EFFECTS OF CLIMATE CHANGE ON COASTAL FRESH GROUNDWATER RESOURCES

Transcription

1 EFFECTS OF CLIMATE CHANGE ON COASTAL FRESH GROUNDWATER RESOURCES PRIYANTHA RANJAN 1*, SO KAZAMA 1, MASAKI SAWAMOTO 1 Graduate Scool of Environmental Studies, Tooku University Aoba yama, Sendai, 9-579, Japan. Department of Civil Engineering, Graduate Scool of Engineering, Tooku University Aoba yama, Sendai, 9-579, Japan. * Corresponding Autor Dr. PRIYANTHA RANJAN Graduate Scool of Environmental Studies, Tooku University Aoba yama, Sendai, 9-579, Japan. Tel: (office) (residence) Fax: ranjan@kaigan.civil.tooku.ac.jp 1

2 EFFECTS OF CLIMATE CHANGE ON COASTAL FRESH GROUNDWATER RESOURCES Abstract Tis study evaluates te impacts of climate cange on fres groundwater resources specifically salinity intrusion in water resources stressed coastal aquifers. Our assessment used te Hadley Centre climate model, HadCM3 wit ig and low emission scenarios (SRES A and B) for years -99. In bot scenarios, te annual fres groundwater resources losses indicate an increasing long-term trend in all stressed areas, except in te nortern Africa / Saara region. We also found tat precipitation and temperature individually did not sow good correlations wit fres groundwater loss. However, te relationsip between te aridity index and fres groundwater loss exibited a strong negative correlation. We also discuss te impacts of loss of fres groundwater resources on socio-economic activities, mainly population growt and per capita fres groundwater resources. Abbreviated Title Climate cange and coastal fres groundwater resources

3 1. Introduction Atmosperic carbon dioxide levels ave continually increased since te 195s. Te continuation of tis penomenon may significantly alter global and local climate caracteristics, including temperature and precipitation. Climate cange can ave profound effects on te ydrologic cycle troug precipitation, evapotranspiration, and soil moisture wit increasing temperatures. Te ydrologic cycle will be intensified, wit more evaporation and more precipitation. However, te extra precipitation will be unequally distributed around te globe. Some parts of te world may see significant reductions in precipitation, or major alterations in te timing of wet and dry seasons (IPCC, 1). Information on te local or regional impacts of climate cange on ydrological processes and water resources is becoming more important. Te effects of global warming and climatic cange require multidisciplinary researc, especially wen considering ydrology and global water resources (Hugo et al, 199; Arnell, 1999, ; Hulme, 1999; Eckardta and Ulbricb, 3; Gertena et al, ; Hitz and Smit, ; Labat et al, ). Previous studies ave typically coupled climate cange scenarios wit ydrological models, and ave generally investigated te impact of climate cange on water resources in different areas (e.g.gleick, 197; Loaiciga et al, 199, Bobba et al, 1997; DETR, 1997; Arnell,1999; Najjar,1999 and Mimikou et al,). DETR (1997) and Arnel (1999) indicated areas experiencing water resources stress due to future climate cange. Tey suggested tat stressed countries would be concentrated in soutern and nortern Africa, around te and in te Middle East, soutern Asia and te Indian subcontinent, Central America, and large parts of Europe. Tey predicted tat as of 5, tese countries will be adversely affected by 3

4 climate cange and sow an increase in water resources stress. By te 5s and s, additional countries in soutern Africa will move into te stressed class and pressures will increase most rapidly in Africa and parts of soutern Asia and Eastern Europe (Arnell, 1999). Wen considering water resources in coastal zones, coastal aquifers are important sources of freswater. However, salinity intrusion can be a major problem in tese zones. Salinity intrusion refers to replacement of freswater in coastal aquifers by saltwater. It leads to a reduction of available fres groundwater resources. Canges in climatic variables can significantly alter groundwater recarge rates for major aquifer systems and tus affect te availability of fres groundwater. Salinisation of coastal aquifers is a function of te reduction of groundwater recarge and results in a reduction of fres groundwater resources. Estimates of global warming are generally based on te application of general circulation models (GCMs), wic attempt to predict te impact of increased atmosperic CO concentrations on weater variables. Te GCM scenarios indicate tat global warming as been increasing during recent decades and tat te trend may become stronger in te future. Because of te complex mecanics and te model structure, different GCMs produce different predictions (Semmler and Jacob, ). However, despite differing predictions, trends in climate variables were consistent (IPCC, ). In tis study we focus on results from te Hadley center GCM (HadCM3) and include two SRES scenarios; a ig emission scenario (SRES- A) and a low emission scenario (SRES-B). Te selected emission scenarios express different results of climate cange and population growt. Te A scenario

5 can be generally described as a pessimistic future. It gives ig CO emissions and related extreme climate impacts. A scenario also describes a world wit ig population growt, relatively slow economic growt and tecnological cange. Te B scenario may be described as a more optimistic future. Population growt is intermediate and CO emissions are steadily reduced, resulting in lower levels of climate forcing (IPCC, 1). To assess te impacts of potential climate cange on fres groundwater resources, we focus on canges in groundwater recarge and sea level rise on te loss of fres groundwater resources in water resources stressed coastal aquifers.. Metodology and data sources.1 Numerical modeling of salinity intrusion Many models ave been developed to study saltwater intrusion. Tey range from relatively simple analytical solutions to complex numerical models. Studies involving te movement of freswater and saltwater in coastal aquifer systems are classically studied using two different approaces (Reilly and Godman, 195). In te first approac, freswater and saltwater are assumed completely immiscible and a sarp interface exists between two zones. In te second approac, freswater and saltwater are assumed to be in a dynamic equilibrium resulting from advection and dispersion witin te aquifer. Since te main aim of tis study is to evaluate longterm beavior of coastal groundwater systems, te sarp interface model is more appropriate for tis study. 5

6 Sarp interface models couple freswater and saltwater flow based on te continuity of flux and pressure. In tis approac, togeter wit Dupuit approximation for eac flow domain, te equation of continuity may be integrated vertically to develop te following system of differential equations (Bear, 1999). ( ) ( ) ( ) t t t t S q y K y x K x f f s f f f f i f fy f i f fx + + = + + αθ δ δ θ 1 (1) ( ) ( ) ( ) + + = + + t t t S q y z K y x z K x f s s s s s b i sy s b i sx δ δ θ 1 () Te location of te interface elevation ( i ) is given by f f s f s f s s i ρ ρ ρ ρ ρ ρ = (3) were K f and, K s represent te ydraulic conductivity in fres and salt water regions. f and, s are te piezometric eads of freswater and saltwater regions and q f, q s are flow rate in fres and salt water respectively. Storage coefficients in fres and salt water regions are given by S f and S s. θ is te porosity of te aquifer media. f ρ and s ρ are specific weigt in fres and salt water. α is a parameter wic can take te value of eiter 1 or according to te type of te aquifer; α = 1 for unconfined aquifers and α = for confined aquifers. From equations (1) and (), it is possible to derive a numerical model using implicit finite difference tecniques. To solve te two simultaneous linear algebraic

7 difference equations, te Strongly Implicit Procedure -SIP (Remson et al, 1971) was used as a suitable numerical tecnique. Empirical evidence suggests tat for cases of flow in eterogeneous or anisotropic media, te strongly implicit procedure is muc more efficient tan oter metods and it does not depend upon te complexity of te problem (Essaid 19) Fres groundwater loss due to salinisation Te concept of freswater - saltwater interface can be used to estimate fres groundwater resources in coastal aquifers. Increases in groundwater recarge sifts te salinity interface seaward and decreases in recarge sift it landward (interface ). Tis interface movement canges te available fres groundwater resources in te aquifer. As illustrated in Fig. 1, wen te aquifer is totally filled wit freswater (interface 1), te freswater loss is zero. Te landward movement of te salinity interface leads to reduce freswater amount in te aquifer. Wen te salinity interface coincides wit piezometric ead (interface 3), te wole aquifer fills wit saltwater and indicates tat freswater loss is 1%. Approximate location of Fig. 1.3 regions wit water resources stress Based on te water resources stressed regions described by DETR (1997) and Arnel, (1999), five major regions experiencing water stress ave been considered. Te countries in tese regions likely would experience te greatest adversewater resources impact of climate cange. Coastal zones in Central America (CAM) Soutern Africa (SAF) and Nortern Africa / Saara (SAH), around te 7

8 (MED) and in te Soutern Asia (SAS) ave been selected to assess te cange in fres groundwater resources due to salinity intrusion (Fig a). Tese regions can be categorized into different climate groups. Based on te Koppen s climate classification (FAO) te five regions locate in monsoon/subtropical, mediterranean, tropical, and desert climates. Te ydro-geological properties of te regions were obtained from te Global Groundwater Information System, developed by IGRAC (). To simulate te effect of sea level rise, data was obtained from te PSMSL data base (Permanent Service for Mean Sea Level) based on istorical trends and future predictions. Table 1 summarizes te climatic, sea level rise, and ydro-geological caracteristics of te five selected regions. Approximate location of Fig. Approximate location of Table 1.. Estimation of loss of fres groundwater resources to due climate canges Tis study uses climate cange scenarios developed from Hadley Centre climate simulations (HadCM3). Te climate data is available at a spatial resolution of.5 degree of latitude by 3.75 degree of longitude, wic is equivalent to a surface resolution of approximately 17 km 7 km at te equator. Te study areas are represented by only a few grid points at tis coarse resolution ( ). Te grid cells along te seaward boarder were selected to represent coastal areas (Fig b). We calculate te water balance on montly time-step for eac cell, treating eac cell as a separate catcment. Evaporation was calculated using te Penman Monteit formula using CROPWAT computer model. Assuming tat te groundwater is replenised wen precipitation exceeds actual evapotranspiration, groundwater

9 recarge was estimated as te difference between precipitation and evapotranspiration. Te estimated montly recarge was used to model te saltwater-freswater interface and calculate te average annual fluctuation of te interface for five selected coastal aquifers. Te model was run to simulate canges in interface on a montly time step, altoug annually averaged values were saved as output. Tis average annual cange in te freswater-saltwater interface was used to estimate te annual cange in fres groundwater resources due to salinity intrusion over a 1 year period. 3. Results and Discussions 3.1 Loss of fres groundwater resources over te next century In selected five regions; Central America (CAM), Soutern Africa (SAF), Nortern Africa / Saara (SAH), Soutern Asia (SAS) and te (MED). Te average annual cange in fres groundwater resources igligts te complexity, consequences, and feedbacks of potential climate cange. Tese results demonstrate tat future fluctuations in fres groundwater resources clearly indicate a long-term trend of increasing loss, except in nortern Africa / Saara region for bot ig and low emission scenarios. Approximate location of Fig. 3 and Fig. Over te next century, te linear regressions between loss of fres groundwater resources (percentage loss) versus time sow different relationsips for five selected regions. As sown in Fig 3a and Fig a, te annual fluctuation of loss of fres groundwater resources display sort term trends in and soutern Asian regions, wile sowing long term trends in oter tree regions for bot 9

10 scenarios. Te region expresses an overall long term increments of. percent annual fres groundwater loss in SRES A and.5 percent annual fres groundwater loss in SRES B scenario. Te soutern Asian region also sowed sort term variations and long term trends wit.75 and.7 percent fres groundwater loss per year for A and B scenarios respectively. One outlier is te sligt negative gradient in te nortern Africa / Saara region. It sows a. percent decrease in fres groundwater loss per year in ig emission scenario and.1 percent decrease in fres groundwater loss per year in low emission scenario. Tis suggests an increase in availability of fres groundwater resources in coastal zones of te nortern Africa and te Saara region in te future. Table summarizes te long term canges in loss of fres groundwater resources for ig and low emission scenarios. Approximate location of Table 3. Comparison of sort term variation of loss of fres groundwater resources and precipitation cange Considering te selected five water resources stressed regions and te selected climate cange scenarios, tis study provides new insigt into temporal canges in coastal fres groundwater resources in response to climatic factors, especially precipitation. Precipitation, te primary source of groundwater recarge, is te largest term in te water balance, and is variable in bot time and space. Terefore, future precipitation canges control te fres groundwater resources variability. Annual fres groundwater losses are well matced canges in precipitation for bot ig and low emission scenarios (Fig 3b, b). Te increases in precipitation are 1

11 followed by increases in fres groundwater resources (reduction in loss of fres groundwater resources). Decrease in precipitation decreases in fres groundwater resources (increase in loss of fres groundwater). Tese canges were clearly identified in te soutern Asian region and te region, wic bot sowed large fluctuations in annual precipitation in bot scenarios. Tere were sligt increases and decreases in precipitation projected in te Central America and soutern African regions, but te annual precipitation is less and does not lead to sort term cange in fres groundwater resources in te A scenario. However, te Sout African region sows sort term variations for te B scenario. Te nortern Africa / Saara region sows tat te precipitation will be increased in te future. Te long term trends in fres groundwater resources also indicate an increase in fres groundwater resources in tis region for bot scenarios. Projected temperature increases are likely to lead increases in open water and soil/plant evaporation. Increased evapotranspiration affects to groundwater recarge and terefore, te rate of fres groundwater loss is lower relative to precipitation in te Nort Africa region. Te projected temperature canges sow trends of increasing annual mean temperature in all five areas (Fig 3c and c). However, we did not find a clear relationsip between increased temperature and fres groundwater resources. 3.3 Correlations between climate and loss of fres groundwater resources To evaluate te correlation between climate factors and fres groundwater resources in eac region, te predicted cange in precipitation and temperature under two SRES scenarios were considered. Te correlation coefficient was calculated as follows; 11

12 Correlation Coefficient ( r ) = 1 1 i ( x i ( x x) x)( y 1 i= 1 i= 1 i y) ( y y) () were x is te climate factor and y is te loss of fres groundwater resources. i represents te time in steps in years. Te variation of climate variables and fres groundwater resources are sown in Fig 5 and Fig. Fig 5a sows te correlation between te cange in precipitation and canges in loss of fres groundwater resources, wile Fig. 5b sows te correlation between te cange in temperature and canges in loss of fres groundwater resources for te A scenario. Also Fig a sows te correlation between te cange in precipitation and canges in loss of fres groundwater resources, wile Fig. b sows te correlation between te cange in temperature and canges in loss of fres groundwater resources for te B scenario. Approximate location of Fig. 5 and Fig. Tey indicate tat only te and te soutern Asian regions sow ig correlation between te precipitation and loss of fres groundwater resources, wit a correlation coefficient of.9 and.7 for ig emission scenario and.5 and. for low emission scenario respectively. However, te canges in precipitation and related canges in fres groundwater loss were small in te Nort African region. Te and soutern Asian regions sow wide ranges of precipitation (over ± mm/year) would lead to widely varying fres groundwater loss. Te region precipitation is primarily widely varying mid-latitude cyclones 1

13 during te winter season. Te soutern Asian region wic as a tropical and monsoon climate, te main source of precipitation is widely varying monsoon rains. In te soutern Asian area, te Himalayas play a critical role in te continental monsoon and te corresponding seasonal variability in precipitation. Jons et al (3) sowed tat te spatial patterns of cange in bot temperature and precipitation are very similar in similar latitudes. He also found tat annual precipitation increased in ig latitudes wile precipitation increased in winter across most mid-latitude regions and temperature increases were greater at ig latitudes. Te correlation between temperature and cange in fres groundwater resources was very poor in all regions for bot scenarios. In general, te correlation coefficients empasize tat te precipitation and temperature individually do not sow a good correlation to canges in fres groundwater resources. Terefore te combined effects of precipitation and temperature were considered for tis analysis. 3. Variation of loss of fres groundwater resources wit respect to an aridity index Aridity indexes are quantitative indicators of te degree of water deficiency present at a given location and typically express te combined effects of water and energy in te region. Our aridity index was te ratio between mean annual precipitation and mean annual temperature (Lang's index) and a widely used modified version developed by E. de Martonne in 195. Tis index also called as Martonne index (Oliver and Fairbridge, 197, Paari and Murai, 1999). 13

14 Martonne s aridity index is defined by; P AI = (5) T +1 Were T is mean annual temperature ( C) and P is mean annual precipitation (mm) Te aridity index versus loss of fres groundwater relationsip exibits a global negative correlation for bot SRES scenarios (Fig 5c and Fig c). Te correlation coefficients imply tat te canges in climate and loss of fres groundwater resources are correlated in te Central American,, soutern African and soutern Asian regions, wit correlation coefficients iger tan.5 for bot scenarios. In te nortern Africa / Saara region, we found poor correlations between climate and fres groundwater resources. However te variations in climate and impacts on fres groundwater resources were small and it migt lead to inconsistent estimates of regional precipitation and temperature in te nortern Africa / Saara region. 3.5 Regional distribution of te aridity index and fres groundwater resources Te regional variation in te absolute value of te aridity index and loss of fres groundwater resources under bot ig and low emission scenarios are sown in Fig. 7. It sow tat te scatter plots of aridity index versus loss of fres groundwater resources results in clusters according to climatic zone. In bot cases, te low aridity index, wic corresponds to iger loss of fres groundwater resource, was evident in te Central America and nortern Africa / Saara regions. For bot scenarios, tese regions sow te aridity index less tan 15 and are associated wit te igest loss of fres groundwater resources. Te Central American region is located in 1

15 tropical climate zone and te nortern Africa / Saara region as desert areas. Te precipitation is low in bot regions, aving average annual precipitation less tan mm/year. In te nort African / Saara region, te rainfall migt become more intense, but tere likely will be more extreme events wit average montly temperatures above 3 C during te warmest monts and extremes above 5 C (IPCC,199). Desert air is very dry wit intense incoming solar and outgoing terrestrial radiation resulting in large daily temperature fluctuations and ig potential evaporation. Te increase in evapotranspiration controls te groundwater recarge and leads to loss of fres groundwater resources. Approximate location of Fig. 7 Te aridity index varies troug wide range in te, Sout Asia, and Sout Africa regions for bot scenarios. Sout Africa sows comparatively ig fres groundwater loss tan oter two regions. As compared to te sout Asia, loss of fres groundwater resources is also less in te region. Te canges in precipitation across different regions of te can be classified according to te current and projected distributions of rainfall. Tere could be an increase in summer drougt consistent wit observed decreases in summer precipitation an increases in winter and spring (IPCC,199). Suc canges are likely to affect mainly areas tat already are sensitive to drougt. However in te region, te mean precipitation is iger and te mean temperature is lower compared to sout Asian region. Terefore te mean loss of fres groundwater resources in te region is also lower tan in sout Asian region. Table 15

16 3 presents te statistical properties of te fres groundwater losses in eac region, wic confirms te average impact on fres groundwater resources in eac area. Approximate location of Table 3 3. Impact of loss of fres groundwater resources on socioeconomic activities Te reduction in fres groundwater resources could be detrimental to various waterdependent socioeconomic activities. Te assessment of impacts of loss of fres groundwater resources on society requires estimation of future water witdrawals. Tese witdrawals will depend not only on future population but also water demand and water use efficiency. Te loss of fres groundwater resources affects local people in water-stressed areas via increased demand resulting from population growt, water use patterns linked wit life style, and degradation of waterseds caused by land-use canges. Projections of future water use vary widely, reflecting different potential population growt and water use efficiencies (Seckler et al, 199). SRES emission scenarios provide multiple predictions of future global and regional population and oter important global-scale quantitative and qualitative socioeconomic indicators. Scenario A, in contrast, predicts a world of lower economic development and weak globalization. Population growt in A is te igest (15 billion by 1), because of te reduced financial resources available to address uman welfare, cild and reproductive ealt and education. B scenario assume lower population growt rate and it leads to a more optimistic world (IPCC, ). 1

17 Te regional SRES population projections were downscaled to te nation scale by CIESIN (). We estimate future population density based on te Gridded Population of te World (GPW) regional projections by CIESIN for te two selected SRES scenarios (Gaffin et al, 3). Table sows te comparison of population, fres groundwater availability, and comparison of per capita fres groundwater resources availability per unit aquifer tickness (1m) in year 1 and 1, for two selected SRES scenarios. It sows tat scenario A predicts te igest population and te resultant igest fres water demand. Table furter sows tat SRES A represents te lowest per capita fres groundwater resources in all regions. Te interesting finding is tat, even toug te availability of fres groundwater resources in te and sout Asia is less under te B scenario, te per capita fres groundwater resources is less for te A scenario. It empasizes critical socio-economic problems in ig populated water resources stressed areas. Approximate location of Table Regionally, Sout Asian countries ave te igest overall population density in te world and increasing demands in te domestic, agricultural, and industrial sectors will place additional stresses on water resources. Even toug te availability of fres groundwater is relatively iger in te sout Asian region, te per capita fres water resources availability in one square kilometer of coastal area is less ( m 3 per person in 1 under A scenario) and will likely decrease in future (115 m 3 per person in 1). Present and future per capita fres water resources sow relatively iger values in African continent, wic ave low population density. Te available fres groundwater resources are less in te African region, but te population is also 17

18 less and per capita resources are iger. In Africa, poverty is linked to te environment in complex ways, particularly in economies tat are based on exploitation of natural resources suc as groundwater resources. Hence, even toug it seems tat te per capita fres groundwater resources are iger on te African continent, te poor communities do not ave capacity and tecnology to access te deep groundwater and extract tem for use. Degradation of tese resources reduces te productivity of poor persons wo most rely on natural resources, and makes poor communities even more susceptible to extreme events. Nicolls et al () sows tat te global coastal population is already large and expanding rapidly (coastal population growt is twice te national population growt). Hence, te socioeconomic problems can be more critical in coastal regions. 5. Conclusions Te investigation of te relationsip between climate cange and loss of fres groundwater resources is important for understanding te caracteristics of te different regions. Among te five selected water resources stressed areas, bot ig and low emission scenarios provide more likely impacts on fres groundwater resources. Te relationsips between precipitation, temperature, and loss of fres groundwater resources igligt te complexity of ydrological consequences, but still indicate an increase in loss of fres groundwater resources in all studied regions except te nortern Africa/Saara region for bot ig and low emission scenarios. Over te 1 years, te loss of fres groundwater resources sows a positive 1

19 correlation for bot scenarios in Central America, te, Sout Asia, and Sout Africa. Te correlation between te cange in climatic parameters and canges in loss of fres groundwater resources were evaluated for bot ig and low emission scenarios. Tese correlations migt be positive or negative at te continental scale, wic indicates complexity in te feedback between precipitation and temperature canges in te ydrological cycle at te regional scale. In general, te correlation coefficients empasize tat te precipitation and temperature individually do not sow a good correlation to canges in fres groundwater resources. Te combined effects of precipitation and temperature ave been considered for te furter analysis. Te aridity index as been used to evaluate te combined effect of regional precipitation and temperature. Te evaluation concludes tat aridity index versus loss of fres groundwater resources exibits a global negative correlation. Te scatter plots of aridity index versus loss of fres groundwater resources sow regional clustering in te scatter points. Low aridity index, wic corresponds to iger loss of fres groundwater resource, was evident in te Central America and nortern Africa / Saara regions. A wide range of aridity index appears and te loss of fres groundwater resources is comparatively less in region. Furtermore, te loss of fres groundwater resources as a result of global warming impacts socio-economic activities in te water stressed regions. Future population growt and increased demand in te domestic, agricultural, and industrial sectors for fres groundwater will make te situation more complex in te sout Asian region. SRES A scenario wic sows igest population growt, predicts te igest 19

20 impact on per capita fres groundwater resources. Even toug te available per capita fres groundwater resources are iger in African region, te poor communities do not ave capacity and tecnology to access natural resources suc as groundwater. Te degradation of tese resources furter reduces te productivity of poor public. Acknowledgements Tis work was made possible troug te grants from Grant-in-Aid for Scientific Researc for JSPS (Prof. Nobuo Mimura) and Global Environment Researc Fund. Te autors would like to acknowledge te generosity of tose grants. We also tank Dr. Brian McGlynn and two anonymous referees for providing detailed and constructive suggestions and comments on te manuscript.

21 References Arnell,N,W,. 1999, Climate cange and global water resources, Global Environmental Cange 9, S31-S9. Arnell,N,W,., Climate cange and global water resources: SRES emissions and socio-economic scenarios, Global Environmental Cange 1, Arnell, N,W.,, Livermore, M.J.L., Kovatsc, S., Levy, P.E., Nicolls, R., Parry, M.L., Climate and socio-economic scenarios for global-scale climate cange impacts assessments: caracterising te SRES storylines, Global Environmental Cange 1, 3. Bear, J., Ceng, A.H.D., Sorek, S., Ouazar, D., and Herrera, I., 1999, Seawater Intrusion in Coastal Aquifers- Concepts, Metods and Practices, Kluwer Academic Publisers. Bobba, A.G., Sing, V.P., Jeffries, D.S., and Bengtsson, L., 1997, Application of a watersed runoff model to nort-east pond river, Newfoundland: to study water balance and ydrological caracteristics owing to atmosperic cange. Hydrological processes 1, CIESIN,, Center for International Eart Science Information Network, Columbia University,. Gridded Population of te World (GPW), CIESIN, Columbia University, Palisades, NY. ( ttp://sedac.ciesin.org/plue/gpw) DETR, 1997, Climate cange and its impacts, Department for Environment, Transport, and te Regions, Te UK Programme,. HMSO, Te Met Office, London. 1

22 Eckardta, K., Ulbricb, U., 3, Potential impacts of climate cange on groundwater recarge and stream flow in a central European low mountain range, Journal of Hydrology, 5. Essaid, H.I., 199, A Multilayered sarp interface Model of coupled freswater and saltwater flow in coastal systems: Model development and application, Water Resources Researc,, FAO-SDRN Agrometeorology Group Köeppen climate zones, FAO - Environment and Natural Resources Service (SDRN), Global climate maps series. (ttp:// Gaffin, S.R., Xing, X., Yetman, G.,, Downscaling and geo-spatial gridding of socio-economic projections from te IPCC Special Report on Emissions Scenarios (SRES). Global Environmental Cange, 1, Gertena, D., Scapoffa, S., Haberlandtb, U., Lucta, W., Sitca, S.,, Terrestrial vegetation and water balance-ydrological evaluation of a dynamic global vegetation model, Journal of Hydrology, 9 7. Gleick, P.H., 19, Metods for evaluating te regional ydrologic impacts of global climatic cange. Journal of Hydrology, Hitz, S., Smit, J.,, Estimating global impacts from climate cange, Global Environmental Cange 1, 1 1. Hulme, M., Mitcell, J., Ingram, W., Lowe, J., Viner, D., 1999, Climate cange scenarios for global impacts studies, Global Environmental Cange 9, S3-S19. IPCC, 199, Te Regional Impacts of Climate Cange: An Assessment of Vulnerability,199, Watson, R.T., Zinyowera, M.C., Moss, R.H. and Dokken, D.J., (Eds), Intergovernmental Panel on Climate Cange, Cambridge University Press, Cambridge, UK.

23 IPCC,, Emissions Scenarios: A Special Report of Working Group II of te Intergovernmental Panel on Climate Cange, Nakicenovic, N., Swart, R. (Eds.), Cambridge Univ. Press, Cambridge, Cambridge, UK. IPCC, 1, Climate Cange 1: Te Scientific Basis. Contributions of Working Group 1 to te Tird Assessment Report of te Intergovernmental Panel on Climate Cange, Hougton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K., and Jonson, C.A., (Eds), Cambridge University Press, Cambridge, UK. Jons, T.C., Gregory, J.M., Ingram, W.J., Jonson, C.E., Jones, A.,Lowe, J.A., Mitcell, J.F.B., Roberts, D.L., Sexton, D.M.H., Stevenson, D.S., Tett, S.F.B., Woodage, M.J., 3, Antropogenic climate cange for 1 to 1 simulated wit te HadCM3 model under up-dated emissions scenarios. Climate Dynamics, Labat, D., Godd, Y., Probst, J.L., Guyot, J.L.,, Evidence for global runoff increase related to climate warming, Advances in Water Resources 7, 31. Loaiciga, H.A., Valdes, J.B., Vogel, R., Garvey, J., Scwarz, H., 199, Global warming and te ydrologic cycle, Journal of Hydrology 17, 3-17 Mimikou, M.A., Baltas, E., Varanou E., and Pantazis, K.,, Regional impacts of climate cange on water resources quantity and quality indicators. Journal of Hydrology 3, Najjar, R.G., 1999, Te water balance of te Susqueanna River Basin and its response to climate cange. Journal of Hydrology 19, Nicolls, R.J., Lowe, J.A.,, Benefits of mitigation of climate cange for coastal areas, Global Environmental Cange 1, 9 3

24 Oliver J. E., Fairbridge R W., (Eds), 195, Te encyclopedia of Climatology, Encyclopedia of eart sciences, Volume XI, Van Nostrand Reinold Publisers, New York. Paari, K., and Murai, S., 1999, Modelling for prediction of global deforestation based on te growt of uman population, Journal of Potogrammetry and Remote Sensing, 5, PSMSL : Permanent Service for Mean Sea Level data set, Proudman Oceanograpic Laboratory (POL), (ttp:// Ranjan S.P., Kazama, S., Sawamoto, M.,, Effects of climate and land use canges on groundwater resources in coastal aquifers, Journal of Environmental Management, Submitted for publication Reilly, T.E., Goodman, A.S., 195, Quantitative analysis of fres-salt water relationsip in groundwater systems- A istorical perspective, Journal of Hydrology,, Remson, I., Hornberger, G.M., Molz, F.I., 1971, Numerical metods in subsurface ydrology: wit an introduction to te finite element metod, pp39. Seckler, D., Amarasinge, U., Molden, D., de Silva, R., Barker, R., 199. World Water Demand and Supply, 199 to 5: Scenarios and Issues. International Water Management Institute Researc Report 19, Sri Lanka. Semmler, T., and Daniela Jacob,, Modeling extreme precipitation events; A climate cange simulation for Europe, Global and Planetary Cange,

25 Tables Table 1. Caracteristics of selected water resources stressed regions Region Climate region a groundwater Related region b Central America (CAM) Soutern Africa (SAF) Nortern Africa / Saara (HAS) (MED) Soutern Asia (SAS) Tropical / subtropical Desert Tropical / Monsoon Nort and Central America Sub Saaran basins Nort African Basins Soutern Europe and Atlas Mountains Peninsular India Caracteristics of coastal aquifers b Geological Hydraulic formation conductivity Sedimentary and alluvial deposits m/s Crystalline rocks and sedimentary deposits Marine sediments, Eolian sand and alluvial deposits Limestone, sandstone and unconsolidated sediments Sedimentary rocks and unconsolidated sediments m/s m/s m/s m/s Source: (a) Koeppen s Climate Classification (FAO) ; (b) Global Groundwater regions, IGRAC ; (c) Permanent Service for Mean Sea Level (PSMSL) Sea level rise c. mm/year 1.5 mm/year 1.5 mm/year.5 mm/year.9 mm/year 5

26 Table Long-term trends of fres groundwater loss for ig and low emission scenarios Rate of cange of fres groundwater loss (% per year) Hig emission scenario ( A) low emission scenario ( B) Central America Nort Africa Sout Africa.7. Sout Asia.75.7 Table 3. Statistical beavior of fres groundwater resources Fres groundwater loss (m 3 km - ) Statistical Hig emission scenario (A) Low emission scenario (B) parameters CAM MED SAH SAF SAS CAM MED SAH SAF SAS Mean Standard Deviation Table. Population and fres groundwater availability in water resource stressed regions Emission scenario A B Region Population density 1 1 Fres Per capita Population groundwater resources density availability (m 3 /cap/year) (m 3 per km ) Fres Per capita groundwater resources availability (m 3 /cap/year) (m 3 per km ) Central 1 America Nort Africa Sout Africa Sout Asia Central 11 America Nort Africa Sout Africa Sout Asia

27 Illustrations Fig 1. Scematic representing te loss of fres groundwater resources due to salinity intrusion in coastal aquifers Fig. (a) Selected water resource stressed areas (Aller et al, 1999); (b) Representation of coastal boundary in HadCM3 resolution Fig 3. Average annual variation of (a) loss of fres groundwater resource, (b) precipitation and (c) temperature for SRES A scenario Fig. Average annual variation of (a) loss of fres groundwater resource, (b) precipitation and (c) temperature for SRES B scenario Fig. 5. Correlation between (a) precipitation versus loss of fres groundwater resource (b) temperature versus loss of fres groundwater resource and (c) aridity index versus loss of fres groundwater resource for A scenario Fig.. Correlation between (a) precipitation versus loss of fres groundwater resource (b) temperature versus loss of fres groundwater resource and (c) aridity index versus loss of fres groundwater resource for B scenario Fig. 7. Scatter plot of te cange in aridity index and loss of fres groundwater resource in different climate regions (a) for A scenario, (b) for B scenario 7

28 P ET Sea level Salt water Interface 1 Freswater level Interface 3 Fres water Interface Freswater loss Recarge Groundwater flow Fig 1. Scematic representing te loss of fres groundwater resources due to salinity intrusion in coastal aquifers Central America Nort Africa / Saara Sout Asia Sout Africa (a) (b) Fig. (a) Selected water resource stressed areas (Aller et al, 1999) ; (b) Representation of coastal boundary in HadCM3 resolution

29 Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Gradient =.15 America Central America Mediteranean NortAfrica Africa/Saara / Sout SouternAfrica Africa Sout Asia Sout Asia Gradient =. Gradient = -. Gradient =.7 Gradient = Time (years) Precipitation (mm/year) Precipitation (mm/year) Precipitation (mm/year) Precipiation (mm/year) Precipitation (mm/year) Central America America Nort Africa Africa/Saara / SouternAfrica Sout Africa SoutAsia 1 Asia Time(year) Temperature (C) Temperature (C) Temperature (C) Temperature (C) Temperature (C) Central America America NortAfrica/ Africa/Saara Sout NortAfrica/ Africa Saara SoutAsia Asia Time (Year) Fig 3. Average annual variation of (a) loss of fres groundwater resource, (b) precipitation and (c) temperature for SRES A scenario 9

30 Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Fres groundwater loss (%) Central America America Mediteranean NortAfrica / Africa/Saara SouternAfrica Sout Africa SoutAsia Sout Asia Gradient =. Gradient =.5 Gradient = -.1 Gradient =. Gradient = Time (years) Precipitation (mm/year) Precipitation (mm/year) Precipitation (mm/year) Precipiation (mm/year) Precipitation (mm/year) Central America America Africa / Nort Africa/Saara SouternAfrica Sout Africa SoutAsia Asia Time (year) Temperature (C) Temperature (C) Temperature (C) Temperature (C) Temperature (C) Central America America NortAfrica/ Africa/Saara NortAfrica/ Saara Sout Africa SoutAsia Asia Time (Year) Fig. Average annual variation of (a) loss of fres groundwater resource, (b) precipitation and (c) temperature for SRES B scenario 3

31 Fres gro undwater loss (% ) Fres grou nd water lo ss (% ) Fres grou nd water lo ss (%) Fres g rou ndw ater loss (%) Fres g roundw ater loss (%) Central America Nort Africa Sout Africa Sout Asia r =. y=-.11x Can ge in precipitation (mm/ year) r =.9 y= -.1x r =. y = -.x r =.5 y = -.11x r =.7 y =-.13x Fres groundwater loss (% ) Fres groundwater loss (% ) Fres groundwater loss (%) Fres groundwater loss (%) Fres g roundw ater loss (%) Central America Nort Africa Sout Africa Sout Asia Cange in temperature (C) r =.13 y =.15x r =.1 y =.3x r =.9 y =.3x r =.5 y =.73x r =.73 y = 1.55x Fres gro und water loss (% ) Fres grou nd water lo ss (% ) Fres grou nd water lo ss (%) Fres g rou ndw ater loss (%) Fres groun dw ater loss (%) Central America Nort Africa Sout Africa Sout Asia r =.5 y = -.3x r =.7 y =-.7x r =.1 y =-.5x r =. y =-.3x Can ge in arid ity index r =. y =-.5x Fig. 5. Correlation between (a) precipitation versus loss of fres groundwater resource (b) temperature versus loss of fres groundwater resource and (c) aridity index versus loss of fres groundwater resource for A scenario 31

32 Fres gro und water loss (% ) Fres grou nd water lo ss (% ) Fres grou nd water lo ss (%) Fres g rou ndw ater loss (%) Fres groun dw ater loss (%) Central America r =.5 y=-.x Nort Africa r =.15 y = -.x Sout Africa r =.5 y = -.3x Sout Asia r =. y = -.15x r =. y = -.5x Cange in precipitation (mm/ year) Fres groundwater loss (% ) Fres groundwater loss (% ) Fres groundwater loss (%) Fres groundw ater loss (%) Fres groundw ater loss (%) Nort Africa Sout Africa Sout Asia Central America r =.1 y =.5x r =.1 y =.5x r =.1 y =.5x r =. y =.75x Cange in temperature (C) r =.33 y =.75x Fres groundwater loss (% ) Fres groundwater loss (% ) Fres groundwater loss (%) Fres groundwater loss (%) Fres groundw ater loss (%) Central America r =.5 y = -.7x r =.7 y = -.3x Nort Africa r =. y = -.3x Sout Africa r =.51 y = -.15x Sout Asia r =.5 y = -.3x Cange in aridity index Fig.. Correlation between (a) precipitation versus loss of fres groundwater resource (b) temperature versus loss of fres groundwater resource and (c) aridity index versus loss of fres groundwater resource for B scenario 3

33 loss of fres groundwater (m 3 / km ) Central America Nort Africa Sout Africa Sout Asia SAS CAM MED SAH SAF Aridity index (a) loss of fres groundwater (m 3 / km ) Central America Nort Africa Sout Africa Sout Asia SAS CAM MED SAH SAF Aridity index Fig. 7. Scatter plot of te cange in aridity index and loss of fres groundwater resource in different climate regions (a) for A scenario, (b) for B scenario (b) 33