International Water Management Institute. 2000. World water supply and demand. Colombo, Sri Lanka: International Water Management Institute. We are grateful to the Consultative Group on International Agricultural Research (CGIAR) and the World Water Commission for providing funding from the World Bank and other donors for this research. Core funding to IWMI is provided by the Governments of Australia, Canada, China, Denmark, France, Germany, the Netherlands, and the United States of America; and by the Ford Foundation and the World Bank. This monograph has been prepared as part of a long-term research program in IWMI and as a contribution to the World Water Vision of the World Water Commission (WWC 2000). / water resources / water supply / water demand / water storage / irrigation water / water use efficiency / models / data collection / analysis / uncertainties / water scarcity / productivity / ISBN: 92-9090-400-3 Please direct inquiries and comments to: International Water Management Institute P. O. Box 2075 Colombo Sri Lanka Copyright © 2000 by IWMI. Responsibility for the contents of this publication rests with the authors. All rights reserved. The International Irrigation Management Institute, one of sixteen centers supported by the Consultative Group on International Agricultural Research (CGIAR), was incorporated by an Act of Parliament in Sri Lanka. The Act is currently under amendment to read as International Water Management Institute.

World Water Vision: Its Origin and Purpose Over the past decades it has become gradually evident for those directly involved that there is a chronic, pernicious crisis in the water world. The participants in the first World Water Forum in Marrakech in 1997 called for a World Water Vision to increase awareness of the water crises throughout the population and develop a widely shared view of how to bring about sustainable use and development of water resources. The WORLD WATER VISION draws on accumulated experience of the water sector, particularly through sector visions and consultation for Water for People (or Vision 21), Water for Food and Rural Development, Water and Nature, and Water in Rivers. It draws on the contributions of regional groups of professionals and stakeholders from different subsectors that have developed integrated regional Visions through regional and national consultations in more that 15 regions. As the Vision developed and evolved, more and more networks of civil society groups, NGOs, women, and environmental groups joined the consultations to contribute to the document. The participatory process that led to the WORLD WATER VISION makes it special. Since 1998, about 15,000 women and men at local, district, national, regional and global levels have shared their aspirations and developed strategies for practical action towards sustainable use and development of water resources. The recent availability of Internet communications made such a consultation possible in the short time frame. This is not an academic exercise. It is the start of a movement. Over the coming months and years stakeholders will develop action plans to implement the recommendations of the World Water Commission and the strategies presented herein. The WORLD WATER VISION aspires to be an inspiration to women and men to overcome obstacles and achieve fundamental changes. Its message is for everybody, particularly for the leaders and professionals who have the power and knowledge to help people to turn visions into reality. It challenges those directly impacted by the water crisis to initiate action on their own and to call on their leaders to bring about sustainable water resources use and development. The Vision recognizes that if sustainable water resources use and development are to be achieved, people’s roles must change. The main actors will be individuals and groups in households and communities with new responsibilities for their use of water and water-related services, as part of a collective strategy. Public authorities will need to empower and support them, and carry out the work that households and communities cannot manage for themselves. Water-sector professionals and environmentalists will provide these stakeholders with the information they need to participate in decision making and help implement their decisions. All these groups working together can achieve this Vision.

1

CHAPTER 1

Introduction

David Seckler Over the past three years, the International Water Management Institute (IWMI) has been developing scenarios of water supply and demand for 2025. Since the first report on this subject (Seckler et al. 1998), the analysis and data have been refined through the development of PODIUM, the Policy Interactive Dialogue Model. PODIUM provides a user-friendly means for policy makers, scientists and others to interact in developing alternative water scenarios for countries in a systematic framework of data and analysis. PODIUM is produced in two versions: the country model, which is used for the study of individual countries in a dialogue mode and the global model, which is a spreadsheet used for aggregating countries into global or regional groups. As the results from the country models become available from dialogue with national experts, they are entered into the global model. The global model discussed here includes 45 countries that represent the major regions of the world and over 80 percent of its population. Thus these countries are a good sample of the world as a whole. In addition, a less detailed analysis of 80 countries has been made and the results used in the preparation of the map (figure 1). It is important to understand that PODIUM differs from many other models in that it does not make predictions, strictly speaking. Rather, it is designed to explore the technical, social and economic aspects of alternative visions of the future. The essence of PODIUM is captured in the above quotation from Walter Lippman. Ideally, the country visions would be specified by policy makers and other representative of the countries, using PODIUM as a basis for dialogue. In fact, PODIUM now is being widely used for this purpose and IWMI is encouraging this use by making it freely available and conducting training courses in its use. However, to estimate the world’s water supply and demand situation in 2025, we have had to assume country visions, based on a set of values and objectives that most people (we believe and hope) share. The resulting vision and analysis presented in this paper constitute what we call the basic scenario. The basic scenario can perhaps best be explained by saying that if countries achieved these results in 2025 they could, in our opinion, feel satisfied that they have done about as well as possible given their circumstances. Once the basic scenario is established, other scenarios can easily be generated. We have used it, for example, to create IWMI’s contribution to the three major scenarios of the World Water Council’s Vision 2025 program (Cosgrove and Rijsberman forthcoming). One of the most important factors to keep in mind in reading this or similar studies is the large amount of uncertainty involved. To illustrate this problem, we will mention one example where the degree of uncertainty can be quantified. The actual evapotranspiration of crops (Eta) is one of the most important factors in all studies of agriculture, irrigation and water resources. It is a purely physical phenomenon, carefully defined by an elaborate set of equations. Yet using the most sophisticated measuring instruments available, the mean Eta of a crop can only be measured (at 95% confidence) within a range of plus or minus 25 percent (Richard Allen 1999 personal communication)!

2

Figure 1. Projected water scarcity.

3

Given the intrinsic uncertainty of the data, the complete lack of reliable data on key factors discussed in this paper and the general propensity of human beings to underachieve or overachieve any target, no matter how reasonable the target may seem to be, any study of this sort would be fortunate if its scenarios for 2025 were accurate within plus or minus 25 percent—and, in some cases the error could be even larger. Thus, for example, when we say, as we do, that world irrigation will require 17 percent more water supply in 2025 we mean that it is not likely to need less than 13 percent more water, nor more than 23 percent more water—and even this range could be too small. The problem, of course, is that nobody knows what the range actually is and it will take several years of intensive data collection and analysis to obtain better estimates. While we do not bother the reader with confidence intervals in the tables and prose, we do try to give an idea of the magnitude of uncertainty surrounding the various numbers. In the future, we hope to reduce the amount of uncertainty by collaboration with national experts. We have already started this process in collaboration with the International Commission on Irrigation and Drainage. As a result of this collaboration, we have adjusted the internationally available data on India, Pakistan, Mexico, and Morocco. We are still in a process of dialogue with national and international experts about China, where there is considerable national and international debate over some of the basic data on agriculture and water. In the meantime, we use international data and IWMI estimates for China. This monograph is divided into six chapters that have been written by a closely interacting team of IWMI staff and Senior Associates. While all the members of the team have made contributions to all of the chapters, we have assigned authorship for each chapter to acknowledge the distribution of the work and responsibility. Since the research underlying this monograph has resulted in several concepts and terms that require definition we have included a glossary as an appendix. Then, the discussion proceeds through the following chapters: • Chapter 2 provides a map (figure 1) and a set of tables, with a brief discussion of each one,

that describes the major data and findings. • Chapter 3 briefly outlines some of the major issues that need to be understood in

understanding the results of the analysis. • Chapter 4 discusses some of the major concepts, technologies and management techniques to

be considered in increasing the effective efficiency and productivity of water used in irrigation.

• Chapter 5 presents a description of the PODIUM model. • Chapter 6 discusses water scarcity and the role of storage in development. Before closing this introduction we would also like to express our appreciation to the World Water Commission and to William J. Cosgrove and Frank Rijsberman of its staff for affording us the opportunity to participate in the important and exciting project of the World Water Vision. We have especially gained through interaction with other researchers in this project.

4

3

CHAPTER 2

Overview of the Data and Analysis

David Seckler, David Molden, Upali Amarasinghe and Charlotte de Fraiture Introduction In this chapter we first provide a brief overview of the major results of the research and then discuss some of the major issues underlying these results. The PODIUM model is designed to simulate alternative scenarios of the future. The results presented here are based on what we call the basic scenario. The basic scenario is rather optimistic. Within an overall framework of social, technical and economic feasibility, it relies on substantial investments and changes in policies, institutions and management systems intended to achieve four major objectives: • Achieve an adequate level of per capita food consumption, partly through increased

irrigation, to substantially reduce malnutrition and the most extreme forms of poverty. • Provide sufficient water to the domestic and industrial sectors to meet basic needs and

economic demands for water in 2025. • Increase food security and rural income in countries where a large percentage of poor people

depend on agriculture for their livelihoods through agricultural development and protection from excessive (and often highly subsidized) agricultural imports.

• Introduce and enforce strong policies and programs to increase water quality and support environmental uses of water.

Realizing these objectives requires three major actions in the field of water resources and irrigation management in water-scarce countries: • Greatly increase the productivity of water resources use. • After productivity is increased, there generally remains a need for substantial increases in the

amount of developed water supplies. • Water resources development must be done with substantially reduced social and

environmental costs than in the past—and people must be willing to pay the increased financial costs this policy necessarily entails.

Water Scarcity As shown in figure 1, we have grouped the forty-five countries into three basic categories of water scarcity. • Group I represents countries that face physical water scarcity in 2025. This means that, even

with the highest feasible efficiency and productivity of water use, these countries do not have sufficient water resources to meet their agricultural, domestic, industrial and environmental needs in 2025. Indeed, many of these countries cannot even meet their present needs. The

4

only options available for these countries are to invest in expensive desalinization plants and/or reduce the amount of water used in agriculture, transfer it to the other sectors and import more food.

• Group II represents countries that face economic water scarcity in 2025. These countries

have sufficient water resources to meet 2025 needs but which will have to increase water supplies through additional storage, conveyance and regulation systems by 25 percent or more over 1995 levels to meet their 2025 needs. Many of these countries face severe financial and development capacity problems in meeting their water needs.

• Group III consists of countries that have no physical water scarcity and that will need to

develop less than 25 percent more water supplies to meet their 2025 needs. In most cases, this will not pose a substantial problem for them. In fact, several countries in this group could actually decrease their 2025 water supplies from 1995 levels because of increased water productivity.

• The crosshatched countries on this map are countries that are projected to import over 10

percent of their total cereal consumption in 2025. The correlation between this set of countries and Group I is clear; the more mixed relationships for the other groups are discussed in chapter 3.

The PODIUM model operates at the country level. Therefore, it generally ignores the substantial differences in water scarcity within countries, at the levels of regions or river basins. For example, about one-half of the population of China lives in the wet region of southern China, mainly in the Yangtse basin, while the other one-half lives in the arid north, mainly in the Yellow river basin. This is also true of India, where about one-half of the population lives in the arid northwest and southeast, while the remaining one-half lives in fairly wet areas. Much the same is true of many other countries. A particularly vivid example is Mauritania, which is mostly desert, but falls in Group II. The reason is that 90 percent of the total population lives along the southern border, along the Senegal River. This geographic issue is discussed in more detail later; for now, it is sufficient to say that we have ignored regional differences in the group classifications for all countries except India and China—because of their huge size in terms of population and water use. Figure1 shows a rough picture of the regional differences in these two countries. In the future, we plan to make further regional distinctions for countries, like Mexico, that have large regional disparities in water and agriculture. Interested readers should consult the maps and analysis in Alcamo, Henrichs, and Rosch 1999, who have made separate estimates of water scarcity for over one thousand major river basins in the world. Data and Results There is an enormous amount of data and calculations in the PODIUM global model. A full printout of the data sheets would probably be three or four meters long. In the confident belief that many readers do not want to wade through this amount of numbers, we have restricted the presentation here to only eight basic tables. To simplify the tables and discussion the following conventions are used:

5

• All the percentages are compound rates. • The “Period” is 1995-2025. • After table 1, the countries are listed alphabetically for easy reference, and

∗ for countries in Group I, physical scarcity is shown in bold. ∗ for countries in Group II, economic scarcity is shown as underlined (some countries may

be both in this format) ∗ countries in Group III, with little or no scarcity, are unmarked.

Table 1 Table 1a provides the 1995 population data for the forty-five countries and the world, together with the United Nations Medium and Low projections for 2025. There are substantial differences between these projections. The Medium one projects a 38 percent increase in population over the period, to 7.8 billion people in 2025, whereas the Low one projects an increase of 28 percent, to 7.3 billion. For reasons explained in Part II, we believe that the UN Low projection is the most likely one. However, we have used the average Medium and Low projections for 2025. Including one-half of the population of India and China in each of Group I and Group II, it is seen that by 2025. • forty-five countries, with 33 percent of their population, will be Group 1, with physical water

scarcity. • countries with 45 percent of their population will have substantially underdeveloped water

resources, requiring 25 percent or more development of additional water supplies. • countries with 22 percent of the population, mainly developed countries, will have little or no

water scarcity. Together, Groups I and II contain of 78 percent of the population in 2025. Of course, this does not mean that everyone in these countries will directly be experiencing water scarcity. As usual, the economically better-off members of most countries will have enough water and food, while poor and weak people will suffer the major part of the burden. An overview of the water scarcity situation for the forty-five countries is shown in table 1b. The definition of the terms is included in the Glossary, but a brief description is included here. • First, we take data on the Renewable Water Resources (RWR) of countries from various

sources (mainly Shiklimanov 1999, WRI 1998 and national estimates). RWR is essentially the average annual runoff of past and present precipitation.

• Second, we assume a potential utilization factor for RWR for each country. This factor depends mainly on the seasonal and interannual variability of precipitation and potential storage facilities. As shown in the table, the weighted average utilization factor is only 36 percent. But this is primarily due to the influence of large rivers in Bangladesh, Brazil and monsoon floods in countries such as India. Most of the countries have utilization factors of around 60 percent and dry countries up to 85 percent.

• Third, multiplying the RWR by the utilization factor results in the (potentially) utilizable water supply (UWS) of a country. UWS is the amount of water that can ultimately be utilized, with full development of surface and subsurface storage and conveyance facilities. In principle, this value includes the sustainable use of groundwater supplies (which is

6

essentially the amount of natural and artificial recharge of aquifers), but there is very little data on this.

• Fourth, we consider the primary water supply (PWS). This is the annual average amount of UWS that is presently being utilized through water storage and conveyance facilities. The primary water supply differs from water diversions to users (or “withdrawals”) in that diversions include both PWS and the amount of PWS that is recycled through various uses in the system. Thus the total amount of diversions can be much larger than PWS. This extremely important distinction between PWS and diversions, first emphasized in the preceding analysis by IWMI (Seckler et al. 1998) has generally been neglected in other studies of water supply and demand—often with substantial errors as a result.

An important constraint on the analysis, which is not shown in this table, should also be mentioned. We have assumed that total evaporation cannot exceed 75 percent of PWS. This constraint essentially leaves 25 percent of PWS to cover evaporation from storage and conveyance systems and for environmental benefits such as provisions of environmental supplies to estuaries and coastal areas plus flushing of salts and other pollutants. Some people believe this figure is too low, that more of this expensive water should be utilized. But we have used the 1995 percentage value for Israel, of 75 percent, as the “Gold Standard” for this constraint. It would be generally agreed (except perhaps among Israelis!) that Israel is in many ways the most careful and efficient (and high-cost) manager of water in the world. If Israel cannot exceed 75 percent, who can? Now, with these basic water concepts in mind, we turn to the Water scarcity criteria of this table. • The first criterion (A), the Degree of Development, defines the countries in Group I, a state of

physical water scarcity: this is simply the percentage of PWS to UWS. Again, partly using Israel as a gold standard, we assume that when this ratio exceeds 60 percent, the country is experiencing physical water scarcity. It is notable that some countries in this group exceed this percentage by several times. This is presumably because of unsustainable groundwater mining. There is no reliable international data on groundwater resources—in terms of amounts, depletions, or natural and artificial recharge. Thus we simply do not know how long it will be before the well runs dry, as it were and the world comes crashing down on these highly depleting countries. As will be seen later, we have substantially reduced the rate of irrigation development to zero or negative levels (as indicated in criterion B) and the growth of crop yields for several countries in this group.

• The second criterion (B), the Growth of PWS, defines the countries in Group II, a state of economic water scarcity. This is the rate at which PWS must be developed over the 1995-2025 period to attain the objectives of the base scenario. We believe that a rate of growth of PWS equal to or greater than 25 percent over this period would be stressful in terms of the financial and development capacity for most countries. If this is true, then all of the countries in Group II will require extraordinary efforts to achieve the objectives of the base scenario. (And many of the countries in Group I can do so only by large cereal imports.)

The last column of this table shows the projected net exports of cereals (net imports are shown as negative values) as a percentage of total cereal consumption in 2025 (for details see table 6). With the sole and highly fortunate exception of South Africa, all of the countries of

7

Group I will be forced to import a large percentage of their cereal consumption, largely because of water scarcity. Group II includes both importers and exporters. Some are importers because they have not developed their UWS for irrigation in the past. Many of the largest exporters in Group III, such as the USA, Canada and France have exceptionally favorable conditions for rain-fed agriculture—along with irrigation in the case of the USA. Also, some of the major importers in these groups, such as Japan, are so partly because of land constraints and partly because of the economic comparative advantages of specializing labor (and land) in other, non-agricultural, sectors of the economy. Table 2 This table summarizes the water picture for the countries in terms of the three major sectors of irrigation, domestic and the industrial users of water. Unfortunately, no quantitative information is available on the water requirements of the important environmental sector. All we know is that it is large and growing rapidly. We have tried to handle the requirements of the environmental sector through “commitments” of water outflow above and beyond the water uses of the three other sectors; but that is obviously subject to revision with further research. For the forty-five countries in 2025. • Total diversions for the three sectors are 4,124 km3, increasing by 29 percent over 1995. • Total PWS is considerably lower, at 2,718 km3, increasing by 22 percent. The difference

between these two figures is due to the recycling and reuse of PWS in meeting needs for diversions.

• It should be noted that several countries can reduce total diversions over the 1995-2025 period if they wish to do so, but PWS cannot be reduced. Total 2025 diversions include these reductions, which is appropriate for global sums—but perhaps the percentage changes should exclude these reductions since reductions in, say, Japan do not help Pakistan meet its diversion requirements. However, this adjustment is not done here.

• China, India and the USA are by far the greatest water users. Together, they account for (504+477+458=) 1,439 km3 of PWS, or 53 percent of the 2025 total.

• Irrigation is and will remain as, the largest single user of water, accounting for about 69 percent of both diversions and PWS in 2025. (Irrigation diversions and PWS are nearly the same because of the upper limit of 75 percent on the system efficiency of irrigation at the basin or country level, as explained in connection with table 2 mentioned above.) However, irrigation is also the most slowly growing water sector, with both diversions and PWS increasing only 17 percent over the period.

• Domestic uses of water, on the other hand, represent the smallest sector, at only 11 percent of diversions, but they are the most rapidly growing sector, with an increase of 84 percent over the period. The 2025 projections for this sector are explained in Seckler et al. 1998. They are designed to provide a minimum level of water to everyone, as a basic need of 20 M3 per year, as defined by Gleick (1996).

• Industrial diversions of water depend on per capita income projections, as described in Seckler et al. 1998. Industrial diversions comprise 22 percent of total 2025 diversions, growing 60 percent over the period. These estimates may be too high because diversions to industry can be dramatically reduced by internal water treatment and recycling. However, since only 10 percent or so of industrial diversions is lost to evaporation, with most of the

8

balance recycled in the system, reduced diversions have only a small effect on needs for additional PWS.

• It is interesting to note that in per capita terms Canada and the USA have the largest industrial diversions. The reason for this is presumably (we have not been able to find clear documentation) that water used for hydroelectric power is considered a diversion. This is appropriate so long as it is understood that almost none of this water is lost to evaporation in the use itself (excluding evaporation losses from reservoirs and losses to sinks due to mismatches of downstream water supply and demand.)

The world’s water needs in 2025 can be summarized as follows: • To meet increased 2025 needs, the world’s primary water supply will need to increase by 22

percent, from around 2,120 km3 per year in 1995 to 2718 km3 per year in 2025. This amounts to an increase in surface and subsurface water storage sufficient to release roughly 600 km3 of water per year. To put this value in perspective, it may be noted that the annual average release of water from one of the largest dams in the world, the High Aswan Dam (HAD) of Egypt, is about 55 km3 per year. Thus the additional storage required is equivalent to about twelve new HADs over the period, or nearly one every 2 years.

• In addition, it is estimated that 1 percent of the existing live storage capacity of the world (6,000 km3) is lost every year because of sedimentation (Keller, Sakthivadivel, and Seckler 2000). If this is true, then an additional 60 km3 of new storage, slightly more than one HAD, will have to be created each year simply to maintain the 1995 base.

• As if this were not enough, perhaps about 10 percent of the world’s total supply of primary water, some 200 km3 per year, is provided through unsustainable overdraft of groundwater resources (Postel 1999). If so, about 200 km3 per year, or four HADs, would be needed to replace this overdraft to achieve sustainable water use.

• The total amount of additional water storage and conveyance required by 2025 thus is around 860 km3 per year. This represents a total requirement in 2025 of nearly 3,000 km3, or effectively a 41 percent increase of PWS. For social, environmental and other reasons, it is hoped that a large part of the additional storage requirement can be met through conjunctive use of groundwater, recharging and drawing down aquifers rather than through construction of surface storage facilities—even at additional cost.

Costs of the various methods of storage and conveyance are discussed in Keller, Sakthivadevel, and Seckler 2000. Unfortunately, the cost per unit of water supply is generally much higher for the alternatives to medium and large reservoirs—whether small tanks, conjunctive use, recycling, or wastewater treatment—than the cost of water supply from these reservoirs. John Briscoe (2000) estimates that “The overall level of investment in water-related infrastructure in developing countries is of the order of $65 billion annually, with the respective shares about $15 billion for hydro, $25 billion for water and sanitation and $25 billion for irrigation and drainage.” We hope to do cost estimates for the basic scenario in the near future. Now that we have shown what the results of the analysis in terms of water are, we shall briefly explain how we got those results. Most of the background discussion of methodology and major issues is in Part II, and here we will concentrate mainly on the results of the data analysis.

9

Table 3 This table shows food consumption in 2025 in terms of calories (kcal). The 1995 data are from FAO 1996. The projections for 2025 are from the IFPRI Impact model 1996 (personal communication). The PODIUM model is designed so that the user can change these projections, but we believe that the IFPRI projections are very good and are used here without alteration. • Total calorie (kcal) consumption per day for all the countries increases from 2,771 kcal in

1995 to 3,037 kcal in 2025, or by about 10 percent. It is generally believed that between 2,700 and 3,200 kcal/capita/day are sufficient, depending on variations in the distribution of income. Thus the 2025 level is sufficient to reduce undernutrition in the world to negligible proportions. The USA, the largest per capita consumer in 1995, even increases by 2025 to 3,888 kcal per capita per day—which may be a dangerous level of consumption. The lowest 1995 consumption is in Ethiopia and will remain below adequate levels in 2025, unless major changes are made.

• The composition of the diet changes over the period, with cereal foods decreasing in importance and animal products and non-cereal foods increasing. There are, however, variations, depending mainly on the per capita income of the countries.

• Total growth of calories, the result of multiplying the per capita value by population growth in table 1, is 41 percent for all the countries. Cereal foods grow only 32 percent, while animal products grow by 58 percent and other foods by 47 percent.

Table 4 This table is partly derived from table 3. It shows consumption of cereals in terms of million metric tons per year. One of the major problems, discussed in Part II, is deriving the demand for feed cereals from the demand for animal products. The conversion ratio, expressed as kg of feed to kg of animal products, is used for this purpose. The 1995 conversion ratios are derived from FAOSTAT Statistical Database (FAO 1997). IWMI’s estimates of changes in conversion ratios over the period are shown in the third column of the table. Animal products in developed countries are mainly derived from intensive feeding systems; in these cases, the conversion ratios will decrease because of technological change. However, in most of the developing countries, a large percentage of the animal feed is from grasses, bushes and trees on wastelands, garbage and insects. Since these feeds are in short supply, conversion ratios for the additional animal products are likely to increase in the future. Combining food and feed cereals, annual cereal consumption is projected to increase by 38 percent by 2025. Table 5 These tables (5a to 5c) summarize the basic production factors of production in agriculture over the period. Since these tables are self-explanatory, only a few observations are made here. • Since the gross irrigated area (the irrigated area that is cropped more than once per year, or

crop intensity) is not reported in the international data, we have made our own 1995 estimates (for some countries, in cooperation with FAO and national experts) for 125 countries.

10

• Next, we assume the growth of net irrigated area and irrigation crop intensity over the period and calculate both for 2025.

• It is seen that net irrigated area increases 22 percent, by 52 million hectares. Gross irrigated area increases 29 percent, by nearly 100 million hectares.

• It is important to understand that the growth of irrigation is for all crops, not just cereal crops. We use cereals as an indicator of the required growth of irrigation for all crops.

• The difference between the 29 percent growth in gross irrigated area here and the 17 percent growth of PWS for irrigation in table 2a shows the substantial increase in irrigation basin efficiency at the country level on the additional irrigated area. Much of the water for the additional area will have to be squeezed out of the 1995 irrigated area through increased basin efficiency at all levels.

• Table 5b shows cereal area. The irrigated area in cereal is projected to grow by 30 percent, compared to a 10 percent reduction in rain-fed cereal area. The reason for this is that total cereal area increases by only 3 percent and therefore most of the growth in net irrigated land is at the expense of rain-fed land. Some countries, however, grow in all three components of cereal area.

• Table 5c shows the yields for cereals, by irrigated and rain-fed areas. Cereal yield for irrigated areas is projected to grow by 40 percent over the period, compared to 12 percent for rain-fed yield.

• Last, because of these differences in growth of areas and yields, there is the remarkable role reversal in contribution to total cereal production. In 1995, irrigation produced 43 percent of the cereals, while rain-fed area produced 57 percent; in 2025 these percentages are exactly reversed.

Table 6 This final table summarizes both the growth of cereal production over the period and production as a percentage of projected cereal consumption (table 4), as a surplus or deficit, in 2025. These surpluses or deficits do not necessarily translate into exports or imports; this depends on policies, prices and other factors outside this analysis. But as an indicator of the 1995 position of countries in this regard, the net exports for the countries are shown in the last two columns. It is important to understand that from a policy standpoint, it is the total balance of trade and the balance of trade in total agricultural products that count. Thus some countries may import more cereals, while exporting more of other agricultural products, while other countries do the reverse. Thus, this table does not give a full picture of the trade situation of countries, even of trade in agricultural products. Nevertheless, since trade in cereals is a very large component of agricultural trade, especially with respect to feed cereals, this table is not without significance. It is seen that most countries can satisfy the cereal requirements of their 2025 populations without major deviations from the 1995 percentage levels. There are, however, a few exceptions to this rule and a few notes to be made. • Altogether, the forty-five countries are net cereal exporters. The amount of net cereal exports

from these countries will decrease from 52 million tonnes in 1995 to 39 million tonnes in 2025. This is primarily a result of lower population growth in the importing countries and the consequent growth in ability to feed themselves—even at slower growth of cereal area and

11

yield than in the past. This is presumably also true for the countries outside this set that are net importers, so the reduced level of net exports should not cause a problem.

• As noted before, in the discussion of table 1, many of the countries that import the most cereal are in group I.

• Perhaps the greatest and most important surprise is the large increase in cereal imports projected for Pakistan. This is due mainly to water scarcity (and salinity), combined with a high rate of population growth.

• A highly controversial issue, noted before, is the future trade situation of China—especially in terms of imports of feed cereals (and/or animal products). Most studies project moderate to massive increases in cereal imports, mainly feed cereals, for China. However, we believe that the combination of low population growth and the creative dynamism of the agricultural economy of China will result in a trade balance for cereals—indeed, perhaps, in a slight surplus.

• Some of the major exporters like Canada and France are projected to reduce cereal exports as a percentage of domestic consumption and slightly in total amounts, while the USA will continue at 1995 levels as a percentage, increasing the amount of net cereal exports by 10 percent or so.

In conclusion, thanks partly to reduced population growth, it is possible to satisfy the world’s food and water requirements in 2025 at reasonable cost and in socially and environmentally acceptable ways through substantial changes in policies, technologies and management systems. If this is done properly, 2025 will witness an entirely new era in human history—an era in which food and water scarcity have virtually been eliminated for everyone! The one black cloud on this otherwise rosy horizon is made by countries that will need large amounts of net food imports, whether because of water scarcity or other reasons, but cannot afford them. Unfortunately, this is likely to be the situation in many of the countries of sub-Saharan Africa.

23

IWMI Base Scenario

Table 1a. Population, water scarcity, and international trade in cereals.

Population Population (UN Med) Percent Avg. UN Low & Medium UN Medium UN Low Country rural population Growth Growth Growth 1995 1995 2025 1995-2025 2025 1995-2025 2025 1995-2025 Million % Million % Million % Million % World 5666 7549 33% 7824 38% 7275 28% All countries (45) 4716 55% 6056 28% 6279 28% 5834 24% % of World 83% 80% 80% 80% Group I 1 1476 14% 2014 36% 2092 42% 1937 31% % of 45 countries 31% 33% 33% 33% Group II 1 2011 24% 2719 35% 2821 40% 2616 30% % of 45 countries 43% 45% 45% 45% Group III 1229 31% 1324 8% 1366 11% 1281 4% % of 45 countries 26% 22% 22% 22% India 934 73% 1273 36% 1330 36% 1216 30% % of 45 countries 20% 21% 17% 17% China 1221 70% 1437 18% 1480 18% 1394 14% % of 45 countries 26% 24% 19% 19%

Group I - Physical water scarcity Saudi Arabia 18 20% 39 112% 40 112% 38 106% Pakistan 136 65% 254 87% 263 87% 246 80% Jordan 6 29% 12 102% 12 102% 11 93% Iran 62 41% 90 44% 94 44% 86 37% Syria 14 48% 25 78% 26 78% 24 71% Tunisia 9 43% 12 38% 13 38% 12 32% Egypt 62 55% 91 47% 96 47% 87 40% Iraq 20 25% 40 98% 41 98% 39 92% Israel 6 9% 7 24% 8 24% 6 0% South Africa 37 49% 44 17% 46 17% 42 12% Algeria 28 66% 45 59% 47 59% 43 53% Group II - Economic water scarcity Brazil 159 22% 208 31% 218 31% 198 25% Turkey 61 31% 84 36% 88 36% 79 29% Mexico 91 25% 125 37% 130 37% 120 32% Philippines 68 46% 104 52% 108 52% 100 47% Thailand 59 80% 70 20% 73 20% 68 16% Ethiopia 55 87% 112 102% 115 102% 108 96% Australia 18 15% 22 23% 23 23% 21 18% Myanmar 43 74% 56 29% 58 29% 53 23% Nigeria 99 61% 179 81% 183 81% 174 76% Bangladesh 119 82% 171 44% 179 44% 163 38% Argentina 35 12% 45 30% 47 30% 43 25% Viet Nam 74 66% 106 43% 108 43% 103 39% Sudan 27 75% 45 69% 46 69% 44 64% Morocco 26 52% 37 42% 39 42% 35 36% Group III - Little or No Water Scarcity Kyrgyzstan 5 30% 8 67% 10 67% 6 22% Canada 30 23% 37 24% 38 24% 35 19% USA 267 24% 315 18% 326 18% 304 14% Indonesia 197 65% 260 32% 273 32% 247 25% Poland 39 35% 39 0% 39 0% 38 -1% Spain 40 24% 36 -8% 37 -8% 36 -9% France 58 27% 60 4% 62 4% 59 1% UK 58 11% 58 0% 60 0% 57 -3% Italy 57 33% 51 -12% 51 -12% 50 -13% Germany 82 13% 79 -4% 80 -4% 77 -5% Uzbekistan 22 34% 32 43% 33 43% 31 37% Turkmenistan 4 36% 6 48% 6 48% 6 43% Romania 23 45% 20 -13% 20 -13% 20 -14% Tajikistan 6 38% 8 47% 9 47% 8 41% Japan 125 22% 119 -5% 121 -5% 116 -7% Kazakhstan 17 20% 17 1% 18 1% 16 -4% Ukraine 51 30% 45 -12% 46 -12% 44 -14% Russian Fed. 148 24% 134 -9% 138 -9% 131 -12% 1. Includes one-half the population of India and China.

24

IWMI Base Scenario

Table 1b. Water resources.

Renewable resources Water scarcity criteria Constraint Int. Trade Renewable Factor Utilizable Primary Water Supply (PWS) Net Export Water Potentially water A. Percent of B. Increase Evaporation % of % of Country Resources Utilizable resources Utilizable Water Supply PWS Primary water supply t. cereal (RWR) RWR of RWR <= 75% cons 1995 2025 1995-2025 1995 2025 2025 km3 % km3 % % % % % % World All countries (45) 34486 36% 12478 18% 22% 22% 64% 68% 2% % of World Group I 1 608 61% 371 97% 104% 7% 68% 73% -12% % of 45 countries Group II 1 14230 25% 3602 7% 12% 64% 62% 66% 0% % of 45 countries Group III 14911 46% 6920 12% 13% 7% 57% 60% 15% % of 45 countries India 2037 38% 774 49% 62% 25% 73% 75% -1% % of 45 countries China 2700 30% 810 44% 62% 40% 70% 75% 3% % of 45 countries

Group I - Physical water scarcity Saudi Arabia 2 80% 2 678% 666% -2% 67% 60% -82% Pakistan 226 55% 124 141% 156% 11% 70% 74% -25% Jordan 1 80% 1 113% 113% 0% 63% 75% -97% Iran 138 60% 83 91% 80% -12% 62% 72% -23% Syria 26 60% 16 65% 80% 23% 70% 75% -13% Tunisia 4 60% 2 72% 77% 6% 68% 74% -65% Egypt 69 85% 58 66% 75% 15% 72% 75% -51% Iraq 75 60% 45 64% 67% 5% 72% 74% -42% Israel 2 80% 2 61% 64% 5% 75% 76% -95% South Africa 50 60% 30 47% 63% 33% 60% 65% 9% Algeria 14 60% 9 32% 61% 88% 58% 63% -79%

Group II - Economic water scarcity Brazil 8120 10% 812 3% 7% 98% 52% 60% -28% Turkey 215 60% 129 21% 41% 94% 66% 63% 0% Mexico 348 60% 209 28% 50% 82% 65% 74% -6% Philippines 320 60% 192 11% 19% 74% 46% 45% -31% Thailand 318 30% 95 17% 29% 74% 76% 71% 14% Ethiopia 88 60% 53 2% 3% 49% 71% 75% -31% Australia 352 60% 211 12% 17% 47% 57% 64% 198% Myanmar 807 60% 484 0% 1% 43% 68% 73% 11% Nigeria 319 30% 96 5% 7% 40% 60% 66% -11% Bangladesh 1390 20% 278 7% 10% 39% 63% 72% -5% Argentina 893 60% 536 3% 4% 35% 69% 71% 84% Viet Nam 866 45% 390 6% 7% 32% 56% 58% 21% Sudan 165 60% 99 16% 20% 26% 69% 74% -17% Morocco 30 65% 20 41% 51% 25% 68% 74% -27%

Group III - Little or No Water Scarcity Kyrgyzstan 49 80% 39 24% 29% 23% 68% 75% -25% Canada 3420 20% 684 6% 8% 22% 50% 50% 59% USA 3048 60% 1829 22% 25% 16% 60% 61% 36% Indonesia 2080 30% 624 6% 6% 15% 62% 67% -4% Poland 56 60% 33 10% 11% 13% 41% 47% 9% Spain 94 60% 57 50% 52% 4% 58% 59% -9% France 195 60% 117 29% 30% 3% 54% 54% 53% UK 108 60% 65 15% 15% 0% 44% 44% 7% Italy 160 60% 96 37% 37% 0% 55% 61% -12% Germany 169 60% 101 40% 38% -3% 50% 50% 20% Uzbekistan 104 80% 83 42% 39% -6% 68% 74% -42% Turkmenistan 71 80% 57 20% 18% -7% 69% 74% -26% Romania 204 60% 123 16% 14% -8% 64% 71% 29% Tajikistan 94 80% 75 8% 8% -8% 68% 75% -72% Japan 450 33% 147 41% 37% -10% 42% 43% -77% Kazakhstan 124 80% 99 18% 16% -12% 64% 73% 8% Ukraine 210 60% 126 19% 16% -14% 56% 64% 4% Russian Fed. 4275 60% 2565 2% 2% -15% 55% 62% -3%

25

IWMI Base Scenario

Table 2a. Summary water use by sectors.

All Sectors Irrigation

Calc. Calc. Calc. Calc

Water Scarcity Total Change Total Change Total % Primary Change % Physical = I Diversions Primary irrigation Total water total water diversions diversions supply PWS

Economic = II supply 2025 Period1 2025 Period1 2025 2025 2025 Period1 2025 km3 % km3 % km3 % km3 % %

World

All countries (45) 4,124 29% 2,718 22% 2,786 68% 1,872 17% 69% % of the world

Algeria 7.6 90% 5.2 88% 4.3 57% 3.3 103% 63% Argentina 34.2 37% 22.9 35% 19.2 56% 13.6 14% 59% Australia 39.6 45% 35.8 47% 27.2 69% 23.4 63% 65%

Bangladesh 35.7 48% 29.0 39% 33.1 93% 27.3 35% 94% Brazil 102.4 113% 55.9 98% 60.2 59% 36.7 117% 66%

Canada 53.0 21% 52.5 22% 2.2 4% 1.8 -8% 3%

China 917.4 58% 503.7 40% 627.9 68% 405.8 25% 81% Egypt 65.2 18% 43.8 15% 47.9 74% 37.0 7% 85% Ethiopia PDR 2.4 89% 1.4 49% 1.2 50% 1.0 22% 72%

France 38.4 3% 34.6 3% 6.4 17% 5.6 0% 16%

Germany 42.4 -3% 38.8 -3% 1.4 3% 1.3 0% 3% India 811.4 31% 476.9 25% 702.0 87% 432.7 18% 91%

Indonesia 81.9 26% 39.9 15% 67.6 83% 33.9 3% 85% Iran, Islamic Rep of 93.5 -3% 66.2 -12% 83.2 89% 62.6 -14% 95% Iraq 38.3 4% 30.3 5% 34.5 90% 28.8 5% 95%

Israel 1.6 9% 1.1 5% 1.1 68% 0.8 -4% 71% Italy 42.7 -3% 35.6 0% 22.3 52% 19.3 12% 54% Japan 90.2 -17% 54.6 -10% 47.1 52% 14.4 -22% 26%

Jordan 1.4 34% 0.8 0% 0.7 49% 0.6 -21% 68% Kazakhstan 23.6 -6% 16.0 -12% 17.2 73% 13.8 -13% 86%

Kyrgyzstan 14.3 25% 11.3 23% 13.2 93% 11.0 23% 97%

Mexico 137.4 91% 104.2 82% 115.3 84% 92.0 81% 88% Morocco 15.6 36% 9.9 25% 13.5 87% 8.7 18% 88% Myanmar 4.2 57% 3.0 43% 3.2 77% 2.6 35% 86%

Nigeria 13.6 65% 6.9 40% 8.3 61% 5.1 16% 73%

Pakistan 293.0 15% 193.5 11% 275.0 94% 186.4 8% 96% Philippines 68.9 102% 37.2 74% 28.8 42% 16.3 18% 44%

Poland 13.3 36% 3.8 13% 0.3 3% 0.3 -13% 7% Romania 31.7 7% 17.6 -8% 17.2 54% 13.7 -13% 78% Russian Federation 82.6 -9% 44.3 -15% 25.8 31% 20.0 -10% 45%

Saudi Arabia 13.0 -9% 12.8 -2% 8.2 63% 7.5 -34% 59% South Africa 21.3 29% 18.8 33% 15.8 74% 13.5 34% 72% Spain 33.2 3% 29.6 4% 19.9 60% 17.6 0% 60%

Sudan 26.6 34% 20.1 26% 24.1 91% 19.4 23% 96% Syrian Arab Republic 16.3 29% 12.6 23% 14.3 88% 11.9 20% 94%

Tajikistan 7.8 -4% 5.8 -8% 6.6 85% 5.4 -10% 93%

Thailand 39.7 92% 27.8 74% 23.6 59% 19.3 38% 69% Tunisia 2.8 23% 1.8 6% 2.0 72% 1.6 0% 86% Turkey 92.9 81% 52.5 94% 66.4 71% 35.3 64% 67%

Turkmenistan 13.2 -3% 10.3 -7% 12.5 95% 10.1 -8% 98%

UK 11.7 0% 9.5 0% 0.2 2% 0.2 -9% 2% Ukraine 31.2 -11% 20.0 -14% 15.2 49% 12.4 -10% 62%

USA 522.8 12% 457.8 16% 200.9 38% 149.2 13% 33% Uzbekistan 43.3 -1% 32.7 -6% 38.5 89% 31.0 -8% 95% Viet Nam 50.5 49% 29.2 32% 30.3 60% 18.0 0% 62%

26

IWMI Base Scenario

Table 2b. Summary water use by sectors.

Domestic Industrial

Estimate Calc. Estimate Calc.

Water Scarcity Per cap Gross Change % Per cap Total Change %

Physical = I domestic domestic Total industrial industrial Total diversion diversions diversions diversions diversions diversions

Economic = II (rr19) (rr19) (rr19) (rr19)

2025 2025 Period1 2025 2025 2025 Period1 2025 m3 km3 % % m3 km3 % %

World

All countries (45) 74 446 84% 11% 147 891 60% 22% % of the world

Algeria 44.3 2.0 77% 26% 28.1 1.3 87% 17%

Argentina 99.7 4.5 80% 13% 231.6 10.5 109% 31% Australia 540.9 12.0 23% 30% 20.0 0.4 48% 1%

Bangladesh 11.4 2.0 188% 5% 3.8 0.7 188% 2%

Brazil 138.6 28.9 83% 28% 63.9 13.3 113% 13%

Canada 167.6 6.1 24% 12% 1218.9 44.6 24% 84%

China 76.3 109.6 291% 12% 125.1 179.8 450% 20%

Egypt 74.1 6.8 104% 10% 115.0 10.5 137% 16% Ethiopia PDR 8.3 0.9 304% 39% 2.3 0.3 304% 11%

France 100.3 6.0 4% 16% 432.6 26.0 4% 68%

Germany 88.9 7.0 -4% 16% 432.0 34.0 -4% 80% India 30.8 39.2 237% 5% 55.2 70.2 353% 9%

Indonesia 36.5 9.5 337% 12% 18.6 4.8 164% 6%

Iran, Islamic Rep of 82.7 7.4 93% 8% 31.6 2.8 122% 3% Iraq 42.8 1.7 35% 4% 51.9 2.1 -2% 5%

Israel 54.7 0.4 28% 24% 20.0 0.1 50% 9%

Italy 137.8 7.0 -12% 16% 265.8 13.4 -12% 31% Japan 123.5 14.7 -5% 16% 239.7 28.5 -5% 32%

Jordan 52.7 0.6 159% 42% 11.2 0.1 303% 9%

Kazakhstan 40.0 0.7 1% 3% 339.7 5.7 1% 24%

Kyrgyzstan 67.5 0.5 67% 4% 67.5 0.5 67% 4%

Mexico 69.6 8.7 86% 6% 106.9 13.4 115% 10%

Morocco 32.9 1.2 127% 8% 24.0 0.9 176% 6% Myanmar 12.3 0.7 159% 16% 5.3 0.3 159% 7%

Nigeria 20.0 3.6 259% 26% 9.7 1.7 261% 13%

Pakistan 32.5 8.3 167% 3% 38.2 9.7 213% 3% Philippines 154.2 16.1 197% 23% 230.8 24.1 281% 35%

Poland 64.3 2.5 26% 19% 269.7 10.4 41% 79%

Romania 126.4 2.5 20% 8% 605.6 12.0 40% 38% Russian Federation 99.2 13.3 -9% 16% 323.6 43.5 -9% 53%

Saudi Arabia 106.7 4.1 175% 32% 18.3 0.7 325% 5%

South Africa 78.2 3.4 29% 16% 47.5 2.1 36% 10% Spain 110.2 4.0 5% 12% 255.8 9.3 12% 28%

Sudan 41.3 1.9 162% 7% 13.3 0.6 238% 2%

Syrian Arab Republic 50.4 1.3 121% 8% 28.0 0.7 146% 4%

Tajikistan 59.3 0.5 47% 6% 79.0 0.7 47% 9%

Thailand 73.7 5.2 299% 13% 155.4 10.9 460% 28%

Tunisia 42.8 0.5 91% 19% 20.0 0.2 167% 9% Turkey 167.8 14.0 175% 15% 149.2 12.5 256% 13%

Turkmenistan 56.5 0.3 48% 3% 56.5 0.3 48% 3%

UK 40.5 2.4 0% 20% 156.0 9.1 0% 78% Ukraine 91.3 4.1 -12% 13% 263.6 11.9 -12% 38%

USA 211.4 66.6 18% 13% 810.5 255.3 18% 49%

Uzbekistan 98.8 3.2 43% 7% 49.4 1.6 43% 4% Viet Nam 101.6 10.7 184% 21% 90.6 9.6 265% 19%

27

IWMI Base Scenario

Table 3. Food consumption (calories).

1995 Calories per capita per day 2025 Calories per capita per day Total growth (Kcal) 1995-2025 (FAO data) (IFPRI projections) (Per capita times pop. growth) Water Scarcity Total Total

Physical = I calorie % % % calorie % % % Cereal Animal Other intake from from from intake from from from foods products foods Economic = II per cap cereal animal other per cap cereal animal other

per day foods prod. foods per day foods prod. foods Period1 Period1 Period1

Kcal % % % Kcal % % % % % %

World 2720 50% 16% 34% 2930 47% 17% 36% 36% 53% 51% All countries (45) 2771 51% 16% 33% 3037 48% 18% 34% 32% 58% 47%

% of the world

Algeria 3019 59% 9% 32% 3156 54% 11% 35% 53% 96% 83%

Argentina 3117 30% 30% 41% 3471 28% 29% 43% 35% 41% 55% Australia 2975 23% 35% 42% 3135 23% 34% 43% 29% 25% 35% Bangladesh 2062 82% 3% 14% 2301 78% 5% 17% 52% 158% 92%

Brazil 2878 32% 18% 50% 3378 28% 21% 51% 34% 76% 57%

Canada 3105 22% 29% 49% 3226 22% 27% 51% 27% 22% 33% China 2766 60% 17% 24% 3112 50% 24% 26% 12% 88% 45%

Egypt 3277 66% 6% 27% 3441 64% 8% 28% 48% 102% 58% Ethiopia PDR 1781 72% 6% 22% 2035 72% 6% 22% 128% 161% 133% France 3550 24% 38% 38% 3889 24% 37% 40% 11% 10% 19%

Germany 3296 21% 32% 47% 3611 21% 31% 48% 3% 2% 9% India 2394 63% 7% 30% 2812 58% 11% 32% 47% 135% 69% Indonesia 2880 63% 4% 32% 3067 61% 6% 33% 34% 101% 44%

Iran, Islamic Rep of 2885 58% 9% 33% 3015 54% 10% 36% 39% 78% 65% Iraq 2262 54% 4% 43% 2364 49% 5% 46% 90% 143% 125%

Israel 3254 34% 19% 47% 3383 34% 18% 48% 28% 23% 33%

Italy 3476 32% 26% 42% 3808 31% 25% 43% -5% -6% 0% Japan 2898 41% 21% 38% 3169 39% 22% 40% -3% 8% 8% Jordan 2729 47% 11% 42% 2852 44% 13% 43% 94% 147% 120%

Kazakhstan 3117 59% 21% 20% 3284 57% 22% 21% 3% 11% 13%

Kyrgyzstan 2398 59% 22% 19% 2526 57% 23% 20% 70% 83% 86% Mexico 3137 46% 16% 37% 3474 42% 19% 39% 38% 81% 57%

Morocco 3177 63% 7% 31% 3321 58% 8% 35% 37% 75% 67% Myanmar 2711 79% 4% 18% 2929 74% 6% 20% 31% 119% 60% Nigeria 2554 42% 3% 54% 2822 43% 4% 53% 103% 132% 95%

Pakistan 2393 54% 15% 31% 2705 50% 16% 33% 97% 135% 124% Philippines 2366 49% 15% 36% 2688 44% 21% 36% 53% 147% 71% Poland 3309 36% 27% 37% 3401 33% 29% 38% -5% 12% 4%

Romania 2927 44% 24% 31% 3008 41% 26% 33% -17% -3% -7% Russian Federation 2814 41% 27% 32% 2894 39% 28% 33% -11% -3% -4%

Saudi Arabia 2730 50% 14% 36% 2853 46% 16% 38% 104% 161% 132%

South Africa 2882 54% 13% 34% 2996 53% 12% 35% 21% 16% 26% Spain 3291 23% 26% 51% 3605 23% 25% 52% -1% -3% 3% Sudan 2354 58% 18% 24% 2704 57% 19% 23% 91% 113% 86%

Syrian Arab Republic 3302 55% 11% 34% 3451 51% 13% 36% 71% 118% 100%

Tajikistan 2231 68% 11% 21% 2351 66% 12% 23% 50% 62% 69% Thailand 2330 52% 11% 37% 2473 45% 17% 38% 9% 100% 32%

Tunisia 3186 53% 9% 38% 3330 49% 10% 41% 32% 69% 54% Turkey 3563 49% 12% 40% 3501 47% 12% 41% 28% 39% 40% Turkmenistan 2583 56% 18% 26% 2721 54% 19% 27% 51% 63% 63%

UK 3210 22% 33% 45% 3516 22% 32% 46% 7% 6% 13% Ukraine 2880 43% 22% 34% 2962 41% 23% 36% -14% -7% -7% USA 3624 23% 28% 49% 3888 24% 27% 50% 27% 22% 29%

Uzbekistan 2565 62% 18% 21% 2703 60% 19% 22% 45% 56% 60% Viet Nam 2449 73% 9% 18% 2715 69% 14% 18% 49% 129% 61%

28

IWMI Base Scenario

Table 4. Cereal consumption (2025).

(IWMI) (IWMI) Water Scarcity Required Required Change Growth Growth Required Required Growth Physical = I total total feed conv. total total total total required food feed ratio food feed cons cons t. cons Economic = II cereals cereals (kg/kcal) cereals cereals non-food cereals cereals Period1 Period1 Period1 cereal Period1

M MT M MT % % % M MT M MT % World All countries (45) 1021 865 8% 31% 50% 277 2163 38% % of the world Algeria 10.0 3.9 20% 53% 135% 1.3 15.1 69% Argentina 6.0 8.3 10% 35% 55% 3.8 18.2 44% Australia 2.0 6.3 10% 29% 37% 2.8 11.1 33% Bangladesh 31.2 0.0 25% 52% 223% 2.2 33.4 52% Brazil 23.3 54.7 10% 34% 94% 10.0 88.0 69% Canada 3.6 25.0 -10% 27% 10% 4.9 33.5 13% China 259.3 241.2 20% 12% 125% 53.3 553.8 46% Egypt 22.8 9.1 10% 48% 122% 4.0 35.9 62% Ethiopia PDR 18.5 0.5 20% 128% 214% 2.2 21.1 130% France 7.3 26.3 20% 11% 32% 4.0 37.6 25% Germany 7.6 26.1 20% 3% 23% 6.0 39.7 15% India 223.9 7.8 25% 47% 193% 27.2 258.9 50% Indonesia 53.5 5.3 25% 34% 151% 6.0 64.8 40% Iran, Islamic Rep of 19.5 11.0 15% 39% 104% 2.2 32.7 56% Iraq 5.3 2.1 15% 90% 179% 1.0 8.4 107% Israel 1.0 2.3 10% 28% 35% 0.6 3.9 32% Italy 8.4 12.4 10% -5% 3% 1.6 22.4 -1% Japan 16.0 20.3 10% -3% 18% 4.3 40.6 7% Jordan 1.3 2.3 10% 94% 172% 0.2 3.8 136% Kazakhstan 4.1 5.2 10% 3% 22% 3.9 13.2 11% Kyrgyzstan 1.3 1.1 10% 70% 101% 0.4 2.8 81% Mexico 22.3 25.9 15% 38% 108% -2.5 45.7 70% Morocco 9.5 1.3 20% 37% 110% 1.9 12.7 42% Myanmar 13.0 1.6 25% 31% 173% 1.4 16.1 39% Nigeria 28.8 5.0 10% 103% 155% 9.9 43.7 108% Pakistan 37.7 2.2 25% 97% 194% 3.9 43.9 101% Philippines 13.3 10.7 10% 53% 172% 1.9 25.9 89% Poland 5.7 19.7 10% -5% 23% 3.9 29.4 13% Romania 3.2 10.2 0% -17% -3% 2.5 15.9 -8% Russian Federation 19.8 45.7 10% -11% 6% 17.6 83.1 -1% Saudi Arabia 6.1 14.9 -15% 104% 122% 1.0 22.0 117% South Africa 9.2 4.8 15% 21% 33% 1.2 15.2 24% Spain 4.0 14.6 10% -1% 7% 2.5 21.2 4% Sudan 8.2 0.3 20% 91% 155% 1.0 9.4 93% Syrian Arab Republic 5.6 3.7 20% 71% 162% 1.8 11.0 95% Tajikistan 1.7 0.1 10% 50% 78% 0.1 1.8 51% Thailand 8.0 11.2 10% 9% 120% 2.2 21.4 51% Tunisia 2.5 1.7 10% 32% 85% 0.5 4.7 48% Turkey 17.7 10.2 15% 28% 60% 12.4 40.3 35% Turkmenistan 1.1 0.8 10% 51% 79% 0.3 2.2 61% UK 5.9 9.6 0% 7% 6% 4.7 20.2 7% Ukraine 7.2 16.2 0% -14% -7% 5.4 28.8 -9% USA 38.8 181.5 -5% 27% 15% 57.0 277.3 19% Uzbekistan 6.2 0.9 10% 45% 72% 0.9 8.0 48% Viet Nam 19.8 1.2 20% 49% 175% 3.2 24.1 53%

29

IWMI Base Scenario

Table 5a. Production factors (irrigation).

(All crops) Data Assumed Estimate Calc Calc Calc Water Scarcity Net Irrigation Gross Net Total Irrigation Total Gross Total Physical = I irrigated crop irrigated irrigated growth crop growth irrigated growth area intensity area area intensity area Economic = II 1995 1995 1995 2025 Period 2025 Period1 2025 Period1

M ha M ha M ha M ha % % % M ha % World 259 All countries (45) 233 144% 336 285 22% 152% 8% 434 29% % of the world 90% Algeria 0.6 100% 0.6 1.0 60% 115% 15% 1.2 108% Argentina 1.7 160% 2.7 2.0 15% 175% 15% 3.5 27% Australia 2.3 148% 3.4 3.8 50% 163% 15% 6.2 80% Bangladesh 3.4 165% 5.7 5.2 42% 170% 5% 8.9 56% Brazil 3.1 155% 4.8 6.9 81% 170% 15% 11.8 144% Canada 0.7 100% 0.7 0.7 0% 100% 0% 0.7 0% China 49.7 175% 87.0 67.0 30% 180% 5% 120.6 39% Egypt 3.3 180% 5.9 3.7 12% 185% 5% 6.8 16% Ethiopia PDR 0.2 180% 0.3 0.2 23% 195% 15% 0.5 36% France 1.6 100% 1.6 1.6 0% 100% 0% 1.6 0% Germany 0.5 100% 0.5 0.5 0% 100% 0% 0.5 0% India 54.3 133% 72.3 63.1 15% 143% 10% 90.2 25% Indonesia 4.6 165% 7.6 5.0 9% 175% 10% 8.8 16% Iran, Islamic Rep of 7.3 115% 8.4 7.3 0% 115% 0% 8.4 0% Iraq 3.5 140% 4.9 3.7 5% 145% 5% 5.3 8%

Israel 0.2 125% 0.2 0.2 0% 125% 0% 0.2 0% Italy 2.7 190% 5.1 3.3 21% 190% 0% 6.3 23% Japan 2.7 100% 2.7 2.4 -15% 100% 0% 2.4 -14% Jordan 0.1 140% 0.1 0.1 0% 140% 0% 0.1 0% Kazakhstan 2.3 100% 2.3 2.3 0% 100% 0% 2.3 0% Kyrgyzstan 1.1 100% 1.1 1.4 30% 100% 0% 1.4 35% Mexico 5.0 104% 5.2 9.1 60% 119% 15% 10.8 107% Morocco 1.3 150% 1.9 1.5 20% 165% 15% 2.5 34% Myanmar 1.6 110% 1.8 2.2 30% 120% 10% 2.6 47% Nigeria 1.0 140% 1.3 1.2 20% 155% 15% 1.8 34%

Pakistan 17.3 150% 26.0 20.1 15% 150% 0% 30.1 16% Philippines 1.6 165% 2.6 1.9 21% 180% 15% 3.5 34% Poland 0.1 100% 0.1 0.1 0% 100% 0% 0.1 0% Romania 3.1 100% 3.1 3.1 0% 100% 0% 3.1 0% Russian Federation 5.4 100% 5.4 5.4 0% 100% 0% 5.4 0%

Saudi Arabia 1.5 130% 1.9 1.1 -30% 130% 0% 1.4 -26% South Africa 1.3 155% 2.0 1.7 30% 170% 15% 2.9 48% Spain 3.6 135% 4.8 3.6 0% 140% 5% 5.0 4% Sudan 1.9 100% 1.9 2.3 15% 115% 15% 2.6 34% Syrian Arab Republic 1.1 150% 1.6 1.4 21% 155% 5% 2.1 27% Tajikistan 0.7 100% 0.7 0.7 0% 100% 0% 0.7 0% Thailand 4.7 145% 6.9 6.4 30% 160% 15% 10.2 49% Tunisia 0.4 110% 0.4 0.4 11% 110% 0% 0.4 11% Turkey 4.2 110% 4.6 6.6 45% 120% 10% 7.9 71% Turkmenistan 1.3 100% 1.3 1.3 0% 100% 0% 1.3 0% UK 0.1 100% 0.1 0.1 0% 100% 0% 0.1 0% Ukraine 2.6 100% 2.6 2.6 0% 100% 0% 2.6 0% USA 21.4 163% 34.9 24.9 15% 168% 5% 41.8 20% Uzbekistan 4.0 100% 4.0 4.0 0% 100% 0% 4.0 0% Viet Nam 2.0 140% 2.8 2.3 12% 150% 10% 3.4 21%

30

IWMI Base Scenario

Table 5b. Cereal area.

Cereal area

Data Data Data Calc Calc Calc Water Scarcity Total Irrigated Rain-fed Total Total Irrigated Total Rainfed Total Physical = I cereal cereal cereal cereal growth cereal growth cereal growth

harvested area area harvested area area Economic = II area area 1995 1995 1995 2025 Period1 2025 Period1 2025 Period1

M ha M ha M ha M ha % M ha % M ha %

World 702

All countries (45) 633 208 425 654 3% 269 30% 384 -10% % of the world 90% Algeria 2.5 0.0 2.5 2.5 0% 0.1 108% 2.5 -1%

Argentina 8.9 1.0 8.0 10.3 16% 1.2 27% 9.1 15% Australia 14.1 0.6 13.4 14.1 0% 1.1 80% 12.9 -4% Bangladesh 11.3 5.7 5.7 11.3 0% 8.9 56% 2.5 -56%

Brazil 19.3 1.2 18.1 19.3 0% 2.9 144% 16.3 -10%

Canada 19.3 0.4 18.9 16.6 -14% 0.4 0% 16.2 -14% China 113.3 73.9 39.4 113.3 0% 102.5 39% 10.8 -73%

Egypt 2.5 2.5 0.0 2.9 16% 2.9 16% 0.0 na Ethiopia PDR 6.8 0.1 6.8 9.2 35% 0.1 36% 9.1 35% France 8.6 0.0 8.6 7.4 -14% 0.0 0% 7.4 -14%

Germany 6.7 0.0 6.7 6.7 0% 0.0 0% 6.7 0% India 100.8 47.0 53.8 100.8 0% 58.7 25% 42.2 -22% Indonesia 14.2 7.2 7.0 19.2 35% 8.3 16% 10.8 54%

Iran, Islamic Rep of 9.0 4.2 4.8 10.5 16% 4.2 0% 6.3 30% Iraq 3.4 3.4 0.0 3.4 0% 3.4 0% 0.0 na

Israel 0.1 0.1 0.0 0.1 -26% 0.1 -26% 0.0 na

Italy 4.1 1.3 2.8 3.6 -14% 1.6 23% 2.0 -31% Japan 2.2 2.2 0.0 1.9 -14% 1.9 -14% 0.0 na Jordan 0.1 0.0 0.1 0.1 -14% 0.0 0% 0.1 -16%

Kazakhstan 18.9 0.6 18.2 18.9 0% 0.6 0% 18.2 0%

Kyrgyzstan 0.6 0.4 0.2 0.8 35% 0.5 35% 0.2 35% Mexico 10.7 2.5 8.2 12.4 16% 5.2 107% 7.2 -12%

Morocco 4.7 0.7 3.9 5.4 16% 1.0 34% 4.4 13% Myanmar 6.1 1.4 4.6 7.6 25% 2.1 47% 5.5 18% Nigeria 17.8 1.1 16.7 27.8 56% 1.4 34% 26.4 58%

Pakistan 11.6 10.4 1.3 13.5 16% 12.1 16% 1.5 16% Philippines 6.4 2.6 3.8 6.4 0% 3.5 34% 2.9 -24% Poland 8.4 0.1 8.3 8.4 0% 0.1 0% 8.3 0%

Romania 6.3 0.8 5.5 6.3 0% 0.8 0% 5.5 0% Russian Federation 53.1 1.6 51.5 53.1 0% 1.6 0% 51.5 0%

Saudi Arabia 1.2 1.2 0.0 0.9 -26% 0.9 -26% 0.0 na

South Africa 6.6 0.4 6.2 6.6 0% 0.6 48% 6.0 -3% Spain 6.6 2.4 4.2 5.7 -14% 2.5 4% 3.2 -24% Sudan 8.8 0.9 7.9 11.9 35% 1.2 34% 10.7 35%

Syrian Arab Republic 3.4 0.8 2.6 4.6 35% 1.1 27% 3.5 37%

Tajikistan 0.3 0.1 0.2 0.4 35% 0.1 0% 0.3 47% Thailand 10.9 3.4 7.4 9.4 -14% 5.1 49% 4.2 -43%

Tunisia 1.1 0.0 1.0 1.1 0% 0.0 11% 1.0 0% Turkey 14.7 0.7 14.0 15.3 4% 1.2 71% 14.1 0% Turkmenistan 0.6 0.1 0.4 0.9 56% 0.1 0% 0.8 73%

UK 3.1 0.0 3.1 2.7 -14% 0.0 0% 2.7 -14% Ukraine 12.2 0.9 11.4 10.5 -14% 0.9 0% 9.7 -15% USA 62.8 20.9 41.8 58.2 -7% 25.1 20% 33.2 -21%

Uzbekistan 1.7 0.4 1.2 2.2 35% 0.4 0% 1.8 47% Viet Nam 7.3 2.7 4.6 9.8 35% 3.2 21% 6.6 43%

31

IWMI Base Scenario

Table 5c. Cereal yield and production.

Cereal yield Percent cereal production

Data Data Calc Calc Water Scarcity Irrigated Rain-fed Irrigated Total Rain-fed Total Irrigated Rain-fed Irrigated Rain-fed

Physical = I cereal cereal cereal growth cereal growth cereal cereal cereal cereal yield yield yield yield production production production production

Economic = II % of total % of total % of total % of total 1995 1995 2025 Period1 2025 Period1 1995 1995 2025 1995

ton/ha ton/ha ton/ha % ton/ha % % % % %

World All countries (45) 3.34 2.19 4.7 40% 2.46 12% 43% 57% 57% 43%

% of the world

Algeria 4.90 1.02 5.0 3% 1.18 16% 5% 95% 9% 91%

Argentina 4.82 2.52 5.6 16% 2.92 16% 19% 81% 20% 80% Australia 4.50 1.63 6.1 35% 2.05 25% 11% 89% 20% 80% Bangladesh 2.86 0.70 3.3 16% 0.95 35% 80% 20% 93% 7%

Brazil 2.95 2.16 5.3 81% 2.91 35% 8% 92% 25% 75%

Canada 4.00 2.70 6.3 56% 3.13 16% 3% 97% 4% 96% China 3.34 2.96 5.2 56% 3.44 16% 68% 32% 94% 6%

Egypt 5.48 0.00 6.0 9% 0.00 0% 100% 0% 100% 0% Ethiopia PDR 2.20 1.33 4.0 81% 1.54 16% 2% 98% 3% 97% France 7.50 6.70 8.7 16% 7.78 16% 0% 100% 0% 100%

Germany 7.10 6.08 8.2 16% 7.06 16% 0% 100% 0% 100% India 2.66 0.95 3.6 35% 1.10 16% 71% 29% 82% 18% Indonesia 3.47 2.06 4.3 25% 2.40 16% 63% 37% 58% 42%

Iran, Islamic Rep of 3.04 0.58 5.0 66% 0.67 16% 82% 18% 83% 17% Iraq 0.80 0.00 1.4 81% 0.00 0% 100% 0% 100% 0%

Israel 2.02 0.00 2.7 35% 0.00 0% 100% 0% 100% 0%

Italy 7.20 3.50 7.4 3% 4.06 16% 48% 52% 60% 40% Japan 4.25 0.00 4.9 16% 0.00 0% 100% 0% 100% 0% Jordan 5.20 0.57 6.0 16% 0.66 16% 56% 44% 60% 40%

Kazakhstan 1.80 0.61 2.1 16% 0.71 16% 9% 91% 9% 91%

Kyrgyzstan 2.50 0.86 3.4 35% 1.00 16% 87% 13% 88% 12% Mexico 4.66 1.82 5.4 16% 2.11 16% 44% 56% 65% 35%

Morocco 3.10 1.02 4.2 35% 1.18 16% 36% 64% 43% 57% Myanmar 2.75 1.76 3.2 16% 2.05 16% 32% 68% 37% 63% Nigeria 3.07 1.00 4.1 35% 1.25 25% 16% 84% 15% 85%

Pakistan 1.98 0.43 2.7 35% 0.50 16% 97% 3% 98% 2% Philippines 2.18 1.47 3.6 66% 1.84 25% 51% 49% 71% 29% Poland 2.80 2.83 3.8 35% 3.82 35% 1% 99% 1% 99%

Romania 4.00 2.64 4.6 16% 3.07 16% 18% 82% 18% 82% Russian Federation 2.70 1.26 3.1 16% 1.46 16% 6% 94% 6% 94%

Saudi Arabia 3.80 0.00 4.4 16% 0.00 0% 100% 0% 100% 0%

South Africa 2.50 1.92 3.4 35% 2.41 25% 8% 92% 13% 87% Spain 4.00 1.56 5.4 35% 1.81 16% 60% 40% 70% 30% Sudan 1.49 0.40 2.3 56% 0.47 16% 30% 70% 36% 64%

Syrian Arab Republic 4.76 0.66 6.4 35% 0.77 16% 70% 30% 72% 28%

Tajikistan 1.68 0.93 2.6 56% 1.08 16% 38% 62% 36% 64% Thailand 2.68 1.32 3.6 35% 1.42 8% 48% 52% 75% 25%

Tunisia 4.30 1.16 5.8 35% 1.34 16% 12% 88% 16% 84% Turkey 4.82 1.87 6.0 25% 2.34 25% 11% 89% 18% 82% Turkmenistan 2.36 1.35 3.2 35% 1.57 16% 34% 66% 25% 75%

UK 8.00 6.93 8.2 3% 8.05 16% 0% 100% 0% 100% Ukraine 4.60 2.26 5.3 16% 2.62 16% 13% 87% 15% 85% USA 5.50 4.91 6.9 25% 6.14 25% 36% 64% 46% 54%

Uzbekistan 2.50 1.52 3.4 35% 1.77 16% 37% 63% 32% 68% Viet Nam 3.17 2.02 4.3 35% 2.34 16% 47% 53% 47% 53%

32

IWMI Base Scenario

Table 6. Production and trade.

Cereal production International trade in cereals

Water Scarcity Total Calc Calc Surplus/ S/D Net Net Ex/imp

Physical = I cereal cereal total deficit percent export/ percent production production growth cereal of cereal import (-) of cereal Economic = II cereal production cons cereals cons

1995 2025 production 2025 2025 1995 1995 M MT M MT % M MT % M MT %

World 1792

All countries (45) 1623 2202 36% 39 1.8% 52 3.3% % of the world 91% Ex: 224 187

Algeria 2.7 3 20% -11.9 -79% -6.29 -70% Argentina 24.6 33 36% 15.3 84% 12.01 95% Australia 24.7 33 34% 22.1 198% 16.36 195%

Bangladesh 20.2 32 57% -1.6 -5% -1.79 -8% Brazil 42.6 63 49% -24.8 -28% -9.65 -18%

Canada 52.5 53 1% 19.6 59% 22.93 78%

China 363.5 572 57% 18.7 3% -16.35 -4% Egypt 13.8 18 27% -18.3 -51% -8.33 -38% Ethiopia PDR 9.2 15 58% -6.6 -31% -0.03 0%

France 57.6 58 0% 20.0 53% 27.50 91%

Germany 40.8 47 16% 7.8 20% 6.40 19% India 175.8 257 46% -2.4 -1% 3.74 2%

Indonesia 39.4 62 58% -2.6 -4% -6.84 -15% Iran, Islamic Rep of 15.5 25 63% -7.4 -23% -5.49 -26% Iraq 2.7 5 81% -3.5 -42% -1.37 -34%

Israel 0.2 0 0% -3.7 -95% -2.75 -94% Italy 19.2 20 3% -2.6 -12% -3.39 -15% Japan 9.3 9 0% -31.4 -77% -28.62 -76%

Jordan 0.1 0 8% -3.7 -97% -1.50 -93% Kazakhstan 12.3 14 16% 1.1 8% 0.36 3%

Kyrgyzstan 1.2 2 78% -0.7 -25% -0.37 -24%

Mexico 26.4 43 63% -2.5 -6% -0.46 -2% Morocco 6.2 9 49% -3.5 -27% -2.70 -30% Myanmar 12.1 18 48% 1.8 11% 0.49 4%

Nigeria 19.9 39 95% -4.9 -11% -1.04 -5%

Pakistan 21.1 33 56% -11.0 -25% -0.72 -3% Philippines 11.2 18 60% -7.9 -31% -2.49 -18%

Poland 23.8 32 35% 2.7 9% -2.12 -8% Romania 17.7 21 16% 4.7 29% 0.51 3% Russian Federation 69.2 80 16% -2.8 -3% -14.99 -18%

Saudi Arabia 4.7 4 -14% -18.0 -82% -5.47 -54% South Africa 13.0 17 27% 1.3 9% 0.75 6% Spain 16.2 19 19% -1.9 -9% -4.11 -20%

Sudan 4.6 8 72% -1.6 -17% -0.32 -7% Syrian Arab Republic 5.7 10 68% -1.5 -13% 0.04 1%

Tajikistan 0.3 1 65% -1.3 -72% -0.90 -74%

Thailand 19.0 25 29% 3.1 14% 4.79 34% Tunisia 1.4 2 20% -3.1 -65% -1.80 -56% Turkey 29.7 40 35% -0.2 0% -0.16 -1%

Turkmenistan 0.9 2 78% -0.6 -26% -0.46 -34%

UK 21.6 22 0% 1.3 7% 2.61 14% Ukraine 29.6 30 1% 1.1 4% -2.23 -7%

USA 320.3 376 17% 98.7 36% 86.51 37% Uzbekistan 3.0 5 58% -3.3 -42% -2.42 -45% Viet Nam 17.8 29 64% 5.1 21% 2.05 13%

23

CHAPTER 3

Water Supply and Demand, 1995 to 2025: Water Scarcity and Major Issues1

David Seckler and Upali Amarasinghe

Prophecy is a good line of business, but it is full of risks.

Mark Twain, Following the Equator

In this chapter, we discuss some of the major issues underlying the analysis and data. Some of this material is necessarily a product of subjective judgement, values and opinions. We have tried to make it clear when this is so. Population Projections2 The United Nations publishes high, medium and low population projections for all the countries of the world for the next century (UN 1999). Most people would agree that the high projection can be ignored. Of the other two projections, shown in table 1a3 of Part I, which is the most reasonable? This question is extremely important in thinking about the twenty-first century. According to the medium projection, world population will reach about 7.8 billion in 2025 and continue to grow indefinitely. But according to the low projection, world population growth will reach 7.3 billion in 2025 and cease to grow around 2040, at a level of about 7.5 billion people. The difference between the two projections in 2025 amounts to over one-half of a billion people. Perhaps even more important is the rapid decrease in the rate of population growth under the low projection. Remarkable as it sounds, by 2025 population growth for the world as a whole will no longer be major issue. While the UN (and the World Bank) maintains that the medium projection is the “most likely” one, we believe that the low population projection is better (Seckler and Rock 1995). For reasons that have never to our knowledge been adequately explained, the medium projection is based on the assumption that the total fertility rate (TFR) of a country cannot remain below the replacement level of 2.2 children per woman of child-bearing age. Under this assumption, countries that now are below that level, of which there are many, will increase to that level; and countries where TFR is falling to that level will never fall below it. The low projection, on the other hand, is based on statistical extrapolations of the trend of fertility rates, often going below the replacement level. This obviously is a more valid scientific procedure. The tragedy of AIDS has caused the UN to substantially reduce population growth projections in many countries, especially in sub-Saharan Africa. But obviously, this is the worst possible way to achieve this result. Similarly, it is also likely that populations will be reduced in some of the most water-scarce countries by the adverse health effects of water scarcity itself. Which projections should researchers and policy makers use? We have decided to use the average of the medium and low projections as a reasonable compromise. Of course, many developing countries will continue to have high rates of population growth through the first half

1 This is a revised version of a pervious paper circulated in 1999. 2Most of the items discussed in this and the following section are discussed in detail in Seckler and Rock 1995. 3All table references, which are part of chapter 3, are located at the end of chapter 2.

24

of the twenty-first century even under this average scenario. Unfortunately most of the Group I and II countries of figure 1 belong in this category, largely because of high population growth. But for other countries, it is important to keep the low population growth in mind when considering water strategies. For example, large-scale water development projects, such as large dams, require 20 years or more to bring on line. If population growth is very low at the time these projects come on line, they may not be needed. Also, the countries that will not have high rates of population growth may be in a better position to export food to other countries that do and food import prices may decrease. To avoid the dangers of costly overdevelopment, water development should be concentrated on smaller projects in the near term, until the population outlook clarifies. There are major exceptions to this rule. For example, India clearly will need large-scale water development projects even under the low projection. Food Demand and Supply Food and water are two of the basic human needs—and the latter, in the form of irrigation, is necessary to produce much of the former. FAO provides excellent data on food production and consumption in the world, conveniently entered on a CD-ROM. These data are used extensively in PODIUM and we are grateful to FAO for this and other data on agriculture and irrigation. The IMPACT model of the International Food Policy Research Institute (IFPRI) provides projections of food demands for 16 major countries and 22 intercountry regions in 2025. PODIUM uses the food demand projections of the IMPACT model in PODIUM, adjusted for the average population projection (and much of the water part of the PODIUM has been incorporated into the Impact model). However, the food supply projections are done independently in PODIUM. PODIUM also provides a means for policy makers to change the projections of food demand in order to target the nutritional standards they wish to achieve for their countries in 2025. Once these targets are set, the model provides a means of testing the feasibility of these targets in terms of agricultural and water constraints and the actions needed to achieve the targets. The single most important component of nutrition is calorie consumption per capita. The average for developing countries is around 2,200 kcal/person/day. With reasonably varied diets, if people satisfy their calorie requirements, they will also satisfy their requirements for protein, minerals and vitamins. A major exception to this rule is when a very high percentage of total calories is from rice, which is low in protein. Other exceptions occur with low vegetable consumption, which may cause vitamin and mineral deficiencies. But, on the whole, the principal target is adequate calorie consumption. But even if the average calorie intake of a country is 2,200 kcal/capita/day, this is not enough to assure that everyone in the country is actually obtaining enough. People with relatively high incomes tend to overconsume calories, mainly from animal products. Therefore, it is necessary to get substantially higher average calorie consumption in a country to attempt to achieve the minimum for poor people. How much higher this amount must be is largely a function of the distribution of income in a country. As a rule of thumb, something in the range of 2,700–3,200 kcal/day is adequate for most countries to satisfy basic food needs, depending on the distribution of income and other factors in individual countries. One of the most difficult issues in projecting the demand for food and related agricultural products in 2025 is consumption of animal products—meats, milk, cheese, etc. In most countries, the total calories consumed and the percentage of calories from animal products increase with income, even at high-income levels. However, because of a variety of causes including urbanization, health concerns and costs, it is likely that there will be:

25

• a reduction in excessive per capita calorie consumption by higher-income groups • a rapid growth in consumption per capita of meat products in developing countries, such as

China and India, as incomes increase, combined with a tendency to plateau at lower levels of consumption than in the traditional meat-consuming countries of the west

• a shift toward more vegetarian, or “Mediterranean,” diets, away from meats • a shift from red meats, notably beef, to white meats, notably chicken These changes on the food demand side will be accompanied by major changes on the food supply side. Traditional forms of animal husbandry produce most of the animal products in developing countries, where animal feeds are mainly from pastures and other lands not suitable for crops and from waste products. But the carrying capacity of these traditional feed resources is reaching its practical limit and most of the additional production of animal products will be from modern, commercial production units that depend on animal feeds (e.g., maize, barley, soybean meal, etc.). However, the carrying capacity of pastures in developing countries could be greatly increased with better seeds and application of inorganic fertilizers (and this would increase the supply of organic fertilizers for crops). Given the propensity to consume more animal products and the conversion from traditional to commercial production of these products, it is reasonable to assume that the production of feedstuffs will have to increase dramatically by 2025. However, it is a remarkable fact that while world consumption of animal products has increased rapidly, consumption of feed cereals has increased very slowly, at an annual rate of only 0.5 percent since 1985. Somehow, the world has received a “free lunch” in the production of animal products. Part of the reason for this is shown in figure 2. Developed countries barely increased consumption of feed cereals at all since 1985; developing countries nearly doubled their consumption from a comparatively small base, but this was offset by the decreased consumption in transitioning economies. Underlying these data are changes in the production of animal products. Three important factors relating to the conversion ratio—the kg of feed required to produce one kg of animal products—have played an important role. • The conversion ratios for red meats are about twice as high as those for white meats; thus, as

consumption shifts from the former to the latter, feedstuffs are freed to produce more animal products.

• The conversion ratios of all animal products have been decreasing rapidly due to technological change in terms of animal breeding, health and nutrition. This frees up more feed to support additional consumption.

• There has been some substitution of feed cereals by other feeds, like oil meals and cassava.4

These factors all tend to decrease the conversion ratio for feed cereals. However, there is an offsetting factor. As developing countries move from traditional to commercial sources of feed for production, their incremental conversion ratios will increase. These effects have been incorporated in table 4 where conversion ratios decrease between 1995 in most developed countries that are already under commercial forms of production of animal products, but, increase in most developing countries that have traditional forms of production. Last, it should be noted that the rapid growth in consumption of vegetables and fruits increases the demand for irrigation. The reason is that highly productive vegetable and fruit production is not possible without very good irrigation and drainage systems. 4We are grateful to Alexandros Nicharos of FAO for pointing this out.

26

Figure 2. Total feed grains, domestic consumption.

Agricultural Policies, Food Production and Rural Livelihoods The issues to be discussed in this section may be introduced by the following statement from the World Bank 1997:

Irrigated farmland provides 60 percent of the world’s grain production. Of the near doubling of world grain production that took place between 1966 and 1990, irrigated land (working synergistically with high-yielding seed varieties and fertilizer) was responsible for 92 percent of the total. Irrigation is the key to developing high-value cash crops. By helping guarantee consistent production, irrigation spawns agro-industry. Finally, irrigation creates significant rural employment. The Bank has been a major actor in the expansion of irrigation systems… More than 46 million farming families have benefited directly from the Bank’s irrigation activities.5

5Rural Development: From Vision to Action. 1997. Environmentally and Socially Sustainable Development Studies and Monographs Series 12.

27

While the exact values cited here may be debated,6 the importance of the central issues is beyond question. These are: • the near doubling of cereal production, which kept food prices low for poor people in face of

rapid population growth • the crucial role of irrigation, working synergistically with the other factors, in achieving this

result • the importance of sustaining and improving rural livelihoods As this report also notes, the policies under which these accomplishments were achieved have been substantially changed by the World Bank and other donors—and, to a lesser extent, by developing countries. Since agricultural policies are major variables in projections to 2025, these changes present an especially difficult problem because the policies under which past results were created are changing in a way that is difficult to predict. A brief and somewhat stylized historical review may be helpful in understanding this issue. From the very beginning of articulated development policies in the 1960s to the mid-1980s, there was a broad consensus on agricultural development policy. There were three major goals of this policy: • Create rapid growth of agriculture as a precondition to general economic growth and social

and political stability. • Produce food supplies in excess of market equilibrium amounts to lower food prices for poor

people (who commonly spend in excess of 70 percent of their total income on food). • Maintain farm prices at profitable levels through price support systems to induce private

investment by farmers in agriculture and, thereby, provide rural employment for a rapidly growing rural labor force.

To achieve these goals, a variety of specific policies was implemented. • Domestic agriculture was protected from excessive food imports to maintain producer prices.

Many countries followed the policy of “trend self-sufficiency”: maintaining production at the trend of demand—importing and drawing down on stores in exceptionally bad harvest years, while exporting and adding to stores in exceptionally good years.

• Large public investments were made in rural infrastructure such as roads, irrigation, market facilities and agricultural research and extension services.

• Productive inputs were subsidized, notably: irrigation, inorganic fertilizers, agricultural research and the production and dissemination of improved germplasm.

This “Development Consensus,” as it may be called, was attacked and substantially revised in the mid-1980s as part of what has been described as the “Washington Consensus” of liberalization, free trade and privatization. The Washington Consensus was actively implemented by “conditionalities” on loans by the World Bank, the IMF and many donor agencies through the rubric of Structural Adjustment Policies (SAPs). One of the major objectives of SAPs was to dismantle nearly all of the elements of the Development Consensus. A few countries willingly implemented the SAPs, some used them selectively to correct abuses in their systems and most adapted them reluctantly and only partially, in order to get grants and loans. Now the

6Specifically, in Table 5C, we estimate that in 1995 irrigated area produced 43 percent of world cereals, with this figure increasing to 57 percent in 2025. However, the World Bank figures are correct for the developing countries.

28

Washington Consensus is itself under attack, partly because of the financial difficulties of the Southeast Asian countries. There is a return to the emphasis on poverty alleviation by donor agencies and rejection of the “trickle down” theory (sometimes derisively called the “trickle on” theory). These policy debates pose two major problems in creating projections and a vision for 2025. The first problem concerns the projections. Most of the historical evidence on agricultural growth and productivity was created during the Development Consensus. If the present attack on the Washington Consensus is successful, the future will develop under yet another, presently unknown, policy regime. For example, the rapid growth in cereal production achieved during the Green Revolution was under the Development Consensus. It is hard to believe that such things as eliminating price support systems and fertilizer subsidies, lowering producer prices through increased import competition and reduced investments in irrigation and agricultural research would have any effect other than reducing the growth of food production, thereby causing increased food prices and poverty. The second problem is how these policy changes affect rural employment and livelihoods. In many developing countries, a high percentage of people live in rural areas and obtain their living from agriculture (table 1a). In the long run, this is undesirable, as agricultural wages and returns to self-employment are too low. But it is better than the alternative of no employment at all. And it is better than the alternative of being forced out of rural areas into urban slums in search of meager employment opportunities under appalling conditions of housing, access to drinking water, sanitation, lawlessness and all the other familiar attributes of this human disaster. The history of developing countries has shown that we cannot simply assume that displaced rural labor will find remunerative employment in the industrial-service sectors. Indeed, these sectors are undergoing their own process of rationalization and modernization and in many cases are reducing employment. Further, human capital resources are becoming progressively less fungible as the electronic revolution and other technological advances create more demand for highly educated and skilled white collar workers but less demand for common labor. As these sectors grow and actually offer remunerative employment, they will naturally attract young people off the farms. But until this actually happens, public investments, subsidies and other basic features of the Development Consensus in agriculture should, in our view, be retained, but made more effective and efficient. In sum, the question of appropriate policy is one of specific situations at specific times. The policies of the Washington Consensus are perhaps appropriate for the USA, with 3 percent of its population in agriculture, but not for India with 73 percent. And, someday, it will perhaps be appropriate for India. But there were many abuses of the Development Consensus in India and other countries and these abuses are in the process of being corrected. Whatever one may think of the other features of Marxian theory, the dialectic of thesis, antithesis and synthesis (or progress through excesses and errors) certainly rings a bell. International Trade in Food Tony Allan (1998) has coined the valuable term “trade in virtual water” to show how international trade can help alleviate water scarcity and other problems in many countries. Countries with plentiful water should export water-intensive crops, like rice, to water-scarce countries. This is a natural application of the principle of comparative advantage in international trade. It happens today with rice, which is exported mainly from wet countries like Vietnam, Thailand and Myanmar and the USA (which has excess irrigation and food production capacity). Wheat is exported from Canada, the USA and Europe where it can be grown in cool seasons, with low-water requirements. Maize is exported from the USA largely because it can be grown

29

without irrigation due to the exceptionally favorable agroclimatic conditions of the “corn belt.” This principle also pertains to trade within countries. Egypt could save nearly 10 percent of its scarce water supplies, for example, by replacing sugarcane production in the very hot south with cool season sugar beet production in the north. Food imports are essential where countries cannot grow enough food because of water or other constraints, as in many countries of the Middle East and sub-Saharan Africa. This is also true for some countries in Southeast Asia, like Malaysia, where the expanding industrial-service sectors are creating severe labor resource constraints in agriculture. In some countries in sub-Saharan Africa, the costs of inland transportation make it better to feed coastal cities through imports than through domestic production—at least in the near term, until the rural infrastructure can be created. A major problem with trade, of course, is that food imports must be paid for in foreign exchange, earned from exports or by grants and loans. This fact is somewhat hidden by large amounts of donor assistance in hard currency and historically heavily subsidized exports from the USA and Europe. In the theory of comparative advantage, every country should be able to export enough to cover imports. But in practice, this does not happen. Many of the most needy countries, such as those of sub-Saharan Africa, do not have sufficient exports to pay for imports. Economic consultants frequently have the revelation that water-scarce countries should devote their irrigation water only to high-valued crops, like flowers, fruits and vegetables, export them and then buy the cereals they need on international markets. A team from a famous American university recently found that the Middle East North African (MENA) countries would not have a water problem if they did this! The problem, of course, is that high-valued crops constitute very narrow and highly competitive markets, where only a modest increase in supply drives prices virtually to zero. Even within India, there are times when apples and potatoes are given away free in the producing regions. Every country, developing and developed, is already trying its best to produce and export high-valued crops. Another advantage of international trade is that imports help to build local markets, tastes and skills that can result in new domestic industries through import substitution. For example, we expect a substantial shift toward import substitution in terms of domestic meat and feedgrain production in countries like India and China as local entrepreneurs catch up in these markets. On the export side of the developed countries, it seems evident that there will be significant environmental and financial constraints on EU exports. (We have heard that EU policy is to achieve self-sufficiency in food, but not to encourage food exports outside of the EU itself.) In the USA and Canada, the ultimate results of the boom and bust cycles that the newly freed agricultural markets have been experiencing are not yet known, but they are currently encouraging an exodus from agriculture. Environmental pressures against irrigation and restoring water quality are also building in these countries. The end result of these considerations is that we believe developing countries with a high percentage of their populations in rural areas will attempt to be as self-sufficient in agriculture as they reasonably can in order to conserve foreign exchange and provide rural livelihoods. They will gradually relax this objective over time as exports grow, the growth of the labor force slows and employment opportunities in other sectors improve. Of course, many countries cannot achieve this objective because of water and other constraints and will need to import considerably more food by 2025. As shown in table 6, it appears that the production potential of the exporting countries will be sufficient to meet needs for increased cereal imports without severe financial or environmental damage. While the trade positions of many countries will change, net cereal exports, as a percentage of total cereal consumption in the world as a whole, will decrease from about 3.3

30

percent in 1995 to 1.8 percent in 2025. This means that total exports of the countries will increase from 187 M Mt in 1995 to 224 M Mt in 2025. Agricultural Production, Area and Yield Agriculture produces so many different kinds of products—from rice and cotton to sugar and kumquats—that it is very difficult to aggregate agriculture into an overall picture. It is, indeed, an “apples and oranges problem.” Traditionally, this problem has been addressed by focusing on the cereals, which are the primary and secondary ingredients of a large group of food and meat products. But, as shown in table 3, with changes in tastes and calorie needs, the demand for food cereals will grow more slowly over the 1995 to 2025 period (36%), than the growth of animal products (53%) and other, non-cereal foods such as vegetables and fruits (51%). Also, in many countries—especially in Latin America and sub-Saharan Africa, cereals do not play a predominate role in nutrition—either directly or as feed cereals through animal products. As the market responds to these changes, investments in agricultural production are shifting out of cereals to other higher-valued agricultural products. For example, as shown directly below, while the total harvested area in cereals has decreased to 1960 levels, total harvested area continues to grow slightly. And in some countries, yields of other crops are now growing more rapidly than cereal yields, which reverses the historical trend. As an irrigation manager in a developing country said recently, “Farmers only irrigate their cereals when they have water left over from the other crops.” As a result of these changes, cereals are progressively becoming a less valid indicator of food production and the agricultural world as a whole. Yet the aggregation problem remains. One way to address this problem is through the use of indices of agricultural production, area and yield, which use prices in a base year as weights to add disparate quantities of products. However, there are some difficulties with this approach that will not be gone into here. To provide an approximate solution to this problem, pending further research, we analyze both the cereal trends and the trends of the general indices in projecting cereal area, yield and production to 2025. This makes the cereal projections a quasi-indicator of total crop production over the period—and, on the whole, results in a somewhat more optimistic projection than would be obtained from projections based on the recent trends of cereals alone. Area While more land may be brought under agricultural production in places such as the large Cerrado area of Brazil, significant amounts of agricultural land will go out of production in many other areas because of urban-industrial sprawl, land degradation through salinization and erosion and diversion of irrigation water to urban-industrial needs. Because of these factors, the net gain in irrigated area may be negative in some countries—even though the statistics show an increase. With the major exception of sub-Saharan Africa, which has a large amount of underutilized rain-fed land, most of the increased food production for 2025 will continue to be (as in the World Bank statement above) from expansion of irrigated area and increased yields on irrigated land. As shown in figure 3 (and contrary to what we and many others believed before we did this graph) while growth in the irrigated area in the world as a whole has decreased, growth of irrigated area has remained nearly constant in Asia and in the developing countries as a whole. It should be noted that the 2025 projections of irrigated and rain-fed areas are for the total crop areas, not just that part of the areas devoted to cereals.

31

Figure 3. Net irrigated area of the world, Asia, DCs and LDCs, 1961–1997.

Source: FAOSTAT database. Yields It is very difficult to project crop yields. In the PODIUM model, yield projections are made through a combination of objective and subjective considerations in a process that may be briefly described as follows: 1. The international dataset does not distinguish between yields on irrigated and rain-fed area;

they are just lumped together in average yields. To estimate these different yields, IWMI has developed a new technique based on agroclimatic variables and the yield-water relationship discussed in chapter 5. These estimates have been checked, where possible, with national and international experts.

2. For each country, the 1985 to 1995 trends of the growth of agricultural area and cereal yields

and agricultural indices are estimated.7 The 1995 yields are examined and compared to yields obtained in other countries with similar agroclimatic conditions, fertilizer use, percentage irrigated area, etc., to get an idea of the yield gap between the technically and economically potential yields and the actual 1995 yields.

7This step uses the valuable Time Series (TS) data and graphics program of the US Economic Research Service, (USDA 1997). IWMI is creating a similar but improved program for this kind of analysis that will be available in 2000.

0

50

100

150

200

250

300

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Are

a (M

ha)

World Asia Developed countries Lessd-developed countries

IWMI December 1999

32

3. Where the 1995 yield gap is narrow, it is assumed that future yield growth will slow, below the trend rate. Where the yield gap is wide, it is assumed that the future growth rate will be the same as, or may even exceed, the trend rate. This usually provides the preliminary estimates of yield growth.

4. These preliminary estimates are extrapolated to 2025 and, together with the projection of harvested area, are used to estimate total production in 2025. This normally results in a surplus or deficit between projected food production and consumption for the country in 2025.

5. The situation of the country regarding international trade in food is then considered in terms of the following factors:

• If a substantial food deficit is projected and it is believed that the country has a basic policy

of trend food self-sufficiency and that it has the capability to implement that policy within reason, it is assumed that effective actions will be taken to increase yields and production by various means—additional research, fertilizer, irrigation, etc.—to reduce the trade deficit, up to economic and technical limits. India and China are in this category, for example.

• If a country does not satisfy these conditions, the food trade deficit is allowed to widen. • Some countries satisfy the other conditions but water limitations prohibit them from

achieving the self-sufficiency policy. Other countries may have the resources to be self-sufficient but do not have the institutional or financial capacity to achieve it, as in much of sub-Saharan Africa. Or some countries would rather invest their funds elsewhere and import more food.

• Last, some countries, such as the USA, Argentina, Canada and Australia are major food exporters and may increase their surpluses. The countries of the EU have historically followed this policy, but for environmental and other reasons may be changing it.

• Of course in the end, world food trade must balance between imports and exports. In sum, the yield projections in PODIUM, as with all its other projections, depend on a combination of objective data analysis and a more subjective analysis based on the Toynbeean process of challenge and response.8

While this is a simple procedure conceptually, it is extremely difficult to implement in practice. First, there are formidable data problems. As noted before, there are no international statistics on yields by rain-fed and irrigated areas. Nor are there data on cropping intensity by seasons for rain-fed or irrigated areas. Nor is there a good map of the irrigated areas of the world. IWMI is trying to alleviate these and other data problems, but it will take time. Another extremely important but generally ignored factor in projecting future yield growth is the interaction between growth of irrigation and yields. In arid areas, yields on irrigated land are three to four times greater than those on marginal rain-fed land. This is partly due to the direct effect of adequate water supplies increasing plant growth and partly due to the indirect effect that the lower risk of irrigated agriculture induces investment in improved seeds, fertilizers, labor and other yield-increasing inputs (Hyami and Ruttan 1985). Also, expansion of irrigated land normally occurs on rain-fed agricultural land. This creates two effects. First, the net gain in yield from irrigation is not quite as large as it would otherwise be because the lost rain-fed yield must be subtracted from the gain in irrigated yield. Second, counteracting this, low-yielding agricultural land is taken out of the computation of average yield and replaced by higher-yielding irrigated land, thereby increasing average yields in two ways. The combination of these direct and indirect effects makes it extremely difficult to answer the question of how rapidly future 8In the sense used by the great historian, Arnold J. Toynbee. We are grateful to John Briscoe for this phrase.

33

yields would grow if the growth of irrigation were reduced. Yet, this is one of the fundamentally important questions in creating a water vision for 2025. Biotechnology Last, there is the bewitching question of the effects of modern biotechnology on future yields. This is a subject about which we know so little that we can only refer to impressions gained from cursory reading in the field. These impressions are well summarized in a recent article in Science, 15 January 1999, which notes that the optimism of economists and biotechnology firms regarding miracle crops is greeted with deep reservations by the scientists who actually have to produce these crops. As Norman Borlaug has observed, there is no gene, or any known set of genes, that increases yields. Only comparatively small advances in disease and pest resistance are on the near horizon. The overall effect of these advances on food production may be about 5 percent to 10 percent of total food production. While this represents a huge opportunity for farmers and biotechnology firms to increase profits, it makes little difference in prospects for food production in 2025. In any case, the real and apparent dangers of biotechnology will certainly slow large-scale adoption of these technologies. However, all such prognostications must be qualified by the “surprise factor” in technological forecasting. Any day there could be a genuinely massive breakthrough in crop technology comparable to the Green Revolution. Another article in the same issue of Science illustrates the grounds for such a revolution. For unknown reasons, the catalytic molecule that regulates photosynthesis has the slowest processing rate, by a factor of thousands, of any other catalyst. If it could be speeded up, the results could be dramatic. Scientists have been working on this problem for decades, with very little progress. Now there may be a dim light at the end of the tunnel because of the discovery of a red algae that has a more rapid catalyst. But who knows when and if such a breakthrough might happen? It could be next year, or never. As the wise ditty says,

A trend is a trend is a trend but the question is will it bend? Will it alter its course through some unseen force and come to a premature end?

Irrigation Efficiency and Productivity It is often contended that water resources are so wastefully used, especially in agriculture, that there is little, if any, need to develop additional water supplies. All that is needed is to increase the efficiency and productivity of the water already being used through better management and technology. In many cases, largely in developed countries, this is true. In many other cases, largely in developing countries, this is not true. And where per capita water use is currently far below the levels and qualities needed to sustain a healthy and productive life, as in much of sub-Saharan Africa, it is absurd. This subject is so important, complex and oftentimes misleading that we have devoted chapter 4 to it.

34

Irrigated and Rain-fed Agriculture A popular idea is to concentrate food production in rain-fed, rather than in irrigated areas. The total cultivated area of the world is about one billion hectares, of which only about one-third is irrigated. Thus, a 10 percent increase in the productivity of rain-fed agriculture would have twice the impact as the same increase in irrigated agriculture. As the beneficial impact would be largely on poor farmers in marginal areas, this is an enormously attractive idea. It should be recognized from the start, however, that this is by no means a new idea. The goal of increasing productivity of marginal rain-fed areas has been energetically pursued, using all the tools of agronomic science, for at least a century, with highly disappointing results. We believe that the sciences and technologies of agronomy and water management have now advanced to the point where there are grounds for optimism in this field—and, indeed, there are notable cases of success on the ground. But before solutions can be found, the depth and extent of the problems must be thoroughly understood. A major part of the problem is shown in figure 4 (Hargreaves and Christiansen 1974). The vertical axis represents the relative yield; this is the actual yield obtained divided by the potential yield and all other factors such as seeds and fertilizers, are at their optimum levels. The horizontal axis shows the relative water supply; this is the actual water supplied divided by the physically optimal water supply. While the absolute values of yield and water will differ among specific agroclimatic areas, the relative value for most crops is roughly the same everywhere. Figure 4. Moisture adequacy and yield function.

This curve helps to explain the great diversity of rain-fed yields in the world. On the one hand, there are vast areas of the most favored rain-fed areas—such as those in the American mid-west and central and northern Europe—which have adequate and reliable water supplies (in the 80% range of figure 4) and thus are close to the optimum conditions for high yields. But most of

35

these highly favored areas have already been fully exploited. Most of the underexploited rain-fed areas of the world are, unfortunately, characterized by the lower one-third of the relative water supply axis, where yields are 25 percent to 35 percent of potential. These areas can be described as marginal rain-fed areas. There are three central problems of agriculture in marginal rain-fed areas.

• Most of a farmer’s costs are the fixed costs of cultivating land, independent of yield. Thus as yields decrease, net returns to farmers decrease even faster. For example, if costs represent 2 Metric tons per hectare (MT/ha), the farmer earns a net of 3 MT/ha at an economic maximum yield of 5 MT/ha, with optimal water supply. But the farmer makes only 1 MT/ha if yield is reduced to 3 MT/ha due to deficient water supply.

• In most cases, rainfall is highly unreliable. Farmers rationally minimize their investments in labor, improved seeds, fertilizers, soil and water management and the like to minimize losses due to drought. But this lack of investment in productive inputs means that even when good rainfall occurs, the yield is not as large as it should be.

• Since rainfall affects large areas, prices rise dramatically in times of drought, when there is nothing to sell and collapse in periods of good rainfall, when harvests exceed subsistence needs and there is a lot to sell.

These problems have been partly overcome in marginal rain-fed areas of developed countries such as the USA, Canada and Australia by large-scale, well-capitalized and highly mechanized farming. With several hundreds, if not thousands, of hectares per farm unit, large tractors and other equipment and sufficient capital to tide them over drought years, marginal rain-fed areas can be profitably farmed. Mechanization provides the ability to practice a variety of water and soil conservation practices—such as land leveling, terracing, fallowing, low-till agriculture, etc.—that are difficult and costly, if not altogether impossible, with only human and animal power. Because of their financial resources, these large farms can survive one out of three or four drought years. We believe that much of the future production of rain-fed farming in marginal areas will depend on the ability to bring these advantages of large-scale farming to small-scale producers through various methods of collective action (see Seckler 1992). But the history of such institutional innovations in developing countries has not been encouraging, to say the least. It is hoped that advances in biotechnology will result in drought-resistant and more water-efficient crops. One problem with this idea is that, hitherto, drought-resistant crops and varieties are, for that very reason, low yielding. Such a crop may produce a more stable yield over varying climatic conditions but at such a low yield potential that it is uneconomical or unable to respond to favorable conditions. As with yield, there is no single gene, or any known set of genes, that determines drought resistance. While there are likely to be advances in this field, largely through classical selective breeding, there is little likelihood of a substantial breakthrough. Last, it is important to guard against the common assumption that rain-fed agriculture somehow uses less water in food production than irrigated agriculture. Several effects of rain-fed agriculture should be understood: • Rain-fed crops are almost always planted on lands that previously supported low-valued

grasses or trees. These plants consume all the water that enters the soil, through evapotranspiration, just as do the crops. Thus there may be a gain in the value per unit of water consumed, or “crop per drop”—if the crops are more valuable than the previous plants. In terms of environmental values, they may not be.

• Rain-fed crops are usually planted using various kinds of soil moisture conservation techniques, such as tillage, mulching, bunding, terracing, etc., to reduce non-beneficial

36

evaporation from the land and runoff from the fields. When non-beneficial evaporation is reduced, there is a real gain in water productivity; but utilizing runoff may simply represent water that does not flow into other surface and subsurface areas where it may have a higher- valued use—such as domestic water supplies or, indeed, downstream irrigated areas.

• The marginal (dy/dx) curve in figure 4 implies that the productivity of water used in agriculture, the “crop per drop,” is highest when the relative water supply is low, at around 0.35. But this finding must be treated cautiously, because it only optimizes returns to water—not to all the other factors of production that affect farmer’s income. Also, the degree of water management precision required to attain this optimum is available only in the most sophisticated irrigation technologies and management systems.

• Another problem in rain-fed agriculture without inorganic fertilizers is that plant density is typically low in order to extract nutrients from the soil. Consequently, the crop canopy is open and non-beneficial evaporation from the soil surface increases. A study in Africa, for example, showed that only 5 percent of the water entering a field was beneficially used for crop production; the balance was runoff and non-beneficial evaporation (Wallace and Batchelor 1980); this study also shows the potential for improving the situation because of this high water loss).

For these and related reasons, contrary to what is commonly thought, a large shift to rain-fed agriculture in many marginal areas could result in reduced productivity per unit of water consumed in agriculture. However, under specific agroclimatic conditions, small-scale farming can be productive in marginal rain-fed areas through supplemental irrigation. Of course, all irrigation is supplemental irrigation because it is designed only to “top up” effective precipitation on the crops. But supplemental irrigation is a technique specifically designed for water-scarce regions, where scarce water is stored and used only in limited quantities at the critical growth stages of crops. In many areas, for example, there is sufficient average rainfall over the crop season to obtain good yields, but yields are greatly reduced by short-term, 15- to 30-day, droughts at critical growth stages of the plant. Water stress at the flowering stage of maize, for example, will reduce yields by 60 percent, even if water is adequate during all the rest of the crop season. If there is a way to store surplus water before these critical stages and apply it if the rain fails in these critical stages, crop production would increase dramatically. There are many ideas for water conservation and supplemental irrigation for smallholders. This is a long and complex subject that cannot be gone into here other than to say that most of these ideas have failed in practice because of two important factors: • They do not adequately consider the need to actually have and store surplus water before the

drought episode. • They fail to consider the economic costs, relative to benefits—which is all the farmer cares

about. One of the single most promising technologies in this field, that has gained wide adoption in India, is “percolation tanks.” These are small reservoirs that capture runoff and hold the water for percolation into shallow water tables. The water is then pumped up onto fields when and only when, it is most needed. Groundwater storage avoids the high evaporation losses of surface storage; with pumps, the water table provides a cost-free water distribution system to farms; and percolation losses from irrigation are automatically captured by the water table for reuse. These percolation tanks can be combined with highly efficient sprinkler and drip irrigation conveyance systems to provide just the right amount of water when it is needed most.

37

In order to evaluate the agricultural potential for marginal rain-fed areas it is necessary to have rather detailed climatic maps of countries. IWMI’s Climatic and Agriculture Atlas of the World (on IWMI’s website: http://www.iwmi.org), the final version of which will be available for Asia and Africa within the next 6 months, will be of enormous help in addressing this issue. In sum, for all these reasons, it is likely that an increasing proportion of the world’s food supply will have to be from irrigation. An important need is supplemental irrigation, in marginal rain-fed areas such as in sub-Saharan Africa, using advanced irrigation technologies. In fact, this absolutely has to happen if sub-Saharan Africa is to produce enough food to feed its rapidly growing population without an unacceptably high level of food dependence and provide remunerative rural employment. Groundwater Depletion Another area of intensive competition for water and rapidly increasing water scarcity is the use of groundwater in irrigation. In terms of impact on food production, one of the greatest technical revolutions in irrigation has been the development of the small-scale pump. Tens of millions of small pumps are currently drawing water out of aquifers to irrigate crops. Over one-half of the irrigated area of India is now supplied by groundwater. Because pump irrigation provides water on demand, yields from pump irrigation can be two to three times those of canal irrigation. Since irrigation supplies about one-half of the total food production of India, one-third or more of India’s food production depends on these humble devices and the aquifers that feed them. Much the same is true in other arid countries. Yet, almost everywhere in the world, groundwater tables in areas that depend on irrigation from groundwater are falling at alarming rates. In many of the most pump-intensive areas of India and Pakistan, water tables are falling at rates of 2 to 3 meters per year. This is not surprising when one considers that the evaporation losses of a typical crop is around 0.5 m of depth and the yield of water in an aquifer is about 0.1 m per meter of depth. Without recharge, groundwater tables would fall by about 5 m per crop per year. Most of these areas receive sufficient average rainfall to recharge the aquifers, but most of the rainfall goes to runoff, not to recharge. We desperately need to change that relationship. It is no exaggeration to say that the food security of India, Pakistan, China and many other countries in 2025 will largely depend on how they manage this groundwater problem. Reducing the amount of pump irrigation is no answer; this simply reduces the most productive agriculture. The answer has to be in groundwater recharge. But this is not an easy solution. Indeed, to our knowledge, no one has devised a cost-effective way to do it on the large scale required. About the only idea that we in IWMI have been able to think of is to encourage, through subsidies if necessary, flooded paddy (rice) cultivation in lands above the most threatened aquifers in the wet season. Paddy irrigation has high percolation losses and is thus a very inefficient form of irrigation from a traditional point of view. But from the point of view of groundwater recharge, this is just what the doctor ordered.9 Of course, one has to be careful not to pollute groundwater through leaching nitrates and other chemicals. Restrictions on fertilizer, pesticides and other chemicals in these recharge areas would be required. Ideally, the recharge areas would not be used for any other purpose—except, possibly fish production—but in densely population areas, land is too valuable to simply be set aside for this single use. IWMI and others are conducting research on ways of maximizing the rate of recharge and controlling pollution effects in this exceptionally important area of water resources management.

9After this statement was written, attracting some criticisms from colleagues, we found that India has been doing precisely this in a 180,000-hectare area for the past 10 years. We will do a study of the results of this important project soon.

38

Competition for Water among the Agricultural, Urban, Industrial and Environmental Sectors By 2025, most of the world’s population will live in urban and peri-urban areas. The people and industries in these areas will demand an increasingly large share of the total water available (table 2b) and much of this will be taken from irrigated agriculture. Already, in India, the Philippines and many other countries, large irrigation areas are literally shut down, either permanently or in times of drought, by cities taking water from farmers, with no compensation paid to them for loss of their livelihoods. Urbanization is creating an enormous pollution load on freshwater supplies and estuaries. The amount of pollutants thrown into the waterways is increasing rapidly and, at the same time, the flows of freshwater are decreasing as more water is evaporated more through intensive use. Thus the concentration of pollutants is increasing even more rapidly than growth of urban populations and industries would indicate. It is only recently that we have begun to appreciate the economic value of the waterways as waste disposal systems. As this previously free service ends and we have to treat water discharges, the costs will run to tens of billions of dollars. But the human health costs of the alternative of doing nothing would be even greater. Already, most vegetables grown in developing countries are irrigated with untreated sewage water from the nearby market towns. As if this were not bad enough, most of the urban population will be concentrated in coastal areas where sewage water, whether treated or not, is discharged into the seas. In addition to the pollution problem, this greatly increases the consumptive use of water and prevents water recycling, thereby contributing disproportionately to water scarcity. One of the most important although generally ignored water-using sectors is the environmental sector. More water is allocated in California to wetlands, free flowing rivers, estuaries and the like than to agriculture. The environmental sector has a strong impact on water scarcity because it can have high consumptive use. Exposed water surfaces evaporate rapidly and naturally flowing rivers generally end in the seas. This is a particularly acute problem in water-scarce developing countries. For example, the wetlands of the Delta Central on the Niger River in West Africa and the Sud on the White Nile in East Africa provide highly valuable wildlife sanctuaries and homes for migratory birds. But both of these wetlands evaporate around 50 percent of the water flow of their rivers. Both of these wetlands are under intensive pressure to redirect the water to lower evaporation losses and provide the water for human use downstream. No one knows how large the water demands for the environmental sector actually are. Historically, water for this sector has been a naturally occurring free good. But now that water is becoming more scarce, deliberate policies and water allocations to this sector have to be made. And it should be noted that, a decision not to develop water supplies for the other sectors on environmental grounds is a de facto allocation of water to the environmental sector. Here is another important area for future research. Concluding Observations on Water Pricing and Institutions It is one thing to estimate the potential for increased water productivity and quite another to achieve it. It is precisely because water is such an important economic good that, ironically, powerful forces do not want to treat it as such. As Mark Twain said, “Water flows uphill, toward power.” Some economists advocate pricing water at full marginal cost, both to achieve economic efficiency and to induce institutional change (see the discussion of these issues in Perry, Seckler, and Rock 1997). But water resources management is subject to failure of not only the public

39

sector but also the private sector: in economic terms, it is subject to “market failure.” Technically, water-recycling effects create massive external benefits and costs that violate the optimizing conditions of free market systems. The intensity of external effects in water use is perhaps greater than in any other sector of the economy; that is why water resources always have been a publicly managed or regulated resource. There are many advantages to pricing water, if it is properly regulated. First and foremost, it provides a means of financing water service agencies and, since they are being paid by their clients, of holding their feet to the fire of performance. Second, entitlements to water provide a means of forcing compensation to users who are harmed by unregulated public and private systems. In many countries, water is being arbitrarily reallocated from farmers to cities (India and Philippines) and for environmental purposes (USA) with no compensation for the loss of livelihoods this creates. Entitlements to return flows also would force payment of compensation to downstream losers created by upstream changes in use (as, it appears, happens under the unregulated market system in Chile). Of course, as economists point out, pricing water can induce water use efficiency and allocative efficiency. But in many developing countries with hosts of small farmers to deal with, the transaction costs of marginal cost pricing are likely to be greater than the benefits.10 Pricing water is a good way to regulate the external costs of water use—for example, in water pollution. This is because the higher the price, the lesser the water that will be used and thus, other things remaining equal, the lesser the pollution. But it is very difficult to regulate the external benefits of water use through pricing. For example, the external benefits of field-to-field irrigation in paddy systems, or the recharge of aquifers from irrigation systems would require a negative price, or subsidy, to reach the optimum level of water use. While this can and is being done, it is not usually considered by the advocates of (positive) water pricing—and all one can say is that this omission is evidence of poor economic training and analysis. Socially, a minimum supply of safe water is one of the essentials of life and most people would agree that everyone should be entitled to receiving that minimal amount. Market systems, on their own, may not have sufficient incentives to achieve that social objective; it depends on technical conditions of the demand and supply curves. But free water supplies to poor people have sometimes resulted in bankrupt water supply agencies, massive subsidies and preferred services to the rich. Thus the introduction of water pricing and the need to manage water at the river-basin level means more and better, not less, public management. But a major problem in water management in developing countries is the large and growing disparity between the remuneration of public and private sector staff. Bloated bureaucracies can only be supported by low wages. This inevitably leads to corruption and brain drains to the private sector—just as the needs for well-trained and dedicated people in the public sector are becoming increasingly acute. The first step in beginning the revolution in water management is to provide generous redundancy payments to marginal staff in public agencies, as a onetime write-off and then use the future savings to upgrade the civil service. The lessons of Hong Kong, Singapore and other countries where public servants are remunerated at rates comparable to the private sector indicate very high rates of return to such policies, if they are effectively implemented. Given failure on both sides of the private-public table, it would appear that a partnership between the two is the only way out of the dilemma. There are innumerable experiments going on all over the world in designing and implementing these public-private sector partnerships. One of the most important research tasks for the future is to carefully and objectively monitor and evaluate these experiments so that everyone can learn from the experience. But until much 10C. J. Perry 1996, IWMI Research Report 2, page 22.

40

more information is developed, it will remain exceptionally difficult to forecast the extent to which the potential gains in water productivity will become real gains.

23

CHAPTER 4

Major Paths to Increasing the Productivity of Irrigation Water1

David Molden and Charlotte de Fraiture Clearly, increasing the productivity of water is a central concern in using water to produce food, fight poverty, reduce competition for water and to make sure there is enough water to sustain ecosystems. To meet the level of increase in productivity of water in IWMI’s base scenario will require great effort—it will not happen automatically. To move beyond these levels is possible, especially in developing countries where levels of water productivity are generally far below potential; however to meet this challenge will require a far greater effort and significant changes in how water is managed. The first task is to understand what needs to be changed and where improvements are required. Then we can target policies, institutions and research efforts to address these problems. The purpose of this section is to outline means to increase the productivity of water in agriculture. But before doing this we need to review concepts about how water is used in agriculture and how PODIUM handles water use. Means of saving water are discussed, and then means of getting the most out of water used by agriculture are presented. First we deal with a major issue—that of efficiency of irrigation. How Efficient is Irrigation? A common perception is that irrigation wastes enormous quantities of water. If we could just be more efficient with irrigation, water would be made available for more agriculture and other water uses and we would not have to develop more water infrastructure. Unfortunately, this perception is in many cases not true and the opportunity for real water savings is much less than thought. The major reason that people believe enormous amounts of water are wastefully used in agriculture is a narrow and technical use of the word “efficiency” in irrigation. Briefly, this word is used to define the amount of water that needs to be withdrawn or diverted (D) from a source—such as a river, aquifer or reservoir—and delivered to the field, through canals, pipes and channels, to meet the water needs of crops. The water needs are defined as the amount of evapotranspiration (Eta) by crops, minus the amount of water supplied by effective precipitation to the crop (Pe)—or net evapotranspiration (NET). Irrigation efficiency (Ei) is commonly defined as Ei = NET/D. This efficiency ratio can vary between 90 percent in the case of drip irrigation systems to as low as 20 percent in the case of traditional paddy (rice) irrigation systems. Thus, it is reasonable to think that large quantities of water can be saved by increasing irrigation efficiency. In some cases this is true, in others it is false. Whether it is true or false depends on what happens to the drainage water, the amount of water that is delivered to the field but not used by the crop for evapotranspiration. Drainage water is mainly from: • seepage from conveyance systems • deep-percolation below the root-zone of plants in fields • surface runoff from fields

1Some of this borrows from written material provided by David Seckler. Comments of R. Sakthivadivel, Ian Makin, Douglas Merrey, and Anisa Divine are appreciated.

24

Drainage water may flow to saline areas or to the oceans where the water is effectively lost to further human uses. In this case, increasing irrigation efficiency (reducing drainage) can result in real water savings. On the other hand, drainage water may flow to other surface and subsurface areas where it can be captured and beneficially reused. This is the return flow of water. Return flow, for example, is a major source of recharge for aquifers in irrigated areas. Or, in paddy irrigation systems, drainage water flows from field to field. In many areas, return flows enter rivers, reservoirs and lakes, from where the water is again diverted for use. Thus, because of recycling, one person’s drainage may be another person’s water supply. In these cases, if an inefficient irrigation system is replaced by a more efficient one, usually at substantial cost, the result may be a zero-sum game where apparent gains at the beginning of the water cycle are off-set by losses of return flows in the rest of the water cycle. The concept of basin efficiency is introduced to take account of return flows and the water recycling effect. For example, in Egypt, the typical efficiency of irrigation at the farm level is only 40 percent to 50 percent. But for the Egyptian irrigation sector as a whole, at the basin level it is close to 80 percent, because of recycling. And much of the remaining 20 percent is beneficially used in other sectors, including the environmental benefits of discharges to wetlands and the Mediterranean. All things considered, there is not much real water saving to be made through improved irrigation efficiency in Egypt, even though it appears to be inefficient at first glance. Another example is in estimating the industrial use of water; the amount of water flowing through a hydroelectric plant is a diversion or application of water to a use, but the consumptive use of that water from evaporation may be virtually zero. Indeed, in many cases, the return flow from one hydroelectric plant on a river supplies a downstream plant. The amount of consumptive use as a percentage of diversions varies enormously among the water using sectors—from nearly zero in the case of hydroelectric plants, to 10 percent to 20 percent in most industrial uses, to 30 percent to 70 percent in irrigation. These diversions thus generate large amounts of drainage, much of which is available for recycling. This means that there is much more water actually available for use in a river basin than is implied in the commonly used concept of water withdrawals, which implicitly assumes that 100 percent of the water withdrawn is consumptively used. A problem with the concept of efficiency, even with basin efficiency, is that it refers only to physical quantities of water, both in the denominator and the numerator. It does not capture differences in the value of water in alternative uses. This is the subject of the more general concept of water productivity. Water productivity can be increased by obtaining more production with the same amount of water or by reallocating water from lower- to higher-valued crops—or from agriculture to other sectors where the marginal value of water is higher. Indeed, the greatest increases in the productivity of water in irrigation have not been from better irrigation technology or management, but rather from increased crop yields due to better seeds and fertilizers. As yields on irrigated land have doubled or tripled over the past three decades, the productivity of water, the “crop per drop” has increased accordingly. For these reasons it is generally best to use the term water productivity, rather than efficiency—understanding that gains in basin efficiency can make an important contribution to gains in productivity. In the PODIUM model, for example, we project a 38 percent increase in food production from irrigated area with only a 17 percent increase in water diversions to agriculture. This large gain in water productivity is due both to yield increases from agronomic factors and to water management factors, including water saving technologies and more timely supplies.

25

Development and Use of Basin Water Resources To further illustrate these concepts, consider development of water resources in a hypothetical basin. Before humans use water in a basin, the basin’s renewable resource could be measured by the amount of water discharging into the sea. With human settlement came more demand for water. A river diversion structure to deliver water for their agricultural and domestic needs was constructed. The amount of water diverted in this case is the primary water supply. Some of this primary water is consumed through evaporation, while the remainder drains back into the river. As time passed by in our basin, users upstream constructed a dam and reservoir upstream of the first diversion structure. Part of the water released from the dam was consumed while the remainder drained back into the river as return flow. River water mixed with return flows was again diverted at the original diversion structure. By constructing the reservoir, more primary water was developed and the downstream diversion structure served the purpose of recycling water. Total deliveries are deliveries from the reservoir plus the downstream structure and are equal to the primary deliveries plus the recycled water. Development of the basin continued until nearly all of the utilizable water resource was exploited. Water consumption increased with expansion of irrigated land and increasing supplies to domestic and industrial uses. Even at full development, the full amount of primary water was not evaporated and there was still outflow to the sea. Deterioration of water quality became a major concern with increasing recycling and flushing of pollutants. Basin development is illustrated in figure 5. Over time, more primary water supply is developed through construction activities. The amount of consumption of primary water steadily rises over time as more water is diverted and consumed by various uses. At full development, consumption approaches the primary water supply. A high degree of consumption of primary water is rarely desirable because of the need to flush salts from agriculture, to allow diverted water to serve environmental purposes and to wash out pollutants. In most cases, some water supply should be considered committed to environmental or other uses. Figure 5. Water resources development in a basin.

Time

Utilizable

Renewable

EvaporatedPrimary

Committed

26

Increasing Production and Productivity of Water Resources A key to mitigating problems of water scarcity is increasing productivity of water. Let us focus on agriculture, as there is a tremendous need to produce more food, mainly grains, from our water resources. There are generally three paths of generating more agricultural output per unit of country-level utilizable water resources: 1. Develop and consume more primary water2 (increase in development of facilities). 2. Consume more of the developed primary water for beneficial purposes (increase in water

savings and “basin efficiency”). 3. Produce more output3 per unit of water consumed (increase in productivity). The first two cases are illustrated in the basin example. More water resources infrastructure was developed so that more water could be consumed by uses deemed beneficial. The third case is dealt with in this section—that of increasing the unit productivity of water. In agriculture, water is consumed by crop evapotranspiration (ET), essential for crop production. We can define a Physical Water Productivity (PWP) term as: PWP = kg/ET For cereal grains, productivity ranges between 0.2 kg/m3 and 1.5 kg/m3. As a rule of thumb, a reasonable level of water productivity for grains is about 1 kg/m3. If the demand for grains grows by 50 percent in a country by 2025, one way to match this increased demand is to increase water productivity by 50 percent. Over time, improved agricultural practices have increased water productivity. Yield (kg/ha) increases that do not require additional supplies increase productivity of water. For example, the use of shorter duration varieties can produce the same yield with a lower total evapotranspiration and thus higher water productivity. Reducing non-beneficial evaporation through drip irrigation or mulching can also increase water productivity. Again, consider our hypothetical basin example, this time including the effects of water productivity. Over time, with increasing water productivity in agriculture, overall physical production grows (figure 6). At full development and at situations where country consumptive use levels are quite high, the only feasible means of increasing overall production is to increase water productivity in agriculture. When water consumed approaches the utilizable water resource, we say the basin is closing. The situation for agriculture is more complicated in a “closing basin” as competition from other sectors increases. Domestic, industrial and environmental uses are likely to increase consumption and agriculture’s share of consumption is likely to decrease. Another very likely situation is that consumption by water for nature will decrease with priority given to human uses. An additional complication during basin closing is the buildup of salinity and pollutants. It is possible that water productivity in agriculture will decrease due to these factors. In situations of scarcity there are complications of decreasing levels of consumption by agriculture, compounded by pollution and salinity buildup. The global challenge is to find ways to increase productivity of water (see box 1) in spite of increasing competition.

2This results in increases in overall production, but not necessarily in productivity per unit of water consumed. 3Output could be measured in mass of produce or in economic terms as net value.

27

Figure 6. Increase in productivity of water consumed overtime.

Time

Utilizable

Renewable

ConsumedPrimary

Production

Increasing Productivity of Water through Improved Water Management The most significant increases in productivity of water have come from improved plant varieties and agronomic practices. More crop per drop is achieved by introducing shorter-duration and higher-yielding crop varieties. Increased use of fertilizers has pushed up yield and the corresponding water productivity. Combined with a stable irrigation water supply, productivity in agriculture has risen dramatically over the last 50 years. There remains scope for improvement. In many areas, potential productivity is not realized and this is in large part due to poor irrigation management. Without stable irrigation, farmers cannot take advantage of the production potential. Figure 7 shows yield and water productivity in three different locations with somewhat similar environments. India’s Bhakra irrigation system constitutes a major part of India’s breadbasket and is situated across the border from the Pakistani Punjab. The Imperial Valley in California is situated in a desert environment like the Punjab. The spread in wheat yields in these areas is from 2 to 6 tons per hectare, with a corresponding spread in water productivity from 0.5 kg/ET to 1.3 kg/ET. Within the Pakistani Punjab, there is great variability in yield with some farmers achieving productivity levels as high as those in California and some farmers well below the average. Of course, production levels are dependent on environmental, market, soil and other conditions that are not equal across sites. In spite of this there appears to be scope to manage resources to achieve greater productivity. Raising the yield potential is a problem that is relatively more important for developed countries. For many developing countries, reaching the existing potential yield is a more pressing problem. How water is managed within irrigation and river basins is one major constraining factor for many farmers in the developing countries.

28

This section outlines means to improve water management to increase productivity. Water management improvements can increase water availability, enhance the environment and directly influence the unit productivity of water. Water Recycling. While recycling occurs naturally in all river basins, increasing attention is being paid to recycling as an integral part of water management. For example, farmers in Egypt and other countries place small pumps in drainage ditches to recycle water and the irrigation agency blends drainage water with freshwater to increase the useable supplies. Millions of shallow tube wells have been developed in the Indo-Gangetic plains that are recycling water effectively, capturing and using water before it flows out of the basin.

Box 1. Increasing productivity of water consumed in agriculture • Agronomic Practices Crop varietal improvement. Plant breeding plays an important role in developing varieties that yield more mass per unit transpiration. For example, by reducing growth period while keeping the same yield, production per unit evapotranspiration increases. Changing the partitioning between green matter and grain yield increases water productivity when evapotranspiration is kept constant. Crop substitution. Switching from a more to a less water-consuming crop or switching to a crop with higher-economic or higher-physical productivity per unit of water consumed by evapotranspiration improves productivity of water. Improved cultural practices. Better soil management, fertilization, pest and weed control often leads to more productivity of land and often water consumed. • Water Management Practices Improved water supply management. Better timing of water supplies can reduce stress at critical crop-growth periods, leading to increased yields. By increasing reliability of supply, farmers tend to invest more in other agricultural inputs leading to higher output per unit of water. Improved on-farm water management. Controlling salinity through water management practices at project or field level can prevent reductions in water productivity. Mulching, direct seeding of rice and greenhouses reduce evaporation. Stopping water deliveries during rainy periods, or raising field bunds in rice areas captures more rain that may otherwise flow to sinks. Deficit, supplemental and precision irrigation. With sufficient water control, it is possible to utilize more productive on-farm practices. Deficit irrigation is aimed at increasing productivity per unit of water by irrigation strategies that do not meet full evaporative requirements. Irrigation supplementing rainfall can improve productivity of water when a limited supply of water is made available to crops at critical periods. Precision irrigation, including drip, sprinkler and level basins, reduces non-beneficial evaporation, applies water uniformly to crops, reduces stress and thus can lead to increases in water productivity. Reallocating of water from lower- to higher-value uses, for example from agriculture to municipal and industrial uses or from low-value crops to high-value crops, can increase the economic productivity or value of water. As a result of such reallocation, downstream commitments may change. In turn, reallocation of water can have serious legal, equity and other social considerations that must be addressed.

29

Another example is the finding of one of IWMI’s scientists who developed a rather surprising theory, based on water balance studies, that the amount of irrigated land in the downstream portion of Kirindi Oya river basin of Sri Lanka could be greatly increased. This could be done with no more water and positive environmental benefits, by installing simple water recycling structures in the drainage system (Renault 1999). The Irrigation Department installed the structure and it indeed happened! One season of agricultural production has justified the costs. Through recycling plus better supply management, 6,000 poor farm families have seen their lives improve over the past 2 years. We believe that there are many opportunities for similar achievements elsewhere where there is scope for water savings. Figure 7. Yields and water productivity of wheat in California, Pakistan and India.

Sources: Molden et al. 1999; Bastiaanssen et al. 1999; Mayberry et al. 1996; IID 1996. Reliability in Supplies. One basic principle in irrigation is to deliver a reliable supply of water. If farmers do not have a reliable supply, they do not know when the next irrigation is coming, they do not know how much water will come, and they do not know if there will be enough water for their crops. In this uncertain environment, farmers will not make investments in seeds, fertilizers, and land preparation, and consequently yields will suffer. At Mahi Kadana in Gujarat, India, changes in the way water was delivered increased the water productivity by almost 100 percent (Sakhtivadivel and Gulati 1997). Before the change, a situation existed where water was supposed to be available at all times in the canal. But in reality, the amount of water in canals was so low, that farmers could not get their supply. A change was made to a rotational system with full supply flow, where farmers would receive water reliably at regular intervals. As another example, one of the reasons that the Bhakra system in India is more productive than many other irrigation systems is said to be its reliable water delivery program.

Wheat yields and water productivity

0

2

4

6

8

Imperial Valley Bhakra, India Punjab, Pakistan

ton/

ha

0.00

0.50

1.00

1.50

kg/m

3ton/ha yield kg/m3ET crop per drop

30

Poor reliability is often linked to weak institutions, incapable of operating infrastructure. A means of improving irrigation, even in the most basic irrigation systems, is to find ways of increasing reliability. To do this requires the right mix of manageable technologies and organizational skills required for water management using these technologies. Precision Irrigation. Another important means of increasing the productivity of water is “precision irrigation,” as we call it. As everyone knows, it is important to deliver the right amount of water to the crop at the right time. But it has only been in the past two decades or so that we have begun to see just how important it is to do this precisely: exactly in the right amount and at the right time. The various forms of precision irrigation—mainly sprinkler, drip irrigation systems and dead-level basins—increase yields over good but ordinary irrigation systems by 20 to 70 percent, depending on the crop and other conditions and they do so with much less water diverted to the crop. Clearly, if we could convert the world to precision irrigation systems, the need for additional water for irrigation would decrease dramatically. There are many opportunities and difficulties on this path that need to be explored. For example, one of the simplest and most effective kinds of precision irrigation is the “dead-level basin.” Here laser-levelers are used to make fields as flat as possible. The results of dead-level basins, in terms of yields and water savings, are about the same as those of sprinkler systems. Yet, they can be installed at low cost, around US$300 per hectare, on even the smallest fields. The reason that they are so effective seems to be that because of poor water distribution over ordinary fields about one-third of the crop suffers frequent drought, another one-third is often waterlogged and only the remaining one-third has something like the right amount of water most of the time! Feedback Versus Feed-Forward Systems. At various times in a crop’s growth cycle, water stress can be particularly damaging. For this reason, the cybernetic principle of feedback, responding quickly just before the crop is in danger of reaching a critical water supply state is very important. Many irrigation systems operate on the “feed-forward” principle of trying to deliver water in anticipation of crop water needs, which is fraught with errors. This is why tube-well irrigation systems in India typically produce yields that are two to three times greater than canal irrigation systems. Tube well water is available virtually on the farmer’s demand while in most Indian canal systems farmers must wait for their turn. When farmers have tanks or ponds close to their field, they can store water and apply it when it is needed. This is how the Chinese “melons-on-the-vine system” of canals feeding small tanks operates. Similarly, groundwater storage and pumping provide an ideal feedback system. Many Egyptian farmers pump from drains for an additional source of water to be more responsive to crop’s needs. Successful feedback systems have proven more difficult to obtain in larger canal systems, as very tight management and infrastructure that provide flexibility in delivery are required. Without adequate institutional capability, these systems often fail. However, there are several successful examples of “on-demand” canal delivery systems throughout the world. Success requires the right blend of infrastructure and institutions. Most of the potential gains of precision irrigation systems are lost without a supply of water when it is needed. With some major exceptions, most of the canal irrigation systems of India and other countries are about as reliable as rainfall. Even small tank systems, which could be managed precisely, often are not. In a case in Sri Lanka, the releases from the tank are highly and positively correlated with rainfall, with the result of oversupply during periods of rainfall and unnecessary shortages during subsequent dry seasons. As in most cases of poor management, one cannot believe how bad it can be until one actually sees it in action—or inaction, as the case may be.

31

Low-Cost Precision Irrigation Technologies for Rain-fed Areas. There is considerable scope for increasing the productivity of, otherwise rain-fed agriculture by the application of supplemental irrigation at critical stages in the crop cycle. Such interventions will rely on the use of precision irrigation technologies combined with water harvesting or groundwater use. Providing a limited supply of water at the right time can save harvests and dramatically increases yields. Low-cost versions of precision technologies, based on those used in commercial large-scale agriculture, provide an opportunity for fighting poverty and increasing productivity. In South Asia and Africa, very low-cost bucket and drip sets are becoming increasingly popular with farmers. In areas where shallow groundwater is plentiful, thousands of poor farmers in Bangladesh have used low-cost treadle pumps to supply water for crops for their own food security and additional income (Shah 1999). But we do not yet understand the potential, or the mechanisms, for large-scale uptake of these technologies. Policies, Institutions and Incentives. Various water management practices we have discussed are appropriate for various settings. For any of these practices to work requires the right set of incentives for farmers—a function of policies and institutions. One difficulty faced is that, as competition for water becomes more intense, how water is used in one part of a basin impinges on how it is used elsewhere in the basin. This requires that a set of laws, regulations and organizations be coordinated to basin-wide water resources. This is a topic of immense importance and complexity, but let it be said that we do not have ready-made solutions to change institutions for managing water in more productive ways. Water Productivity and Development of Water Resources There is a direct relation between increase in water productivity and the need for future water developments. The more productive agriculture becomes, the less the need for water resources development. To illustrate, consider water needs for India in 2025 (figure 8). In 1995, average grain yields were 2.7 tons per hectare. About 600 cubic kilometers of water were diverted to irrigation uses. Considering the growth in population and improvements in diet, diversion requirements in 2025 were calculated for different settings. If there is no increase in grain yield, India will have to double diversions to irrigation. In the IWMI base case, yield increases by 3.6 tons per hectare and associated diversion requirements increase by 14 percent to about 820 cubic kilometers. If average grain yields increase by 70 percent, no more increases in water diverted to irrigation will be required. With the high yield increase, there will be no more need to expand area or intensity and no more need for infrastructure to develop new water for irrigation. There would be a need to improve and add infrastructure to provide for more water control. Achieving this yield level would require major improvements in agronomic and water management practices. This would require significant shifts in agricultural practices and policies. Conclusions The extent to which water constrains agricultural production and the ability of the world to feed itself are functions of productivity of water. As we produce more with the same amount of water, the need for new developments of water resources infrastructure and competition for water will become less, local food security can be enhanced and there will be more water available for nature, domestic and industrial uses.

32

Figure 8. Relationship between yields and increase in diversions to agriculture for India under different scenarios.

There is potential for increases in water productivity. In many developing countries, the problem is to reach the physically potential yield. There appears scope to double water productivity in many cases, but some serious water management constraints have to be overcome. There are many opportunities for improving the management of water leading to increased productivity. The first task is to understand where these are applicable. For example, with increased consumption of water in basins, the scope for water savings through increases in efficiency in irrigation systems may be much more limited than thought. Recycling is widely prevalent and still has potential for saving water. Gains are possible by providing more reliable supplies, through precision technology and through management systems that include feedback. Supplemental irrigation with low-cost precision technologies offers a means for poor farmers to produce more. With increased and increasing competition for water, implementation of solutions will require significant changes in the institutions that are responsible for accomplishing the business of water management.

0

200

400

600

800

1000

1200

1400

1995 2025- No YieldIncrease

2025 - IWMI BaseCase

2025 - High YieldIncrease

Div

ersi

on

s to

Irrr

igat

ion

(km

3)

2.7 tons/ha

3.6 tons/ha

2.7 tons/ha 4.6 tons/ha

23

CHAPTER 5

The PODIUM Model

Charlotte de Fraiture and Upali Amarasinghe 1. Background In 1997, IWMI developed a model for estimating future water supply and demand for the year 2025. Comparing water supply and demand at country level, the model predicted water scarcity based on two factors: the percent increase in water diversions over the period 1990 and 2025; and the projected water diversions expressed as a percentage of annual renewable water resources. Building on these experiences IWMI upgraded the model concepts by refining algorithms and data inputs, resulting in the PODIUM model. The PODIUM model is now produced in two versions: the country model and the global model (refer to box 2). In this report, we focus at the global level model and its applications for the Vision 2025 exercise. At present data from some 100 major countries covering 96 percent of the world’s population are entered in the spreadsheet. In this application of PODIUM the model is used to test the three scenarios formulated by the Vision 2025 Scenario Panel. In all scenarios year 1995 is taken to represent the actual situation while year 2025 stands for the future situation. Box 2. The PODIUM model One of the main features of the PODIUM model is that all main variables and assumptions are made explicit and can be changed easily by the user. This feature makes the model an excellent tool for scenario testing. PODIUM works at any scale —a country, basin, subbasin, or the globe. For global projections, PODIUM aggregates results country by country to the global scale . The country is considered the most appropriate unit of analysis for the Vision 2025 exercise because policies about food and trade are made on a country basis. The PODIUM model is produced in two versions: the country-level model and the global-level model. Country model Analyzes food and water issues at country level. Its objective is to raise policy makers’ awareness concerning future food and water requirements and water scarcity. The model assists policy makers in analyzing the quantitative aspects of future food and water demands and allows them to test user-defined scenarios, which could form the basis of policy formulation. Being developed for decision makers—not necessarily experts in computer programming—the model has a very user-friendly interface and a transparent structure. The country model also proved useful for educational purposes. The country model is available at IWMI’s website (at http:\\www.iwmi.org). Global model The global level model is an Excel spreadsheet used for aggregating country-level analyses into global or regional groups. The results from the country model are entered into the global model as they become available.

24

In the following paragraphs the main model components will be described, followed by a section discussing scenarios and modeling results. Water Demand Variables Water demand computations include the agriculture, domestic and industrial use sectors. Actual and future water uses in agriculture—being IWMI’s expertise and accounting for 80 percent of actual water deliveries worldwide—are analyzed in detail. The projections for domestic and industrial uses are described only briefly. Agriculture The computation process comprises the following three steps.

• The first step estimates national food requirements based on assumptions concerning population growth, daily calorie intake and composition of diets.

• In the second step, the projected cereal production is estimated based on the expected yields and cultivated area, under both irrigated and rain-fed conditions. Projected cereal production is then compared to the required production as computed in step one.

• Finally, the third step converts the projected food production into water demand. The computed water demand is compared to actual water diversions in the base year (1995) and the available renewable water resources.

Step one: Food demand The main driving forces behind food demand are population growth and diet composition. In this application of the PODIUM model we use the UN medium projection (revision 1998) for population growth, unless otherwise specified in the scenarios. Diet composition is described by daily calorie intake per capita, and the consumption of cereal and animal products expressed as percentage of the total calorie intake. Figure 9. Cereal requirement variables in the PODIUM.

Cereal requirements

Population

Diets - Daily calorie intake - % Cereal products - % Animal products

25

The diet composition for the base year, 1995 is derived from the FAOSTAT database, taking the average over the years 1994–1996. Projections regarding calorie intake and diet composition are based on the results obtained from IFPRI’s IMPACT model, using the Vision 2025 scenarios. In the present version of the PODIUM model cereals are used as a proxy for overall food consumption and production patterns. In the Asian context this gives satisfactory results because the major part of the diet consists of cereals and most irrigated land is devoted to cereal crops. In Middle Eastern and African countries, having different consumption and cropping patterns, this may cause some distortions. IWMI is currently expanding the model to allow for a variety of food crops. Step two: Food production PODIUM considers food production under both irrigated and rain-fed conditions. The main driving forces behind agricultural production include developments in total harvested area, expansion of the irrigated area and increasing crop yields under rain-fed and irrigated conditions. For the base year the total harvested area is derived from FAOSTAT database taking the average of 1994–1996. For most countries, irrigated yields and areas were taken from FAO’s study “Agriculture towards 2015-2030” (FAO work in progress 1999). In some cases, we used data from other local sources. Rain-fed yields and areas are computed by subtracting irrigated production from the total production as given in the FAOSTAT database. As a check the computed values for rain-fed production were compared with available data. In case of major discrepancies the irrigated production variables were checked and revised as necessary. Projections of yield and areas under both rain-fed and irrigated conditions follow the guidelines as described in the Vision scenarios for different country groups (table). Growth rates of yields for individual countries, were adapted taking into account observed trends in the previous 30 years—analyzing time-series in the FAOSTAT database—and absolute yield levels. In countries where actual yield levels are low it is assumed that yields will grow faster than in countries where yields are already quite high. The expansion of the irrigated area broadly follows the guidelines describing the Vision 2025 scenarios for the different country groups (table 1). For individual countries growth rates are adapted taking into account observed trends over the last 30 years and the irrigation potential. Irrigation potential is limited by availability of suitable land and/or water resources. For some countries, estimated irrigation potential is reported in the AQUASTAT database, and for other countries, we made use of IWMI’s and other specialists’ knowledge of individual countries. Projections of total food production—the sum of rain-fed and irrigated production—are compared to the required amount of food as determined in the first computation step. Countries facing a deficit are regarded as food importing countries while countries with surplus production are food exporters. A schematic overview of variables is given in figure 10. In the previous step, the irrigated area in the base year and for 2025 under different scenarios were determined. The net irrigation water requirements are calculated by subtracting the effective rainfall from the crop evaporation, for two different seasons. Effective rainfall is approximated as a percentage of the 75 percent probable rainfall. This percentage varies between 65 percent and 85 percent depending on the rainfall pattern. Crop evaporation is determined by multiplying the reference ET with the crop factor (Kc) as given by FAO (Doorenbos and Pruitt 1977). The reference ET is derived from the IWMI atlas for those countries where data grids are available; in the other countries data published by Hargreaves were used. The net irrigation

26

Figure 10. Cereal production variables in the PODIUM. Step three: Water for food production requirements (NET) are converted in volume by multiplying with the irrigated area. It is assumed that NET in the base year is the same as in 2025.1 The model takes into consideration evaporation by crops plus ‘non-beneficial’ uses such as evaporation from bare soil and from water bodies (canals and reservoirs). This is accounted for by a factor expressing non-beneficial evaporation as percentage of NET. This factor varies between 10 and 20 percent. Net irrigation requirements are converted into total irrigation diversions by dividing NET plus percolation by irrigation efficiency. Irrigation efficiencies are estimated from available global databases (for detailed technical descriptions of the estimation process please refer to IWMI Research Report 19, annex B). The growth in irrigation efficiency is described in the different scenarios. We added the condition that irrigation efficiency cannot exceed 75 percent (this is the maximum achievable efficiency). For paddy areas an allocation is made for percolation losses. For most countries percolation losses are estimated at 200 mm per season, unless other information is available. Percolation in paddy is assumed to remain at a constant level in 2025. Note that the percolation losses are not regarded as ‘consumptive use’ because generally they can be reused later by tube-well irrigation or downstream uses. The PODIUM model distinguishes between total diversions and primary water supply. The difference between the two is the recycling of return flow (drainage water and recharge to groundwater). Drainage and seepage water is not ‘lost’ but is potentially reusable for other uses. Although this principle—the IWMI paradigm as described in Seckler 1997 and Perry 1999—is widely accepted it is difficult to quantify the amount of water that is actually recycled. In the PODIUM model the recycling factor is derived by estimating the amount of water that flows out of the system, expressed as a percentage of the total diversions (TD). Primary water supply (PWS) is computed by adding the amount of water evaporated by crops plus non-beneficial evaporation plus the outflow. The amount of water recycled is then calculated by subtracting total diversions from the primary water supply (refer to formulas in annex). 1The PODIUM model allows for different values of NET in 2025, for example due to global climate change. But since projections regarding climate change are full of uncertainties and are considered unlikely to have major impacts in the coming 25 years, these effects were not taken into account in this application of the PODIUM model.

Irrigated production

Total cereal production

Irrigated area Irrigated yield Rain-fed area Rain-fed yield

Rain-fed production Irrigated production

27

Finally, the PODIUM model distinguishes a third category of water: evaporatively used water or evaporated water (E). This is the amount of water depleted by crop evaporation plus non-beneficial evaporation. The concepts are schematically depicted in figure 11. Formulas are given in box 3. Domestic and Industrial Water Use Domestic and industrial diversions for the base year 1995 are taken from international databases published by the World Resources Institute. The growth in domestic and industrial diversions is computed as follows: First the relationships of 1995 per capita GDP and 1995 per capita domestic and industrial withdrawals are established. The per capita GDP projections of each country are calculated using the GDP growth rates of the 18 regions for the World Water Vision scenarios (Strzepek, Polestar model). The projections of per capita GDP and the per capita diversions and GDP relationships of 1995 are used in estimating the 2025 domestic and industrial diversions. Figure 11a. Water requirement variables in the PODIUM.

NET

Crop evaporation Effective rain

Percolation rate in paddy areas

Non-beneficial evaporation

Classical efficiency

Total irrigation diversions (TD)

minus Recycling

Diversions for domestic

Diversions for industrial

Primary water supply (PWS)

28

Figure 11b. Recycling. A distinction is made between diversions and evaporative use: not all water diverted is actually consumed but a large portion flows back into the system and is potentially available for reuse. This is accounted for by the ‘evaporation factor’ (EFuse). For domestic uses this factor (EFdom) is around 20 percent (i.e., 20% of the water diverted is evaporated while 80% flows back to the system). For industrial use this is 10 percent. It is assumed that these factors remain constant in the year 2025. Renewable and Utilizable Water Resources The PODIUM model is not a hydrological model that can simulate water availability at water basin or country level. Instead it relies on data available from international databases such as made available by Shiklomanov (1999). Not all renewable water resources can be exploited. A part may come in floods that drain off directly to the ocean before it can be captured and a part may percolate to saline groundwater that renders it unsuitable for use. The PODIUM model distinguishes between renewable water resources (RWR) and utilizable water resources (UWR) using the ‘potential utilization factor’ (PUF). This is the percentage of the annually renewable water resources that is potentially utilizable. For some countries this amount is reported in the AQUASTAT database or other data sources while for other countries we made use of IWMI’s and others’ knowledge of individual countries. In this exercise it is assumed that annually renewable and potentially utilizable water resources in the year 2025 remain at 1995 levels. Indices of Water Use and Scarcity Primary water supply (PWS) is the amount of water that can be evaporated given the state of development of a country’s water resources. PWS represents the first-time diversions of water and represents an upper bound to how much of the available water resources can be consumed.

Total diversions

Agriculture, domestic, industry Evaporative use

Return flow

Agriculture, domestic, industry Evaporative use

Return flow

etc.

Outflow to sea or downstream countries

Total use (E)

29

Water diverted to a use that is not consumed either flows to a sink, or reenters the river or man-made network and is recycled. Total diversions, often reported as withdrawals comprised primary water plus recycled water. Total water deliveries are dependent on how much water is recycled. Evaporative use is defined in PODIUM as the part of water delivered to a use that is evaporated. A country evaporative use factor is defined as the consumption divided by the primary water supply. PODIUM considers evaporative use in irrigation, domestic and industrial uses separately. From the concepts explained in this chapter two important factors related to water scarcity are identified, i.e., evaporation factor for a country and degree of development. The first factor is when consumption approaches primary water supply. PODIUM uses a country consumptive use factor to indicate consumption of primary supplies.

a. Evaporation Factor (EFcountry) is defined by total evaporation (E) divided by primary water supply (PWS) for all sectors.

EFcountry = E/PWS Total evaporation includes crop evaporation, non-beneficial evaporation and evaporative use in the domestic and industrial sectors. Primary diversions include diversions for agriculture, industry and domestic uses. It is advisable that the country evaporation factor does not exceed 75 percent. An allowance of 25 percent is made to cover water needed to meet leaching requirements and to replenish groundwater. If the country evaporation factor exceeds 75 percent environmental problems such as salinization and groundwater table decline are anticipated. When the evaporation factor is high, one solution to scarcity is to construct more diversion and storage facilities. Development of primary water can continue until it reaches the limit set by utilizable water resources. PODIUM uses the degree of development to indicate the state of water resources development in a country.

b. Degree of Development (DoD) is defined as the primary water diversions for all sectors divided by utilizable water resources. Primary diversions include diversions for agriculture, industry and domestic uses.

DoD = PWS/UWR Absolute water scarcity occurs when the degree of development is greater than 60 percent, in other words more than 60 percent of the utilizable water resources are diverted. The cut-off value is set at 60 percent to make adequate allowance for water for environmental uses (wetlands, minimum required flow in rivers). Unfortunately, little or no information is available on water requirements for environmental uses. To be on the safe side, 40 percent of the utilizable water resources is allocated for those uses; however this cut-off value can be modified as more information becomes available2.

2 The value of 60 percent is chosen based on experiences in Israel. According to PODIUM estimates, in 1995 Israel used 61 percent of the utilizable water resources. Israel can be seen as an extreme water-short country where huge water conserving efforts and investments took place. It is set as standard of maximum achievable water development.

30

Some countries have achieved only a small degree of development, but do not have the financial resources to develop more. We consider these countries as economically water scarce. Other countries have developed to their maximum extent. One situation where water is considered scarce is when water resources development has reached an upper limit, and there is insufficient water to meet basic food and domestic needs. This is a condition of absolute scarcity. Another important condition of scarcity is when there is insufficient development of utilizable water. In this case, basic needs of people are not met, but there is water that could be developed because of lack of capacity and financial resources. Increasing Evaporative Use of Developed Primary Supplies At a use or service level of analysis, we consider an irrigation system, domestic water supply, or an industry. We can define a use level Evaporation Factor (EFuse) as the ratio of water consumed by evaporation to the total water delivered to the use. In irrigation, this is known as irrigation efficiency and values typically range between 20 and 70 percent. Irrigation with pipes, sprinklers and lined canals, could reach the higher level. In other cases, use of surface irrigation, with unlined canals, result in an EF at the lower range say between 20 and 40 percent3. The Evaporation Factor for cities and industry is often between 10 to 20 percent. In some cases this can be considerably higher. Where industries are required to keep water clean the factor can be quite high as industries recycle the water internally, and require less deliveries of water. If much city water is used by vegetation such as lawns or trees, the factor can be higher. What happens to the water that is not consumed? In some cases, it cannot be reused because it flows directly to a sea or other sink. In other cases, it is available for reuse downstream. Diversion structures along rivers downstream of major reservoirs often capture drainage return flows. Cascades of reservoirs are common in irrigated areas. Groundwater pumping and conjunctive use are increasingly an extremely important mechanism for recycling. Water “inefficiently” used in irrigation recharges aquifers where it is stored and pumped later. To capture this relation between country and project-level evaporative use, we relate country evaporative use to use level consumptive use by: EFcountry = M x average EFuse, where M is a multiplier that is a function of how water is reused. The multiplier is defined as: M = EFcountry /average EFuse = Total Diversions/Primary Water Supply To increase country evaporative use of primary water, either the multiplier or the use level consumption factor must be increased. To increase the multiplier, downstream diversion, pumping and storage facilities are required. Pumping return flows from drains or groundwater increases the multiplier. Increasing EFcountry requires decreasing deliveries to uses, or increasing evaporation by intended uses. In irrigation, project-scale efficiency increases require investments in lining, pipes and precision on farm application by sprinkler, drip and level basin irrigation. In summary, there are two ways of increasing country evaporative use with existing primary facilities:

3 A lower value of EF is not necessarily a worse situation as we shall see in the discussion of recycling. In some rice areas this factor can be as low as 10 percent, but irrigation can be quite effective. In these cases, irrigators deliver water to meet deep percolation and seepage requirements. We have avoided using project or application efficiency as the implication is that a larger efficiency is better.

31

1. Increase use level evaporative use factor 2. Increase recycling

Increasing efficiency at irrigation systems could save water only if there is limited recycling. Based on observations by IWMI and others, the degree of recycling of water resources by all sectors is already high, much more than commonly perceived. This means that the country evaporative use factor is higher than thought, and the scope for water savings smaller than anticipated. A possible explanation for this is that investments in recycling water (increasing M) are typically less expensive than increases in project efficiency; can be done with smaller, more divisible technologies that do not necessarily require centralized decisions on investments; are easier to manage; and in many cases quite productive. PODIUM takes into consideration use level evaporation factor, and recycling. The evaporative use factors for different uses are inputs to the model. Recycling is handled by a factor referred to as the multiplier. The user adjusts the multiplier by estimating the percentage of water diverted to each use that is available for reuse or flows to sinks.

23

CHAPTER 6

Water Scarcity and the Role of Storage in Development1

Andrew Keller, R. Sakthivadivel and David Seckler2 The annual precipitation for most of India occurs in just 100 hours. The other 8,660 hours of the year are dry.3 Introduction By 2025, one-third of the population of the developing world will face severe water shortages (Seckler, Molden, and Barker 1998). Yet, even in many water-scarce regions, large amounts of water annually flood out to the sea. Some of this floodwater is committed flow to flush salt and other harmful products out of the system and maintain the ecological aspects of estuaries and coastal areas (Molden 1997). But in many cases, the floodwater is not fully utilized; and, of course, the floods themselves can do a great deal of harm. This problem is epitomized in India where, as noted above, annual precipitation is concentrated in the four monsoonal months and then in only a few hours of those months. Because of the sporadic spatial and temporal distribution of precipitation, the only way water supply can be controlled to match demand is through storage. This is true whether the demand is for natural processes or human needs. In natural systems, precipitation may be intercepted by vegetation and temporarily stored on plant surfaces and on the soil surface. When water infiltrates the ground, it is stored in the soil and may percolate to groundwater storage. On land, surface water is stored in watercourses, lakes and other water bodies and in frozen form as snow and ice. Man can create and enhance water storage by such activities as water conservation tillage, constructing dams and dikes to impound water and artificially recharging groundwater. Regardless of the method or type of storage, the purpose is to capture water when and where its marginal value is low—or, as in the case of floods, even negative—and reallocate it to times and places where its marginal value is high. Here, “marginal value” includes all of the economic, social and environmental values of water. As competition for water increases in many regions of the world, an increasingly higher proportion of normal flow of water is likely to be consumed and the risk of shortages in periods of low flow will increase. For this reason, the need for additional storage as a proportion of the total water consumed will increase in the future. In evaluating various kinds of water storage systems, it is useful to think in terms of three distinct hydrological situations in river basins (Seckler 1996; Keller, Keller, and Seckler 1996; and Perry 1998): • Open basins are those that have an excess of water, over and above all committed ecological

and environmental requirements, flowing to the seas, saline aquifers, or similar sinks4 during 1Andrew Keller originally presented the subject of this paper at the 1998 World Bank Water Week Conference, December 15, 1998, Annapolis, Maryland in a session on dams. The title of that presentation was "Water Scarcity and the Role of Dams in Development." For this paper, we changed the title, substituting the broader term "storage" for "dams," to reflect the importance of increasing storage, regardless of type, to address water scarcity. 2Fellow, Senior Irrigation and Water Management Specialist, and Director General of IWMI, respectively. We are grateful to Chris Perry, David Molden and Jeremy Bird for comments on earlier drafts of this paper and to Asit Biswas for review of the final draft. 3Anil Agarwal, Centre for Science and the Environment, New Delhi, India, 1998. 4Flow to sinks is an economic as well as a physical concept. Sometimes it can be prevented, but at an unacceptably high cost.

24

the low-flow season of the year. In open basins, the excess water has no opportunity cost, its marginal value is zero or negative and all that needs to be considered is the cost of utilizing more of the water relative to the benefits of doing so.

• Closed basins are at the opposite end of the spectrum. Here there is no excess water flowing into sinks at any time of the year. This case represents a zero-sum game in physical terms. Additional water use by one party means reduced use by another party. Here the only options are to: reduce nonproductive evaporation and transpiration losses out of the basin, for example by reducing the non-beneficial uses by weeds, shrubs and trees (as is being done in South Africa); increase the total productivity of water by reallocating water from lower- to higher-valued uses; minimize effective water depletion due to salinization and pollution and losses to sinks; and augment water supply with trans-basin diversions or desalinization (Keller, Keller, and Davids 1998).

• Semi-closed basins represent the major opportunities for adding value to water through storage. In these basins, there is no excess outflow to sinks during the low-flow season, but there is excess outflow during the high-flow season. Thus, storing water and reallocating it between seasons can achieve potentially large increases in the value of water.

Two of the largest river basins, the Amazon and the Zaire, are open basins, but there are very few open basins left in the highly populated arid regions of the world. In China, for example, the Yangtze River, in the wet south, is open but the Yellow River in the arid north is closed. China is now creating trans-basin diversions from the Yangtze to the Yellow River basin to alleviate the problem of water shortage in the north. Other examples of completely closed rivers are the Colorado River in the US and Mexico and the Cauvery in south India. Such large and important rivers as the Indus in Pakistan, the Narmada in India and the Ganges of Nepal, India and Bangladesh are semi-closed. The Ganges represents a classical problem of international waters, with the catchment areas and major new dam sites largely in Nepal, a major need for water in the low-flow season in Bangladesh, flood control storages in the head reaches and a large demand for hydropower and irrigation in India. Development of the Narmada River (the Sardar Sarovar project) in India has become an international cause celebre because of concerns over resettlement and environmental issues, notwithstanding its enormous economic and social benefits (Seckler 1992). There are four major ways of storing water—in the soil profile, in underground aquifers, in small reservoirs5 and in large reservoirs behind large dams. Storage in the soil profile is extremely important for crop production, but it is relatively short-term storage, often only sufficient for a period of days. Here we concentrate on the three kinds of technologies that store water for periods of months, in small reservoirs, or years, in aquifers and large reservoirs. These three technologies are compared from the hydrological, operational and economic standpoints. Some of the environmental aspects of these options are also mentioned, but these aspects are very location-specific and are not discussed in detail. The two principle conclusions of this analysis are: • Aquifers and small and large reservoirs all serve an indispensable role in water storage and

each technology has strong comparative advantages under specific conditions of time and place.

• Where it is possible to do so, substantial gains can be achieved by combining all three storage technologies in an integrated system.

5Structures with height less than 15 m and volume less than 0.75 mcm.

25

Kinds of Storage Table 7 summarizes the comparative advantages, limitations and essential issues associated with aquifers, small reservoirs and large reservoirs. Table 7. Comparative advantages, limitations and key issues associated with groundwater, small reservoir and large dam water storage.

Groundwater Storage Small Surface Water Reservoirs Large Dam Reservoirs Advantages: Advantages: Advantages: Little evaporation loss Ubiquitous distribution Operational efficiency Available on demand Water quality

Ease of operation Responsive to rainfall Multiple use Groundwater recharge

Large, reliable yield Carryover capacity Low cost per m3 water stored Multipurpose Flood control and hydropower Groundwater recharge

Limitations: Limitations: Limitations: Slow recharge rate Groundwater contamination Cost for extracting Recoverable fraction

High evaporation loss fraction Relatively high unit cost Absence of over-year storage

Complexity of operations Siting High initial investment cost Time needed to plan and construct

Key Issues: Key Issues: Key Issues: Declining water levels Rising water levels Management of access and use Groundwater salinization Groundwater pollution

Sedimentation Adequate design Dam safety Environmental impacts

Social and environmental impacts Sedimentation Dam safety

Groundwater Storage One of the major advantages of storing water in underground aquifers is that it can be stored for years, with little or no evaporation loss, to be used in drought years as a supplementary source of water supply. It also has the advantage that storage can be near or directly under the point of use and is immediately available, through pumping, on demand. The tube well revolution that has swept through agriculture capitalizes on these advantages. For example, crop yields under tube well irrigation in India are frequently two to three times greater than those from irrigation by canal systems alone (see table 8). Table 8. Average food grain yields in tons per unirrigated hectare and per net irrigated hectare by irrigation source (groundwater, canal and tank) in four Indian States (after Dhawan 1986). State Year(s) Unirrigated Groundwater Canal Tank Punjab 77-79 1.08 5.46 3.24 - 63-65 0.75 3.06 1.18 - 50-51 0.37 1.75 0.94 - Haryana 76-77 78-79 0.38 5.74 2.39 -

Andhra Pradesh 77-79 0.42 5.69 3.43 1.96 57-59 0.47 3.11 2.27 1.35 Tamilnadu 77-79 0.49 6.53 2.60 2.33 64-66 0.61 4.00 2.14 2.08 56-68 0.66 3.78 1.69 1.86 Source: Robert Chambers 1988.

26

Another great advantage of groundwater is that as water slowly percolates down into the aquifer it is usually purified of biological pollutants. Thus, groundwater is usually the best source of drinking water, especially in rural areas of developing countries where water treatment facilities are not available. The critical issue facing many groundwater aquifers today is that the volume of water withdrawal exceeds long-term recharge, resulting in rapidly declining groundwater levels in many areas. Closely related to this is the key issue of managing groundwater access and utilization, since groundwater is a common property concern with individual benefits and collective costs. Declining groundwater levels are often about 2 m/yr.6 The extraction of water from aquifers in some districts of India (North Gujarat, Southern Rajasthan, Sawrashtra, Coimbatore and Madurai districts in Tamilnadu, Kolar District in Karnataka and the whole of Rayalasema Region in Andhra Pradesh) exceeds recharge by a factor of two or more. As these aquifers are depleted, the resulting cutbacks in irrigation could reduce India’s harvest by 25 percent or more (Seckler, Molden, and Barker 1998; Shah 1993). Groundwater levels in the Pishin Lara Basin, Pakistan, have steadily declined approximately 2m/yr. since 1987 (Prathapar 1998). In China, almost wherever there is pump irrigation the groundwater levels are declining. Under much of the north China Plain, where nearly 40 percent of China’s grain is harvested, water levels are dropping at roughly 1.5 m/yr. (Worldwatch Institute 1999). Groundwater depletion also has serious equity implications since falling water tables take the resource out of reach of small and marginal farmers. Falling water tables can make wells for domestic water supply run dry. An especially dangerous aspect of falling groundwater tables is illustrated in Bangladesh, where toxic levels of arsenic are being found in the drinking water of millions of people. One theory is that falling groundwater tables have permitted oxidization and mobilizations of natural deposits of arsenic in these areas. Other important problems of groundwater storage are water quality, the pumping costs to extract groundwater and the recoverable fraction of recharge. From a basin-wide perspective, nearly all of the groundwater recharge may be recoverable, but, from a more local perspective, there are some losses. Typically, groundwater recovery under artificial recharge averages 75 percent of the recharge volume (OAS 1997). While falling groundwater tables are a major problem in many areas, many other areas suffer from the opposite problem of rising water levels, with waterlogging and salinization as consequences (Prathapar 1998). Rising water tables also prevent effective sewage disposal in rural villages, with latrines overflowing and polluting the drinking water from wells. The problems of rising and falling water tables are among the most important issues in water policy. Declining groundwater levels in many metropolitan cities such as Mexico, Bangkok and many parts of Japan cause land subsidence. It is commonly thought that groundwater withdrawal should be decreased to the sustainable rate of natural recharge. In some cases, this is correct, but the problem is that this reduces production from this valuable resource. It is much better to artificially recharge the aquifers with excess water wherever possible. However, much more research and development are needed in the field of artificial recharge before this will be a widely used technology. The problem of rising water tables is more tractable in most cases, using well-known, but often expensive, drainage techniques. Care should be taken to prevent the problem of rising water tables in the first place, for example, by not irrigating highly saline areas unless an acceptable drainage plan is in place.

6For aquifers with specific storage capacities of 10 percent, a typical value, a 2-meter decline in water level represents about 200 mm of actual water. Thus, where groundwater levels are falling 2 m/yr. extractions are exceeding recharge by approximately 200 m/yr.

27

Small Reservoirs Here we use the standard definition of small dams as structures less than 15 meters high and with an embankment volume generally less than 0.75 million cubic meters (BOR 1987 and ICOLD 1998). Included within our discussion of small surface reservoirs are small tanks and micro-storage facilities such as dug cisterns and farm ponds. Small reservoirs have the advantage of being operationally efficient. They are flexible, close to the point of use and require relatively few parties for management. Because of these attributes, they can be responsive to demands, the supply to demand mismatch can be small and managerial and institutional issues are easier to handle. Because of their limited storage capacity, small reservoirs respond rapidly to precipitation runoff, often refilling several times a year. Thus, the actual amount of water delivery from a small reservoir can be several times its one-time storage capacity. The great operational benefit of small storages is their rapid response times. Like groundwater systems, they can respond to rainfall on fields, thus maximizing effective rainfall and minimizing operational losses. Small reservoirs often serve multiple uses such as bathing, washing, livestock and aquaculture in addition to irrigation. Small reservoir storage is ideal from the standpoint of operational efficiency, but generally less effective than groundwater or large dams for water conservation. The ratio of high surface area to volume of small reservoirs leads to high evaporation loss. Micro-storage facilities lose, on average, 50 percent of their impoundments to evaporation in arid and semiarid areas (Gleick 1993; Sakthivadivel, Fernando, and Brewer 1997). Other limitations are that their small storage volume does not allow for seasonal or annual carryover and, in addition, there are the cost and safety problems of handling overflow during extreme storm events. The seepage and percolation “losses” from small tanks in Sri Lanka account for 20 percent of reservoir volume (Tasumi 1999) against 5 percent of reservoir volume in large dams. These small reservoirs can act as percolation tanks, recharging aquifers and retarding runoff. Since seepage “loss” can be both an advantage and a disadvantage of small reservoirs, depending on perspective, it is not listed in table 7, but, from a basin-wide hydrologic standpoint, it is generally an advantage. In fact, in India small reservoirs that have high percolation rates, “percolation tanks,” are often preferred because of their contribution to groundwater recharge. Perhaps the greatest threat facing existing reservoirs, both large and small, is sedimentation. While highly variable, it is estimated that 1 percent of the total global freshwater surface storage capacity is lost each year to sediment (Palmieri 1998).7 This does not seem like much until it is realized that the world needs to increase the amount of storage by 25 percent just to stay where we are over the next 25 years! Often, small dams are built without adequate climate and hydrologic analysis. Due to small catchment areas and large variation in rainfall some small tanks in Sri Lanka, for example, do not get sufficient water in 3 out of 10 years. Inadequate hydrologic analysis can also result in insufficient spillway capacity and lead to dam failure due to breaching of the embankment. An issue facing large and small dams alike, but primarily small dams, is dam safety. Of 8,818 high-hazard, non-Federal dams inspected by the US Army Corps of Engineers, one-third (2,925) were determined to be unsafe (FEMA 1996). It is unknown whether the fraction of dams outside of the US with safety problems is greater or lesser than this.8 7We note that some have serious reservations about the validity of this sedimentation value. While we were unable to validate the value, we believe that if correct, it is alarming and important to point out. 8 Many of the “unsafe” dams in the US were rendered so by changes in the applicable design standards—especially the switch to probable maximum flood (PMF) for spillway capacity. In addition, many dams in the US were built privately with less control of standards than is often the case outside the US.

28

Contrary to common opinion, it is very difficult to construct safe small dams. First, in order for them to store as much water as possible, it is desirable to have a large catchment area. But large catchment areas have large runoff, in excess of storage capacity, in extreme storm events. The water must therefore be spilled over or around the dam. However, it is very expensive to build concrete and steel spillways and many small dams, especially in developing countries, do not have them. Consequently, water spillage can breach the dam. In addition, small dams often are constructed in the dry season, when there is inadequate soil moisture and water to properly compact soil during construction. Consequently, water seeps through the dam, creating “pipes” that can breach a small dam from within. Large Reservoirs By 1997, there were an estimated 800,000 dams in the world, 45,000 of which qualify as large dams. More than half of the large dams have been constructed in the past 35 years. In 1997, an estimated additional 1,700 large dams were under construction (WCD 1998). The aggregate design storage capacity of the world’s large dams is about 6,000 km3 (LeCornu 1998). This compares with total water withdrawals of 3,410 km3 (WRI). Considering loss of storage due to sedimentation, lack of filling, etc., perhaps one-half of the design storage is actually achieved, or about one year’s total withdrawals. However, given that a large percentage of withdrawals is from recycled water, the aggregate design storage capacity of 6,000 km3 seems to us to be incredibly high. It is interesting to note that of all the registered large dams in the world only 5 percent are in Africa where most of the severe economic water-scarce countries are located. Fifty-five percent of the large dams are in North America and Europe, where, largely because of this, there are not likely to be severe shortages (LeCornu 1998; and Seckler, Molden, and Barker 1998). Large surface water reservoirs have the advantage of greater yield relative to the available inflow than small reservoirs, and their yield is generally more reliable. This is because of lower evaporation loss fractions in large reservoirs due to their greater depth. Because of their depth, many large reservoirs can store water for multiyear carryover to weather droughts. In monsoonal climates, large reservoirs store excess flows in the wet season for use in the dry season. Other advantages of large surface storage facilities include their relatively low cost per unit of utilizable water (see table 9) and multipurpose qualities—e.g., hydropower and irrigation. According to the Secretary General of ICOLD 30 percent of the world’s registered large dams are multipurpose (LeCornu 1998).9 An emerging new use of large reservoirs in the USA is “mimicry” of the natural hydrograph to mitigate environmental impacts associated with water development. By releasing artificial flood flows from large storage dams, the hydraulics of the natural river system can be imitated and the dynamic conditions of an environmentally healthy system recreated with less water than under virgin conditions.

9We believe this number is an anomaly of the reporting in the ICOLD dam registry and that the ratio of multipurpose large dams is likely much greater than 30 percent.

29

Table 9. Water supply costs10 (1998 dollars)* Storage Capital Costs

($/1,000 m3) Lifetime Delivery Costs

($/1,000 m3) Source

Technology Low Median High Low Median** High

Large storage projects (storage and conveyance costs only)

110 270 1,600 2 5 32 Keller for this paper

Medium and small storage projects (storage and conveyance costs only) 130 320 2,200 7 17 110 Keller for this paper

Micro-storage projects (storage costs only) 160 390 2,500 7 17 110 Keller for this paper Dug storage 500 800 1,200 22 35 60 OAS 1997 Artificial groundwater recharge 190 210 230 Gleick 1993 Groundwater development and pumping 20 40 110 Keller for this paper Diversion projects (interbasin) 190 200 400 Gleick 1993 Conservation practices 40 105 300 Keller et al. 1998 Recycling wastewater (secondary treatment) 120 170 220 Gleick 1993 Reverse osmosis (for brackish water) 160 350 540 Gleick 1993 Recycling wastewater (advanced water treatment)

260 460 660 Gleick 1993

Desalinization of seawater 600 1200 2000 Keller et al. 1998 *The costs obtained from Gleick 1993 were indexed to 1998 dollars from 1980 dollars using appropriate construction cost trends from BOR (1998) and RSMeans 1998. The small, medium and large storage costs were computed by applying BOR 1969 cost estimating guidelines to dam statistics obtained from the BOR website 1999 and from Gleick 1993. Indexed dam cost figures from the western US (BOR 1969) and the State of Tamilnadu, India (Sakthivadivel 1999) were used as a check. **Median cost is taken as 2.5 times the low-end cost for large, medium and micro projects. Large dam reservoirs are more complex to operate than small reservoirs and groundwater systems, from the standpoint of meeting the needs of the individual user. Because they command large areas, they are often far from the points of use. This distance, measured as the water travel time from the dam to the point of use at around 3 km/hr,11 can be weeks long. Therefore, large dam operations cannot be responsive to individual demands that deviate from their expected values, so there is potential for large mismatches of supply to demand. For example, water released from the High Aswan Dam on the Nile in Egypt takes 10 days to reach irrigated areas in the Nile Delta. If there is an unexpected rainfall event in a portion of the Delta that temporarily reduces the demand for irrigation, the water released at Aswan will likely be spilled directly into the Mediterranean Sea unused. On the other hand, unexpected rises in demand may not be met, causing water stress to crops. The flexibility of large storage structures is further reduced when they are multipurpose and potentially conflicting demands (for example, hydropower generation and irrigation) exist. Other factors limiting the flexibility of large dam operations are the many parties and levels involved in their management and countless institutional prerequisites. Reservoirs that are sited upstream of major demands have maximum operational flexibility to shift water among competing uses, for example, taking advantage of rainfall in one area to conserve water for use at another location or time. Where reservoirs are too far downstream in relation to basin demands, surplus flows may become unusable. The Oum er R’bia in Morocco is a case in point. There, sufficient storage is unavailable in the upper catchment to meet the demands of irrigation facilities in the area, while excess water accumulates in large downstream reservoirs with limited potential uses further downstream.

10Costs of technologies related to table 9 are elaborated in pp. 72-74. 11Based on an allowable critical flow velocity in earthen channels of around 1m/sec. Lined sections may have twice this flow velocity. Where gradients are shallow, such as in Egypt’s Nile Delta, flow rates are much slower.

30

An important general issue facing large dams is their social and environmental impacts. The intense social and environmental debate over large dams led to the establishment of the World Commission on Dams, which started its work in May 1998. The Commission’s report is due by June 2000 (WDC 1998). Many of the negative impacts associated with large dams occur because they are constructed on-stream where they obstruct fish passage, inundate important aquatic and riparian habitat, dislocate historic communities, etc. Consequently, many of the new dams being planned and constructed, particularly in the US, are for off-stream storage, for example the recently completed Los Vaqueros Dam in California. In the US, beginning in 1997, decommissioning of large dams has exceeded their construction rate.12 Between June 1997 and July 1998, the US Interior Secretary, Bruce Babbitt, symbolically took his sledgehammer to six large dams (Babbitt 1998). In response to the dam- busting furor, the International Commission on Large Dams (ICOLD) has prepared a position paper on dams and the environment (ICOLD 1998). The ICOLD paper discusses sustainable development of water resources and the role of dams and reservoirs. Comparison of Large and Small Reservoirs Both large and small reservoirs are appropriate technologies under specific conditions of time and place. Table 10 provides a means of examining these issues by comparing the massive High Aswan Dam (HAD) and its reservoir, Lake Nasser, on the Nile in Egypt with the more than 17,000 small tanks in Sri Lanka.13 Several observations may be made from the values in table 10: Table 10. Contrast of characteristics of the High Aswan Dam and reservoir, Lake Nasser, with a typical minor tank in Sri Lanka.

Characteristic High Aswan Dam Typical minor tank in Sri Lank Storage capacity 168.9 km3 (16.89 million ha-m) 4.1 ha-m Surface area 6,500km2 (650,000 ha) 5.0 ha Net irrigated area 2,648,000 ha 5.0 ha Storage fraction of area times depth 0.29 0.4 Annual evaporation loss 14 km3 (1.4 million ha-m) 2.0 ha-m Annual evaporation depth 2.7 m 1.0 m Dam height 111 m 2 m Crest length 3,830 m 170 m Embankment volume 44,300,000 m3 2,600 m3

Travel time to command area 10-days to 60 percent of total command Few hours Command area 3.4 million irrigated hectors <10 ha • The storage capacity behind HAD is 168.9 km3, three times Egypt’s annual allocation from

the Nile and sufficient to meet 3 years of total water needs for all of Egypt. HAD literally saved Egypt from the disasters that afflicted most of Africa during the great drought of the late 1980s.

• In comparison, small reservoirs are used primarily to meet water demands within a period of a few months. The storage capacity behind HAD is over 240 times the aggregate capacity of all 17,000 minor tanks in Sri Lanka. HAD commands 3.4 million irrigated hectares compared

12Most of the recently decommissioned dams in the US are hydropower dams, which, besides having adverse environmental impacts, suffer from dam safety and other issues and are not economical to repair or upgrade. 13HAD statistics are from Gleick 1993 and the minor tank numbers are derived from Sakthivadivel, Fernando, and Brewer 1997.

31

to around 700,000 hectares in Sri Lanka and supplies water to meet the domestic and industrial needs of 60 million people.

• The ratio of HAD’s Lake Nasser surface area to Egypt’s irrigated area is about 5:1; the ratio with small tanks is near 1:1. This means that the evaporation from the small tanks exceeds that of the area they irrigate.

• The dispersion of area inundated by small tanks may be better, in terms of environmental impact, than the concentrated inundation that occurs with large reservoirs. On the other hand, small tanks often submerge the best agricultural lands.14

• The high operational flexibility of small tanks and high overall effectiveness of cascade systems (noted below) can provide substantial benefits over large reservoirs.

The point in comparing these two surface storage systems is that they are both very different yet appropriate technologies in their respective settings. Small dams could not collectively capture the surplus flows of the Nile as effectively as the High Aswan Dam. Nor, on the other hand, would a single large water impoundment with the combined capacity of all the small tanks in Sri Lanka be effective in servicing all the associated small irrigation systems.15 This is not to say that there is no room for improvement in either case or that either is optimally designed or sited to maximize the capture of flows that would otherwise be lost to the sea. Complementarities It is important to consider complementary opportunities among different types of storage systems to improve conservation and productivity of water. Water conservation per se may not increase water productivity because of inefficient operation and mismatches with crop water requirements. Table 11 presents the characteristics of storage types for providing the needed conservation and operational efficacies. Among the alternatives available, combinations of storage systems are most likely to produce superior results. The suitable combinations of storage types depend on a number of factors including topography, hydrology and the existence of suitable aquifers. Table 11. Characteristics of storage structures.

Storage type Conservation on potential

Operational flexibility

Adequacy Reliability

Large reservoir H L H L Small reservoir L H L L Groundwater storage H H L H Large and small reservoir combined H H H L Large and small reservoirs combined with groundwater storage

H H H H

H = High; L = Low; Adequacy = Sufficiency of yield to meet needs of command area; Reliability = Assuredness of water deliveries. Several combinations already exist and work satisfactorily. The combination of small and large reservoirs is nicely demonstrated by the "melons on a vine" irrigation schemes in China, Sri Lanka and other countries. Here, a few large storage facilities supply water to numerous small

14Small tanks (by definition) only submerge a few feet up the sides of a valley—the rest is valley floor. Large dams flood a lot of non-valley floor area that is usually less-productive land. 15Note that there are several large storage facilities in Sri Lanka.

32

tanks within a river basin. In this manner, small reservoirs act to dampen supply and demand mismatches from large reservoirs. In the Imperial Irrigation District in southern California, small regulator reservoirs of 500,000 m3 save more than 12 million m3 annually of canal flows that otherwise spill to the Salton Sea; this results in an annual water conservation to storage volume ratio of 25:1. In Southern Sri Lanka, construction and linking of a large storage reservoir at Lunugamvehera with five small, existing, cascading reservoirs resulted in a 400 percent increase in crop production. In fact, cascading small reservoirs can significantly increase crop water use by capturing drainage, return flow and surpluses from upstream reservoirs. Complementarities also occur where surface storage, particularly in the form of micro-reservoirs, retards runoff and enhances groundwater recharge. With improved tube well technology now available and within reach of small farmers, many storage reservoirs, which were previously used as irrigation tanks in the arid and semiarid tracts of India, have now been converted to recharge ponds, and tube wells have taken the place of irrigation canals. These successful experiments indicate that combinations of big and small reservoirs along with effective aquifer management can provide efficient solutions for conserving water and increasing its productivity. Hitherto, this concept has not been effectively put into practice from the planning stage, although it has been practiced in many areas of the world. With water becoming scarce, the use of such integrated planning for conserving water could lead to higher water productivity while maintaining an environmental and ecological balance. Costs The typical low, medium and high costs (in 1998 dollars) of various water supply technologies are presented in table 9. The lifetime delivery costs reflect the present value capital and operation and maintenance costs over the economic life of the technologies divided by the total volume of water they produce and deliver. The surface storage capital costs differ greatly between the low and high ends. This is due to the wide variability in dam construction costs associated with site conditions, dam types, construction methods, spillway requirements, etc. We found some large dams that cost as little as $1 per 1,000 m3 of storage16 and others that were more than $15,000 per 1,000 m3. Rather than 16These extremely low-cost dams are generally concrete arch or gravity dams. An example of such a low-cost ($1/1,000 m3) dam is the Kariba Dam in Zambia and Zimbabwe, which became one of the world’s largest reservoirs (by volume right behind HAD’s

Box 4. Conjunctive use of groundwater and small reservoir water Oosambadi Peria Eri is situated 10 km from Thiruvannamali in Tamilnadu, India. This small reservoir has an 80-hectare command area, 53 farmer beneficiaries and 60 wells, mostly dug. Prior to 1986, only one crop was grown. Even this crop could not be successfully irrigated without supplemental well water, because reservoir water, when directly used for irrigation, is sufficient only for about 70 days when the reservoir is full.

In 1986, only four farmers in the command area did not own wells. It was decided by the Water User Association that these four farmers would be provided with water at the common cost and that the reservoir water would be used only for recharging the aquifer. In 1986, the sluices of the reservoir were permanently closed. From then on, farmers have grown two crops, one paddy and one other crop. Conjunctive use of surface water and subsurface water has been practiced for the la st 14 years. Similar switching over to conjunctive use has taken place in more than 16 minor irrigation reservoirs in the dry district of Coimbatore, Tamilnadu, India.

33

give these extremes, we present in table 9 typical low- and high-end costs. As expected, the distribution of surface storage costs is positively skewed; that is, the average cost of storage is greater than the median. The median cost of storage in table 9 is estimated at 2.5 times the typical low-end cost, whereas the average cost of storage is closer to four times the low end. It very well could be that dams can no longer be built for the low-end costs listed in table 9 given current dam safety requirements and the costs of mitigating negative environmental impacts. The cost of the Los Vaqueros Dam, a key component in the first major water project to be built in California in the past decade, was $346 per 1,000 m3 of storage (commensurate with the medium cost of large storage projects given in table 9). The dam cost, however, represented only 10 percent of the total project costs, which were large because of efforts to minimize environmental impacts.17 The storage capital costs in table 9 are the cost of the storage facility per 1,000 m3 of gross reservoir capacity plus the associated cost of the conveyance system. For the storage capital costs to be comparable to the other water supply technologies listed in table 9, these costs must be adjusted to account for all the usable water a reservoir will “produce” over its life. To do this, one has to divide the storage costs in table 9 by an estimate of how many times its capacity a reservoir will release water over its life. This is affected by the mean annual inflow to a reservoir relative to its storage capacity, the rate of sedimentation and the evaporative loss fraction. For example, the combined large dam storage capacity on the Colorado River system in the western US is approximately four times the mean annual flow of the river. Thus, on average, the Colorado River storage cycles once every 4 years. So, in the course of 100 years, a reservoir on the Colorado River will regulate and release 25 times its storage capacity. However, losses in capacity due to sedimentation18 and losses of water due to evaporation reduce the total release volume to approximate 20 times the storage capacity over 100 years. In monsoonal climates, storage is only a small fraction of the mean annual flow. (For example, the Tarbela Dam on the Indus River in Pakistan has a live storage capacity less than 10 percent of its mean annual inflow.) Since these reservoirs are filling and releasing during the wet season, they may realize 1.5 times their usable storage capacity per year. However, sedimentation will reduce their relative yield by 50 percent over a 100-year life of a typical reservoir. Thus, the yield of a large dam reservoir in a monsoonal climate might be 75 times its usable storage capacity over the course of 100 years. If we assume that, on average, medium and large dam storage projects deliver 50 times their storage capacity and small dams 20 times (due to shorter lives and greater evaporation fractions than larger reservoirs), the effective cost per 1,000 m3 delivered to a 50-hectare command is $2 to $32 for large dams and $7 to $110 for medium and small dams. We have also added cost estimates for alternative sources of water supply. The conservation practices listed in table 9 include programs targeted at real water savings. Water saved by such activities is transferable to other uses or available for expanded use within the project area without adverse consequences downstream. An example of a water conservation practice is canal lining where the seepage from the unlined canal is lost to non-beneficial evaporation or to a saline sink. The median cost for conservation practices listed in table 9 was derived from the water conservation agreement between the Imperial Irrigation District (IID) and the Metropolitan

Lake Nasser). However, we note that Kariba had high environmental and social costs associated with it, which are not reflected in the dollar cost of the dam. (Kariba is a case study of the International Commission on Dams.) 17The Los Vaqueros Project was winner of the 1999 Outstanding Civil Engineering Achievement award, largely because of the way it addressed environmental concerns (Hunt 1999). 18Note that the effect of sedimentation on total reservoir yield for reservoirs with capacities twice or more times their mean annual inflow is relatively small compared to losses due to sedimentation in reservoirs with capacities smaller than their mean annual inflow.

34

Water District (MWD) of Southern California. Under the terms of this agreement, MWD paid for conservation at IID in exchange for the water saved (Keller, Keller, and Davids 1998). The costs for desalination of seawater given at the bottom of table 9 include all approaches, of which multistage flash distillation and reverse-osmosis are the most common (Gleick 1998). The low-end desalination costs are engineering estimates for cogeneration/reverse-osmosis desalination (Keller, Keller, and Davids 1998). The technologies presented in table 9 vary widely in scale, so, to make the costs comparable, we have included the conveyance cost associated with delivering water from the facility to within 50 hectares of its point of use in the storage capital cost. For large surface storage projects, we estimate the cost of conveyance facilities to range between $100 and $600 per 1,000 m3 of gross reservoir capacity. For medium and small-scale surface storage projects, the costs in table 9 include $80 to $500 per 1,000 m3 of gross reservoir capacity. For wastewater recycling, interbasin diversion, reverse osmosis and desalinization, the costs include $0 to $10 for conveyance of 1,000 m3 of developed water. The delivery costs in table 9 include no conveyance costs for micro- and dug storage, groundwater development and recharge, and conservation practices and distillation, as these technologies are generally scaled to the 50-hectare command area or are additions to existing systems. Conclusion Under all but the most optimistic scenarios, there is a dearth of freshwater storage. If climate change, as a result of global warming, manifests, the need for freshwater storage will become even more acute. Increasing storage through a combination of groundwater and large and small surface water facilities is critical to meeting the water demands of the twenty-first century. This is especially so in monsoonal Asia and the developing countries in the tropics and semitropics. As an immediate first step, we must assess the major river basins of the world, whether they are open, closed or semi-closed. The productivity of water as presently used must also be assessed to determine the extent to which increased demands for irrigated agricultural production can be met by increasing water productivity and the extent to which increased demands will require increased consumption of water. The uncommitted discharges from those basins that are open or semi-closed must then be determined and plans made to effectively capture and put this water to use. Combinations of small and large storage and surface water and groundwater recharge are generally the best systems where they are feasible. In monsoonal Asia, research and development are needed on how to manage water under monsoonal conditions.

23

Appendix

Glossary Classical efficiency (CE): That part of the total diversions to an irrigation scheme that is evaporated by crop evapotranspiration. Here, it is applied at irrigation scheme level. Committed water: Water reserved for use by the environment, downstream countries, or other downstream uses that have a right to the water. Crop coefficient (Kc): Factor representing the relationship between reference and crop evaporation. Crop water requirements (ETcrop): The depth of water needed to meet the evapotranspiration requirements of a disease-free crop, growing in large fields under nonrestricting soil conditions and achieving full production potential under the given growing environment. Effective efficiency: Within irrigation, this is the ratio of net irrigation requirements (NET) (defined below) to the diversions less beneficially reusable outflow. Effective efficiency differs from classical efficiency in that it includes recyclable water. Effective efficiency can be applied at any scale within a basin. This is adapted from terms called effective efficiency introduced by Jensen (1977) and Keller and Keller (1996) (this is the same as the Kellers’ equation except that it does not discount for pollution). At the basin scale, it reduces to the ratio of NET to the primary water supply (PWS) (defined next page). Effective rainfall (Peff): That part of the total rainfall that can be beneficially used by crops. Evaporation (E): The amount of water that leaves the basin or country as vapor. Evaporation can be beneficial or non-beneficial. Non-beneficial (Enb) includes evaporation from open water bodies (reservoirs, canals) and from bare soil. Evaporation factor (EF): Factor expressing the percentage of total diversions actually evaporated. Net irrigation requirements (NET): Amount of irrigation water needed to supplement rainfall to meet crop evapotranspiration requirements, defined as ETcrop less effective rainfall. This excludes irrigation water needed to cover conveyance and application losses, non-beneficial evaporation and leaching requirements. Outflow to sinks or downstream countries (O): Amount of water flowing out of the basin or country. This includes the amount of water committed to downstream countries and the minimum flow to the sea to ensure navigation and avoid salt intrusion. Percolation (P): Downward flow past the crop root zone in paddy cultivation. Note that percolation is not regarded as wholly consumptive’ use since part of the seepage replenishes the groundwater and can potentially be reused later.

24

Potential utilization factor (PUF): Factor expressing the percentage of the annual renewable water resources that could be controlled and managed with full development of storage and conveyance facilities. Primary water supply (PWS): The total diversions to all sectors, less the amount of water recycled. Primary water is also equal to evaporation plus flows to sinks within the unit of analysis. Primary water is an important concept in that it provides an upper limit to the amount of water that can be depleted by various uses and is a good indicator of the amount of water resources development. Recycled water (R): Water that has already been diverted at least once upstream. The difference between total diversions and primary water is the amount of water that is recycled. Recycling takes place, for example, by reusing drainage water or pumping groundwater. Reference crop evapotranspiration (ETo): The rate of evapotranspiration from an extensive surface of 8 to 15 cm tall, green grass cover of uniform height, actively growing, completely shading the ground and not short of water. Renewable water resources (RWR): Average annual flow of rivers and recharge of groundwater generated from endogenous precipitation plus incoming flow originating outside the country, taking into consideration the quantity of flows committed to upstream and downstream countries through formal and informal agreements or treaties. This gives the maximum theoretical amount of water available for the country (definition by FAO, data by WRI). Total diversions (TD) (Total water supply): This is the amount of water diverted from its natural courses to various uses. Typically, in water resource systems, water is recycled. Total diversions equal PWS plus recycled water. Thus, total diversions are often larger than primary water supply and could be larger than potential utilizable water resources. Utilizable water resources (UWR): That part of the water resources, which is considered to be available for development. This figure considers factors such as the dependability of the flow floods, extractable groundwater, minimum flow required for nonconsumptive use, etc. Also called water development potential or manageable water resources. Seventy-five percent (75%) probable rainfall (P75): The amount of rainfall that occurs at least 3 out of 4 years.

25

Bibliography Agarwal, Anil. 1998. Coping with dry spells and drought years—India’s traditions in water harvesting. Stockholm Water Front, No. 1–2, May 1998. Stockholm, Sweden. Alcamo, J.; T. Henrichs; and T. Rosch. 1999. World water in 2025. Global Modeling and scenario analysis for the World Commission on Water for the 21st Century. Duplicated. Allan, Richard. 1999. Email communication to USCID group, 15 March 1999. Re: Slash and burn versus intensive irrigation. Allan, T. 1998. Moving water to satisfy uneven global needs: "Trading" water as an alternative to engineering it. ICID Journal 47(2):1–8. Babbitt, Bruce. 1998. Dams are not forever. Remarks of the United States Department of Interior Secretary to the Ecological Society of America, August 4, 1998, Baltimore, Maryland, USA: Bastiaanssen, W. G. M.; D. J. Molden; S. Thiruvengadachari; A. M. F. R. Smit; L. Mutuwatte; and G. Jayasinghe. 1999. Remote sensing and hydrological models for performance assessment in Sirsa Irrigation Circle, India. Research Report 27. Colombo, Sri Lanka: International Water Management Institute. BOR (Bureau of Reclamation). 1969. Reclamation instructions: Series 150. Denver, Colorado, USA: Bureau of Reclamation, United States Department of Interior. BOR. 1987. Design of small dams. Denver, Colorado, USA: United States Department of Interior. BOR. 1998. Bureau of Reclamation construction costs trends. Denver, Colorado, USA: Bureau of Reclamation, United States Department of Interior. BOR. 1999. Reclamation’s dams. http://www.usbr.gov/dsis/dso_imap.html. Denver, Colorado, USA: United States Department of Interior. Briscoe, J. 1999. The financing of hydropower, irrigation and water supply infrastructure in developing countries. International Journal of Water Resources Development 15(4):459–491. Buddruss, Georg. 1993. On artificial glaciers in the Gilgit Karakorum. Studien zur Indologie und Iranistik 18: Germany: Chambers, Robert. 1988. Managing canal irrigation. Oxford, England: Oxford and IBH Publishing Company, Pvt. Ltd. New Delhi, India. Cosgrove, W. J.; and F. R. Rijsberman, eds. Forthcoming. World water vision: Report. World Water Commission Staff Report. Dhawan, B. D. 1986. Output impact according to main irrigation sources: Evidence from four selected states. Paper to Seminar on Water Management–The Key to Development of

26

Agriculture, New Delhi 28-30 April 1986, Indian National Science Academy, New Delhi 110 002. Doorenbos, J.; and W.O. Pruitt. 1977. Guidelines for predicting crop water requirements. FAO Irrigation and Drainage Paper No. 24. Rome, Italy: FAO. FAO (Food and Agriculture Organization). 1997. FAOSTAT statistical database in CD-ROM. Rome: FAO. FEMA (Federal Emergency Management Agency). 1996. National dam safety. USA: Federal Emergency Management Agency. http://www.fema.gov/home/mit/damprgm.htm. Gleick, Peter. 1998. The world’s water 1998–1999: The biennial report on freshwater resources. Washington, D.C., USA: Island Press. Gleick, Peter, ed. 1993. Water in crisis: A guide to the world’s fresh water resources. Oxford, England: Oxford University Press. Gleick, P. H. 1996. Basic water requirements for human activities: Meeting basic needs. Water International 21(2):83–92. Hargreaves, G. H.; and J. E. Christiansen. 1974. Production as a function of moisture availability. ITCC Review III(9):179–189. Association of Engineers and Architects in Israel. Hyami, Y.; and Vernon Ruttan. 1985. Agricultural development: An international perspective. Baltimore, USA: Johns Hopkins University. Hunt, Craig S. 1999. Putting nature first. American Society of Civil Engineers. ICOLD (International Commission on Large Dams). 1998. Position paper on dams and environment. PP: International Commission on Large Dams http://genepi.louis-jean.com/cigb/chartean.html. IID (Imperial Irrigation District). 1999. Personal communication about irrigation records. Jensen, M.E. 1977. Water conservation and irrigation systems: Climate. In Technical Seminar Proceedings, 208–250. Columbia, Missouri: Keller Andrew; and Jack Keller. 1995. Effective efficiency: A water use concept for allocating freshwater resources. Water Resources and Irrigation Division Discussion Paper 22. Arlington, Virginia, USA: Winrock International. Keller Andrew; Jack Keller; and David Seckler. 1996. Integrated water resource systems: Theory and policy implications. IWMI Research Report 3. Colombo, Sri Lanka: International Irrigation Management Institute. Keller, Jack; Andrew Keller; and Grant Davids. 1998. River basin development phases and implications of closure. Journal of Applied Irrigation Science 33(2). October 1998: Frankfurt, Germany.

27

Keller, Jack; and O.A.A. Fadl. 1993. Water resource development phases, costs and dynamics in arid regions. Proceedings of the Regional Seminar on Freshwater Quality and Efficiency held in Cairo, Egypt, October 17–18, 1993. Cairo, Egypt: Center for Environment and Development in Arid Regions and Europe (CEDARE). LeCornu, Jacques. 1998. Dams and water management. Report of ICOLD Secretary General to the Conférence Internationale Eau et Développement Durable, March 19-21, 1998. Paris, France. http://genepi.louis-jean.com/cigb/article-barrages-an.html. Mayberry, Keith S.; and Gerald J. Holmes. 1996. Guidelines to production costs and practices for imperial county field crops 1996–1997. Circular 104-F. Holtville, CA, USA: University of California Cooperative Extension. Molden, David. 1997. Accounting for water use and productivity. SWIM Paper 1. Colombo, Sri Lanka: International Water Management Institute. Molden, D. J.; R. Sakthivadivel; and Z. Habib. 1999. Basin water use, conservation and productivity: Examples from South Asia, Transactions of the 17th Congress of the International Commission on Irrigation and Drainage, Granada, Spain: International Commission n Irrigation and Drainage. OAS (Organization of American States). 1997. Source book of alternative technologies for freshwater augmentation in Latin America and the Caribbean. Washington, D.C., USA: Unit of Sustainable Development and Environment, General Secretariat, Organization of American States (OAS). http://www.oas.org/EN/PROG/srcecont.htm Palmieri, Alessandro. 1998. Reservoir sedimentation and the sustainability of dams. 1998 World Bank Water Week Conference, December 15, 1998. Annapolis, Maryland, USA: World Bank. Perry, Chris. 1998. The IWMI water resources paradigm—definitions and implications. Colombo, Sri Lanka: International Water Management Institute. Perry, C. J. 1996. Alternative approaches to cost sharing for water services to agriculture in Egypt. Research Report 2. Colombo, Sri Lanka: International Irrigation Management Institute. Perry, C. J. 1999. The IWMI water resources paradigm. Definitions and implications. Agricultural Water Management 40(1):45–50. Perry, C. J.; D. Seckler; and M. Rock 1997. Water as an economic good: A solution, or a problem? Research Report 14. Colombo, Sri Lanka: International Irrigation Management Institute. Postel, S. 1999. Pillars of sand: Can the irrigation miracle last? New York, NY, USA: W.W. Norton & Company (Worldwatch book). Prathapar, S.A. 1998. Summary of activities. IWMI Program Planning Meeting, November 30–December 3, 1998. Colombo, Sri Lanka: International Water Management Institute. Renault, D. 1999. Field report from Kirindi Oya. Duplicated.

28

RSMeans Company. 1998. RSMeans 1999 heavy construction cost data. Kingston, Massachusetts: RSMeans Co. Sakthivadivel, R. 1999. Personal e-mail communication. February 8 – May 31, 1999. Sakthivadivel. R.; and O. T. Gulati. 1997. Assessing management interventions impact with IIMI's minimum set of indicators: The case of Mahi-Kadana Project, India. In International Seminar: Research Program on Irrigation Performance (RPIR), Mendoza, Argentina, 3–7. November 1997. Sakthivadivel, R.; Nihal Fernando; and Jeffrey Brewer. 1997. Rehabilitation planning for small tanks in cascades: A methodology based on rapid assessment. IIMI Research Report 13. Colombo, Sri Lanka: International Irrigation Management Institute. Science. 1999. Crop scientists seek a new revolution. 15 January 1999, Vol. 283. Seckler, D. 1992. Irrigation policy, management and monitoring in developing countries. In Roundtable on Egyptian water policy: Conference Proceedings, ed. M. Abu-Zeid and D. Seckler, Cairo: Water Research Center, Ministry of Public Works and Water Resources. Seckler, David. 1996. The new era of water resources management: From “dry” to “wet” water savings. IWMI Research Report 1. Colombo, Sri Lanka: International Water Management Institute. Seckler, David. 1992. The Sardar Sarovar Project in India: A commentary on the report of the independent review. Discussion Paper No. 8, July 1992. Center for Economic Policy Studies, Water Resources and Irrigation Policy Program. Arlington, Virginia, USA: Winrock International Institute for Agricultural Development. Seckler, D.; U. Amarasinghe; D. Molden; R. De Silva; and R. Barker. 1998. World water demand and supply, 1990 to 2025: Scenarios and issues. Research Report 19. Colombo, Sri Lanka: International Water Management Institute. Seckler, David; David Molden; and Randolph Barker. 1998. Water scarcity in the twenty-first century. IWMI Water Brief 1. Colombo, Sri Lanka: International Water Management Institute. Seckler, David; and Michael Rock. 1995. World population growth and food demand to 2025. Water resources and irrigation division. Discussion Paper. Arlington Virginia, USA: Winrock International. Seckler, David; Upali Amarasinghe; David Molden; Radhika de Silva; and Randolph Barker. 1998. World water demand and supply, 1990 to 2025: Scenarios and issues. IWMI Research Report 19. Colombo, Sri Lanka: International Water Management Institute. Shah, Tushaar. 1993. Groundwater markets and irrigation development. Political economy and practical policy. Chapter 7. Bombay: Oxford University Press.

29

Shiklimanov. 1999. Personal communication. Data prepared for the World Commission on Water. Duplicated. Tasumi, Masahiro. 1999. Water balance and return flow in reservoir cascade irrigation system of Sri Lanka. Submitted to Department of Agriculture, Grad. School of Tottori University, Tottori, Japan, January 1999. Tushaar Shah; M. Alam; M. Dinesh Kumar; R. K. Nagar; and Mahendra Singh. 1999. Pedal pump and the poor: Social impact of a manual irrigation technology in South Asia, Colombo: International Water Management Institute (draft paper). Shah, Tushaar; Dinesh Kumar; R. K. Nagar; and Mahendra Singh. 1999. Pedal pumps and the poor: Social impact of a manual irrigation technology in South Asia. Colombo, Sri Lanka: International Water Management Institute. Duplicated. UN (United Nations). 1999. World population prospects, 1998 Revision. New York: UN Department for Policy Coordination and Sustainable Development. USDA (US Department of Agriculture). 1997. Time series data viewer. USDA Economic Research Services. Washington D.C.: United States Department of Agriculture. Wallace, J. S.; and C. H. Batchelor. 1980. Managing water resources for crop production. Wallingford, Oxon OX10 8BB, UK: Institute of Hydrology. WCD (World Commission on Dams). 1998. World Commission on Dams: Strategy and objectives. http://www.dams.org. Wolcott, Barbara. 1999. Aquifer recharge: A natural solution. Water Engineering and Management. World Bank. 1997. From vision to action. Environmentally and Socially Sustainable Development Studies and Monograph Series 12. Washington, D.C.: World Bank. Worldwatch Institute. 1999. State of the world 1999. New York, NY, USA: W.W. Norton and Company. WRI (World Resources Institute). 1998. World resources 1998-99. A guide to the global environment. New York: Oxford University Press.