Introduction

Water resources occupy a specific place among other natural resources. Water is the most distributed substance on our planet; in different amounts it is available everywhere and plays an important role in the surrounding environment and human life.
Of most importance is fresh water. No activities of human being and life itself is impossible without it because it can be substituted by nothing. Human being always consumed fresh water and used for his purposes water objects, however during many hundreds of years man’s impact on water resources was insignificant and  of local character. The magnificent properties of natural waters -- their renovation during water cycle and ability for self-purification -- allowed retaining for a long time a relative purity, quantity and quality of fresh waters. This gave birth to a known illusion of unalteration and inexhaustibility of water resources that were considered as a free-of-charge gift of the natural environment. Under these conditions, historically the tradition appeared of careless attitude to using water resources, concept of minimum expenses for waste water purification  and water objects' protection.
The situation has drastically changed during the recent decades. In many regions and countries of the world, the unfavourable results of long-term, often unreasonable, man’s activities, were discovered. This concerned the direct use of water resources and the surface transformations on river watersheds.
To a large extent this was due to a drastic increase in global water withdrawal since the 1950s. In turn this increase was caused by the intensive development of productive forces in all spheres of the world economy in the course of scientific and technological revolution. As compared with the previous decades, during 1951-1960, annual water withdrawal increased fourfold. This occurred because of the drastic expansion of irrigated areas, the growth of industrial and heat- power- engineering water consumption, and the intensive construction of reservoirs on all the continents.
  During recent 25-30 years all over the world there is an especially intensive anthropogenic change of hydrologic cycle of rivers and lakes, their water quality, water resources and water budget. The values of water resources, their dynamics with time and distribution over the territory are now determined by not only natural climate variations, as it has been previously, but also the man’s economic activities. In many regions and countries of the world, water resources are quantitatively depleted and much contaminated. Therefore, they cannot already meet the ever increasing demand in them and become the factor impeding the subsequent development of economy and the growth of population commonwealth.
Especially drastic problems of fresh water arise in arid regions characterised by very limited water resources, high degree of their use and very fast demography growth. In 1978 in Argentina, the First World Conference on Water Resources was held by the UNO. This conference discussed the unfavourable situation with fresh water in one third of the world  located primarily in the zones with insufficient moisture. It was also mentioned that the same problem would arise by the end of our century in the most countries of the world.
The Conference contributed to strengthening and co-ordination of international co-operation in studying and assessing water resources. It also stimulated developing national investigations and attracted attention of the general public, governments, planning organs and decision-makers to water problems, water management and hydrological science development.
    Problems of water resources on regional and global scales are in the sphere of activities of many governmental and non-governmental organisations, primarily, UNESCO, WMO, UNEP, FAO, IAHS, IAWR, etc.  Numerous scientific conferences and symposia focused on these problems, and enormous quantities of studies were published in different countries of the world. On the basis of close international co-operation the most detailed and complete estimates of water balance and water resources for all the continents and Earth's natural zones were published in 1974. They were obtained by Russian scientists under the guidance of the State Hydrological Institute (SHI) and cited in the monograph “World Water Balance and Water Resources of the Earth”. Baumgartner and Reichel (Germany) also presented such estimates in 1975 in the monograph “World Water Balance”. So far the data placed into these monographs have been widely used by scientists as the most overwhelming and reliable. All subsequent publications also dealing with data on water resources on a global scale give no new information as compared with the above studies. To a full measure this also refers to the data on water resources and water use by the countries of the world well known to scientists. The Institute for World Resources, Washington, D.C., periodically issues their publications. However, the reliability of the data is not high because of their different sources, including obsolete, and obtaining methods. It is also important that they were obtained for different years and long-term periods.
In this connection UNESCO established a new project within the International Hydrological Programme. Its goal was to generalise the world data on water resources and their use as well as to prepare the monograph “World Water Resources by the Beginning of the 21st Century”. The State Hydrological Institute (St. Petersburg, Russia) was charged to fulfil the project during 1991-1996. Scientists of the institute guided by the author of the present summary successfully completed this work and prepared the monograph for publication. The summary suggests the reader the basic results and inferences in a general shape obtained when fulfilling the project and preparing the monograph.
However, taking into account the sufficiently reliable new data which are obtained for some countries during the last time, in the following tables for individual countries and regions more completed and up-dated values of water resources, water use and water availability are presented. They can be differ from the data given in the monograph, but they don’t change the made conclusions and common tendency of the water resources and water use distribution over the world and their temporal changes.
The author considers it his duty to express his sincere gratitude to his colleagues, workers of SHI. Due to their great efforts it became possible to obtain new data on the dynamics of water resources of the Earth and their use. Among them are heads of laboratories of SHI: Prof. Vladimir I. Babkin and Dr. Vladimir Yu. Georgiyevsky; senior scientists: Dr. Tamara Ye. Grigorkina, Dr. Natalia V. Penkova, Dr. Irena P. Zaretskaya; scientists: Anna V. Izmailova, Valentina P. Yunitsyna, Alexander I. Moiseenkov; leading engineers: Tatiana V. Grube, Irena A. Nikiforova, Tatiana I. Printseva, Elena L. Skoryatina and Valentina G. Yanuta.
On behalf of SHI scientists the author expresses his particular gratitude to Dr. V. Grabs, Director of GRDC, Germany. Also he thanks heads of hydrological services of the countries on different continents who kindly presented the necessary observation river runoff data used to assess global water resources.

2. Water storage on the Earth and hydrological cycle.

Water is the most spread substance in the natural environment. Water exists in three states: liquid, solid, and invisible vapour. It forms oceans, seas, lakes, rivers, and ground waters in the top layers of Earth’s crust and soil cover. In a solid state, it exists as ice and snow cover in polar and alpine regions. Some amount of water is contained in the air as water vapour, water droplets, and ice crystals as well as in the biosphere. Huge amounts of water in a bounded shape enter the composition of different minerals of Earth’s crust and core.
To reliably assess water storage on the Earth is a complicated problem because water is very dynamic. It is in a permanent motion converting from liquid to solid or gaseous phase, or vice versa. In addition to the quantitative estimation of water storage, it is necessary to determine the shape (free or bounded) and the volume (sphere) on our planet the water is taken into account. Usually water of the so-called hydrosphere is estimated. This is free water being in state in liquid, solid or gaseous state in the atmosphere, on the Earth’s surface and in crust down to the depth of 2000 meters. By modern estimates the Earth’s hydrosphere contains a huge amount of water of about 1386 million cubic kilometres. However, 97.5% of this amount are saline water and only 2.5% fresh water. The greater portion of the fresh water (68.7%) is in the shape of ice and permanent snow cover in the Antarctic, the Arctic, and mountainous regions. Next 29.9% are fresh ground waters. Only 0.26% of the total amount of fresh waters on the Earth are concentrated in lakes, reservoirs, and river systems. They are most accessible for economic needs and very important for water ecosystems (Fig. 1).
The above values characterise the so-called natural static water storage in the hydrosphere. It is the amount of  water simultaneously contained, on the average for a long-term period, in water bodies, aquifers, and the atmosphere. For the shorter time intervals (years, seasons, months) the values of water storage in the hydrosphere permanently vary during water exchange among the ocean, land, and the atmosphere. This exchange is usually called the turnover of water on the Earth, or the global hydrological cycle. The scheme of global hydrological cycle is depicted in Fig. 2.
 Solar heat evaporates water into the air from the Earth's surface. Land, lakes, streams, and oceans send up a steady stream of water vapour. It spreads over the surface of our planet falling down as precipitation. Precipitation falling down on land is the main source of the formation of land waters: rivers, lakes, groundwater, glaciers. A portion of atmospheric precipitation evaporates, some part penetrates and charge groundwater, and the rest water as river flow returns to the ocean where again evaporates, and the process repeats again and again. A considerable portion of river flow does not reach the ocean having evaporated in the so-called drainless regions, or the regions of endorheic runoff. On the other hand, some part of groundwater passing by the river systems, directly goes to the ocean or evaporates. Quantitative indices of different components of the global hydrological cycle obtained by Russian hydrologists are shown in Fig. 2 . Every year the water turnover on Earth involves 577,000 km3 of water. It is the water that evaporates from the oceanic surface (502,800 km3 ) and from land (74,200 km3 ). The same water amount falls  as atmospheric precipitation (on the ocean 458,000 km3 and on land 119,000 km3). The difference between precipitation and evaporation from land surface (119,000 - 74,200 = 44,800 km3/year) represents the total runoff of Earth’s rivers (42,600 km3/year), and a direct groundwater runoff to the ocean (2200 km3/year). These are the principal sources of fresh water to supply life necessities and man's economic activities.
River water is of great importance in the global hydrological cycle and supply of humankind with fresh water. This is due to the fact that the role of individual components in Earth's  water turnover depends both on the value of water storage and its dynamics. Usually the latter is  estimated by the period of full replenishment. Hydrospheric water of different kinds is fully replenished during this period in the process of hydrological cycle. Its value varies in a very large range (table 1). For instance, the period of full recharge of oceanic waters occurs during about 2500 years, permafrost and polar ice for 10,000 years, deep groundwater and mountainous glaciers for 1500 years. Water storage in lakes is fully replenished for 17 years and in rivers for 16 days. In hydrology and water management, basing on water exchange characteristics, the two conceptions are often used to assess water resources in a region: the static, or secular, freshwater storage and renewable water resources. The static, or secular, storage includes conventionally the freshwater with the period of full renewal of many years or decades (large lakes, groundwater, glaciers, etc.). Its intensive use unavoidably results in depleting the storage and unfavourable ecological consequences. It also disturbs the nature equilibrium established for centuries, whose restoration would require tens or hundreds of years.
The renewable water resources include the water yearly replenished in the process of water turnover on the earth. It is  mainly the river runoff estimated in the volume referred to a unit of  time (m3/s, km3/year, etc.) and formed in the region at issue or incoming from outside, including the groundwater inflow to the river network. This kind of water resources includes also the  yearly renewable upper aquifer groundwater not drained by the river systems. However, on the global scale their volumes are not large as compared with the river runoff volume and of considerable importance only for individual specific regions

Table 1.   Periods of water resources renewal on the Earth
 
 

Water of Hydrosphere	Period of renewal
World Ocean	2500 years
Ground water	1400 years
Polar ice	9700 years
Mountain glaciers	1600 years
Ground ice of the permafrost zone	10000 years
Lakes	17 years
Bogs	5 years
Soil moisture	1 years
Channel network	16 days
Atmospheric moisture	8 days
Biological water	several hours

In the process of turnover the river runoff is not only recharged quantitatively, but its quality is also restored. If it were so that man could suddenly stop to contaminate rivers, then with time water could return its natural purity. So the river runoff, actually representing the renewable water resources, is the most important component of hydrological cycle. It exerts a pronounced effect on earth's surface ecology and humankind economic development. It is the river runoff that is most widely distributed over the territory and provides the major volume of  water consumption in the world. In practice, by the value of river runoff an estimation is made of the water availability and deficit of water resources in this or that region. Therefore, the monograph mainly deal with assessing river runoff, its spatial-temporal dynamics on the global scale, and analysing its use for different economic needs nowadays and in the future.

3.  Renewable water resources: Time and space variability.

3.1.  Initial data and methodological approaches.

Quantitative characteristics of renewable water resources of a region or basin can be determined by the two methodological approaches: by meteorological data and observational materials on river runoff.
The first methodological approach is widely applied in case of insufficient hydrological observations and with the availability of significantly detailed and overwhelming meteorological data. In practice it is usually realised by using the simplest equation of long-term water balance of a territory. According to this equation the value of renewable water resources averaged over a long period of time is determined by the difference between precipitation and evaporation from land. Precipitation is calculated by observational data and evaporation by various calculation formulae. This estimation technique is quite simple. It has been widely used in many countries in the beginning of our century, when the hydrological network was developed but insufficiently, and meteorological data was much more in amount. Quite frequently this technique is presently applied. For instance, in 1996 under the guidance of the South-American Branch of UNESCO, an estimation of water resources for the Latin-American countries was made in this way.
The method for estimating renewable water resources by meteorological data, though being very simple, has a number of essential disadvantages. They cannot allow recommendations for its use in detailed calculations, especially for countries and regions with limited water resources. First , because of low accuracy, this technique is inapplicable to arid and semiarid regions, where river runoff is very small. By the absolute value, it is close to an error of determination of evaporation and precipitation. Second , by this method it is impossible to reliably estimate water resources for each certain year (moreover, for season and month). These data are extremely necessary for modern planning of water management. Third , this technique is inapplicable to estimate water resources of the countries and regions located in the basins of international rivers. In this case a larger volume of river runoff comes from outside rather than is formed in the territory at issue.
In this connection, to estimate global water resources (for the continents, regions and countries located in various physiographic conditions), the second methodological approach has been applied. It is based on observational materials from the world hydrological network, and the meteorological information serves as subsidiary. This approach was successfully applied by Russian scientists as early as the 1960-70s to prepare the abovementioned monograph “World Water Balance and Water Resources of the Earth”. At the present time the longer series of hydrological observations have been collected. Thus, it was possible to obtain the data, previously inaccessible, on many regions of Africa, Asia, and Latin America  poorly covered with hydrological information. Therefore, there are more grounds to apply this approach.
By the WMO data, not long ago and at present 64,000 stations have measured river runoff in the world. They are very unevenly spread over the territories. The duration of their observations varies from a few months to more than 180 years. To this or that extent, at our disposal there have been data on river runoff from about 40,000 stations. It is quite a different in quality information about river runoff of different countries, including fragmentary and not systematised data for individual years and even months. Included is also the observation data that served the basis for constructing national or regional charts of mean annual river runoff. Using all these data to estimate the dynamics of water resources on the global scale is impossible, and no need to do this. Therefore, for subsequent analysis the selection was made of hydrological sites basing on the following principal conditions:

•  the availability of the most long series of continuous observations;

•  the location of the sites on big and medium rivers, possibly evenly distributed over the territory;
•  observations should reflect the natural (or close to natural) river runoff regime.

3.2. Continents, natural-economic regions, and countries.

The mean value of renewable global water resources is estimated at 42,750 km3 per year, and it is very variable with space and time. Table 2 presents the distribution of water resources and water availability by the Earth’s continents. By an absolute value the largest water resources are characteristic of Asia and South America (respectively, 13,500 and 12,000 km3 per year). The smallest are typical for Europe and Australia with Oceania (respectively, 2900 and 2400 km3 per year). For individual years the values of water resources can vary in the range of ±15-25% of their average values. Absolute values reflect but not to a full measure the water availability of the continents, as they differ very much in area and especially by the population number. Specific water availability of the continents in cubic metres of water per year that fall on 1 km2 of area and per one person is presented in table 2 . Due to a rapid Earth’s population growth since 1970 to 1994 the potential water availability of Earth’s population decreased from 12.9 down to 7.6 thousand m3 per year per person. The greatest reduction of population water supply  took place in Africa (by 2.8 times), Asia (by two times), and South America (by 1.7 times). The water supply of European population decreased for that period only by 16 %.
Studies have shown (ref. Fig. 4 ) that variations in the total river runoff on the continent and Earth as a whole are of cyclic nature. As seen in Fig. 4 , the cycles of wet and

3.3. River basins, continental slopes, and the inflow to the World Ocean.

An assessment of renewable water resources for countries, regions, and the continents is based on the runoff of river basins calculated by observation data from the hydrological network.
Table 6 presents information about water resources of the principal world rivers. Their average long-term runoff volume is above 100 km3. The greatest river of the world Amazon gives 16% of annual global river runoff. Twenty seven per cent of the world water resources are formed by the five largest river systems: Amazon, Ganges with Brahmaputra, Congo, Yangtze, and Orinoko.  The rivers, whose data are presented in table 6 , are located on all the Earth's continents, except for Australia. The total water resources of all these rivers comprise 52% of the world water resources.
Due to generalising river runoff data from the global hydrological network it was possible to assess the freshwater inflow to the World Ocean. This inflow is very important for studying its freshwater balance and dynamic processes.
It is worth mentioning that the freshwater inflow to the World Ocean cannot be identified by its value with global river runoff for two reasons. First , river watersheds mostly belong to the so-called endorheic (drainless) runoff regions that are not connected to the World Ocean. The total area of endorheic runoff regions is about 30 million km2  (20% of the total land area). However, only 2.3% (about 1000 km3/year) of annual global river runoff are formed in these regions. This can be attributed to occupying the most part of the territory of drainless regions by deserts and semideserts with a very low precipitation. The largest drainless regions include  the Caspian Sea basin in the territory of Europe and Asia, the greater part of Central Asia, North-Eastern China, and Australia as well as the Arabian Peninsula, and North Africa. Small drainless regions are also available in other regions including North and South America. Second , in the regions of exorheic drainage directly connected to the World Ocean, water resources of river basins not always coincide  with river mouth runoff. This especially pertains to the regions with hot climate. There, basin water resources are formed in mountainous areas with a large precipitation amount. As moving towards the mouth, runoff is lost in much for evaporation in the plain and low-land parts of the basin. Ganges and Indus in Asia, Niger and Zambezi in Africa, Mississippi and Colorado in North America are examples of the above river basins. In exorheic regions, about 1100 km3 of runoff per year are lost for evaporation and do not reach river mouths. The volume of 380 km3 out of this amount falls on Asia, 300 km3 on Africa, and 340 km3 on North America.
Thus the total river water inflow to the World Ocean will be somewhat less than the value of renewable water resources of the Earth’s continents. The values of an average long-term river water inflow to the oceans and their year-to-year variability for 1921-1985 is shown in Fig. 6 .
Approximately half the total river water inflow to the World Ocean falls on the Atlantic Ocean where four of six largest rivers of the world go to (Amazon, Congo, Orinoko, and Parana). The smallest amount of river water (5000 km3 per year) inflows to the Arctic Ocean, however river waters are of most importance for the regime of this ocean. There is a very simple explanation. The Arctic Ocean contains as much as 1.2 % of total water storage in the World Ocean. At the same time it accepts 12.5 % of  global river runoff.
To simulate the dynamic processes in the oceans, it is very important to take into account not only the volume of river water inflow, but also its distribution over the World Ocean. River runoff enters the World Ocean very unevenly. The data presented in Fig. 7 demonstrates this very clearly; it presents the distribution of inflow to the World Ocean by latitudinal zones. On the average, about 42% of total river runoff enter the ocean in the equatorial region between 10єN and 10єS. This observation data for individual years or seasons obtained at the global hydrological network seem to be of great interest to oceanologists, who develop water circulation  models for the ocean.

3.4. River runoff and underground water.

The above values of renewable water resources are estimated by the total river runoff. As known, river runoff includes the water, directly incoming to the hydrographic network during rainfalls or snow melt, and groundwater of the upper aquifers feeding rivers  more or less evenly throughout a year. A part of groundwater also referring to the renewable water resources does not enter the rivers but goes directly to seas and oceans or is expended for evaporation. In this case an estimation of renewable water resources by only river runoff data produces underestimated results.
Of great practical importance is a reliable information about renewable groundwater volumes not entering the river network. These volumes are compared with river runoff values. It is useful to know in what regions they are of most importance. Obviously, these are the regions with a weakly developed hydrographic network primarily including plains with arid and semiarid climate. There the river runoff is small by value, and upper aquifer groundwater can play a great role in the total volume of renewable water resources.
To reliably estimate this groundwater for all regions of the world is a very complicated problem and so far impracticable because of lack of necessary data. Nevertheless, this estimation was made for a number of regions and countries. It allows drawing certain conclusions for different physiographic conditions on the global scale.
In particular, the most detailed estimation of renewable water resources, including river runoff and not related groundwater, was made in 1995 by FAO. It is given for all countries of the African continent, where arid and semiarid regions occupy more than half of the territory. By the FAO data, the total volume of renewable groundwater resources, not related to river runoff, is 188 km3 per year for the continent as a whole, or as much as 5% of total river runoff volume. However, these values play a great role in the total volume of renewable water resources for individual countries located in arid regions including Egypt, Libya, Tunis, and Morocco. Certainly, they should be taken into account.
The estimations made give an idea that almost the same situation takes place on the other Earth’s continents.
Thus, the values of renewable groundwater resources, not related to river runoff, could be neglected when estimating renewable water resources on the global scale, i.e., for the continents, large natural-economic regions, countries, and river systems. Exception is small countries located in arid regions with a weakly developed hydrographic network. Among these countries are primarily those located in Northern Africa and on the Arabian Peninsula. For these countries,  the values of renewable groundwater resources are important for the total  renewable water resources. They even can be above (for many countries by manifold) the total river runoff volume.

4. The use of water resources.

4.1. Background.

To reliably assess in detail the future water resources and water availability at the present time, it is insufficiently to have data on values and natural variations of river runoff. It is necessary to take into account their changes due to human activities. During the recent decades natural variations in river runoff and quantitative and qualitative characteristics of renewable water resources have been much affected by many factors. Among them are an intensive development of industry and irrigated landuse, population growth, urbanisation and related drastic water consumption, and Earth’s surface transformations.
Renewable water resources, available in large river watersheds, countries, and selected natural-economic regions, are being affected by a complex of anthropogenic factors. They include those related directly to water intake from river systems for irrigation, industrial and domestic water use. Among them are also a reservoir runoff control, river watershed surface transformations, including a-and-deforestation, field management, urbanisation, and drainage. All these factors differently affect the total values of water resources, river runoff regime, and water quality.
To estimate the role of all anthropogenic factors is likely improbable taking into account possible  anthropogenic global changes in renewable water resources for the continents and natural-economic regions. However, there is no necessity to do this. It is quite possible to neglect taking into account the factors related to transforming the surface of river catchments. These factors exert major effects on the runoff of small and middle rivers, and on monthly and extreme river runoff characteristics rather than annual values as well as on water quality. Under certain physiographic conditions these kinds of human activities can even promote an increase in renewable water resources of small and middle rivers by decreasing the total evaporation in the basins.
An estimation of global anthropogenic effects on water resources is to be made basing, primarily, on considering the role of the factors related to direct water intake from water bodies and reservoir runoff control. These factors, causing the unilateral decrease in surface and groundwater runoff, are widely distributed, most intensively develop and able to exert an especially pronounced effect on water resources condition in large regions.
Speaking of man’s impact on water resources, it is impossible to avoid touching a very sharp modern problem. It is the problem of anthropogenic changes in global climate, the so-called global warming, due to increasing carbon dioxide concentration in the atmosphere and strengthening the “greenhouse effect”. An expected rise of air temperature and change in precipitation could not help affecting the values of renewable water resources and the character of their economic use. However, an insignificant anthropogenic global climate change recorded for recent decades is more or less reflected in observation-based estimates of water resources and water consumption variations. As to calculations for the future, it should be mentioned that the global warming forecasts, so far available for most regions, are very contradictory. This pertains primarily to expected changes in precipitation. Therefore, they are unusable for obtaining anywhere near certain estimates of water resources and water consumption. In addition, according to the recent assessments for the future, the most considerable anthropogenic changes in global climate are to be expected approximately after 2030-2040.
 Thus, a quantitative estimation of using global water resources, for the past years and in the decades to come, was based on water use for public and domestic needs, industrial production, and agriculture (irrigation). Water losses during reservoir construction were taken into account as well. All the estimations for the future have been made with no account of the possible anthropogenic global climate change, i.e., for the stationary climatic situation.

4.2. Principal water users and the tendencies of their development.

The municipal water use is directly related to water withdrawal by population of cities, towns and housing estates, domestic and public services enterprises. The public supply includes also water expenses for the industry that directly provides the needs of urban population and consuming water with high quality from the city water supply system.. In many cities, a considerable water quantity is spent for watering vegetable-gardens and personal houses, and garden plots.
The volume of public water use depends on urban population number and the degree of its equipping with services and utilities. It means the availability or unavailability of water pipelines, canalisation, centralised hot-water supply. Also it much depends on climatic conditions. In many large equipped cities of the world the modern water withdrawal comprises 300-600 l/day per person. By the end of the current century, in industrially developed countries of Europe and North America, the specific per capita urban water withdrawal is expected to increase  up to 500-800 l/day. On the other hand, in developing agricultural countries of Asia, Africa, and Latin America, public water withdrawal is 50 to 100 l/day. In individual regions with insufficient water resources, it is not more than 10 to 40 l/day of fresh water per person.
A greater part of  the water that has been withdrawn from the urban water supply system  is returned back after being used (purified or not) as waste water to the hydrographic network. This occurs if the urban canalisation operates effectively. The principal part of consumption consists of water losses for evaporation with leaking from water supply and canalisation systems, for watering plants, streets, recreation zones, and personal plots. Thus, to a large extent, it depends on climatic conditions. In dry hot regions the losses are certainly larger than in cold and humid.  The water consumption for personal human needs is insignificant as compared with water losses for evaporation.
Relative values of consumption expressed usually in percentage of water intake depend to a considerable extent on the volume of full specific water withdrawal for public supply. So, in modern cities equipped with centralised water supply and efficient canalisation systems,  the specific water withdrawal is  400 to 600 l/day, and consumption is usually not above 5 to 10% of total water intake. Small cities with a great stock of individual building not fully provided with centralised system  have a specific water withdrawal of 100 to 150 l/day. Consumption significantly grows and can reach 40 to 60% of water intake. In this case, the least values  take place in northernmost regions, the largest in dry southernmost.
The modern tendency of public water supply development in all countries of the world is the construction in small and large cities of effective centralised water supply and canalisation systems. Also it is connecting to these systems a greater number of buildings and populated areas. In this connection,  in the future the specific per capita water withdrawal is expected to increase, and the values of water consumption expressed in percentage of water intake are to be considerably decreased.
Water in industry is used for cooling, transportation and wash, as solvent, and enters the composition of finished production. As the principal water user in industry, thermal and atomic power comes forward. It requires a great amount of water to cool assemblies. The volumes of industrial water withdrawal are quite different not only for individual branches of industry but also for making one kind of production depending on the technology of manufacturing process. It depends on climatic conditions as well. As a rule, in the northern regions, industrial water withdrawal seems to be considerably less than in southern regions with higher air temperatures.
In addition to thermal power, the principal industrial water users are chemistry and petroleum chemistry, ferrous and non-ferrous metallurgy, wood pulp and paper industry, and machine building. Major characteristics of industrial water use (the volume of fresh water withdrawal, water consumption, water diversion) depend to a very large extent on the water supply system scheme accepted for use. As known, there are the two basic schemes: an inflow and circulating. With the inflow water supply system  the water extracted  from the source, after being used (purified or not), is discharged into water streams. With the circulating system the used water is cooled, treated, and returned back to the water supply system. Thus, the system of circulating water supply excludes the discharge of used waters back into water bodies or water streams and envisages their multiple use in the production.
The necessary fresh water intake in case of the circulating water supply is insignificant. It is determined by the discharge necessary to restore water consumption spent in production and regeneration processes as well as for periodic  water replenishment in circulating cycles.
The value of industrial water consumption is usually an insignificant fraction of water intake. However, it strongly varies depending on industrial branches, water supply nature, technological process, and climatic conditions. In thermal power, this value is about 0.5 to 3% of water intake. In most industrial branches it is 5 to 20% reaching 30 to 40% in individual branches. With the inflow water supply system the water consumption expressed in percentage of water intake is considerably less than with the circulating system, and fresh water intake -- vice versa.
The development of industrial water withdrawal is one of the main causes of natural water pollution in the world. This is explained by the fast industry growth in different countries, especially the most water-consuming  productions, and the intensive development of thermal and nuclear power  plants. A greater part of water intake, after being used for industrial needs, is discharged as waste water. In most cases, it is not or partly purified, which  intensively pollutes water bodies.
To struggle with the pollution, many countries undertake energetic measures to decrease an industrial water withdrawal, and especially the waste water discharge. Therefore, since the 1970-80s a stable tendency has been observed as to stabilisation and even decrease in industrial water withdrawal.
Estimating the future volume of industrial water use for individual regions, countries or river basins, it should be taken into account that it varies under the influence of different tendencies. On the one hand, this volume should increase due to growing industry and thermal power production. On the other hand, this increase is not to be proportional to industrial growth. In the future, most countries would have the tendency to ever increasing transition to circulating water supply systems.  Many industrial branches will pass to the so-called water-free, or dry, technologies.
In some countries and regions of the world, there is the tendency of a greater use for industrial purposes of marine waters. All other things being equal, the volume of industrial and thermal power water consumption is to be much greater in southern regions with dry hot climate, than in the northern redundantly moisturised regions. In addition, industrial and thermal power water consumption depends much on the water supply system applied. With the inflow system, there are  the least consumption, and with the circulating system, fresh water intake and waste water volume drastically decrease. However, the consumption frequently increases by 1.5 to 3 times.  Therefore, in the future, circulating water supply systems should be developed by all possible ways. This would make it possible to repeatedly use  water in industry. Then, water consumption is to be expected to slightly increase (in percentage of water intake) as a whole for individual countries, regions, and large river basins.
Water use by agriculture is primarily determined by the development of irrigated landuse. In many countries and regions of the world, irrigation is the principal water user.
It causes a deficit of water resources especially during dry years. Land irrigation has been practised for millennia. However, irrigated lands were mainly introduced into use in the world in the 20th century. Before the late 1970s, almost all developed and developing countries on all the continents had preserved the tendency of intensive irrigation development. This intensive irrigation could provide the growth of irrigated areas and guarantee an increased  crop production. In the 1980s, the rate of global increase of irrigated areas much dropped. This occurred at the expense both of developed and developing countries. The cause  was, primarily, a very high cost of irrigation system construction, then  soil salinization, the depletion of irrigation water-supplying sources, and the problems of environmental protection. In a number of developed countries, by the present time the amount of irrigated lands has stabilised or even has a tendency to a reduction.
Considering the problem on the global scale, one should mention that the development of irrigation of dry lands follows from the necessity of food supply of humanity. At the present time about 15% of all cultivated lands are being irrigated. However, the food production from the irrigated areas amounts almost to half the total crop production in terms of value. In the modern world, the population number grows with a great rate. At the same time, there is the sharp food deficit experienced now by almost the two thirds of the world population. Therefore, irrigation is being given a great role in increasing the landuse and cattle-breeding efficiency. Thus, irrigated farming is expected to intensively develop in the future world. Irrigated areas would expand mainly in the countries with an extremely rapid population growth and sufficient water and land resources. The total global irrigated areas seem to increase, although with not the same rate as in the 1970s.  Subsequently the water use for irrigation is expected to grow.
Water expenses for irrigation are determined by  irrigation areas, the values of specific water intake in cubic metres per one hectare per year, and returnable waters in percentage of water intake. They depend on general physiographic conditions, serviceable condition of irrigation systems, watering techniques, and crop composition.
Information about water intakes and irrigated areas available in different countries allow calculation of specific water withdrawal for irrigation under different physiographic conditions. After the adequate analysis and generalisation it can be laid down into the basis of calculations of total water intakes by large natural-economic regions and continents. It is natural that the smallest values of specific water withdrawal are observed in northern countries and regions. For instance, in the north of Europe they are within 300-5000 m3/ha, in the regions of South Europe and East-European countries they amount to 7000-11000 m3/ha. The returnable waters are equal approximately to 20-30% of  water intake. In the USA, specific  water withdrawal for evaporation is estimated by different authors at 8-10 thousand m3/ha, returnable waters at 40-50% of water intake. In the countries of Asia, Africa, Central and South America, there is a great  variety of climatic conditions, crop composition, and  watering techniques. Therefore, the values of specific water withdrawal are very variable: from 5000-6000 m3/ha to 15000-17000 m3/ha, and in individual regions of Africa to 20,000-25,000 m3/ha.
The values of specific water withdrawal usually vary. In the future, they would considerably change depending on advanced irrigation systems, improved watering requirements, regime, and techniques. All these are to be taken into account in obtaining prediction estimates for irrigation water withdrawal in large regions of the Earth. A considerable water economy can be reached with using the most perfect modern engineering methods and means of watering (sprinkling, drip irrigation, etc.) that help to increase the crop productivity and decrease the irrigation water volume.
In terms of water economy the most efficient systems of drip irrigation are still very expensive and not widely used in the world. However, they decrease water expenses by approximately twice as much and increase the productivity. So, in the future they are expected to be more widely used, which should result in decreasing the values of specific water withdrawal. The same is promoted by improving the available irrigation systems, raising their efficiency and general effectiveness.
In agriculture, in addition to irrigation, water is spent on domestic needs of population, in cattle-breeding, and on modernising rural populated areas. The problem of supplying rural population and livestock with a high-quality fresh water is of great importance in many developing countries of the world located in arid regions. However, quantitatively as applied to water resources and water balance of large regions, the total water expenses for agriculture are insignificant as compared to these for irrigation (approximately, 5-8%). Mostly, water expenses for agriculture include these for irrigation.
Reservoirs. The construction of large reservoirs can lead to cardinal transformations in the time-spatial distribution of river runoff and increasing water resources in the regions during the low flow limiting periods and dry years. As a result of flooding vast territories, reservoirs make a considerable contribution into evaporation from water surface in the regions with insufficient moisture. This leads to decreasing total water resources of the regions. Thus reservoirs are one of the great fresh water users. This role of reservoirs is necessary to be taken into account in estimations of total water consumption by countries and continents, although many authors do not do this.
As long ago as millennia, reservoirs also were being constructed. However, as the objects of global scale they appeared only in the second half of the 20th century. For the recent 40 years all the largest reservoirs with total volume of more than 50 km3 have been built. At the present time the total full volume of world reservoirs is about 6000 km3, and the total area of their water surface reaches 500 thousand km2.
In developed countries of the world, the reservoir construction was most intensive during 1950-1970. At that time river runoff was almost fully regulated in many well-developed regions. Subsequently the rates of reservoir construction considerably decreased. However, in the countries with rich natural resources of river runoff they are still high. In developing countries, the highest rates of river runoff regulation were recorded during the 1970-80s. In accordance with the modern tendencies and available plans for future, during the next few decades quite a high rate of reservoir construction would be preserved in different regions of the world. This is caused by an increasing role of hydropower engineering under the conditions of liquid and solid fuel deficit. In addition, reservoirs provide a greater part of water consumption by industry, heat and atomic power stations, and agriculture. They are the basis for large-scale water management systems regulating space-and-time river runoff as well as protecting populated areas from floods and inundations. However, in the future the types of constructed reservoirs are supposed to be modified as well as their destination and territorial location. Reservoirs will be constructed in mountainous,  piedmont, and in weakly developed regions with no flooding vast areas of fertile lands suitable for agricultural use. In developed countries predominantly small and middle-sized reservoirs will be built. Reservoirs’ construction in different regions of the globe results in decreasing fresh water resources. This occurs because of additional losses for evaporation, whose value is very essential in the   total water consumption in individual regions.

4.3. Methodological bases for assessing and forecasting global water use.

There are several basic factors that determine quantitative characteristics of water use in large regions and countries of the world: the social-economic development level, population number, physiographic (including climatic) features, and the area of the territory. Their combination determines the volume and structure of water use, its dynamics and tendencies of development in the future.
For all natural-economic world regions presented in Fig. 3, and selected countries and river basins with reliably determined water resources characteristics, the analysis was made of global water use dynamics with space and time. For every region, country or basin the total water  withdrawal and consumption for urban population needs (domestic water consumption), industry (including thermal power), irrigated farming and agriculture, were estimated. Also an assessment was made of water losses for additional evaporation from reservoir water surfaces. All estimations have been made for different design levels: for the previous years, including  1900, 1940, 1950, 1960, 1970, 1980; for the current period of 1990 to 1995, and for the future for 2000, 2010, and 2025. This approach made it possible to follow the distribution of water use in the world by the territory and the dynamics with time within the current century and by the beginning of the 21st century.
Primarily water use was estimated for the countries of the world. Then the values obtained were generalised for large natural-economic regions and the continents.
Totally to this or that extent the analysis was made of water use data and the determining factors for about 100 countries. Preference was given to the national data on actual or calculated water use in individual countries or groups of countries. At the present time this data, more or less detailed and reliable, is available for many countries on all the continents. With the lack of actual data, estimations were made by using specially developed methodological approaches. They took into account the principal factors determining the value and dynamics of water use (total water withdrawal and consumption). In this case, the analogue method was widely used. As analogues, the use was made of the countries with  reliable water use data. These countries were located in similar physiographic conditions and had the same level and features of economic development.
Water use by population in cities and rural areas was estimated by data on dynamics of population number (urban and rural) and per capita specific water withdrawal. The dynamics of population number for the past years was taken from the statistical hand-books and for the future from the  data of the 1995 UNO forecasts. Per capita specific water withdrawal and the fraction of total water consumption for every country were taken from the published national data or the materials of international organisations. In case of the data unavailability, the above values were based on specific water withdrawal in countries-analogues.
An assessment of water use for irrigation was carried out by analysing the dynamics of some characteristics for the previous 30-40 years. They included population number; the area of irrigated lands by years (by FAO data) including in specific values (in ha per capita); the values of Gross  National Product (GNP), expressed in US dollars per capita. In this case the values of specific water withdrawal and water consumption were taken from the data of national estimations or by countries-analogues.
Calculations of water withdrawal for 2000, 2010, and 2025 were mainly based on forecasting the areas of irrigated lands. For this purpose, the analysis was primarily made for previous years of the tendencies in their changes in combination with the above determining factors. The analysis was carried out separately for every country. As a result, by groups of countries with different GNP-levels clear analogies were found in the tendencies of changing irrigation areas depending on population number and GNP-values. The indicated tendencies served the basis for the forecasts for future irrigation areas basing on the forecasts of population number and GNP-values for every country as initial data. The limiting factors are areas of the lands suitable for irrigation and the values of water resources accessible for use.
To estimate the future specific water withdrawal for irrigation, consideration was given to irrigation tendency to decrease due to improving technological procedure and engineering watering means directed to economising water resources.
Industrial water withdrawal was calculated basing on the dynamics of industrial production in different regions of the world. As analogues, acceptance is made of the available data on industrial water withdrawal in many countries of the world including those with different level of economic development, located in different physiographic conditions. Calculations for the current and future periods were carried out separately for thermal power  and other industrial branches with considerably differing tendencies and rates of development and  water losses. Then they were summed up for every region. Total water consumption by thermal power engineering was assumed to be 1 to 4%.  In other industrial branches, it was taken as 10 to 40% of water intake depending on industry development level, the availability of water circulating supply systems and climatic conditions. An assessment for the future period to 2025 was made for every country taking into account special UNIDO (1996) developments. These developments were based on the analysis of modern situation in the world and the forecast of GNP-values. As a result, for all principal countries an increased industrial water withdrawal was predicted to occur by 2025 as compared with 1990, with considering different scenarios of global development and electric power production. We were based on the most optimistic development scenario (Global Balance), however, for an average growth level of electric power production. In accordance with this UNIDO scenario for developed countries, water withdrawal is expected to increase by 1.4 to 2.9 times and for developing by 3 to 10 times.
 The analysis of industrial water withdrawal growth by UNIDO (by all scenarios) showed its overestimation and not a full correspondence to the modern tendencies of this water withdrawal change. In this connection the UNIDO data for 2025 by the accepted scenario (Global Balance) were reduced by 20 to 30% for developing countries and down to 40 to 60% for developed countries.
Additional water losses for evaporation from reservoirs were calculated for all principal reservoirs of the world with the volume of more than 5 km3 by the difference between average evaporation from water surface and land. In this case the coefficient showing the ratio of an additional area of reservoir water surface to its total area was taken into account.
The initial data on reservoirs (area, volume, location, years of construction and other characteristics) were taken from generalising international monographs as well as from other publications by individual countries and regions. The evaporation norms for water and land surfaces were determined by the charts of the Atlas of World Water Balance.
The future losses for evaporation from reservoirs  were estimated for every region. These estimations took into account the recent decades' tendencies and available long-term plans of building large reservoirs in different countries and regions as well as their physiographic features. The modern tendencies of building reservoirs in developed and developing countries of the world, indicated in Section 4.2, were taken into account.

4.4. The dynamics of water use in the world.

Table 7 and figure 8   shows the dynamics of water use by the continents for the current century and for the future till 2025 obtained on the basis of the above initial data and methodological approaches.
The modern (for 1995) global total water withdrawal comprises about 3790 km3/year, consumption - 2070 km3/year (61% of withdrawal). In the future the total water withdrawal will be growing by about 10-12% for every ten years, and by 2025 it will reach approximately 5240 km3/year (a 1.38-fold increase). Water consumption will be growing somewhat slower and the increase will be by 1.33 times. At the present time about 57% of total water withdrawal and 70% of global water consumption fall on Asia, where the major irrigated lands of the world are located. During the next decades the most intensive growth of water withdrawal is expected to occur in Africa and South America (by 1.5-1.6 times), the smallest in Europe and North America (1.2 times).
Table 8 , figures 9 and 10 show the role of individual water users in the dynamics of global total water withdrawal and consumption as well as of population and irrigated area growth.
At the present time, agriculture receives 66% of total water withdrawal and 85% of consumption in the world. In the future the role of agriculture will slightly decrease mainly at the expense of expected more intensive growth of other water users, primarily the industrial and public water withdrawal. By 2025, agriculture is expected to increase its requirements for water withdrawal by 1.3 times;  industry by 1.5 times, and the global public supply by 1.8 times. An additional evaporation from reservoirs greatly contributes into water losses. They are above the  total water consumption by industry and public water supply together.
By specified data the global irrigation area in 1995 was 253 million ha. By 2010 it is expected to grow up to about 290 million ha, and by 2025 to 330 million ha.
It is very interesting to make a comparison of water withdrawal in the world (table 7 & 8) with the forecasts by many authors previously published by the world press (Doxiodec’s, 1967; M.I. Lvovitch, 1974; Falkenmark, 1974; Richa, 1982, etc.). This comparison shows the forecasts to predict considerably excessive values for the future. This fully refers also to the forecasts published in Russia by the author of this summary as long ago as 1987.
The main cause of these excessive forecasts is due to the two factors. First, all authors of the 1960-80s forecasts  predicted a drastic increase in irrigated lands. This was so because these forecasts were based on very high rates of global irrigation development being at that time ahead of population growth rates (see Fig. 10 ). Second, the forecasts took into consideration but insufficiently the tendencies of stabilisation that appeared in the 1970-80s  and even of decrease in industrial water withdrawal in many countries of the world.
The role of individual water users in the dynamics of water use on all the continents is shown in Tables 9 & 10 . Europe and North America have a similar structure of modern and future water use. Here industry plays a considerable role in the total volume of water withdrawal. In 1995, the European industry used 45% of total water withdrawal. By 2025 it is expected to increase to about 45%. In North America, the current water withdrawal by industry is 41% of total water withdrawal in 1995 and at 2025. As to water consumption, in Europe and North America, the leading role belongs to agriculture that shares more than 70% of water consumption.
In Asia, Africa, and South America, agriculture (irrigation) plays the leading role in the structure of water use. In 1995, irrigation took 60-80% of total water withdrawal and 64-91% of total water withdrawal. These indices will slightly change also by 2025, although by this time industrial water consumption is expected to grow by 2 or 2.5 times on these continents. Nevertheless, the fraction of industry in total water withdrawal would not be above 24% in South America, 15% in Asia, and 6% in Africa. The characteristic feature of water use structure in Africa is a great role of evaporation from reservoirs. At the present time and in the future, it amounts to 34-35% of the total water consumption on the continent.
The dynamics of water withdrawal by natural-economic regions of the world is shown in tables 11 & 12 . The values of water withdrawal are very unevenly distributed by the regions of the continents and do not agree with the values of water resources. For instance, in Europe 95% of water withdrawal fall on southern and central parts of the continent; in North America, the USA take 73% of water withdrawal; in Australia and Oceania 89% of water withdrawal fall on Australia. On the Asian continent the greatest volume of water withdrawal belongs to Southern Asia regions including India, Pakistan, Bangladesh, and South-Eastern Asia with the most part of China irrigated areas.
In Africa, the greatest water withdrawal takes place in the northern part (North Africa); 50% of water withdrawal on the continent fall on this region. In South America, water withdrawal is more or less evenly distributed by the regions of the continent. The dynamics of water use growth till 2025 considerably differs by regions. In developed countries and in the countries with limited water resources, water withdrawal is expected to rise by 15-35%. In the regions with developing countries with sufficient water resources, the water withdrawal growth can be 100 to 200%.

4.5. Water resources and water use.

Of particular interest is comparison of water use with renewable water resources of surface waters. These data by all regions of the world for 1995 and 2025 are presented in Fig. 11 . This figure shows a comparison between the total water withdrawal and the values of local water resources summed up with half the inflow from outside. Thus, it is conventionally anticipated that every region can have at its disposal half the fresh water inflow from neighbour regions.
In accordance with the data presented in Fig. 11   the modern water withdrawal in the world as a whole is not great in total amounting to 8.4% of global water resources. By 2025 this figure is expected to increase up to 12.2%. However, water resources in the world are distributed very unevenly, which is seen even when comparing water withdrawal and river runoff by the continents on the average. Even at the present time in Europe and Asia, water withdrawal comprises 15-17% of water resources, and in the future it will reach 21-23%. At the same time in South America and Oceania only 1.2-1.3% of river runoff are used, and even in the future it is unlikely that this value  will be above 1.6-2.1%.
The distribution of river runoff and water use is especially uneven in the natural-economic regions of the world. Within every continent (except for South America), on the one hand, there are regions with a large extent of using water resources. On the other hand, regions with an insignificant water use (especially water consumption) as compared to water resources (see Fig. 11 ). For instance, in Southern and Central parts of Europe, modern water withdrawal amounts already to as much as 24-30% of water resources. At the same time in the northern part of the continent these values are not above 1.5-3.0%. In the northern part of North America water withdrawal is not above 1% of water resources, and for the US territory this value is 28%. Even greater contrast takes place for Africa and Asia. In the northern part of Africa even at the present time renewable water resources are almost totally withdrawn (water withdrawal is 95% of water resources). In other regions (especially in Central Africa), water withdrawal is negligibly small as compared with the value of water resources. In Asia,  including the regions of Southern, Western, and Central Asia & Kazakhstan the use of water resources is very great (42-84%). At the same time in the region of Siberia and the Far East, this use is not above 1%. Only in South America, in all regions, the value of using water resources is insignificant being not more than 2-4%.
In the future, by 2025, the unevenness in the distribution of water resources and water use will be preserved and even much more increase. At the present time in many regions the use of water resources is already quite great. In the future this use will grow much and reach critical values. By contrast, in northern regions and in the regions with excessive moisture on all the continents water use (especially water consumption) will comprise, as previously, a very insignificant part of water resources.
Analisis of the extent of modern water resources use have been done also for individual countries of the world as the ratio of water withdrawal (for 1995) to water resources (local water resources summed up with half the inflow). The above data show that water resources are fully depleted in many countries. They use not only all local water resources but also a greater part of fresh water inflow incoming from neighbour territories. According to the above results at present about 75% of the Earth’s population live in the countries and regions with the extent of water resources use of more than 20%.

5. Water availability and water resources deficit.

Water resources distribution over the territory of the Earth is uneven. Also they disagree with population spread and economic development. These are very clearly revealed by analysing and comparing the specific water availability for a single period of time for different regions and countries. The specific water availability represents the value of actual per capita renewable water resources.
For every design level the specific water availability is determined by dividing water resources without water consumption by the population number. In this case, water resources are assumed to be  the river runoff formed in the territory of the given region and summed up with half the river water inflow from outside. So, the specific water availability is meant the residual (after use) per capita quantity of fresh water. Obviously, as population and water consumption grow, the value of specific water availability decreases.
The values of specific water availability were obtained for all natural-economic regions and selected countries for the 1950-2025 period. As expected, their analysis revealed a strong unevenness in their distribution over the Earth’s territory.
For instance, the greatest water availability of 170-180 thousand m3 per capita for 1995 is in the regions of Canada  with Alaska and in Oceania. At the same time, in densely populated regions of Asia, Central and South Europe, and Africa the modern water availability is within 1.2-5 thousand m3 per year. In the north of Africa and on the Arabian Peninsula, it is as much as 0.2-0.3 thousand m3 per year. It is worth mentioning that water availability of less than 2 thousand m3 per year per capita is considered to be very low, and less than 1 thousand m3 per year catastrophically low. With these values of water availability, very serious problems arise unavoidably with population life-support, industry and agriculture development.
To see more clearly the distribution of specific water availability values by natural-economic regions of the world, it is presented on global charts for 1950, 1995, and 2025 ( Fig. 12 ). On all these charts the specific water availability of every region is designated with shading by the following gradations: (in thousand m3 per year per capita):
  < 1 - catastrophically low;
 1.1.-2.0 - very low;
 2.1-5.0 - low;
 5.1-10 - average;
 10.1-20 - high;
 > 20 - very high.
In 1950 (Fig. 12a) over the greater part of the Earth’s surface the specific water availability was average or above average, and only in North Africa it was very low. In Central and South Europe, North China and South Asia it was low (from 2.1 to 5.0 thousand m3 per year). The catastrophically low water availability was observed in no regions of the world. By 1995 the situation drastically changed. In many regions of the world (Fig. 12b), population water supply sharply decreased and became catastrophically low in North Africa and on the Arabian Peninsula, very low in North China, Southern  and Western Asia. In seven regions more, a low water availability of 2.1 to 5.0 thousand m3 per year) was recorded. At the present time, totally for 76% of population the specific water availability is less than 5.0 thousand m3 per year per capita (by the accepted gradation low water availability, very low and catastrophic). As this takes place, 35% of Earth’s population have very low or catastrophically low water supply.
The situation will more deteriorate in the beginning of the next century (Fig. 12c). By 2025 the greater part of Earth’s population seems to live under the conditions of low and catastrophically low water supply. Approximately 30-35% of the world population will have catastrophically low  fresh water supply (< 1 thousand m3 per year per capita). At the same time for all design levels including the future, a high specific water availability will take place in North Europe, Canada and Alaska, almost all over South America, Central Africa, Siberia, the Far East, and Oceania.
To discover water resources deficit in the world in the future, it is very important to analyse the tendencies and rates of changing specific water availability depending on social-economic and physiographic conditions. The analysis was made of the data obtained for natural-economic regions of the world. It showed that the rates of lowering water availability depended on the two factors: social-economic development of the countries included into the region and on climatic conditions of the region. The graphs presented in Fig. 13 convincingly confirm this. They show the dynamics of specific water availability since 1950 to 2025 in relative units (as compared with 1950) averaged for the three groups of regions including:
 

•  industrially developed countries;
•  developing countries under the conditions of sufficient and excessive moisture;
•  developing countries in arid and semiarid countries.

6. Anthropogenic changes in global climate and water resources.

Throughout the entire time of existence of hydrometeorology all the methods for estimating water resources, water use, water availability, their temporal and spatial distribution were based on the conception of climate stationarity. It implied that the climatic conditions and subsequent water resources variations in the future would be analogous to those that took place during the past observational period.
In hydrology and water resources this conception is so far to a full extent used all over the world not only to assess water resources and water use but also to calculate extreme river runoff characteristics necessary for construction design.
This conception is laid also down into all the above values of water resources and water availability including for the distant future of 2010-2025.
A long-term experience of design and exploitation of different water management structures in the world showed the correctness and reliability (any case to the present time) of using premises of climate stationarity. In different countries in different periods of time regional and global prediction estimates were obtained for water use and water availability, including the distant future. They were always based on the conception of stationary climatic situation.
As long ago as ten years the rightfulness of this conception arose no doubts at scientists in the field of water problems. However, the situation has cardinally changed in the recent years when the problem was raised of anthropogenic climate change due to atmospheric CO2 increase because of the carbon fuel burning, industry development, and deforestation.
Carbon dioxide (and some other gases) being even in small amounts in the atmosphere is able to considerably attenuate the long-wave radiation thus creating the so-called “greenhouse effect”. This in turn promotes an air temperature rise. By some very authoritative assessments of climatologists, in the decades to come this effect can lead to so drastic global climate change that had never been recorded throughout the history of mankind. In case it is true, this should be taken taken into consideration when estimating water availability for the distant future. In this case the questions arise: How much rightful is the acceptance of the conception of climate stationarity especially when estimating water availability for 2010-2025? What errors are possible to occur and in what regions with ignoring the factors of global warming?
As long ago as the previous century, individual measurements of atmospheric CO2 were carried out. However, they were very unrealistic. The modern systematic measurements were started in 1958. At that time the carbon dioxide concentration was 315 conventional units (0.031% by the volume of the total gas amount). By 1990 it increased to 350 units, or by 11.1%. By using different methods it was quite reliably established that in 1880 (the beginning of the epoch of intensive industrialisation) the carbon dioxide concentration was 285 units. It means that for 110 years carbon dioxide has grown by 22.8%, and 25% of this value fall on the recent 10 years, i.e., the intensity of increase ever grows.
By the present time the carbon dioxide growth has led to about 0.5єC rise in air temperature. As empirical data show, an especially great rise of air temperature occurred beginning with the 1980s. And in the recent 10 to 15 years in some regions of the Earth the air temperature has increased by 1-2єC, which resulted in a noticeable change in renewable water resources and in particular in their distribution during a year. Many authors associate these changes in climatic characteristics and river runoff regime directly with  the processes of anthropogenic global warming.
In the future the carbon dioxide concentration will continue to grow with the intensity determined by the scenarios of power engineering and industry development. Also this will depend on the measures undertaken in different countries to diminish the release into the atmosphere of carbon dioxide and other accompanying gases. Nevertheless, in accordance with the available forecasts the growth rates in the decades to come will be high enough. During the next century Earth’s atmospheric carbon dioxide is expected to double (i.e., to reach the value of 700 conventional units). However, as to the question when it can exactly occur, there are different points of view. The most forecasts for 1984-1988 anticipated that carbon dioxide concentration of 700 units would be reached by 2050 (see Fig. 15 ). However, a more detailed analysis was accomplished quire recently by a group of UNO experts (IPCC). They were basing on different factors of the natural environment and new data on the future use of organic fuel and the measures to be undertaken to limit carbon dioxide emission to the atmosphere. As a result, lower prediction estimates of concentrations were obtained. Figure 15 presents one of the variants of these estimates. It anticipates the most intensive growth of concentrations and predicts them to double only by the end of the next century.
Carbon dioxide concentration growth results in rising the global air temperature, and the value of warming can vary in a wide range depending on the premises accepted and detailed consideration of acting factors. So, by assessments of 1984-88 with atmospheric carbon dioxide doubling the global air temperature was predicted to rise by 2.5-3.5єC (as compared with the 1980s temperature), whereas by assessments of 1994-1996 by 1.6 to 2.2єC ( Fig. 15 ). The recent estimates consider to a larger extent the role of atmospheric aerosol that noticeably attenuates the solar energy income and contributes to air cooling, as well as some other factors.
The above values of potential changes in global air temperature, whatever reliable they would be, are insufficient even for the most approximate estimates of the future water resources. They can be obtained only on the basis of the data on potential regional changes of climatic conditions (primarily, precipitation and air temperature by seasons or months). Unfortunately they are assessed extremely unreliably even for the most large regions and river basins. To forecast changes in regional climate with global warming, the atmospheric general circulation models (GCMs) are widely used as well as the materials of palaeoclimatic reconstruction of the past warm epoch climate. There is a wide variety of GCM-types developed in different countries of the world. They produce the values of monthly variations in air temperature and precipitation for the entire land territory with doubling carbon dioxide concentration, and some of these models even for smaller concentrations. It is necessary to mention that with a 2 to 3єC global warming in individual regions especially in high latitudes the air temperature is expected to rise up to 5 to 6єC, whereas the smallest values (0-1єC) can take place in the subequatorial regions. This general feature of global warming is given by all GCM-types and the use of palaeoclimatic analogues.
The greatest complexity of using in practice GCMs to estimate the regional climate change is disagreement among the results produced for the same regions. In case of estimating air temperature changes the results of using different models agree as if qualitatively. When assessing changes in precipitation for the same region, we can obtain not only drastically differing but even directly contradictory results. This makes it impossible to plan any real events to solve the water supply problem in the future. In these conditions, with global warming numerous estimates were obtained in different countries for possible changes in the characteristics of river runoff, water resources and water requirements. These estimates are to be considered as the analysis of the sensitivity of river catchments  and water management systems to various potential future changes in regional climate rather than the forecasts. Undoubtedly, this analysis is of scientific interest and can be very useful in practice to develop measures for increasing the efficiency of water management systems under the conditions of a drastic climate change. Below cited are some principal conclusions drawn in different countries when studying climate change effects on water resources. These conclusions were presented in the reports of a group of UNO experts (IPCC) that worked since 1988 to 1995. Also they were given by the author of this booklet together with colleagues from the USA at the Second World Climate Conference (1990).
The studies convincingly pointed to the exclusively serious consequences for water resources (especially in arid regions) possible with changing regional climatic characteristics. For instance, with a 1 to 2 єC increase in the annual air temperature and a 10% decrease in precipitation, a 40 to 70% reduction of annual river runoff can be expected in the regions of insufficient moisture. These data are obtained by many authors for individual river catchments located in arid zones of the Northern Hemisphere. The same is true for water resources of large economic regions. The US scientists calculated that with a 2єC climate warming and a 10% precipitation reduction a 1.5 to 2.0-fold decrease in  water resources is possible for water management regions of the country located in the arid climate zone.
The conclusions about strong sensitivity of water resources even to comparatively small changes in climatic characteristics were also obtained for many other world regions. For all physiographic conditions the values of water resources turn out to be more sensitive to changes in precipitation than in air temperature. The results obtained allows very certain inferences to be made. In case the global warming will be accompanied with precipitation reduction,  water resources in arid regions of the world will drastically diminish. As shown above, these regions occupy about 30% of European territory, 60% of Asia, the greater part of Africa, the south-western regions of North America, 30% of South America, and the most part of Australia. In the conditions of modern changing climate they have ever annually increasing deficit of fresh water.
Estimations made for the regions with cold and temperate climates showed that the global climate change could affect the seasonal and monthly river runoff to a considerably larger extent than the mean annual values. Especially great changes in river runoff distribution during a year are to be expected in the regions where the principal water resources are formed during the spring high water period. Among them are the territories of the greater part of the FSU, Europe, North America, and some mountainous regions.
For instance, calculations were made for the three principal rivers of the European part of Russia and the Ukraine: Volga, Dnieper, and Don. The calculations made show that with a 2.0 to 2.5єC global warming,  changes are expected to occur in their annual mean river runoff by 12 to 20% with increasing winter runoff by 2.0 to 3.0 times and decreasing spring high water runoff by 25 to 40%. Analogous inferences have been obtained by researchers for the Scandinavian countries (Finland, Norway, Sweden) as well as for some catchments in Poland, Belgium, and mountainous regions of the USA. This potential very significant change in seasonal river runoff (increasing winter and decreasing spring), as compared with the annual, is physically explained very simply. It occurs at the expense of more intensive snow melt during winter thawing taking place with rising winter air temperatures.
As a result of the studies carried out, a very important scientific and practical inference is drawn. There are the regions where with global warming  river runoff is mainly formed during a spring high water period. Then river runoff distribution during a year is more sensitive to air temperature rather than annual precipitation change.
Absolutely different situation takes place in the wet tropical regions. There the regime of river runoff and the values of water resources practically fully depend on changes in annual precipitation and its distribution during a year.
One of the most important aspects of studying hydrological consequences of global warming is estimating possible changes in extreme characteristics of river runoff: maximum and minimum water discharges in rivers. These estimations were made by analysing empirical data and theoretical calculations by models. As a whole, with enough extent of reliability they demonstrated that on river catchments, especially small and middle, the global warming would lead to more considerable changes in extreme river runoff characteristics rather than mean annual and seasonal. On the one hand, increasing maximum floods can be expected, and on the other hand, more frequent occurrence of severe droughts. Both could result in very serious economic and ecological consequences, especially for urbanised territories and agricultural regions with unstable moisture.
It is well known that the sharpness of water management problems in different regions is determined by the ratio of available water resources and requirements for them. An expected climate change would obviously affect not only the values of natural water resources but also water requirements, water use and availability largely dependable also on climatic conditions.
With predicted future climate change, on the one hand, the plans for future development and location of irrigated areas, the most water-consuming industrial branches, and reservoir constructions are to be revised. On the other hand, the problem may arise with water supply of the available water users. As this takes place, all these changes would, primarily, touch the hot arid regions. These regions, already at present have the difficulties  with water supply, conflicts among different water users and water consumers. At the same time it should be mentioned that water use is quite an inertial value smoothly varying with time. Therefore, with  global warming the regional water availability  would be primarily determined by changing water regime in the territory. Any case, the global warming would contribute  even more uncertainty into the problem of assessing future water consumption and availability. In some regions of the world, water availability could considerably improve. In the others, the water problems would become even more acute.
As a whole, studies on the problem of climate change effects on water management allow making the following important conclusion. With an expected global warming and great uncertainty of regional climate change, the water management systems of river basins are to be complex and maximally flexible, capable to efficiently control over water resources with different climate situation. In this connection, with global warming the regions with a great extent of river runoff regulation (many regions of the USA and Canada, river basins of Europe) would have considerable advantages relative to solving the problem of water supply and flood regulation as compared with the regions with river systems having natural runoff regime (e.g., Southern and South-Eastern Asia and South America).
Approximate estimates have been recently obtained by the author et al. (SHI) for possible change in water resources with global warming for natural -economic  regions, continents, and the Earth as a whole. They showed that the results fully depend on accepted climatic scenarios for the future, based on different GCMs and palaeoclimatic analogues. For instance, basing on palaeoclimatic scenarios an increased global river runoff of up to 16-18% is expected with a 2 to 3єC global warming. In this case the estimates appear to be very favourable for hot arid zones as well as tropical regions. However, these optimistic inferences are not confirmed if we are based on GCM-scenarios with atmospheric carbon dioxide doubling. By these scenarios, there is no noticeable increase in river runoff in arid regions of the Earth and the tropical zone. For these regions GCMs give a very considerable air temperature growth and a slight precipitation increase .
The results of  water resources assessments are to be analysed in more detail. These assessments are based on different climatic scenarios for the five representative world regions located in North America, South Europe, South-Eastern Asia, Sahel, and Australia. For this purpose, by using the same technique the estimates of variations in annual river runoff have been obtained by three most advanced types of GCMs with carbon dioxide doubling and palaeoclimatic scenarios with a 2 to 3C warming. This analysis of the results shows, first, that calculations by the model scenarios can yield the inferences differing manifold and even contradictory in terms of process directionality. Second, for the most regions the model scenarios yield a water resources decrease to 10 to 30%, palaeoclimatic - an increase up to 15 to 40%. The principal cause of these discrepancies is the different estimates of future potential changes in precipitation. This is the weakest place of all the climatic scenarios. Actually, changes in precipitation  mostly affect water resources and create great uncertainties when assessing them for the future.
All the aforementioned allows the following general conclusions to be made about anthropogenic global climate change effects on water resources.
 

•  Water objects and water management systems are very sensitive to changes in climatic characteristics, and the most noticeable changes are expected with a more than 1єC-warming and a more than 10% -precipitation change.
•  The  currently available scenarios of potential anthropogenic changes in regional climate are extremely uncertain. There is disagreement among them. It concerns primarily a precipitation change to be the principal factor determining  water resources and water use. None of currently available scenarios of future climate can produce a reliable basis for estimating global changes in water resources for the large natural-economic regions, the continents and the Earth as a whole.
•  With assessing water resources, water use, and water availability on the global scale till 2025 it is quite admissible to neglect the possible anthropogenic climate change. This is grounded by the data shown in Fig. 15 . According to this data, by the recent estimates the global air temperature is expected to rise by 2025 by as much as 0.6єC. Also it is confirmed by great uncertainties in regional precipitation and air temperature changes.

7. On the ways of eliminating fresh water deficit in the world.

As seen (Figs. 12 - 13 ), with an extremely uneven natural space-and-time water resources distribution, an intensive man’s activities, and a rapid population growth, even at the present time a significant fresh water deficit takes place in many world countries and regions, especially during dry years. Calculations show that in the decades to come the most part of Earth’s population and many tens of countries in the world would have a critical situation with water supply. Water resources deficit becomes a factor deteriorating the living standard of population retarding the economic and social development in most developing countries of the world. It is already clear that in the first half of the 21st century the water problem will be of the most importance even among such global problems of humankind as food  and power production.
The critical situation with water supply will require enormous financial and material expenses for elaboration and realization of the measures to eliminate the deficit of pure fresh water in different physiographic conditions.
At the present time and in the visible future the most realistic and efficient measures would be: an  overall economy and protection of water resources by a drastic decrease in specific water consumption, especially in irrigated landuse and industry, reduction or full cessation of waste water discharge into the hydrographic network, the more full use of local waters as a result of seasonal and long-term river runoff regulation, the use of salt and brackish waters, an active influence on precipitation-forming processes; using secular water storage in lakes, underground aquifers, glaciers; and the territorial re-distribution of water resources.
All these measures require rather great material expenses and have different limitations. Almost all of them exert a pronounced effect on the natural environment. It means that they are far from being harmless in terms of ecological consequences. These consequences could be rather significant and little predictable. An exclusion is the measures for waste water treatment and decrease in specific water withdrawal that are always necessary, desirable and useful for preserving water resources and the natural environment.
    All the methods for obtaining additional water resources seem to find with time a wider usage to solve the water supply problem. Primarily, this pertains to those regions of the world where they would appear to be most appropriate, ecologically admissible, and economically profitable by physiographic conditions and the character of water use.
It is worth mentioning that among the measures for eliminating water resources deficit of especially  long existence is man’s runoff regulation and its territorial re-distribution that are very promising in the future. The measures for a partial transfer of river runoff from one region to another are objectively grounded by the current reality of the water resources formation, their spatial distribution, and the character of use. First, the river runoff resources on the Earth are quite enough as a whole to meet the demands for water requirements for many decades ahead. Second, fresh water resources on the Earth are distributed extremely unevenly: on every continent, there are regions with excessive water resources and the regions with their deficit. Third, man’s economic activities promote the strengthening of natural unevenness in water resources spatial distribution. It means that where water resources are in excess, they are less used, and river runoff practically is not reduced. In the regions with water resources deficit due to anthropogenic factors effect the water deficit becomes with every year more and more tangible. Therefore, the human intention is obvious to work out and implement the measures for water intake from those regions, where it is in excess, and water transfer to the regions with its insufficiency. In the future, as water requirements and  technological and economic possibilities grow, the volumes and scales of runoff transfer seem to increase.  As this takes place, principal difficulties of developing large-scale measures for river runoff diversion in the world are determined by the necessity of detailed estimation of its effects on the natural environment, reliable forecast of possible ecological consequences, development and realisation of effective measures for their elimination rather than financial and technological possibilities.
In the distant future with anthropogenic global climate change and heat-moisture re-distribution over the Earth’s territory, as some scientists believe, it would be necessary to return to the large-scale projects of the territorial river runoff re-distribution. These projects were intensively developed during the 1960-70s. At the same time, our knowledge about possible anthropogenic climatic changes causes great difficulties in the use of large-scale projects of river runoff diversions to solve the water supply problems due to the two reasons. First, unfavourable climatic changes could embrace vast areas including the basins planned for runoff withdrawals. Second, we have very great uncertainties concerning possible regional climatic changes in order to realistically plan different large-scale measures even in the far distant future.

Conclusion

Every time when new data on Earth’s water resources and their use are presented, especially for the future time, the question arises of their reliability and accuracy. These depend in much on many factors and considerably differ for individual countries, regions and even continents. Water renewable resources estimates are based on observation data from hydrological network. Therefore, their reliability is primarily determined by the condition of this network: the number of hydrological sites, the character of their spatial distribution, the duration and continuity of observations, measurement quality and processing. So, basing on the WMO analysis, let us mention that more than half the observational stations for water discharge on world rivers are located in Europe and North America. The countries of these continents have the longest observation series. Most of hydrological sites (70%) equipped with self-recorders allowing most detailed and objective information to be obtained, are located in these countries. Therefore, it is natural that for the regions and countries of these continents the most reliable estimates are obtained for water resources dynamics and water availability. They agree well with the previous estimates. The discrepancies available are almost fully attributed to different periods of averaging of observed data.
Water resources estimates are most erroneous for of a number of regions of Africa (Northern, Eastern and Western Africa), Asia (Southern and South-Eastern Asia), and the islands in the northern part of the North-American continent poorly covered with hydrological data. To more reliably estimate water resources in these regions, it is necessary to develop hydrological network, improve the quality of observations and processing of materials.
In many developing countries, hydrological network is weakly developed, and its number is being reduced. However, the time spans between measurements, processing and data publications and submission to regional and international centres are being increased. There are a lot of countries that have the possibility but not interested in the operational exchange of hydrological information and its timely publication. Any case the situation with hydrometeorological information exchange cannot be considered normal in the world. Actually, in the mid-1990s there is the possibility to analyse global river runoff data  for the period to 1985. This takes place when meteorological information (air temperature and precipitation) is generalised for all over the world with only  a few months delay. This retards not only reliable estimation of water resources but studying the global hydrological cycle and improving GCMs. Urgent acting measures are to be undertaken by authoritative international organisations to cardinally improve the state of global hydrological network, and the situation with collecting, processing and exchanging hydrological information.
The situation is not better with the problem of taking account of fresh water use. For most developing countries reliable systematised materials about water withdrawal and diversion are, as a rule, unavailable. Even the published national data are based on very approximate proxy estimates.
In addition to initial hydrological and water management data some errors in water availability estimates arise because of estimating only river runoff data. At the same region in many regions of the world a considerable part of water withdrawal is being carried out due to underground water. By the recent data generalised by Prof. J. Margat (France) it can be supposed that at the present time about 600-700 km3 of annual water withdrawal are supplied from underground water. A greater part of this water is used for irrigation and municipal needs. For a number of countries with almost no river runoff (e.g., the Countries of Arabian Peninsula) underground water is the main source of population and economy supply with fresh water. Therefore, for these countries, calculations of water availability dynamics have primarily to take into account groundwater. An assessment of water availability there only by river runoff data yields an underestimated result.
The general pattern of global water availability would not change much in case of considering this problem on the global scale for the continents and selected natural-economic regions. First , the above water withdrawal values at the expense of underground water are as much as 15% of total global water withdrawal and approximately the same of water consumption by continents and regions. Second , according to approximate estimates, at least half the used underground water is hydraulically connected to river runoff. In this case underground water withdrawal would directly affect the adequate reduction of river runoff.
Thus taking into account the circumstances mentioned above the indicated values of water consumption at the expense of underground water can be decreased by half. Nevertheless, ignoring underground water account in assessments of water availability dynamics for a number of regions seems to produce a more pessimistic pattern than as it is.
However, it is far from being so because specific water availability was calculated for all regions and countries by mean annual river runoff with no consideration of their variability from year to year and during a year, which is especially important for arid and semiarid regions. In case calculations are based on minimum annual river runoff for the observation period, specific water availability decreases by about 1.2-2.0 times depending on climatic conditions of individual regions and countries. Thus, the inferences made about water availability level and water resources deficit by regions or countries are to be probably considered as one of the most optimistic variants of global water availability dynamics.
That the variant of above estimates is optimistic is confirmed by the fact that they were obtained with no account of qualitative depletion of water resources consisting in ever increasing pollution of natural water. This problem is very acute in industrially developed and densely populated regions of the Earth where no efficient waste water purification takes place. The major sources of intensive pollution of waterways and water bodies are contaminated industrial and municipal waster water as well as returnable water from irrigated massifs. By assessments made, in 1995 waste water volume was 326 km3/year in Europe, 431 km3/year in North America, 590 km3/year in Asia, and 55 km3/year in Africa. Many countries practise discharging a greater part of waste water containing harmful substances  into the hydrographic network. No preliminary purification is carried out. Thus water resources are polluted and their subsequent use becomes unsuitable, especially for population supply.
As known, every cubic metre of contaminated waste water discharged into water bodies and streams make unsuitable 8 to 10 cubic metres of pure water. This means that the most regions and countries of the world already at present are facing the threat of catastrophic qualitative depletion of water resources. Therefore, it is necessary to consider water supply problems for every region in detail. In this case the dynamics of fresh water quantity and quality anthropogenic change is to be taken into account.
 The reliability of prediction estimates for 2010-2025 requires a special attention. They certainly consider the tendencies observed during the past decades. However, to a considerable extent they are based on long-term demographic and global economic development forecasts by countries expressed in GNP-values. It is natural that the reliability of these forecasts would determine in much our future estimations of water withdrawal and water availability estimates. To some extent, additional errors can arise especially for arid and semiarid regions because of no accounting of the expected anthropogenic  global climate change due to increasing atmospheric carbon dioxide and other “greenhouse” gases.
As known, forecasts of  future population, industry and thermal power growth are usually given for different social-economic development variants based on various factors and premises. This information is used to predict water withdrawal and water availability. Nevertheless, we give an average, most realistic, as we think, variant of the process development in the future. An approximate estimation was made for potential departures from the mean water availability indicators in 2025 for different variants of initial data.  The values obtained are + 10-15% for the regions with predominance of developed countries and + 20-25% for the regions with the predominance of developing countries.
More reliable and detailed water availability values for the future can be obtained.  They should consider river runoff variability during a long-term period and a year,  underground water data, the dynamics of  water resources quantity and quality with the stationary climate and anthropogenic global climate change.  All these are to be considered as the goals for subsequent studies on the problem of complex estimation of global water resources.
To solve this problem, it is necessary to provide a close co-operation of scientists from different countries, international organisations dealing with the problems of hydrology, climatology, complex use and protection of water resources.
 

State Hydrological Institute
Email: ishiklom@zb3627.spb.edu