Hydrogeological conditions of a crystalline aquifer: simulation of optimal abstraction rates under scenarios of reduced recharge

Hindawi Publishing CorporationThe Scientific World JournalVolume 2013 Article ID 606375 8 pageshttpdxdoiorg1011552013606375

Research ArticleHydrogeological Conditions of a Crystalline AquiferSimulation of Optimal Abstraction Rates under Scenarios ofReduced Recharge

Sandow Mark Yidana Obed Fiifi Fynn Larry Pax ChegbelehProsper M Nude and Daniel K Asiedu

Department of Earth Science University of Ghana Legon Accra Ghana

Correspondence should be addressed to Sandow Mark Yidana smyidanaugedugh

Received 3 October 2013 Accepted 10 November 2013

Academic Editors Z Dai N Hirao and K Nemeth

Copyright copy 2013 Sandow Mark Yidana et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

A steady state numerical groundwater flow model has been calibrated to characterize the spatial distribution of a key hydraulicparameter in a crystalline aquifer in southwestern Ghana This was to provide an initial basis for characterizing the hydrogeologyof the terrain with a view to assisting in the large scale development of groundwater resources for various uses The results suggestthat the structural entities that control groundwater occurrence in the area are quite heterogeneous in their nature and orientationascribing hydraulic conductivity values in the range of 45md to over 70md to the simulated aquifer Aquifer heterogeneitiescoupled possibly with topographical trends have led to the development of five prominent groundwater flowpaths in the areaEstimated groundwater recharge at calibration ranges between 025 and 913 of the total annual rainfall and appears to holdsignificant promise for large-scale groundwater development to support irrigation schemes However the model suggests that withreduced recharge by up to 30 of the current rates the system can only sustain increased groundwater abstraction by up to 150of the current abstraction rates Prudent management of the resource will require a much more detailed hydrogeological study thatidentifies all the aquifers in the basin for the assessment of sustainable basin yield

1 Introduction

Groundwater resources constitute arguably the most reli-able buffer against the unremitting effects of climatechangevariability and the concomitant ramifications onsustainable agriculture especially in the developing worldThis is so because groundwater is mostly protected from highsurface temperatures and the corresponding high evapotran-spiration rates that affect surface flows and impoundmentsand thus render them ineffective sources of irrigation watersupply to sustain large-scale irrigation activities In thelight of this regional hydrogeological studies that lead tothe development of aquifers for sustainable abstraction andmanagement of groundwater resources is crucial

There are various approaches available for regional hydro-geological investigations and for providing the necessaryinformation required for optimal aquifer and basin yield

management [1]They include the application of remote sens-ing aerial photography surface geological field investigationsapplication of advanced geophysical methods and subse-quent drilling to access aquifers The conventional practice isthe use of all data obtained from remote sensing techniquestogether with field based data and borehole information todevelop a conceptual model which is then converted intoa numerical model [2ndash11] to predict the hydrogeologicalconditions of aquifers Basin wide hydrogeological investi-gations usually culminate in the development of regionallybased numerical groundwater flow models which capturethe essential aspects of the regional hydrogeology and thusprovide decision support systems for the management ofgroundwater resources The use of models will continue toprovide useful leads to the effective management of flowand solute transport in aquifers especially where climatechangevariability and its attendant effects on the spatial

2 The Scientific World Journal

and temporal variations in rainfall patterns place a hugeuncertainty on the sustainability of rain-fed agriculturalactivities and where increasing global temperatures and lowhumidities render surface impoundments unsustainable asirrigation water sources They are unavoidable in large-scalehydrogeological studies that have the objectives of constrain-ing some key aquifer storage and hydraulic parameters [2 1012] for better aquifer characterization

In Ghana erratic rainfall patterns have been noted inrecent times This has affected rain-fed agricultural activitiesin most areas Efforts are being made at phasing groundwaterresources at reasonable depths to meet growing domesticwater needs whilst exploring the possibilities of developingthe same resources for large-scale irrigation activities toaugment agricultural activities This would enhance foodsecurity and ultimately contribute to poverty reduction inthe rural farming communities Towards achieving theseobjectives detailed hydrogeological investigations have beenproposed to effectively characterize aquifers for proper devel-opment of the resource This study forms one of the initialstages of this effort and utilizes a steady state numericalgroundwater flowmodel to characterize the spatial variationsin the hydraulic conductivities of a crystalline basementaquifer in southwestern Ghana In addition to providing aconceptual framework of the groundwater flow system inthe local environment the methodology is of internationalimportance as the utility of groundwater models in suchendeavours can be applied in other terrains

2 The Study Area

The study area (Figure 1) is characterized by two main rainyseasons referred to as major and minor cropping seasons(AprilndashJuly) and (SeptemberndashNovember) It is characterizedby a tropical climate with high temperatures averaging239∘C (75∘F) and a double maxima rainfall pattern [13]Rainfall ranges from an average of 1000mm in the northernparts to 1400mm in the south The entire domain is drainedby tributaries of the Volta river basin [14] The majoreconomic activity there is agriculture which employs themajority (about 66) of the economically active populationCashew coffee teak rubber and tobacco are the main cashcrops in the area [14]

The area is underlain by aquifers of the Birimian Province[15] which is one of the five proposed hydrogeologicalprovinces in Ghana It is characterized by basaltic andandesitic lavas pyroclastic rocks hypabyssal intrusive rockand greywacke [16] The basic volcanic and pyroclasticrocks have been largely altered to chloritised and epidotisedrocks that are sometimes loosely referred to as greenstonesGenerally rocks of the Upper Birimian are intensely foldedwith steep dips and general north-south or north-west strikes[17] In the absence of primary permeabilities the rocksowe their hydrogeological conditions to the presence andpervasiveness of secondary structural entities which arediverse amongst rocks of the Birimian Province Available

District capitalRoads

Phyllite schist tuff and greywacke

Quartzite phyllite grit conglomerate andschist including basic intrusiveMetamorphosed lava and pyroclastic Rock andhypabyssal basic intrusive phyllite and Greywacke

Drobo

2∘30

9984000998400998400W

2∘30

9984000998400998400W2

∘45

9984000998400998400W

2∘45

9984000998400998400W

8∘09984000998400998400

N

8∘09984000998400998400

N

7∘459984000998400998400

N

7∘459984000998400998400

N

7∘309984000998400998400

N

7∘309984000998400998400

N

0 110 220 33055

Jaman

(km)

Figure 1 Map of the study area showing the geology

hydrogeological data suggest that aquifers of the BirimianProvince are amongst themost prolific in the country Depthsof boreholes drilled through rocks of the Province rangebetween 35m and 62mwith an average of 42m [15] Boreholedepths in areas underlain by the granitoids are similar andrange between 35m and 55m with an average of 50m [18] Insome areas the regolith is tapped at relatively shallow depthswith relatively shallow hand dug wells Aquifer transmissivityof the productive zones of the Birimian Province rangesbetween 02m2d and 119m2d with an average of 74m2d[15] In these aquifers storativity ranges between 0003 and0008 Transmissivity within the regolith is slightly higherthan that observed in the integrated aquifer system andranges between 4m2d and 40m2d with an average of about10m2d For the integrated aquifer systems in the provinceborehole yields are generally low and range from 048m3hto 364m3h with a mean yield of 76m3h Differences in thedegree of weathering within the granitoids probably accountfor the lower yields observed in these rocks

3 Materials and Methods

31 Choice of Numerical Code The flow of groundwaterof constant viscosity and density through a heterogeneousanisotropic porous medium under nonequilibrium condi-tions is described by (1) [19ndash21] which is obtained through

The Scientific World Journal 3

the application of the Darcy law and the law of conservationof mass

119870119909

1205972

ℎ

1205971199092+ 119870119910

1205972

ℎ

1205971199102+ 119870119911

1205972

ℎ

1205971199112plusmn119882 = 119878

119904

119889ℎ

119889119905 (1)

where 119870119894 119882 and 119878

119904 respectively refer to the hydraulic

conductivity in the 119894th direction sourcessinks and theaquifer specific storage

This equation has been used in diverse forms to modelgroundwater flow depending on the prevailing conditionsFor instance where it cannot be safely assumed that densityand viscosity are largely constant (1) is modified to accom-modate such a situation Such a situation arises especially incoastal aquifers where salinewater intrusion imposes variablesalinities which in turn result in spatially variable ground-water densities In such a case a finite element code of thekinds of FEMWATER [22] which is based on the form of (1)which incorporates viscosity and density variations attendingchanges in the chemical constitution of groundwater in thespace of a domain

In the current study steady state conditions wereassumed In this respect the time variable nature of thehydraulic head on the right hand side of (1) was regardednegligible because the current sinks are not consideredsignificant enough to cause such changes In addition acursory analysis of the hydrochemical data from the terrainhad suggested that groundwater density and viscosity wouldlargely be the same throughout the domain As such (2) wasused

119870119909

1205972

ℎ

1205971199092+ 119870119910

1205972

ℎ

1205971199102+ 119870119911

1205972

ℎ

1205971199112= 0 (2)

On the basis of this United States Geological SurveyrsquosModular Finite Difference Groundwater Flow Modellingcode MODFLOW-2000 [23] was chosen to simulate thegroundwater flow system in the study area MODFLOW isarguably the most tested finite difference numerical codeand has proven to accurately predict hydrogeological systemswithin the limits of field conceptualization and the availabledata The Groundwater Modelling System GMS [24] whichcontains MODFLOW and has an inbuilt graphical interfacefor MODFLOW output and a versatile GIS system forconceptualization purposes was employed in this study

32 Data Sources and Conceptualization The hydrogeolog-ical data for this study were obtained from the offices ofthe Community Water and Sanitation Agency CWSA theSAL Consult (a hydrogeological consulting firm based inAccra) theWater Resources Commission and theGeologicalSurvey Department of Ghana Data from borehole logsand the historical account of the geology and hydrogeologyassisted in the vertical conceptualization of the terrain Theconceptualization was based on logs from 22 boreholesThe base map (the geological map of the terrain) wasimported and registered in the GMS using the Map toolsSpatial discretization for the aquifer hydraulic parameterswas facilitated by the local geology

On the basis of the available data the terrain wasconceptualized as a single layer model The lower limitsof the aquifer system were conceptualized as a confininglayer to coincide with the impervious material beneath Theupper limit was modelled as a convertible layer to coincidewith semiconfining conditions prevailing in the area In theabsence of physical boundaries to limit flow on all sidesthe vertical walls were all conceptualized as General HeadBoundaries GHB [25] This would enable the simulation ofnet groundwater flow across the boundaries of the terrainThe GHB is a head dependent condition which simulatesflow across a boundary based on the head difference acrossthe boundary and the conductance of the material across theboundary (3) Consider

119876 = 119862119889119889ℎ (3)

where 119876 119862119889 and 119889ℎ are respectively the flow across the

boundary the conductance of the material and the hydraulichead difference across the boundary

An observation coverage was created to accommodatehydraulic data from 12 boreholes for the purpose of cali-bration Coverages were similarly created for the hydraulicconductivity and recharge fields aswell as theGHB conditionData of the upper and lower limits of the aquifer wereimported as text files as part of the conceptualization processA grid system was then automatically developed to overlaythe domain A uniform rectangular grid system was used forthe entire domain as the intended purposewas to general flowwithout focus on any particular location

33 Numerical Simulation The conceptual model was con-verted into a numerical MODFLOW model to begin thesimulationThe thickness of the layerwas defined bymappingthe top and bottom elevations obtained from the boreholelogs into MODFLOW The appropriate solver and flowpackages were then selected to begin the simulation of flowAn elaborate account of documentation of the MODFLOWcode is contained in the US Geological Survey reports [23]However in the GMS system the solver and flow packagesare available in the Global options Three flow packagesare available the Layer Property Flow (LPF) the BlockCentered Flow (BCF) and the Hydrogeologic Unit Flow(HUF) packages There are five solver packages the stronglyimplicit procedure the Pre-Conditioned Conjugate Gradient(PCG) the Slice Successive Over-relaxation (SOR) and theGeometric Multigrid (GMG) approach In this study theLPF and PCG were respectively chosen as the flow andsolver packages to simulate the conditions in the terrainThe LPF package supports two types of layers confinedand convertible In a confined layer the transmissivity isconstant throughout the simulation The value of the trans-missivity is computed from the hydraulic conductivity andcell elevations Similarly under transient conditions thestorage is computed by multiplying the specific storage bythe cell elevation A convertible layer is one in which thetransmissivity varies with the hydraulic head throughout thesimulation The transmissivity is computed based on the cellthickness and the hydraulic conductivity and includes the

4 The Scientific World Journal

possibility of a cell converting to a no flow boundary whenthe water level goes below the bottom elevation of the layerIn transient simulation (which was not simulated in thisstudy) for a convertible layer the storage contribution toflow is computed from confined andor unconfined storagedepending on the level of the hydraulic head compared tothe top elevation of cells [26] The specific storage valuesare multiplied by cell volume to obtain confined storagecapacity whereas the specific yield is multiplied by the cellarea to obtain unconfined storage capacity Two storagecapacities are therefore stored and the program uses theappropriate value depending on head conditions In thisstudy however since the simulation was done under steadystate conditions this aspect of the package was not utilizedas storage parameters are not assigned under steady statesimulations Since the aquifer system being simulated in thisstudy is semiconfined the convertible option was selected

The Preconditioned Conjugate Gradient (PCG) solver[27] uses both outer and inner iterations The user specifiesthe maximum number of iterations so that the simulationterminates when the convergence criteria are not met withinthemaximumnumber of iterationsWhen the solver and flowpackages were chosen the model was set on forward run tobegin the simulation

34 Model Calibration The only observation data used tocalibrate the model were hydraulic head data as there are noknown springs and other drains in the area The calibrationobjective was therefore to ensure that the model computedhydraulic head data closelymatched the observed data withina margin of error established during the conceptualizationprocess Manual calibration was first performed by alteringthe values of the hydraulic conductivity and recharge andrunning the model each time When the model stabilizedthe calibration was switched over to the automatic calibrationmode through the Parameter Estimation PEST [23 24 2728] Parameter limits were defined for each of the coverages ofrecharge and hydraulic conductivity based on previous expe-riences in similar terrains and climate zones The calibrationtarget was set at plusmn25m which means that for each of theobservation wells calibration was said to have been achievedwhenever the observed andmodel computed hydraulic headswere with 25m of each other In the PEST mode the pilotpointmethod [28 29] was used for the hydraulic conductivityparameter so that a smooth interpolation surface would beproduced of the hydraulic conductivity field at calibration

35 Sensitivity Analysis Sensitivity analysis is recommendedafter calibration of every model The objective is to test thestability of the model in the face of subtle variations in someof the key aquifer parameters A highly sensitive model toany of the parameters is regarded as unstable and thereforenot reliable for use in predicting scenarios In this projectsensitivity analysis was performed automatically through thePEST In this way histograms would be generated at the endof the calibration to indicate parameter sensitivities

36 Abstraction Scenarios Since the model was conceptu-alized and calibrated under steady state conditions it is

255 260 265 270 275 280 285 290

255

260

265

270

275

280

285

290

Com

pute

d he

ad (m

)

Observed head (m)

Figure 2 A linear plot showing the relationship between observedand computed hydraulic head

not appropriate for modelling fluctuations in groundwaterstorage However it can be used in a limited fashion toevaluate recharge and abstraction scenarios For each of theobservation wells used in this study the yields estimated dur-ing pumping tests were applied as the initial abstraction ratesIn the scenario analyses abstraction rates were increased by10 20 50 100 and 150 whilst maintaining rechargeat the calibrated rates In the next scenario the recharge wasdeliberately reduced by 10 20 and 30whilst abstractionrates were increased at 10 20 50 100 and 150This scenario was expected to simulate possible reductionin groundwater recharge rates due to decreasing rainfallamounts at the recharge areas

4 Results and Discussions

41 The Groundwater Flow System The model was deemedcalibrated when the computed heads for all the wells werewell within 25m of the observed heads Figure 2 illustratesthe relationship between the observed and model computedhydraulic heads for the 12 wells used for the calibrationA good match between model computed and observedhydraulic heads is obvious in Figure 2 suggesting that themodel is reasonably calibrated within the limits of the dataused and is therefore representative of the hydrogeologicalconditions prevailing in the terrain The sensitivity analysissuggests that the model is largely stable to subtle variationsin the key aquifer parameters of hydraulic conductivity andrecharge Since model calibration was achieved by varyingthe parameters of hydraulic conductivity and recharge theranges of values of these parameters at calibration wereadjudged representatives of the conditions prevailing in theterrain The predicted hydraulic head field is as presented inFigure 3 which suggests that the areas of the highest hydraulichead are in the northern parts of the terrain Howeverthere is no clearly defined direction of flow indicating thatthe local structural entities that control the flow system arenot oriented in any preferred direction Local flow systemsapparently dominate the flow in the area Where the aquiferis highly heterogeneous with respect to the key aquifer

The Scientific World Journal 5

Hydraulic head (m)

29845

28526

27208

25890

24572

23253

21935

20617

19299

17980

16662

15344

14026

12707

11389

0 30 X

Y

Z

Scale (km)

Figure 3The potentiometric surface predicted by themodel for thestudy area

hydraulic parameters flow is largely haphazard and local flowsystems are predominant The same is true when the terrainis of considerably variable topography [20] Heterogeneitiesin the hydraulic conductivity are more likely responsible forthe observed pattern of flow in the current study

Five prominent flow paths have been defined in thestudy area (Figure 4) through detailed particle tracking fromMODPATH [30] There is a conspicuous flow divide in thenorth-central parts of the area This is a hydraulic boundaryto flow attributedmainly to aquifer heterogeneity rather thantopographical complexities in the terrain Although localtopography has an influence on flow systems [1 19] andcan sometimes be used to predict the direction of flow theultimate determinant of the flow is the level of heterogeneityof the material of the aquifer Where the topography isrugged and the aquifer is largely homogeneous and isotropicthe system is deficient in local flow systems and regionalflow systems are dominant On the other hand where thesurface topography is largely flat marked heterogeneitiesof the aquifer can impose several complex local systemsdue to high levels of refraction of groundwater flowpathsthrough materials of contrasting hydraulic properties Thesignificance of this lies in the fact that where the surfacetopography alone is used to predict the trajectory of a particlein groundwater for the purpose of assisting in remediationpurposes the interventions can be misdirected if detailedanalyses are not performed to culminate in the simulation ofthe flow In all cases of groundwater resources development

298452852627208258902457223253219352061719299179801666215344140261270711389

0 30 X

Y

Z

Hydraulic head (m)

Scale (km)

Figure 4 The most prominent groundwater flowpaths in the studyarea

and solute transport studies therefore numerical flow mod-elling is inevitable The detail of the flow system is a relevantprerequisite for the effectual management of the resource

Yidana et al [9] predicted a general NE-SW groundwaterflow pattern amongst similar crystalline basement aquifersin the south of the terrain using a similar methodologyThe obvious departure in the current study suggests that secondary permeabilities in the BirimianProvince are inherentlydiverse and do not necessarily align in a particular directionThus whereas the regional structural geology suggests thedominance of NE-SW trending structural entities [16] localgroundwater flow conditions may considerably deviate fromthis regional pattern especially where extensive weatheringoccurs to enhance the local hydrogeology Previous observa-tions in other parts of Ghana [9ndash11] suggest that where theflow system is dominated by regional systems a prominentNNE-SSW flow system is observed The regional hydroge-ology of Ghana therefore appears to be largely controlledby the regional structural grain This finding implies thatwhere the hydrogeology of an area is based on secondarypermeabilities variable groundwater flow patterns resultingfrom the spatial variabilities of the entities controlling thehydrogeology should be expected A similar model within agranular unconfined aquifer system in the Keta Basin [26]produced a much more smooth flow pattern due to the dom-inance of primary permeabilities and the interconnectivitiesof such entities The use of numerical techniques in thisway is important towards enabling a proper understandingof mechanisms of groundwater flow and will facilitate themanagement of aquifers and groundwater resources

42 The Hydraulic Conductivity Field A smooth map ofthe horizontal hydraulic conductivity field in the domain

6 The Scientific World Journal

has been established through the pilot point method [27ndash29] in this study It suggests a significantly heterogeneoussystem where hydraulic conductivity ranges between 45mdand over 70md (Figure 5) and amply explains the observedgroundwater flow pattern There is no obvious relationshipbetween the hydraulic conductivity field and the generaldistribution in hydraulic head (Figure 3) This suggests that(i) local variations in the vertical hydraulic conductivitieswhich control vertical groundwater recharge in the area aresignificantly different from the patterns of the horizontalhydraulic conductivity andor (ii) subsurface flows are themajor sources of groundwater recharge in the terrain Thecomputed horizontal hydraulic conductivities in this studyare largely consistent with the aquifer transmissivity dataobtained from pump tests within the Birimian Province andreported in other publications [15] and is largely in syncwith the general character of fracture controlled aquifersystems The validity of hydrogeological data on aquiferparameters from pump tests in Ghana has been the subjectof debate on grounds of speculations of partial penetrationof boreholes and the concomitant effects on the results ofpump tests Generally the use of pump-test estimates ofaquifer parameters in regional hydrogeological investiga-tions has been considerably criticized in the literature [31ndash33] Razack and Huntley [31] and Huntley et al [32] forinstance proved that such analytical techniques have thepropensity to underestimate aquifer transmissivity values inheterogeneous alluvial aquifers and overestimate the samein fracture controlled aquifer systems Mace [33] proposesempirical relationships between aquifer transmissivity andspecific capacity for the estimation of the former given data ofthe latter Such empirical relationships have been establishedfor some aquifers in Ghana [34 35] Other researchershave proposed the use of cokriging and the strength ofthe relationship between aquifer transmissivity and specificcapacity to produce smooth prediction maps for the formerHowever since such maps would be the product of datafrom pump tests estimates they suffer the same defects interms of reliability Generally aquifer transmissivity estimatesfrom pump tests provide the character of the aquifer atthe well scale alone and must therefore be used cautiouslyat the regional scale This is because measures of aquiferpermeability are scale dependent and therefore point esti-mates may bemisleading In order to sufficiently characterizeregional aquifers through the conventional pump tests alonea very large number of such tests will have to be conductedThis is highly impracticable due to the prohibitive costs ofsuch endeavours It is on the basis of this fact that regionalnumerical groundwater flow models of the kind used inthe current study have been proposed for such estimatesNumerical models by their nature predict aquifer parametersbased on set of observed data and boundary conditionsThey therefore largely predict the regional balance of theseparameters fairly and accurately

43 Groundwater Recharge and Resource Sustainability Esti-mated vertical groundwater recharge in the area rangesbetween 3mm and 1095mmper annum equivalent to 025and 913 of the average annual precipitation in the basin

0 30 X

Y

Z

7020

6553

6085

5618

5151

4683

4216

3748

3281

2813

2346

1879

1411

944

476

Hydraulic conductivity (md)md)

Scale (km)

Figure 5 Calibrated hydraulic conductivity field for the study area

Apparently a significant proportion of the total groundwaterinput in the terrain results from subsurface flows throughthe GHB condition contributing to over 35000m3day ofthe total groundwater input into the system Contributionsof subsurface flows are largely responsible for the observedgroundwater flow geometry in the study area This much ofrecharge holds significant promise for groundwater resourcesdevelopment in the area The estimated rates are compatiblewith the estimates of Yidana et al [9] in the southern parts ofthe terrain using the ChlorideMass Balance CMB techniqueandmodel calibration It is also in agreement with the generaltrend observed in other parts of the country where rechargehad been estimated through a variety of othermethods [11 1826] However since transient conditions were not simulatedchanges in groundwater storage could not be estimatedThe application of groundwater flow modelling in hydroge-ological studies and groundwater resources assessment hasnot been popular in Ghana and much of the West Africansubregion where inaccessibility to potable water resourcessometimes constitutes one of the major causes of extremedeprivation

44 Scenario Analyses Freeze and Cherry [1] defined theconcepts of sustainable aquifer and basin yields to reflectthe maximum permissible abstraction rates from all wellsthrough the aquifer and basin respectively The need forclear definitions of these concepts stem from the fact thatmost groundwater resources studies aim at determining themaximum abstraction rates that are compatible with the

The Scientific World Journal 7

hydrogeological environment The definition of the exactabstraction rates that are regarded as optimal is dependenton several considerations such as the number of interveningprocesses andor factors that have a bearing on groundwaterresources In this study sustainable aquifer yield is definedas the maximum possible groundwater abstraction rates thatwill lead to minimal effects on the general distribution ofthe hydraulic head At the current calibrated groundwaterrecharge conditions the model suggests that the aquifer cansustain increment in groundwater abstraction rates up to200 of the current rates with minimal (less than 15)depreciation in the hydraulic head throughout the entiredomain This is in keeping with the observed recharge ratesand the apparently relatively pristine groundwater abstrac-tion conditions prevailing in the area The implication is thatwhereas the current recharge rates may not be guaranteeddue to possible climate change impacts the aquifer holdssignificant fortunes for large-scale abstraction to supportirrigation schemesThe second scenario assessed the impactsof increased groundwater abstraction rates whilst recharge isreduced from the current ratesThis study assessed reductionin recharge by 10 20 and 30 of the current levelsThe results of the simulation suggest that there will bean appreciable depreciation in the hydraulic head leadingto possible reversals in the flow pattern when recharge isreduced by up to 30 whilst abstraction rates continue toincrease by up to 150 of the current levels The assumptionof steady state conditions in the current study was basedon the fact that the current abstraction rates are consideredvery minimal and insignificantThis has been amply justifiedthrough the simulation of the various scenarios of groundwa-ter abstraction rates in this study

5 Conclusions

This research proves that model calibration is one of thebest approaches towards constraining the spatial distribu-tion of aquifer hydraulic parameters in large-scale regionalhydrogeological assessmentsTherefore in regional hydroge-ological studies the use of numerical groundwater models isvery much recommended for achieving the goals of aquifercharacterization In this study a hydraulic conductivity fieldhas been developed for a crystalline basement aquifer inSouthern Ghana through model calibration It suggests thataquifer hydraulic conductivity in the area ranges between450md and over 70mdThe heterogeneity in the predicteddataset appears to be dictated by the heterogeneities in thestructural entities that govern the hydrogeological propertiesof the aquifers in the area Local flow systems appear tobe predominant due to the observed heterogeneities in theaquifer properties in the area Five prominent flow lineshave been identified in the study area where groundwaterrecharge rates range between 025 and 913 of the totalannual precipitation in the area A substantial proportionof this recharge appears to accrue from subsurface flowsThis much recharge holds promise for large-scale develop-ment of groundwater resources for irrigation in the areaas suggested by the results of the various scenarios of

groundwater abstraction simulated in the terrain Even witha reduction in recharge by up to 30 of the current rates thedomain can sustain an increase in groundwater abstractionby up to 150 of the current abstraction rates Evaluationof the effects of climate changevariability on groundwaterresources sustainability and livelihood will require the use ofgroundwater flow models especially when such models arecalibrated for transient conditions

References

[1] R A Freeze and J A Cherry Groundwater Prentice Hall Eng-lewood Cliffs NJ USA 1979

[2] D U Ophori ldquoConstraining permeabilities in a large-scale gro-undwater system throughmodel calibrationrdquo Journal of Hydrol-ogy vol 224 no 1-2 pp 1ndash20 1999

[3] Q Zhang R E Volker and D A Lockington ldquoNumerical inv-estigation of seawater intrusion at Gooburrum BundabergQueensland AustraliardquoHydrogeology Journal vol 12 no 6 pp674ndash687 2004

[4] B He K Takase and Y Wang ldquoNumerical simulation of gro-undwater flow for a coastal plain in Japan data collection andmodel calibrationrdquo Environmental Geology vol 55 no 8 pp1745ndash1753 2008

[5] S Khan T RanaH F Gabriel andMKUllah ldquoHydrogeologicassessment of escalating groundwater exploitation in the IndusBasin Pakistanrdquo Hydrogeology Journal vol 16 no 8 pp 1635ndash1654 2008

[6] F Barry D Ophori J Hoffman and R Canace ldquoGroundwaterflow and capture zone analysis of the Central Passaic RiverBasin New Jerseyrdquo Environmental Geology vol 56 no 8 pp1593ndash1603 2009

[7] S Jusseret V T Tam and A Dassargues ldquoGroundwater flowmodelling in the central zone of Hanoi VietnamrdquoHydrogeologyJournal vol 17 no 4 pp 915ndash934 2009

[8] V Shankar P Eckert C Ojha and C M Konig ldquoTransientthree-dimensional modeling of riverbank filtration at Grindwell field GermanyrdquoHydrogeology Journal vol 17 no 2 pp 321ndash326 2009

[9] S M Yidana S Ganyaglo B Banoeng-Yakubo and T AkabzaaldquoA conceptual framework of groundwater flow in some crys-talline aquifers in SoutheasternGhanardquo Journal of African EarthSciences vol 59 no 2-3 pp 185ndash194 2011

[10] S M Yidana ldquoGroundwater flow modeling and particle track-ing for chemical transport in the Southern Voltaian aquifersrdquoEnvironmental Earth Sciences vol 63 no 4 pp 709ndash721 2011

[11] N Attandoh S M Yidana A Abdul-Samed P A Sakyi BBanoeng-Yakubo and P M Nude ldquoConceptualization ofthe hydrogeological system of some sedimentary aquifers inSavelugu-Nanton and surrounding areas Northern GhanardquoHydrological Processes vol 27 pp 1664ndash1676 2013

[12] M T Pavelko Estimates of Hydraulic Properties from a One-Di-mensional Numerical Model of Vertical Aquifer-System Defor-mation Lorenzi Site LasVegas NevadaWater-Resources Inves-tigations Report 03-4083 US Geological Survey 2004

[13] K A Dickson and G Benneh A New Geography of GhanaLongman London UK 1995

[14] Government of Ghana Brong-Ahafo 2013 httpwwwghanagovghindexphpindexphpItemid=187

[15] B Banoeng-Yakubo S M Yidana J O Ajayi Y Loh and DAsiedu ldquoHydrogeology and groundwater resources of Ghana a

8 The Scientific World Journal

review of the hydrogeology and hydrochemistry of Ghanardquo inPotableWater and Sanitation J MMcMann Ed vol 142 NovaScience New York NY USA 2011

[16] G O Kesse The Mineral and Rocks Resources of Ghana A ABalkema Publishers Rotterdam The Netherlands 1985

[17] N R Junner and T HirstThe Geology and HydroGeology of theVolta Basin Memoir 8 Gold Coast Geological SurveyTheGoldCoast Australia 1946

[18] M A Carrier R Lefebvre J Racicot and E B Asare ldquoGround-water recharge assessment in Northern Ghana using soil mois-ture balance and chloride mass balancerdquo GeoEdmonton vol 8pp 1437ndash1444 2008

[19] P A Domenico and F W Schwartz Physical and Chemical Hy-drogeology John Wiley amp Sons New York NY USA 1990

[20] C W Fetter Applied Hydrogeology Prentice Hall EnglewoodCliffs NJ USA 4th edition 2001

[21] C R Fitts Groundwater Science Academic Press New YorkNY USA 2002

[22] H J Lin D R Richards and C A Talbot FEMWATER AThree-Dimensional Finite Element Computer Model for Simulat-ingDensity-Dependent Flow andTransport inVariably SaturatedMedia US Army Corps of Engineers 1997

[23] A W Harbaugh E R Banta and M G McDonald MOD-FLOW-2000 the US Geological Survey Modular Ground-WaterModel-User Guide to Modularization Concepts and the Ground-Water Flow Process Open-File Report 00-92 US GeologicalSurvey 2000

[24] Aquaveo Groundwater Modeling System Version 71 AquaveoProvo Utah USA 2008

[25] M P Anderson andWWWoessnerApplied Groundwater Mo-deling Simulation of Flow and Advective Transport AcademicPress New York NY USA 2002

[26] S M Yidana and L P Chegbeleh ldquoThe hydraulic conductivityfield and groundwater flow in the unconfined aquifer system ofthe Keta Strip Ghanardquo Journal of African Earth Sciences vol 86pp 45ndash52 2013

[27] M C Hill A Computer Program (MODFLOWP) for EstimatingParameters of a Transient Three-Dimensional Ground-WaterFlow Model Using Nonlinear Regression Open-File Report 91-484 US Geological Survey 1992

[28] M C HillMethods and Guidelines for Effective Model Calibra-tion Water Resources Investigations Report 98-4005 US Geo-logical Survey 1998

[29] M C Hill E R Banta A W Harbaugh and E R AndermanMODFLOW-2000 the US Geological Survey Modular Ground-Water Model-User Guide to the Observation Sensitivity andParameter-Estimation Processes and Three Post-Processing Pro-grams Open-File Report 00-184 US Geological Survey 2000

[30] D W Pollock Userrsquos Guide for MODPATHMODPATHPLOTVersion 3 A Particle Tracking Post-Processing Package for MOD-FLOW theUSGeological Survey Finite-DifferenceGround-WaterFlow Model Open-File Report 94-464 US Geological Survey1994

[31] M Razack andD Huntley ldquoAssessing transmissivity from spec-ific capacity in a large and heterogeneous alluvial aquiferrdquo Gro-und Water vol 29 no 6 pp 856ndash861 1991

[32] D Huntley R Nommensen and D Steffey ldquoThe use of specificcapacity to assess transmissivity in fractured-rock aquifersrdquoGround Water vol 30 no 3 pp 396ndash402 1992

[33] R E Mace ldquoDetermination of transmissivity from specific cap-acity tests in a Karst Aquiferrdquo Ground Water vol 35 no 5 pp738ndash742 1997

[34] S M Yidana D Ophori and B Banoeng-Yakubo ldquoGroundwa-ter availability in the shallow aquifers of the southern voltaiansystem a simulation and chemical analysisrdquo EnvironmentalGeology vol 55 no 8 pp 1647ndash1657 2008

[35] S Y Achcampong and JWHess ldquoHydrogeologic and hydroch-emical framework of the shallow groundwater system in thesouthern Voltaian Sedimentary Basin Ghanardquo HydrogeologyJournal vol 6 no 4 pp 527ndash537 1998