Schultz - Uhlenbrook.Water Security
Schultz - Uhlenbrook.Water Security
B. Schultz
UNESCO-IHE, Delft, The Netherlands and Rijkswaterstaat, Centre for Public Works, Utrecht,
The Netherlands
S. Uhlenbrook
UNESCO-IHE, Delft, The Netherlands
ABSTRACT: Water security involves the sustainable use and protection of water systems, pro-
tection against water related hazards, sustainable development of water resources and the safe-
guarding of water functions and services for humans and the environment. In spite of the cur-
rent wealth of experience, very good know how and technologies, and sometimes high financial
means, problems with respect to water and water management remain in many cases enormous.
These include: water shortages, inefficient water use in irrigation and drinking water supply, in-
adequate sanitation, water logging and salinisation, inadequate operation and maintenance, pol-
lution through urban and industrial wastewater, fertilisers and pesticides, and flooding of culti-
vated, urban and industrial areas. This paper gives a brief overview of certain key issues of
water security, whereby mainly biophysical and engineering aspects are addressed. It illustrates
what these key issues and aspects mean and what they may imply for our societies, at present
and in future.
1 INTRODUCTION
The pace of change in our world is speeding up, accelerating to the point where it threatens to over-
whelm the management capacity of political leaders. This acceleration in history comes not only from
advancing technology, but also from unprecedented world population growth, even faster economic
growth, and the increasingly frequent collisions between expanding human demands and the limits of
the earth’s natural systems.
For more than 6,000 years people have been using water to improve their living conditions, both
employing it to their advantage in many different ways and protecting themselves against its
harmful effects. . Tremendous successes have been obtained, but in many cases water manage-
ment, flood and drought protection, drinking water supply and sanitation schemes are far from
optimal;they function significantly below what could be expected, or are simply absent. Other
functions and uses of water, such as its role in ecosystems, transport, recreation, landscaping,
etc. are often inadequately addressed and the connected processes are still not fully understood.
Therefore, many challenges remain. It is remarkable that although nowadays there is such a
wealth of experience, and we have very good know-how, technologies and high financial
means, the problems in certain cases are still enormous. Some major problems are, in a list that
can easily be extended (Schultz 2001):
• water shortages;
• inefficient water use in irrigation and drinking water supply;
• inadequate sanitation;
• destruction of (aquatic) ecosystems;
• waterlogging and salinisation of soils;
• inadequate operation and maintenance of water management, drinking water supply, sani-
tation and flood protection schemes;
• pollution through urban and industrial wastewater, fertilisers and pesticides; and
• flooding of cultivated, urban and industrial areas.
This paper presents a brief overview of certain key issues connected to water security. They
will be defined and discussed in terms of their implications for our societies, at present and in
future. The paper discusses mainly bio-physical and engineering aspects; issues connected to
water governance and the need for knowledge transfer and capacity building are discussed fur-
ther in an accompanying paper (van Hofwegen 2008).
Water security involves the sustainable use and protection of water systems: the protection
against water related hazards, the sustainable development of water resources, and the safe-
guarding of (access to) water functions and services for humans and the environment. Measures
to increase water security are in this context primarily concerned with human interventions in
water systems and/or a wise utilization of water and water-related environmental services.
These are aimed at the enhancement of the beneficial and sustainable uses of water for various
purposes such as water supply, irrigation, drainage, navigation, hydropower, environmental con-
trol, and the protection against water related disasters such as floods and droughts. Interventions
in water systems are often necessary to meet the needs of society and the environment in its
widest sense and in order to be able to face the challenges of all kinds of global changes (e.g.
climate change, land use change, etc.). Obviously, the negative effects (e.g. environmental or
socio-economic) of such interventions need to be minimized, and also their wider impact on
downstream water users needs to be considered. Effects of climate change and other global
changes are increasingly important factors for identification of the design parameters that define
the location and dimensions of hydraulic structures, water management, flood protection, drink-
ing water supply and sanitation systems to increase water security.
One of the key aspects is the relationship of water security to food security. The agricultural
sector is already the largest water consumer (approximately 70% of the global water supply). At
present about 45% of the global food production is realised without a water management system
and 55% with either an irrigation or a drainage system. However, it is expected that global agri-
cultural yields will have to be doubled in the next 25-35 years and that 80-90% of this increase
will have to be realised on the existing cultivated area. This can only be achieved with more ef-
ficient use of water resources and substantial improvement and extension of water management
systems, in combination with increases in either surface or sub-surface water storage in exten-
sive regions. (van Hofwegen & Svendsen 2000, Schultz et al. 2005). This problem will be ag-
gravated if more agricultural land is used to satisfy the rapidly increasing energy demand (i.e.
production of energy crops, bio fuels etc.).
The hydrological cycle can be divided into atmospheric, surface and subsurface water systems.
Surface water includes all water bodies, which are in direct contact with the atmosphere (i.e.,
streams, lakes, snow and ice as well as water in the biosphere). Subsurface water includes
groundwater and water in the unsaturated zone above the groundwater table (including capillary
fringe). The global reserves of water as a function of storage are shown in Table 1, and the di-
rections of water fluxes are indicated in Figure 1. Table 1 indicates that freshwater resources are
mainly stored in glaciers and permanent snow cover as well as in groundwater.
Table 1. World water resources; renewable periods are calculated as the mean volume divided by the
mean flux (Shiklomanov, in prep.,cited in UNESCO 2006).
__________________________________________________________________________________________________________
Location Volume % of total vol. %of Vol. recycled Renewal
3 3 3
(10 km ) in hydrosphere freshwater annually (km ) period years
__________________________________________________________________________________________________________
Ocean 1,338,000 96.5 -- 505,000 2,500
__________________________________________________________________________________________________________
Groundwater (gravity
& capillary) 23,400* 1.7 16,700 1,400
Predominantly fresh
Water 10,530 0.76 30.1
__________________________________________________________________________________________________________
Soil moisture 16.5 0.001 0.05 16,500 1
__________________________________________________________________________________________________________
Glaciers & pemanent
snow cover 24,064 1.74 68.7
Antarctica 21,600 1.56 61.7
Greenland 2,340 0.17 6.68 2,477 9,700
Arctic Islands 83.5 0.006 0.24
Mountainous
regions 40.6 0.003 0.12 25 1,600
Ground ice
(permafrost) 300 0.022 0.86 30 10,000
__________________________________________________________________________________________________________
Water in Lakes 176.4 0.013 -- 10,376 17
Fresh 91.0 0.007 0.26
Salt 85.4 0.006 --
__________________________________________________________________________________________________________
Marshes and swamps 11.5 0.0008 0.03 2,294 5
__________________________________________________________________________________________________________
River water 2.12 0.0002 0.006 43,000 16 days
__________________________________________________________________________________________________________
Biological water 1.12 0.0001 0.003 --
__________________________________________________________________________________________________________
Water in atmosphere 12.9 0.001 0.04 600,000 8 days
__________________________________________________________________________________________________________
Total volume
in hydrosphere 1,386,000 100 --
__________________________________________________________________________________________________________
Total freshwater 35,029.2 2.53 100
__________________________________________________________________________________________________________
*Excluding groundwater in the Antarctic estimated at 2 million km3, including freshwater of about 1 mil-
lion km3
This table shows great disparities betweeen the huge volume of saltwater and the tiny fraction of freshwa-
ter between the large volumes of water contained by the glaciers and the water stored in the aquifers, and
between the amount of groundwater and the small volumes of water in rivers, lakes and reservoirs.
The proportion of water stored in the atmosphere, soil, and in river channels is very small,
and the dynamic residence times (based on mean renewable period) are short. However, because
of the importance of this water for humans (including agriculture) and ecosystems, understand-
ing the water fluxes and residence times in these different domains is of primary importance.
All the rainwater that falls on the earth and is not evaporated, transpired, or withdrawn artifi-
cially, contributes to the flow of rivers and groundwater. The discharge of a river differs signifi-
cantly in space and time, dependent on different physiographic characteristics, like climate, size
of the river basin, topography (slopes and other terrain parameters) as well as soil and geologi-
cal parameters. Depending on human activities, the quality and quantity of water that enters the
rivers differ as well. Rivers can transport natural or artificial components; the load can differ in
relation to the discharge. Both quantity and quality of the river water will determine if it is use-
ful for irrigation or domestic water supply.
Understanding and being able to predict all processes of the water cycle at different spatial
and temporal scales is critical for the effective management of water resources, including hydro-
logical extremes; thus it is crucial for an optimal water security. However, the large spatial and
temporal variabilities that exist in water storages, fluxes, and residence times of the different
components, as well as the often poorly understood interactions between the different compo-
nents, make the development of hydrological and water resources development models a chal-
lenging task. This becomes more difficult when water quality issues also need to be taken into
account, which is increasingly required especially where human influences, such as modernisa-
tion of water management schemes, land reclamation, or discharge of polluted water, are in
play.
Figure 1. Schematic sketch of the global water cycle. Water storages and fluxes are indicated by boxes
and arrows. (adapted from Oki & Kanae 2006)
Figure 2. Scheme of the hydrological cycle with the branching cycle, expressing human influence
(adapted from Rodda & Matalas 1987)
4 HUMAN INFLUENCE ON THE HYDROLOGICAL CYCLE
Within the hydrological cycle a sub-cycle of water diversion, including water consumption
through irrigation, domestic water supply and drainage exists. This branching cycle – expressing
the human influence – significantly impacts the primary hydrological cycle. The direct human
influence on the hydrological cycle only concerns much less then 1% of the water resources on
earth, as can be derived from Table 1 and is illustrated in Figure 2 (Rodda & Matalas 1987).
However, the side effects of human activities influence almost all accessible waters on earth
and, consequently, the availability of services from water and the aquatic environment.
Table 2 summarizes the available water resources on different continents and presents an out-
look to water resource availability in 2025 (Shiklomanov 2000). It is demonstrates that resource
availability will decrease everywhere at the same time that withdrawals and consumption are
expected to be increasing. The situation appears to be worst for Africa and Asia, in particular in
the emerging and least developed countries.
Table 2. Estimated and projected volumes of available renewable water resources and water use by conti-
nent, 1990 and 2025 (Shiklomanov 2000)
Available renewable water re- Water use in 109 m3
sources
Area 109 m3/person m3/person 1990 1990 2025 2025
3
m /year in 1990 in 2025 withdrawal consumption withdrawal consump-
tion
Africa 4,047 6,180 2,460 199 151 331 216
North America 7,770 17,800 12,500 642 225 836 329
South America 12,030 40,600 24,100 152 91 257 123
Asia 13,508 3,840 2,350 2,067 1,529 3,104 1,971
Europe 2,900 3,990 3,920 491 183 619 217
Australia and 2,400 85,800 61,400 29 16 40 23
Oceania
World 42,655 7,800 4,800 3,580 2,196 5,187 2,879
The water scarcity index, is distributed unevenly globally. This index, Rws, is defined as (W
– S)/Q, where W, S, and Q are respectively the annual water withdrawal by all the sectors, the
water use from desalinated water, and the annual renewable freshwater resource (Oki & Kanae
2006). A region is usually considered highly water stressed if the water scarcity index is higher
than 0.4 (e.g. Falkenmark & Rockstoem 2004). Oki & Kanae (2006) calculated that currently
2.4 billion people live in highly water stressed areas. This number is likely to increase to 3-5 bil-
lion people in 2025, depending on the assumed future scenario for the development of climate,
economics etc. Water stress (scarcity) is currently very high in Northern China, Central Asia, at
the border between India and Pakistan, the Middle East, parts of Europe and Southern and
Northern Africa, Western South-America, and the middle and western areas of the United
States. It should be noted that the given numbers for the continents are averages over a large in-
tra-continental variability, which is demonstrated in the global distribution of the water scarcity
index. Due to population growth, and increase both in water use per person and in standard of
living, in general the situation will become more severe (Figure 3).
On the other hand, an increasing number of people are also expected to live in flood prone ar-
eas, especially in urban conglomerations. These areas are located in river valleys, deltas and
coastal zones. Floods in these areas have diverse causes; they may be caused by excessive rain-
fall (urban – flash – floods), extreme river discharges (flooding caused by the surrounding area)
– either by excessive rainfall or snowmelt – or by storm surges at sea.
Figure 3. Current and future projections of population under high water stress using three business-as-
usual scenarios. (adapted from Oki & Kanae 2006) The threshold values are set to be (A) the water-
crowding indicator Aw = Q/(C G) in 1000 m3 per year and per capita, and (B) the water scarcity index
Rws = (W – S)/Q > 0.4, where Q, C, W, and S are renewable freshwater resources (Q), population (C),
water withdrawal (W), and water generated by desalination (S), respectively. Error bars indicate the
maximum and minimum population under high water stress corresponding to the renewable freshwater
resources RFWR projected by six climate models; climatic conditions averaged for 30 years are used for
the plots at 2025 (averaged for 2010 to 2039), 2055 (averaged for 2040 to 2069), and 2075 (averaged for
2060 to 2089)
5.1 General
This paper utilizes the three categories of countries that have been identified using the criteria
ofleading international organisations (UN Population Reference Bureau 2005, World Bank
2005, UNCTAD 2002). These three are: developed countries, which include most of the coun-
tries of Western and Central Europe, North America, some countries in Central and South
America, the larger countries in Oceania and some countries in Asia; emerging countries, which
include most of the Eastern European countries (including Russia), most of the countries in
Central and South America, most of the countries in Asia (including China, India and Indone-
sia), and several countries in Africa; and least developed countries, which include most of the
countries in Africa, several countries in Asia, one country in Central America and most of the
smaller countries in Oceania. From Figure 4 it can be seen that by far the majority of the
worlds’ population (almost 75%) lives in the emerging countries. It is also evident that the
greatest population growth will be in the least developed and emerging countries. In the devel-
oped countries almost no further growth is expected. The standard of living in the emerging
countries is rapidly rising; however, about 1.2 billion people in the least developed and emerg-
ing countries are still poor (GNI less than 1 US$ per day), and out of these about 70% live in ru-
ral areas (World Bank 2005).
Figure 4. World population and growth in least developed countries, emerging countries and developed
countries (UNDP 2005, International Commission on Irrigation and Drainage 2006)
The estimated and projected global water use by sector in 109 m3/year for 1950, 1990 and
2025 is shown in Table 3 (Shiklomanov 2000). From Tables 2 and 3 it can be derived that on a
global scale this water use requires only a small percentage of the resources, and it seems that
there is still a considerable reserve to meet future needs. However, since water resources are to a
large extent generated by river runoff, only a certain percentage can be used. The remainder has
to be drained off to the sea during floods, and a minimum flow to the sea has to be maintained
during other periods. On the other hand, the possible contributions from groundwater need to be
considered. Table 3 also clearly demonstrates that agriculture is by far the largest water user, al-
though the ratios are changing. Another substantial factor is that water use in agriculture gener-
ally results in a significant increase in evaporation, while other types of water use generally re-
sult in a return flow. However, although this return flow may not result in a significant ‘loss’ of
water, it may result in a significant pollution of water and a change in the hydrological regimes
(e.g. water storage for hydropower generation in seasonal runoff regimes).
Table 3 Estimated and projected global water use by sector in 109 m3/year for 1950, 1990 and 2025
(Shiklomanov 2000)
Item 1950 1990 2025
withdrawal consump- withdrawal consump- withdrawal Consump-
tion tion tion
Agriculture use 1,124 856 2,412 1,907 3,162 2,377
Industrial use 182 14 681 73 1,106 146
Municipal use 53 14 321 53 645 81
Reservoirs 6 164 275
Total 1,365 894 3,580 2,196 5,187 2,879
Table 4 Continents and categories of countries ranked according to the population density with reference
to the arable land (from Schultz et al. 2005).
Continent Total area Cultivated Total Population density
in 106 ha land population in persons/km2
in 106 ha in million with reference to
in 2005
total area arable land
Asia 3,339 547 3,765 113 688
Africa 3,031 201 840 28 418
Europe 2,299 307 732 32 238
Americas 4,016 384 850 21 221
Oceania 856 55 31 4 57
From Table 4 it can easily be observed that the Asian continent has by far the largest population
and the highest population density, both with reference to total area, as well as to arable land. If
one also takes the population growth shown in Figure 4 into account, then it becomes clear that
in the nextdecades most of the activities in the field of water management to increase water se-
curity are to be expected in Asia, with Africa in second place. And the need is very obvious for
water management with the purpose of increasing water and food security.
5.3 Water need for urban and industrial water supply and sanitation
The migration from rural to urban areas is especially strong in the emerging countries; for in-
stance, it is expected that by 2020 more than 85% of the population of Latin America will live
in urban centres. In conjunction there is a rapid rise in the standard of living and domestic water
demand; consumption per capita as well as the need for urban and industrial water supply is
quickly increasing. Whereas in the least developed countries water use is often in the order of
magnitude of 5 l/person/day, in the developed countries it may easily exceed 100-150
l/person/day. In addition we dietary change, generally resulting in an increase in consumption of
meat and a diversification of crops, both resulting in an increase in water use in the production
chain.
The increase in water use per person also has a significant impact on sanitation, especially
that required in urban areas. In general, the development of provisions for sanitation lags behind
the development of urban and industrial water supply, resulting in substantial discharges of un-
treated wastewater with numerous consequences on water quality and water-borne diseases. Es-
pecially in the countries of the European Union this has resulted in far reaching measures to
control the pollution of surface waters, and the adoption of the European Water Framework Di-
rective is a milestone (European Commission 2001). In arid and semi-arid zones we see another
development, which is the reuse of urban wastewater in agriculture. This can increase the effi-
ciency of water use (management of the local water ‘cycle’); however, various health issues can
arise if it is not done correctly. For an interesting overview of developments with respect to this
see Huibers et al. (2005). Because domestic and industrial water supply generally commands a
substantially higher price per unit than agricultural water supply, we also often see in cases
where competition develops that water is shifted from the agricultural sector to the urban and
industrial sector.
5.4 Water need for other uses and their specific requirements
For specific water uses such as transportation, water recreation, etc. special provisions may be
required. For commercial ship transport, for example, large-scale investments are made for the
construction, operation and maintenance of river bed improvements, canalisation, harbour facili-
ties, storm surge protection, sluices, etc. Such investments are generally justified by commercial
and/or political considerations. Although in such cases water is not really consumed, the meas-
ures taken to promote these types of water use mayhave far reaching consequences on other
types of water use and on the environment. Decisions on such measures need to be taken in an
integrated way that also considers the environmental and socio-economic implications.
Hydropower is another important use of water is. According to the database of the Interna-
tional Commission on Large Dams (ICOLD) there exist about 49,000 large dams in the world
and numerous small dams. Most of the reservoirs behind these dams have been developed for
hydropower generation; other uses include irrigation, municipal and industrial water supply,
flood control, navigation, dilution of sewage water, log transport, stabilization of lake water lev-
els and estuary improvement. In many cases reservoirs fulfill multiple objectives. To assure wa-
ter security in projects where reservoirs play a role, due consideration needs to be given to the
environmental impacts and to the required environmental flows in the river system.
Moreover, the water dependency index, considering virtual water, is introduced in contrast to food in-
dependence. The water dependency index as it is proposed and the assessment of water deficit, based
on virtual water, give a perverted idea of the possibility of national food self-sufficiency.
7 CONCLUSION
Emerging and least developed countries are in particular confronted with the need to increase
significantly their food supply in the coming decades. These countries can only do this by in-
creasing their local production, and where inescapable in combination with increased imports.
In the rainfed areas without a water management system, water harvesting and watershed man-
agement can be expected to result in some improvements, especially in the livelihood of poor
farm families. There is, however, no way that the cultivated area can contribute significantly to
the required increase in food production without a water management system, which can
achieve the needed changes. The share of irrigated and drained areas in food production will
have to increase, accompanied by significant improvements in existing irrigation and drainage
systems through modernisation in combination with increased stakeholder involvement. In addi-
tion, new systems will have to be installed in areas where so far agricultural production has been
achieved without a water management system, or in new land reclamations (Schultz et al. 2005).
For the foreseeable future increasing flood management and flood protection provisions will
be required in lowland areas where there is continuing growth in population, standards of living,
urbanisation and industrialization. . Additional complications with respect to this are created by
the effects of global changes (e.g. climate change, land use changes) and land subsidence, which
occur in many of the lowland areas. Such processes make these areas increasingly vulnerable.
This could result in the need to abandon such areas, or change the agricultural use in the me-
dium or long-term (for example, shifting to extensive pasture land). In instances where this be-
comes an actual problem there will be an urgent need for timely and complicated measures with
various implications at different levels (Schultz 2006).
8 REFERENCES
Allan, J.A. 1998. Virtual water: a strategic resource, global situations to regional deficits. Groundwater
36(4):545–546.
Alphen, J. van, Beek, E. Van & Taal, M. 2006. Floods, from defence to management. In Proceedings of
the 3rd International Symposium on Flood Defence, Nijmegen, The Netherlands, 25–27 May 2005.
Leiden: Taylor & Francis / Balkema.
Brown, L.R., et al. 1996. State of the World 1996. The Worldwatch Institute. London: Earthscan Publica-
tions Ltd..
Duivendijk, J. van 2005. Manual on planning of structural approaches to flood management. New Delhi:
International Commission on Irrigation and Drainage.
Dukhovny, V. 2007. Water and globalization: case-study of Central Asia. Irrigation and Drainage 56(4).
European Commission 2000. EU Water Framework Directive, Directive 2000/60/EC. Brussels, Belgium.
Falkenmark and Rockstoem 2004. Balancing water for humans and nature. The new approach in ecohy-
drology. Earthscan.
German National ICID Committee 2005. Proceedings 21st ICID European Regional Conference, Inte-
grated land and water resources management: towards sustainable rural development, 15 – 19 May
2005, Frankfurt (Oder), Germany and Slubice, Poland. (CD-ROM).
Hoekstra, A.Y. & Hung I.Q. 2002. Virtual water trade: quantification of virtual water flows between na-
tions. RRS 11, Delft: UNESCO-IHE.
Hofwegen, P.J.M. van & Svendsen, M. 2000. A vision of water for food and rural development. The
Hague, The Netherlands.
Hofwegen, P.J.M. van 2008. Water security: Everybody’s concern, everybody’s responsibility. This vol-
ume, Delft: UNESCO-IHE.Huibers, F. P., Raschid-Sally, L. ab& Rag Ragab (eds.) 2005. Wastewater
Irrigation. Irrigation and Drainage (Special Issue).
International Commission on Irrigation and Drainage (ICID), 2004. Updated statistics on irrigation and
drainage in the world, www.icid.org, New Delhi, India.
Oki, T. & Kanae, S. 2006. Global Hydrological Cycles and World Water Resources. Science, 313: 1068–
1072.
Oki, T., S.M., Kawamura, A., Miyaka, M., Kanae, S. K., & Musiake K. 2003. Virtual Water trade to Ja-
pan and in the world. In Value of Water Research Report 12: 221–235. Delft: UNESCO-IHE.
Oki, T., Valeo, C., & Heal, K. (eds.) 2006. Hydrology 2020: An Integrating Science to Meet World Water
Challenges. IAHS red book 300.
Rodda, J.C. & Matalas, N.C. 1987 Water for the future. Hydrology in perspective. In Proceedings of the
Rome Symposium, April 1987, International Association of Hydrological Sciences, Wallingford, Great
Britain. IAHS Publication no. 164,.
Schultz, B. 2001. Irrigation, drainage and flood protection in a rapidly changing world. Irrigation and
Drainage. 50(4).
Schultz, B., Thatte, C.D., & Labhsetwar, V.K. 2005. Irrigation and drainage. Main contributors to global
food production. Irrigation and Drainage 54(3).
Schultz, B. 2005. Question 53. Harmonious coexistence with floods. General report. In Proceedings of
the 20th Congress of the International Commission on Irrigation and Drainage (ICID), 10–18 Septem-
ber 2005, Beijing, China.
Schultz, B. 2006. Opportunities and threats for lowland development. Concepts for water management,
flood protection and multifunctional land-use. In Proceedings of the 9th Inter-Regional Conference on
Environment–Water. EnviroWater 2006. Concepts for Water management and Multifunctional Land-
Uses in Lowlands, Delft, 17–19 May, 2006.
Shiklomanov I.A. 2000. Assessment of water resources and water availability in the world. Water Inter-
national. 25(1).
UNCTAD 2002. The Least Developed Countries Report. In United Nations Conference on Trade and
Development, Geneva, Switzerland. www.unctad.org.
UNDP Population Reference Bureau 2005. 2005 world population data sheet. Washington DC, USA.
UNESCO 2006. Water – A Shared Responsibility. The United Nations World Water Development Report
2. Paris: UNESCO.
Working Group on Non-structural Aspects of Flood Management, 1999. Manual on non-structural as-
pects to flood management. New Dellhi: International Commission on Irrigation and Drainage (ICID).
World Bank 2005. Global economic prospects and the developing countries. Washington DC, USA.
World Bank 2003. World Bank Atlas, 35th Edition. Washington DC, USA.