Cover Image
close this bookFreshwater Resources in Arid Lands (UNU, 1997, 94 p.)
close this folder6: Global warming and groundwater resources in arid lands
View the document(introduction...)
View the documentQuaternary climate history
View the documentGlobal warming and desertification
View the documentCharacteristics of groundwater in arid lands
View the documentUtilization and conservation of groundwater
View the documentConclusions
View the documentReferences


Isamu Kayane

Quaternary climate history

Past stable-isotope records from the ice core and the deep ocean sediment core revealed that the earth has experienced a repeated climatic rhythm of glacials and interglacials with a cycle of about 100 ka (thousand years) after 800 ka BP (before the present). According to a climate history since 3.5 Ma (million years) BP reconstructed by a Chinese scientist based on analysis of loess deposits in northeastern China, the climate has experienced alternate periods of desertification and inter-desertification (Geng 1986). In his interpretation, the typical profile of loess deposits consists of a series of alternate depositions of the eolian yellow loess layer transported from the inland desert region during the dry period (desertification), and the brown-reddish palaeo-soil layer formed in situ during the wet period (inter-desertification). Although the glacial and the interglacial do not necessarily correspond to the inter-desertification and the desertification, respectively, it may be said that the inland climate in China was drier during the interglacial, or warm, period than during the glacial, or cold, period. The present climate since 2 ka BP in northern China corresponds to a desertification.

Closed lake level records in the northern middle-latitude zone also suggest a drier climate during the hypsithermal interval (HI) at around 6 ka BP, the warmest period during the present interglacial, and a wetter climate during the last glacial maximum (LGM) at around 18 ka BP (Street-Perrott and Harrison 1985).

The saltiest lake water in the world, in the Dead Sea, is a result of rapid desiccation of the lake after the LGM to the HI, during which the lake level dropped by more than 200 m (Issar et al. 1989). However, in the Indus River basin on the same latitude of 30°N, the climate was drier during the LGM and wetter during the HI compared with the present climate. The Indus civilization flourished during the wetter climate following after the HI, and finally disappeared at around 3.3 ka BP (Khanna 1992) when a dramatic decrease in precipitation occurred during the course of global cooling after the HI (Lamb 1982). Climatic changes near the 30°N latitude were very variable and differed from region to region, as shown in table 1, because a small latitudinal shift of the position of the polar front could result in dramatic increases or decreases in local precipitation. It may be said that the climate in arid lands in the subtropical or middle-latitude zone is highly variable both in time and space, as evidenced by many past climatic records.

Table 1 Climatic Changes in the Zone of Latitude 30°N



LGMa(18 ka BP; cold)

Hlb(8-6 ka BP; warmest)

Global cooling(ca. 4ka BP)

Ca. 650 AD

Global warming

Indus River

Dry R>

Wet R>




Wet R>

Dry R>

Wetting R>

Drying R>


a. Last glacial maximum.
b. Hypsithermal interval.

Global warming and desertification

The record from the polar ice core clearly shows that the increase in greenhouse gas concentrations in the atmosphere started with the Industrial Revolution in the eighteenth century, though the rate of increase accelerated after the middle of the twentieth century. It is the general consensus that global warming has occurred because of the increase in atmospheric concentrations of greenhouse gases due to human activities (Houghton et al. 1996). Human activities may influence local precipitation directly through changes in local vegetation cover, and indirectly through global warming.

The direct influence of deforestation of the Amazonian tropical rain forest on Amazonian climate has been discussed (e.g. Shukla et al. 1990). Almost all research results suggest a positive feedback effect of the decrease in local evapotranspiration by deforestation on local precipitation, i.e. the local precipitation will decrease following deforestation, although quantitative evaluation is still the subject of future research. If desertification caused by overgrazing or by other means were to occur in arid lands, the same kind of feedback effect as in the deforestation of the humid Amazonian forest might appear and the local precipitation would decrease. But it is also necessary to take into account the indirect influence of global warming on local precipitation, as is discussed below. The effect of this indirect influence is, in many regions, greater than that of the direct influence.

The global climate system redistributes energy from lower to higher latitudes. The energy surplus in the lower latitudes, bounded roughly by latitudes 35° North and South, results from the latitudinal gradient of the Earth's energy budget. The total energy distributed by ocean currents is greater than that distributed by the atmosphere.

The mean time during which a water molecule passes through a hydrological system (such as a lake, the Pacific Ocean, or the troposphere) is termed the mean hydrological residence time. The "memory" of a hydrological system increases with a longer residence time. The mean residence time is about 10 days for atmospheric vapour, but about 3,000 years for ocean. The atmosphere has a very short memory compared with the longer memory of the ocean. If we could stop the energy supply to the atmosphere, its motion would cease within a month, but the ocean would continue its circulation for a longer period following such an energy cut-off.

The total heat stored in hydrological systems should also be taken into account in assessing the future evolution of the global environment. A short memory is synonymous with a small heat capacity. Thus, the atmosphere contains insufficient heat to act as the source of future dynamic changes in the global climate. It can only respond to changes in forcing, such as changes in solar irradiation or heat supply from the ocean. Future atmospheric behaviour not only depends on increases in greenhouse gas concentrations and earth orbital changes but also depends heavily on changes in the sea surface temperature (SST), global ocean circulation, and the increased atmospheric turbidity (Kayane 1996).

Figure 1 shows the distribution of the SST trend calculated by the SST database for 1930-1989 released from the United Kingdom Meteorological Office for areas of 5 degrees latitude by 5 degrees longitude. The average rate of the SST increase for the whole ocean is about 0.9°C/100 years, which is greater than the rate of the global air temperature increase. The increase in ground surface temperature in the eastern part of North America since the middle of the nineteenth century also exceeds the increase in the surface air temperature (Deming 1995). It is worth noting that the SST increase is predominant in the tropical ocean, but the SST in the Atlantic Ocean north of 30°N was markedly decreased, probably owing to an increased cloud amount caused by water vapour transported from the low-latitude zone. The effect of global warming on the SST and the air temperature does not appear to follow the same trend: the temperature decreases in some regions and increases in others owing to global warming. This is also true for precipitation changes.

Figure 2 shows possible causal relations between global warming and hydrological processes, although the anthropogenic contribution to the SST remains the subject of ongoing research. Increases in ocean evaporation, ocean precipitation, and global precipitation, corresponding to the intensified global energy and water cycle, are processes directly deducible from the SST increase.

Figure 1 Global Trends of the Sea Surface Temperature (SST) for Each 5 by 5° Change in Latitude and Longitude Calculated Using the United Kingdom Meteorological Office (UKMO)/SST Database (Source: Kayane et al. 1995)

Figure 2 Causal Relationships between Global Warming and Regional Precipitation Changes (GHGs, Greenhouse Gases)

Figure 3 Long-Term Time Series (Circles) and Their Trends (Line) in the Boreal Summer Monsoon Rainfall (June-September) at Colombo (Open Circles and Dashed Line) and Nuwara Eliya (Solid Circles and Solid Line) for 1986 1992 (Source: Kayane et al. 1995)

The SST in the Indian Ocean increased by 0.5-1.0°C during 1930-1989. As a result of increased ocean evaporation due to the SST increase, the rainfall during the SW monsoon season from June to September at Colombo increased by about 30 per cent during 1869-1993. However, the rainfall at Nuwara Eliya, a station at an elevation of 1,895 m in the central high mountains in Sri Lanka, has decreased by about 40 per cent during the same period (fig. 3). An almost linear increase in rainfall during 1870-1970 was also observed at Calicut (Lengerke 1976), a coastal station in south-west India located to the west of the Western Ghats Mountains, where the same orographic effect on rainfall pattern is expected as on the Sri Lankan south-west coast. Such long-term changes in rainfall in Sri Lanka and southwest India can be interpreted as the result of intensified Indian monsoon circulation caused by global warming (Kayane et al. 1995). Changes in local rainfall may take opposite trends within a relatively small island like Sri Lanka, owing to global warming.

If the Hadley (north-south) circulation were intensified by global warming, the subtropical high would also be intensified, resulting in a decrease in precipitation in arid lands in the middle latitudes. One such example has recently been reported by Liu and Zhao (1996) for the Tibetan Plateau. During the last 40 years, the annual precipitation had an increasing trend in south-east Tibet and had no obvious change in north-east Tibet, while it had a decreasing trend by 5-10 per cent in the north-west and central Tibetan Plateau. Because of the combined effects of the increase in evaporation due to temperature increase and the decrease in precipitation, the river discharge in the middle reaches of the Yarlung Zangbo River decreased by 10 per cent, at the Lancang River by 5 per cent, and at the Lhasa River and the Nyang Chu River also by about 5 per cent.

Brenes Vargas and Saborio Trejos (1994) reported that, generally speaking, the increase in rainfall in the windward side and the decrease in the leeward side of the central mountain range in Costa Rica might be a result of the intensified North Atlantic high, which would strengthen the north-east trade winds to Costa Rica. However, a part of the coastal area in the leeward side shows an increasing trend, presumably caused by the intensified local atmospheric circulation from ocean to coast induced by the strengthened trade winds.

In Patagonia, located in the westerly zone of the southern hemisphere, definite increasing trends in annual precipitation during the past 100 years are observed at Rio Colorado, Neuquen, and Paso de los Indios, all located in northern Patagonia, but stations in southern Patagonia show no obvious change and certainly no decreasing trend (Quintela et al. 1995).

There have also been substantial annual precipitation changes in certain latitudes and regions, most notably a decrease in the African Sahel in the middle latitudes after the 1960s, and a fairly steady increase in the former USSR in the polar frontal zone during the past 100 years (Folland et al. 1990).

The observed precipitation trends described above can be interpreted as the result of an intensified global energy and water cycle, i.e. as indirect effects of human activities on local precipitation through global warming. It may be concluded that the precipitation variability has increased globally both in time and space. This may raise serious problems with respect to water resources and food supply in the future, especially in arid lands.

Characteristics of groundwater in arid lands

Groundwater in arid lands is considered to be a stable water resource not influenced directly by year-to-year climatic variation. This may be true if it is used to a renewable degree, but may not be true if the abstraction rate exceeds the natural recharge rate. Deep groundwater being used in most arid lands was recharged under climatic and hydrological conditions in the past that were much wetter and very different from those of the present.

The mean residence time of groundwater in the world is about 1,000 years, which is far longer than the residence time of about 10 days for river water or the water vapour in the atmosphere. The mean residence time may also be defined as the ratio of the total water storage to the annual recharge rate for the hydrological system concerned. Longer residence time of groundwater in a groundwater basin implies a larger amount of groundwater storage in the groundwater basin, and a smaller rate of annual recharge to it. If the abstraction rate from a groundwater basin exceeds the annual recharge rate to it, the water-table will decline, indicating a decrease in total storage as experienced in many groundwater basins around the world.

Figure 4 Changes in the Electric Conductivity (E.C.) of Groundwater in Sri Lanka from Puttalam in the Dry Zone to Kandy in the Wet Zone (Source: Song and Kayane 1996)

The groundwater in the Great Artesian Basin in arid central Australia may be the oldest groundwater currently being used in the world. Its age, dated by the radioactive isotope of chlorine-36, is estimated to be older than 1 Ma (Bentley et al. 1986; Torgersen et al. 1991). The main groundwater resource in the Nubian sandstone in the eastern Sahara dated by carbon-14 was about 25 ka, although the groundwater in some oases was found to be of recent age recharged from the Nile Valley (Munnich and Vogel 1962). The above two cases are typical examples of groundwater resources in arid lands. The deep groundwater in arid lands is "fossil" and being "mined" in the same sense as petroleum.

The water quality of old groundwater is another issue to take into account. Generally speaking, the longer the residence time, the higher the concentration of dissolved ions in groundwater. Groundwater tends to evolve chemically toward the composition of sea water during the course of flow. It was found by Chebotarev in the Great Artesian Basin that this evolution is normally accompanied by the following regional changes in dominant anion species (Freeze and Cherry 1979):

The sequence above is termed the Chebotarev sequence. Therefore, the deeper groundwater is older and more saline than shallow groundwater. However, very saline groundwater is sometimes found in shallow aquifers in arid lands. Groundwater we investigated in the dry zone of Sri Lanka with marked dry and wet seasons is one such example. Figure 4 shows changes in electric conductivity (E.C.), which increases with increasing amounts of dissolved ions, in a groundwater profile from Puttalam in the dry zone to Kandy in the wet zone. The shallow groundwater has higher E.C. than the deep groundwater. Predominant ion species in the shallow groundwater in the dry zone of Sri Lanka are Cl- in anions and Na+ in cations (Song and Kayane 1996).

The dry zone in Sri Lanka is not arid land in the strict climatological sense. Natural vegetation in the dry zone is tropical jungle. The annual rainfall is 9001,500 mm, but the annual evapotranspiration is as high as 1,300 mm, so that the annual run-off is small compared with the annual rainfall. Once the natural forest which had shaded the soil surface had been cleared by humans, the rate of soil evaporation during the dry season became very high. Then the soil water infiltrated during the preceding rainy season is easily evaporated, leaving dissolved salts near the soil surface. When the next rain comes, the infiltrated water dissolves the accumulated salts during the process of percolation to the water-table. These processes have been continuing for a long time, ever since the land was first deforested for agricultural use.

Salinity of shallow groundwater is very high in some localities of arid lands, although its age is young, as in the dry zone of Sri Lanka. The evolution of groundwater there does not start from the initial stage of the Chebotarev sequence, but starts from its last stage: i.e. the groundwater is highly saline there from the beginning. This is an example of groundwater-quality deterioration caused by humans in arid lands. Integrated management of soil and water is essential in arid lands.

Utilization and conservation of groundwater

In the middle-latitude zone, the natural vegetation changes from forests to grasslands near an isohyet of 500 mm/year, and from grasslands to deserts near that of 250 mm/year. The amount of recharge to the groundwater is the difference between the annual precipitation and the annual evaporation, provided that no surface run-off occurs. In the case of a dry climate with precipitation less than 500 mm/year and where the potential evaporation far exceeds the annual precipitation, as in drylands, the amount of groundwater recharge is heavily dependent on the actual evaporation lost from the land surface. Deforestation and desertification result in a decrease of local precipitation due to positive feedback effects, as discussed above. On the other hand, it is quite difficult, if not impossible, to increase the amount of local precipitation anthropogenically.

For the sustainable use of drylands, it is desirable to develop such methods that induce positive feedback effects to increase the amount of available water. The amount of actual evaporation lost from the soil surface is an issue of human intervention in drylands. Dry farming is a technique to maximize the input from meteoric water into the soil, and at the same time to minimize the evaporation loss from the soil surface. Deep cultivation of soil before the rainy or snowy season is a technique to introduce rain or snow-melt water into deep soil layers. The evaporation loss of soil moisture could be decreased by forming a loose surface layer in which capillary continuity of soil pores is discontinued by disturbing or crushing the surface soil.

Yamanaka et al. (1994) conducted an experiment that contributed to the increased understanding of the dynamic behaviour of water vapour in the soil layer and the role of a surface dry layer (SDL) on soil evaporation. When the SDL is formed during the evaporation process at the bare soil surface, the water vapour in the SDL plays an important role by connecting the liquid water in the soil and the water vapour in the atmosphere. The thickness of the SDL is about 4-5 cm, and the lower boundary of the SDL coincides with a surface where soil water vaporizes. The SDL acts as a strong barrier against the transport of the water vapour in the soil evaporated from the evaporation surface below the SDL.

Since the experimental result mentioned above is obtained for the diurnal change in water vapour concentration within a standard sand layer only, its applicability is of limited nature. Future progress in research on the dynamic behaviour of transport processes of heat and water near the soil surface, including the dynamic behaviour on a much longer time-scale of seasons and years, and for different soil types, may contribute to the proper management of soil and groundwater in arid lands.

Groundwater, soil water, river water, lake water, and mountain glaciers are linked through the regional hydrological cycle. However, in such hydrological characteristics as the residence time, water storage, water quality, and recharge and discharge processes, they are quite different from each other. For the proper use of groundwater in arid lands, the conjunctive use of waters with different hydrological characteristics is necessary. The science of hydrology may contribute to an increased understanding of hydrological processes and hydrological characteristics of natural waters, especially for groundwater in arid lands.


1. A history of the climate indicates that the present interglacial period corresponds to a desertification period in arid lands. There is much evidence that the climate of arid lands has been very variable since the last glacial maximum. Climatic changes are different from region to region, as indicated by changes in precipitation in the Dead Sea and the Indus River basin.

2. The SST has increased markedly in the low latitude ocean during the past 60 years. The global energy and water cycle has intensified in response to the increase in SST, resulting in an increase in precipitation in some regions and a decrease in other regions. Generally speaking, the precipitation in arid lands seems to be on a decreasing trend.

3. The residence time of groundwater in arid lands is very long, which suggests a relatively small amount of natural recharge compared with the huge amount of groundwater reserve. High salinity in groundwater is a result of long residence time in the aquifer but, in arid lands, shallow groundwater with shorter residence time is also highly saline in some cases.

4. Basic scientific research is needed for the proper management of soil and water in arid lands. Recently, interesting results about the role of surface dry layer that reduces evaporation loss from the soil surface have been obtained.


Bentley, H.W. , F.M . Phillips, A.N. Davies, M.A. Harbermehl , P.L. Airey, G. E. Calf, D. Elmore, H.E. Gove, and T. Torgersen. 1986. Chlorine 36 dating of very old groundwater. 1. The Great Artesian Basin, Australia. Water Resources Research 22: 1991-2001.

Brenes Vargas, A., and V.F. Saborio Trejos. 1994. Changes in general circulations and its influence on precipitation trends in Central America: Costa Rica. Ambio 23:87-90.

Deming, D. 1995. Climatic warming in North America: Analysis of borehole temperature. Science 268: 1576-1577.

Folland, C.K., T. Karl, and K.Y. Vinnikov. 1990. Observed climate variations and change. In: J.T. Houghton, G.J. Jenkins and J.J. Ephraums (eds.), Climate Change: The IPCC Scientific Assessment, 195 238. Cambridge University Press, Cambridge.

Freeze, A.C., and J.A. Cherry. 1979. Groundwater. Prentice-Hall.

Geng, K.H. (ed.). 1986. Desert Climate in China. Scientific Publishing, Beijing.

Houghton, J.T., L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg, and K. Maskell (eds.). 1996. Climate Change 1995: Contribution of WGI to the Second Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge.

Issar, A.S., H. Tsoar, and D. Levin. 1989. Climatic change in Israel during historical times and their impact on hydrological, pedological and socio-economic systems. In: M. Leinen and M. Sarnthein (eds.), Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmospheric Transport, 525-541.

Khanna, A.N. 1992. Archaeology of India. Clarion Books, New Delhi.

Kayane, I. 1996. Interactions between the biospheric aspects of the hydrological cycle and land use/cover changes. Global Change Newsletter 25: 8-9.

Kayane, I., K. Nakagawa, H. Edagawa, and C.M. Madduma Bandara. 1995. Hydrological consequences of global warming revealed by the century-long climatological data in Sri Lanka and southwest India. An Interim Report of I GBP Activities in Japan 1990-1994, 65-81. Science Council of Japan, Tokyo.

Lamb, H.H. 1982. Climate History and Modern World. Methuen, London.

Lengerke, H.J. von. 1976. Climatological Survey of the Nilgiris Area (South India). Unpublished dissertation, University of Heidelberg.

Liu, T., and C. Zhao. 1996. Preliminary Research on the Response of Regimen to Climate Change in Tibet Plateau. Presented at the 28th International Geographical Congress, The Hague, 5-9 August 1996.

M, K.O., and J.C. Vogel. 1962. Untersuchungen an Pluvialen Wassern der Ost Sahara. Geologische Rundschau 52: 611-624.

Quintela, R.M., O.K. Scarpati, L.B. Spescha, and A.D. Capriolo. 1995. The Probable Impact of Global Change on the Water Resources of Patagonia, Argentina. Presented at the lGU Conference on "Global Change and Geography," Moscow, 14-18 August 1995.

Shukla, J., C. Nobre, and P.J. Sellers. 1990. Amazon deforestation and climate change. Science 247: 1322-1325.

Song, X., and I. Kayane. 1996. Zonal characteristics of groundwater quality controlled by climate in Sri Lanka. Journal of the Japanese Association of Hydrological Sciences 26: 151-165.

Street-Perrott, F.A., and S.P. Harrison. 1985. Lake levels and climate reconstruction. In: A.D. Hecht (ed.), Paleoclimate Analysis and Modeling, 291-340. John Wiley and Sons, New York.

Torgersen, T., M.A. Habermehl, F.M. Phillips, D. Elmore, P. Kubik, B.G. Jones, T. Hemmick, and H.E. Gove. 1991. Chlorine 36 dating of very old groundwater. 3. Further studies in the Great Artesian Basin, Australia. Water Resources Research 27: 3201-3213.

Yamanaka, T., T. Hiyama, and J. Shimada. 1994. Behavior of the water vapor in the sandy soil during the bare soil evaporation. Journal of the Japanese Association of Hydrological Sciences 24: 31-46.