from the flooded fields into the bunds and then down to the water table. Much percolation loss can therefore be prevented by reducing water flow into the bunds: some farmers do this by sealing the bund walls with mud. Decreasing the depth of ponded water also reduces percolation loss.

IRRI’s work shows that maintaining a saturated soil throughout the growing season can save up to 40% of water in clay loam soil, without yield reduction. Weed control is possible through chemical, mechanical, or manual means – with the latter, often without additional cost. Farmers prefer to flood their land continuously, however, as insurance against future water shortages, and to control weeds without manual, mechanical, or chemical inputs.

Research shows that where weed growth is a serious problem, continuous flooding up to formation of the rice plants’ panicles, and continuous saturation after that, uses 30-35% less water than the traditional practice of continuous flooding – and without any increase in weed infestation or reduction in rice yield.

back to top