Beyond drops for crops: a system approach for assessing the values of water in rice-based systems

D. Renault^a and T. Facon ^b
aIrrigation Management Officer, Water Resources, Development and Management Service, FAO, Rome, Italy ^b Water Management Officer, FAO Regional Office for Asia and the Pacific, Bangkok, Thailand

RICE BEYOND THE GRAIN AND THE PLANT

Rice-based production systems are intimately connected with water development. For hundreds of years, natural selection pressures (e.g. drought, submergence, flooding, nutrient stresses and biotic stresses), as well as human manipulation, have contributed to the diversity in rice agro-ecosystems. Whether rainfed lowland or upland, deep-water/flood-prone or irrigated, rice cultivation is intimately linked to water. The water-ponding system of rice cultivation has widely shaped the landscapes of rural areas; a typical rural landscape in Asia is that of paddy fields on mountain slopes developed over the years or of a flat plain with abundant ponds and waterways. Water-ponding also has the advantage of allowing efficient weed control without chemical inputs, which is good for the environment.

The value of paddy field cultivation for the local community is seen not only in food production but also in many other areas, such as:

• social capital - the strong social cohesion needed to deal with the massive labour requirement for building and maintaining the terrace or canal system as well as for synchronizing the cropping pattern;

• soil erosion and landslide protection; and

• flood protection - storage of precipitation within the bunds of the paddy system.

The integration of paddy cultivation in local cultures is thousands of years old in some parts of the world. Religious and social life is tied to rice production and the rice season cycle. However, in many countries these systems are in a critical situation for various reasons, including the widening gap between the incomes they can provide and the potential incomes in other sectors (a situation highlighted by symptoms such as ageing or womanization of the farming population and part-time farming) and decreasing farm sizes.

Since the 1950s, new types of rice system have emerged: the "modern" medium- or large-scale agency-managed systems, new ecosystems resulting from human activity, mostly non-voluntarily but there nevertheless. These systems evolved from supplementary irrigation in the rainy season to double-cropping and year-round irrigation, and more recently to multiple use and conjunctive use.

Rice ecosystems under submergence are endowed with unique features which influence natural resource management. Soil-submergence-related biochemical processes exercise a great deal of influence on ecosystem sustainability and on vital aspects, such as carbon sequestration, nutrient cycling and water quality. Rice soils under sub-merged conditions provide environments which are congenial to soil organic matter build-up, carbon sequestration and biological nitrogen fixation. Soil organic matter serves as a nutrient reservoir, particularly for providing nitrogen to plants. Soil submergence under irrigated low-land rice ecosystems is an important contributor to the long-term sustainability of organic matter resources and to the nutrient-supplying power of soils. On the other hand, paddy fields with their methane emissions are a major contributor of greenhouse gases.

Besides sustaining crop production, the value of water in rice-based ecosystems is multifold. In the agroforestry gardens of Asia it also provides an air-cooled environment for people living below the canopy of trees, such as the coconut trees, which are very often fed by water from the neighbouring paddy fields. To the uninformed, rice cultivation may appear to be a high water-consuming crop, because of spillover and percolation; however, once the recycling process and all the various downstream uses of water are considered, it is clear that rice generates highly water-productive and efficient agro-ecosystems. The impact of rice cultivation on natural resources and aqua-ecosystems is twofold: on the one hand, water withdrawal for paddy systems results in diminishing water availability for natural ecosystems (e.g. wetlands); on the other, rice agro-ecosystems are responsible for the creation of human-made wetland systems (some of which are already considered as such by the Ramsar Convention [Convention on Wetlands of International Importance Especially as Waterfowl Habitat]).

As far as health is concerned, the effects of rice cultivation are both positive and negative. Widespread standing water in rice systems without doubt contributes to the increase of the vector of water-borne diseases (e.g. malaria). However, diverse rice systems utilizing natural predators will help reduce disease hosts, and the economic wealth generated by rice intensification will make for wider use of protection against the mosquito. Thus, the balance between the positive and negative impacts on health is somehow mixed.

Understanding the values of water in a rice-based system requires the need to look beyond crops towards the introduction of a more holistic or system approach. This system approach encompasses various systems (Figure 1):

• nutrition;
• farmer livelihood;
• rural economic; and
• water management.

This paper is a brief introduction to this system approach.

FIGURE 1
Embedded system approach for tackling rice-based systems

WATER UNDER EXAMINATION

In a global world where competition for natural resources is increasingly acute, each human need and activity must be scrutinized carefully in order to improve efficiency and equity in resource allocation. Water for food is by far the main requirement for human beings and it will remain so regardless of past and future increases in agricultural productivity. It is therefore logical to scrutinize water consumption per agricultural product in order to promote the most efficient use of water for food production.

However, the evolution of rice systems will not be guided by considerations only related to water, but also by socio-economic factors. So the system approach aims not only to find out all the valuable uses generated by water management in rice-based systems, but also to find environmentally and economically sustainable ways of managing the rice systems or realistic alternatives that will lead to equivalent values.

Many agricultural and irrigation experts from arid areas and the western world have for a long time accused rice cultivation of using too much water compared to other cereals. At the same time, experts and the entire population of Asia, where rice has been cultivated for so many centuries, praise rice cultivation as the basis of their life.

Only recently have these two points of view been reconciled. Recent investigations by the water management community have documented clearly the grounds on which rice cultivation is felt to be so valuable for Asian societies and it is clear that water management performance in rice systems must look beyond the crop and be assessed on a system basis. In practice, recent accurate water balance studies often show that for rice-based systems:

• the beneficial uses of water are much greater than the crop use itself (multiple beneficial uses in the command area);

• there is still room for performance improvement (reduction of real losses of water); and

• there is often a need to improve water productivity through increased yields.

These last two points may require changes in the traditional water management system. Therefore, one should not confuse the findings that the rice system fulfils multiple roles with an advocacy for maintaining the rice systems in their present state.

In the following context:

• water resources have been developed chiefly for rice irrigation;

• it is necessary to satisfy new water needs - e.g. urban or industrial - from the same resource and to manage irrigation systems for multiple uses; and

• water resources as well as irrigation management are moving towards more participatory and decentralized management.

It is important to understand what roles or uses the systems are fulfilling at present, the opportunities for the systems to fulfil new roles and how the systems can be managed to optimize the total productivity of water.

Beyond crop production, the values of rice agrosystems are many and not always easy to identify and evaluate. Probably the best way to quickly get a sense of what these values are is to consider the alternative to rice: what if rice cultivation no longer existed? What would the consequences be for the ecosystems, the economy and the livelihood of the people?

LIFE... WITHOUT RICE CULTIVATION?

Although it has been a way of life for centuries throughout Asia and in many other parts of the world, the current agricultural economy is not much in favour of rice cultivation. The real price at farmgate has dropped significantly during recent decades and the temptation is high, whenever possible, for farmers to switch to other crops. The issue of continuing rice cultivation is critical in many rice-based systems and not only in the wealthiest Asian countries. In some countries in southern Europe it is becoming a hot issue due to the decline of support from the European Union. At the same time, people are increasingly aware that suppressing rice cultivation will modify dramatically the entire water circulation in sometimes very large command areas; the recognition of the externalities linked to rice cultivation is fuelling the debate. Therefore, the hard way to learn about the values of rice cultivation is to figure out what would happen to the agro-ecosystem without it.

It can be demonstrated that the consequences for many agrocontexts in Asia would be dramatic because:

• there is little or no alternative for staple food cultivation during the rainy and wet season as cereals cannot stand heavy rains (no rice in a monsoon type agrocontext would mean a big reduction in food).

• rice cultivation and related water management practices are de facto sustaining a complex agroecosystem that would simply collapse if deprived of an abundant water support for rice.

A DIVERSIFIED SYSTEM APPROACH FOR RICE CULTIVATION

It must be clearly stated that there is no single system for rice. Rice systems are site-specific and although in many areas there may be similar bioclimatic conditions, farming systems can still be different. Therefore, it is crucial at this stage to not consider one generic system but to embrace rice agro-ecosystems in their full diversity: from monocrop industrial farming systems to fully integrated agrosylvopastoral systems.

Another issue related to rice agro-ecosystems is the seasonality of their behaviour: during the wet season (e.g. monsoon) they are known as "wet systems", while during the other cropping seasons they are "dry systems" requiring an external water supply and which can mean very high crop diversification. In addition, the costs for water may be significantly different for the dry and wet seasons.

Therefore the range of options may vary: from traditional rainy-season management, more or less unchanged; to dry-season management, drastically changed (with a potentially positive impact on rice itself as well as other uses); to bringing significant changes in both the wet and the dry season.

RICE - ONE OF THE MAIN DIETARY ENERGY SUPPLIES

Rice is the staple food for approximately half the world population. It has always been vital for fulfilling human food needs, particularly in Asia where population density is high and per caput availability of arable land very low. Rice is the only staple food crop that can stand heavy rains during the wet season, and this is the reason why the growing of rice has spread so much in tropical countries, particularly in Asia.

The 575 million tonnes of paddy rice produced in 2002 represented the world's single most important source of dietary energy (21 percent). Rice cultivation expanded until 1975, when it virtually stabilized at around 145 million ha; expansion still takes place in some areas, for example, in low-income countries where it recently increased from 75 to 92 million ha. World average yields have more than doubled since 1961, reaching 3.8 tonnes/ha in 2001.

Compared with wheat, rice is effective for energy (kcal/kg), but significantly less for protein and much less for other micronutrients, such as iron, magnesium, phosphorus and manganese. There is therefore a need to overcome rice nutritional deficiency, which may be done in two main ways: improving the nutrient content of the product or diversifying the intakes.

Improving the nutrient content

Many diverse aspects within the rice-based system can be considered in order to improve the micronutrient status of rice:

• selection of varieties with exceptional nutrient content;

• biofortification - a sustainable technological tool that can be used to enhance the micronutrient content of the grain and to reduce grain phytate content which inhibits the bodily absorption of these micronutrients; and

• simple processing techniques, such as promotion of less refined milling of the grain and parboiling, to preserve available nutrients.

The system response: diversifying the products

The diversification of other products in the diet has always been the response of communities to compensate for rice nutrition deficiencies. In rural dominated societies, rice-based systems are usually quite successful in producing food products which are complementary to rice (fisheries, fruits, vegetables, meat and milk) and this should be further encouraged. In modern large rice cultivation systems, perspectives vary. The main strategic options include:

• focusing on the rice crop only, with a view to maximizing rice output (e.g. the granary schemes in Malaysia), especially in systems with large holdings or after some form of consolidation;

• engineering into systems design and management the satisfaction of different uses;

• transforming the irrigation schemes into multiple-use schemes (e.g. rice in the wet season and diversified rice, fish and cash crops in other seasons); and

• supplying water for urban requirements.

The question is how can we preserve or enhance the values while fulfilling the aspirations of a much better life?

BEYOND DROPS FOR CROPS: A COMPLEX AGRO-ECOSYSTEM

An agro-ecosystem is a complex of air, water, soil, plants and animals in a defined area that people have modified for the purposes of agricultural production. It develops in an environmental setting which defines the resources available in a social setting which conditions how farmers and consumers interact with each other and with the environment.

Rice-based systems exhibit some peculiarities that need to be properly accounted for when sustainability is at stake. More than for any other crops, rice agro-ecosystems are profoundly marked by water and land management. For hundreds of years, natural selection pressures (e.g. drought, submergence, flooding, nutrient stresses and biotic stresses) have contributed to the diversity of rice agro-ecosystems.

The marriage of rice and water has generated complex patterns of water circulation within the command area of rice-based systems, be they rainfed or irrigated. This type of system is often considered (both traditionally and in modern management) as a "cascade system": water spills from one field to another, recharges groundwater, supplies nearby land, drains from one system to feed another one, and recycles drainage water through surface ponds and other open water bodies.

In rice-based systems, the quasi-permanent presence of water throughout the command area heavily modifies the environment and affects the conditions required for vegetal and animal life to prosper, allowing the diversification of local production (fisheries, trees, gardening). "Rice is life" also because it mobilizes resources and paves the way to many other important products for human nutrition, wealth and the environment. In some dry zones of South Asia, coconut trees can only grow if associated with paddy fields; the coconut is traditionally known as the "tree of life" because it produces almost everything required for life, from additional food to material for the house, as well as shade for a cooler living environment. The symbiosis between rice and other agricultural productions is a major feature of rice-based systems that needs to be fully recognized, particularly with regard to the livelihood of smallholders. One should not forget of course that some of these rice systems have been created by the drainage of natural wetlands, which may have provided some of these services in the past.

In many places, water delivery systems required to serve rice production are de facto serving many other purposes critical for the rural poor, such as the domestic water supply, and water for fisheries and the environment. Therefore, targeting improved water management in rice-based areas may serve several Millenium Development Goals, namely: food security, poverty alleviation and domestic water supply.

This particular ecosystem (resulting from the marriage of rice and water) presents positive and negative externalities which need to be fully considered in order to move towards sustainable development. Internalizing externalities is both a constraint (cost) and an opportunity for rice-based systems.

WATER-PONDING: A REMARKABLE FEATURE OF TRADITIONAL RICE CULTIVATION

Rice is the only cereal that can stand water submergence and this feature is crucial to understand the long and diversified story of rice and water. The plant's adaptation strategies include: surviving submerged conditions with no damage; elongating stems by several decimetres in one day in order to escape oxygen deficiency due to rising watertables; and withstanding severe drought periods. On the basis of this agroecological diversity, ecologists have proposed several classification systems, and the most widely used distinguishes five water-related categories: rainfed lowland, deepwater, tidal wetlands, rainfed upland and irrigated rice.

For thousands of years, paddy has been associated with fields of standing water in various agro-ecological systems. Water use at field level is thus always high due to infiltration, percolation and spill. Paddy is always blamed, therefore, for being a high consumer of water (between 900 and 2 250 mm per season) compared with other cereals (400 to 600 mm).

The increasing water crisis at global and local level raises questions about the future of paddy. Issues of sustainability force us to consider whether paddy should give way to less water-consuming cereals or crops. This requires further scrutiny and an impartial debate. Competition for water for other uses is increasing rapidly, and there is no doubt that water consumption will come under increasing scrutiny. There is a need for reliable figures on inputs and outputs, and the advantages and disadvantages associated with paddy cultivation. Hence, a new look at paddy cultivation may review the value of the traditional paddy system as well as technical developments.

THE VALUES OF WATER IN RICE-BASED SYSTEMS

Because of the complexity of the water-flows generated by water-ponding, traditional rice-based ecosystems are the habitat for a huge variety of living resources, including both terrestrial and aquatic organisms. Rural people rely on the biodiversity associated with rice-based ecosystems - including the natural predators of rice pests - and often enhance this biodiversity with cultivated plants, domesticated animals and aquaculture, so as to guarantee their daily food supply and income. Fish, frogs, snails, insects and other aquatic organisms derived from these ecosystems are a source of animal protein and essential fatty acids. In Asia, during dry periods, paddy fields are often the only source of water for valuable trees (coconut) and home gardens.

The control of water layers at field level for submerged rice growth has led over the ages to the development of specific water management and cultivation practices that produce some specific beneficial outcomes. The terrace system in mountainous areas is a typical product of the ponding technique which has allowed cultivation even on steep slopes. This technique is instrumental in preventing soil erosion and landslides. Another advantage of this technique is its capacity for flood control: the field bunds generate a significant capacity of water storage which reduces peak flows under heavy rains. The permanent presence of water on the field also generates percolation of water and groundwater recharge, which very often has benefits for other uses of water.

One major advantage of water-ponding in rice cultivation is the fact that it prevents weed development, thus avoiding the use of herbicides or reducing the labour requirement.

Clearly, switching from paddy to other dry cereals implies making a decision with profound consequences for many aspects of rural livelihood.

The paddy ecological environment is based on the peculiarity of water submergence and anaerobic root zones. Submerged rice soils create environments congenial to organic matter build-up, carbon sequestration and nitrogen fixation. Rice ecosystems are important contributors to the long-term sustainability of organic resources and nutrients supplying power to soils. However, anaerobic conditions also favour the release of methane into the atmosphere, which constitutes a global ecological issue related to the greenhouse effect. The traditional irrigated paddy fields are one of the main methane contributors, but improvements in field techniques can drastically reduce emission by drying the field and aerating the soil (intermittent irrigation, drainage once per season). Varietal improvements can also enhance this effect.

SOCIAL CAPITAL AND RICE WATER MANAGEMENT

Historically, rice cultivation is the result of a collective enterprise: investment in and management of rice requires and has stimulated the development of social capital. Investment in the infrastructure and shaping of the landscape for the ponding system (terraces) require collective organization within the community. Water management also relies on a high sense of collective interest: crop and water calendars must be organized for each large block of fields in order to efficiently manage water and organize the field work (land preparation, transplantation, drying-up for harvesting).

RICE SYSTEMS AS WETLANDS

Many rice systems are classified as wetlands and fall into one of the major categories of the Ramsar classification, namely the subcategory of man-made wetlands (for example, the Kirindi Oya project [see Figure 2]). Investigating the value of water in these rice systems is therefore similar to performing an economic valuation of wetlands (their functions, products and attributes) according to the guidelines developed by IUCN (World Conservation Union) and other conservation agencies (Barbier, Acreman and Knowler, 1996). Rice systems perform a variety of services generally attributed to wetlands and varying according to the type of rice agro-ecology.

Field water management for rice cultivation in rice systems

Water takes on a prominent role in rice production. Unlike many other cropping systems - where water is mainly used for a productive purpose (transpiration) - the rice-cropping system uses water in numerous ways, both beneficially and non-beneficially. Estimates of the water needs for productive and beneficial use in rice systems can be divided into three categories:

1. crop water requirement for evapotranspiration (evaporation and transpiration);
2. seepage and percolation; and
3. requirement for specific water management practices, such as land preparation and drainage of water prior to the tillering stage.

The daily dry-season evapotranspiration rate is between 5 and 12 mm per day (500-1 200 mm on the basis of 100 irrigation days). Evapotranspiration needs comprise:

• first, transpiration for maintaining physiological processes that lead to plant development and growth; and

• second, water to compensate for evaporation from the soil or water layer.

Land preparation needs are in the order of 150 to 250 mm per season.

Seepage and percolation needs of 2 to 7 mm per day are required if moisture is to be kept at saturation level (200-700 mm) or if submerged conditions are to be maintained. For drainage prior to the tillering stage (panicle initiation), another 50 to 100 mm must be added. Based on these calculations, the total water requirements of irrigated rice range from 900 to 2 250 mm per season.

The actual water demand of farmers is often much higher in order to account for conventional application efficiencies of less than 50 percent.

The importance of the multiple uses of water in a rice-based system

In many rice-based systems, a great part of the water enters the field as precipitation or surface irrigation, or it spills or percolates from adjacent fields. This is the well-known "cascade" system, which applies at field and subsystem levels. It is therefore critical that a water budget be made on an appropriate scale and not limited to field level. In some systems in Asia, water consumption at field level can go hand in hand with many other important uses of water (see Figure 2): crop consumption (evapotranspiration at field level) accounts for only 23 percent of the total inflow in the command areas, while fisheries take advantage of the multistreams and water bodies in the system, accounting for about 19 percent of the economic activities of rice.

The importance of recirculation in rice-based systems

Water circulation in many rice systems is characterized by a lot of recirculation at various levels: from field to field or from subcommand area to subcommand area. The inference of poor efficiency is often due to a confusion of scales and to a misunderstanding of water circulation (i.e. multiplying conveyance losses at various levels by field losses to estimate overall efficiency, instead of adopting a proper water balance approach with proper spatial and time boundaries, which may lead to much higher values for efficiency itself).

In traditional rice systems, a lot of the services are provided by this recirculation or storage and circulation of losses. In numerous modern systems, these services are generated accidentally or are a result of poor management; they may sometimes even lead to negative rather than positive externalities (e.g. rise of salted ground-water).

Lessons from rice-based systems for water management improvement

In the irrigation domain, improving management performance has for a long time been sought through water-efficiency improvements: mobilizing, conveying and distributing water and ultimately making water available to crops.

TABLE 1
Estimated seasonal consumptive use of water for evapotranspiration and other uses

EFFICIENCY VS PRODUCTIVITY

Irrigation system performance has mostly focused on efficiency in water deliveries and application, the objective being to minimize the cost borne and the water losses generated by water transport and application. The formula below captures rationale of efficiency:

However, using only the indicator with the above formula for the rice-based system has led to many misleading judgments; in the past, paddy field cultivation has often been wrongly accused of very poor water management performance on the basis of low efficiency at field level. To understand the performance of rice-based systems, it is necessary to take into account not only the crops. The focus should be more on productivity, accounting for beneficial outputs (Molden, 1997), as expressed in the following formula:

For example, water duties in Asia are often computed as the sum of the precipitation and the total irrigation supply divided by the net command area (irrigated fields); it is very common in the humid tropics to record very high water duties (>4 000 mm/year). This high water-input value would lead to a very low performance indicator in terms of efficiency. However, when it is considered that the entire gross command area of a rice-based system benefits from water through crops and other uses, real performance shows a high increase. In the Kirindi Oya project (southeast Sri Lanka) the net command area is only 40 percent of the gross command area. Accordingly, water duties (when referring to the gross command area) are more realistic (1 360 mm/year in this project) (Renault, Hemakumara and Molden, 2001).

FIGURE 1
Embedded system approach for tackling rice-based systems

Water management

Whether applied to traditional systems or to new technical paddy systems, improved water management is badly required to supply users with an appropriate and reliable water service, particularly those who are practising new water-saving techniques and promoting stakeholder governance arrangements and economic and environmentally sustainable management at scheme level.

DIRECTIONS FOR THE FUTURE DEVELOPMENT OF RICE SYSTEMS

At field level: rice emancipation from water-ponding or rice an ordinary cereal

Traditionally, rice has been largely developed in monsoon Asia in areas and at a time when water was quite abundant. The more recent developments of rice cultivation have occurred more in areas or in periods of the year which are much dryer than in the past. Furthermore, water abundance in some parts of monsoon Asia is a concept that has been eroded with the huge increase of water requirements for other uses (domestic-industrial-environmental). Therefore, there is a general trend towards the development of water-saving irrigation techniques for rice.

TABLE 2
Opportunities and challenges related to drier rice cultivation

Besides the development of new high-yielding varieties with classical field practices, there are many attempts worldwide to experiment with new types of rice practices at field level, many of which are specifically motivated by water saving. Water consumption for paddy rice is higher than for any other cereal, even though a great part of the field losses are recycled. The increasingly acute competition for water leads many rice-growing countries to search for ways to reduce water consumption per grain. During recent decades, various new techniques of rice have been tested by international and national rice institutes. The following techniques tend to suppress, partially or totally, water-ponding at field level: aerobic rice, alternate wet and dry, and system rice intensification.

These new techniques revolutionize the age-old idea that rice is an aquatic crop - an idea which stems from the fact that rice can develop well in water, giving rice a serious advantage over weed control. Recent developments have demonstrated that rice can also be grown in dry soils like a classic cereal. Less water-consuming techniques are much more sensitive to water stress, however, and they strongly depend on a reliable water supply during the wet season (to compensate for the dry spell) as well as during the dry season itself. This can only be achieved by an irrigated infrastructure.

If these techniques confirm their potential for improving water productivity (more grain per drop), then rice will become much more water-efficient, but in some instances to the detriment of other uses of water in the command area. Therefore, the choice between increasing crop-water efficiency and maintaining water productivity of various uses must be properly taken into consideration.

Environmental protection and improvement

There are a growing number of environmental concerns which challenge sustainable rice production. Indiscriminate use of pesticides and inefficient use of fertilizers, clearing of flooded forests and destructive fishing gear all need to be confronted, as do the emissions of carbon dioxide, methane nitrous oxide and ammonia. Not only is environmental protection of increasing public concern, efforts to increase protection must also comply with a growing number of international agreements such as those negotiated in the World Trade Organization (WTO), in the United Nations Framework Convention on Climate Change and in the Convention on Biological Diversity. Rice development will only be sustainable if approached as an ecosystem that provides the habitat with a variety of organisms (e.g. fish and insects) often used by indigenous people. Yet these important rice-based resources are generally undervalued by governments and the international community; though the consideration of all rice-based resources presents an important future opportunity to alleviate rural malnutrition and poverty.

Adequate approaches have to be taken to protect the environment and the people which depend directly on it.

Paths for water management improvements in rice-based systems

Discovering and recognizing the diversified values of rice ecosystems is no reason for complacency. There is room for improving water management in both technical and economic terms. In fact, this is a very important opportunity for most rice-based systems to find various ways of valuing water, to come up with acceptable solutions for bearing the cost of the operation and maintenance of water management.

As mentioned previously, however, the systems of rice cultivation are very diversified and thus do not require the same path of modernization.

In simple terms, a modern water management service is one adapted to users' demands and to their willingness to pay. More fundamentally, the thinking process behind modern water management involves defining the management objectives as a basis for defining operation requirements and for designing both infrastructure and management systems, adopting the following approach:

• First, acknowledge or identify the various functions or services that the rice system already provides or is expected to provide.

• Second, improve the capacity of stakeholders in: valuing and planning techniques; defining the management objectives related to the various roles and values of water; setting water management strategies; operating rules; management arrangements (including contributions by various stake-holders); and design features.

In a sense, the recognition of these different values of water in the rice systems is a requirement for moving towards integrated water resources management.

In large-holding modern systems (as there are in developed countries), the general path for modernization is to become very efficient in providing water to crops only, i.e. minimizing other uses of water and focusing on yield increases. The crop and the farmers bear the responsibility for and cost of maintaining and operating the water management infrastructure, and part of the productivity gains and yield increases go towards the management costs. However, modernization of the systems is often also motivated by environmental values and the application of stricter environmental standards. In irrigation districts in California, for example, modernization is motivated by the need to keep water ponded in the fields for a sufficient length of time to permit the degradation of chemical pesticides. At the same time, the value of storage ponds and drainage discharges as habitat for wildlife is now well recognized and systems are therefore requested to maintain them.

In large modern systems in developing countries, there are also opportunities for enhancing the multiple roles of water in rice systems by engineering them into system design and operation. For example, the typical design of irrigation schemes can be improved by the introduction of intermediate storage (generally ignored in design standards) in the form of ponds at one or various levels in the system: this, as shown by the "melon on the grapevine" concept in China, provides opportunities not only to significantly improve reliability and flexibility of the systems as well as decentralize their management, but also for fisheries, habitat etc. Another example is the introduction of constructed wetlands at the drainage outlets both to improve water quality and to provide habitat.

In traditional smallholding rice systems presenting many other uses of water within the command area, the path for modernization would be to consider all uses and all users of water, and promote a much more efficient service of water with costs shared among different categories of users (Renault and Montginoul, 2003). This is an important opportunity for traditional systems which most of the time cannot be economically sustainable if relying on farmers to pay for the operating costs. Again, improvements from considering diversified uses go along with improving the productivity of each use (e.g. increased rice and fish yields). A good compromise must be found so as not to undermine the gains recorded for one use by adverse side effects on other uses (pesticide leaching impacting fisheries, for example).

There is also an opportunity to engineer into these systems measures that enhance the multiple roles or mitigate the impact of the intensification of rice systems. For example, it is possible to introduce ponds as dry-season refuges for fish. From an institutional point of view, water user associations (WUAs) (or similar arrangements created by participatory irrigation management reforms) should include all stakeholders (i.e. not only farmers if users are a different group) and specify in their statutes the various roles of the system. In Japan, for example, many land development districts (the equivalent of WUAs) are evolving towards service providers for integrated water resources management.

For traditional systems also, it can be noted that evolution towards the conjunctive use of shallow ground-water during the dry season appears to both preserve the desirable features of the traditional systems and allow for intensification and diversification of crops.

CONCLUSIONS

Rice-based systems highlight the fact that performance is about productivity and efficiency; one criterion is not sufficient to capture the richness of complex agro-ecosystems. The diversified values of water in rice-based systems are a critical asset for many systems to improve performance and promote economically sound management as well as to account for negative externalities.

Rice cultivation and water must be tackled with a system approach, diverse systems leading to diverse options for the modernization of management; rice systems must be properly integrated into land and water resources management.

Finally, it is very important to revise concepts of rice systems management - not only for existing command areas but also for the development of new systems in other contexts, say for Africa. This is the only way to value the full potential of the agroclimatic context and to prevent disappointments resulting from technical as well as economic project failures.

REFERENCES

Barbier, E.B., Acreman, M.C. & Knowler, D. 1996. Economic valuation of wetlands: a guide for policy makers and planners. Gland, Switzerland, Ramsar Convention Bureau.

Molden, D. 1997. Accounting for water use and productivity. SWIM Paper 1. Colombo, Sri Lanka, IWMI.

Renault, D. & Montginoul, M. 2003 Positive externalities and water service management in rice based irrigation systems of the humid tropics. Agric. Water Management, 59: 171-189.

Renault, D., Hemakumara, M.H. & Molden, D.W. 2001. Importance of water consumption by perennial vegetation in irrigated areas of the humid tropics: evidence from Sri Lanka. Agric. Water Management, 46(3): 201-213.