Hydrogeology and groundwater flow in a basalt-capped Mesozoic sedimentary series of the Ethiopian highlands

1 23

Hydrogeology JournalOfficial Journal of theInternational Association ofHydrogeologists ISSN 1431-2174Volume 19Number 3 Hydrogeol J (2011) 19:641-650DOI 10.1007/s10040-010-0667-0

Hydrogeology and groundwater flow ina basalt-capped Mesozoic sedimentaryseries of the Ethiopian highlands

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag. This e-offprint is for personal use only

and shall not be self-archived in electronic

repositories. If you wish to self-archive your

work, please use the accepted author’s

version for posting to your own website or

your institution’s repository. You may further

deposit the accepted author’s version on a

funder’s repository at a funder’s request,

provided it is not made publicly available until

12 months after publication.

Hydrogeology and groundwater flow in a basalt-capped Mesozoicsedimentary series of the Ethiopian highlands

Ine Vandecasteele & Jan Nyssen & Wim Clymans &

Jan Moeyersons & Kristine Martens & Marc Van Camp &

Tesfamichael Gebreyohannes & Florimond Desmedt &Jozef Deckers & Kristine Walraevens

Abstract A hydrogeological study was undertaken in theZenako-Argaka catchment, near Hagere Selam in Tigray,northern Ethiopia, during the rainy season of 2006. Ageological map was produced through geophysical meas-urements and field observations, and a fracture zoneidentified in the north west of the catchment. A perchedwater table was found within the Trap Basalt series abovethe laterized upper Aram Aradam Sandstones. A map ofthis water table was compiled. Water-level variationduring the measurement period was at least 4.5m.Variation in basal flow for the whole catchment for themeasurement period was between 12 and 276m3/day. A

groundwater flow model was produced using VisualMODFLOW, indicating the general direction of flow tobe towards the south, and illustrating that the waterwayshave only a limited influence on groundwater flow. Thesoil water budget was calculated for the period 1995–2006, which showed the important influence of thedistribution of rainfall in time. Although Hagere Selamreceived some 724mm of rainfall per year over this period,the strong seasonal variation in rainfall meant there was awater deficit for on average 10months per year.

Keywords Hydrogeology . Groundwater recharge/waterbudget . Groundwater flow . Ethiopia

Introduction

This study aims at providing a more extensive under-standing of the hydrogeology and hydrodynamics of atypical catchment in the Ethiopian Highlands. NorthernEthiopia is a region which is prone to land degradationand the population has frequently suffered from faminesresulting from poor crop yields in the past. Fieldwork wascarried out in the framework of the Integrated MayZegzeg Watershed Management Project, supported bythe Flemish Interuniversity Council. The project aims atdeveloping effective soil and water conservation techni-ques experimentally within the May Zegzeg catchment,which are applicable to the wider region (Nyssen 2001).Specific aims of the field study were geological mapping,calculation of the water balance, mapping of the watertable and modelling of the groundwater flow. A networkof piezometers was installed in the catchment to allowwater level measurement over the 3-month study periodduring the summer of 2006 (Vandecasteele 2007). Springdischarges were measured as well as discharge at aconcrete dam at the lower limit of the catchment.Geoelectrical measurements were carried out to identifyfractures and support the geological interpretations made.A detailed understanding of the hydrological resourcesavailable is necessary to aid rehabilitation of degradedland in the region through appropriate soil and waterconservation techniques.

Received: 31 August 2009 /Accepted: 13 October 2010Published online: 3 December 2010

* Springer-Verlag 2010

I. Vandecasteele ()) : J. MoeyersonsRoyal Museum for Central Africa,Leuvensesteenweg 13, 3080, Tervuren, Belgiume-mail: [email protected]

I. Vandecasteele :K. Martens :M. Van Camp :K. WalraevensLaboratory for Applied Geology and Hydrogeology,Ghent University,Krijgslaan 281-S8, 9000 Gent, Belgium

J. NyssenDepartment of Geography,Ghent University,Krijgslaan 281-S8, 9000 Gent, Belgium

W. ClymansPhysical and Regional Geography Research Group, KU Leuven,Redingenstraat 16, 3000 Leuven, Belgium

T. GebreyohannesMekelle University, PO Box 231, Mekelle, Ethiopia

F. DesmedtHydrology Department,Free University of Brussels,Pleinlaan 2, 1050 Brussels, Belgium

J. DeckersDepartment of Earth and Environmental Sciences,Division for Land and Water Management, KU Leuven,Celestijnenlaan 200E, 3001 Leuven, Belgium

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

Study areaThe study was carried out in the Zenako-Argaka catch-ment, covering an area of about 4 km2 just east of HagereSelam in the highlands of Tigray, northern Ethiopia(Fig. 1). The catchment has an accentuated topography(Fig. 2), its highest and lowest points lying at respectively2,600 and 2,200 m above sea level (a.s.l.). The MayZegzeg waterway draining the catchment flows into theGeba river system. The main wet season falls betweenJune and September, with some smaller events in March–April. The average annual rainfall for the period 1995–2006 was 724 mm.

GeologyFigure 2 shows the lithologies present in the catchment:the Antalo Limestone, Amba-Aradam Sandstone and theTrap Series Basalts. The Antalo limestone sequence maybe up to 700 m thick, though in the catchment only the top190 m is exposed. These Mesozoic parasequences ofmassive white and pale yellow marine limestone and marlare presumably part of the sub-unit Jtd, as described byArkin et al. (1971). Based on the classification of Boselliniet al. (1997), the sub-units A3 and A4 could be identified(see Fig. 3). The overlying Amba-Aradam formationconsists of alternating banks of shale and coarse-grainedsandstone. The topmost layers show cross bedding,suggesting a fluviatile origin. There is evidence of contactmetamorphism and sub aerial alteration in the upper partof the formation, causing a lateritization and compaction.This is most likely due to intensive weathering and

deposition of the flood basalts directly over these sand-stones. The formation has an approximate thickness of50 m in the study area.

Tertiary volcanics in the study area have beendeposited as flood basalts, and were found to be up to110 m thick in the catchment. The erosion of thisformation has formed highly fertile soils (vertisols,approximately 7 m thick in the study area). Lacustrinesediments, also of Tertiary age, are found interbeddedbetween the volcanic series (e.g. at 4 m depth in boreholeP2). These sediments also form a hill 25 m high just to thenorth of the main road at a height of approximately2,600 ma.s.l.

Materials and methods

Geological fieldworkA geological map of the catchment was made based on aliterature study (Mohr 1962; Bosellini et al. 1997) anddetailed vertical sections observed and described in theeastern and western parts of the study area. Lithologicalcontacts, main waterways, landslides, pathways and thehydrological boundary of the catchment were locatedusing a handheld global positioning system (GPS), and themap (Fig. 3) was plotted using ArcView®.

Geoelectrical prospectionElectrical profiling allows detection of lateral changes in theresistivity of the subsoil, so that the presence of fracture

Fig. 1 The study area, located within the province of Tigray, in northern Ethiopia (adapted from African Studies Center 2007)

642

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

zones and the extent of saturation of the ground can bedetermined (Kirsch 2006). The presence of water in aformation lowers the resistivity of the rock, so that profilingcan also be used to give an indication of water-table depth.Electrical sounding can give an indication of the depth atwhich geological contacts occur. Geoelectrical measure-ments were taken using a Terrameter SAS (ABEM 2009).The Wenner four-pin configuration (Wenner 1916) was usedfor the profiles, in which the distance between electrodes was10m, and the estimated penetration depth 5m. Three verticalsoundings were taken using the Schlumberger configuration(Ernstson and Kirsch 2006); two in the upper and one in thelower part of the catchment. The locations of all measure-ments are shown on Fig. 3.

Precipitation measurementsThe precipitation data discussed here were extrapolatedbased on the Thiessen Polygon method (Croley andHartmann 1985) from daily rainfall measurements recordedby the four pluviometers located on Fig. 3 (MU-IUC 2007).

Water balance approachFigure 4 gives an overview of the hydrological cycle as itoccurs in the May Zegzeg catchment. The estimation ofgroundwater recharge is based on the soi-water budgetmethod, in which storage is limited to the water held inthe soil; precipitation replenishes it and evapotranspirationdepletes it. In this manner, a model can be developed todescribe the soil-water balance based on precipitation,potential evapotranspiration (PET), and the water holdingcapacity (CAP) of the soil. The monthly PET in thecatchment was estimated based on the Thornthwaitemethod (Thornthwaite and Mather 1955). PET values are

extrapolated from monthly temperatures; the PET calcu-lated is representative of a month with 30 days and 1 h ofsunlight daily. In WATBUG (Wilmott 1977), the programused to carry out the calculations, a value for the latitude(14°N) was also given so that the number of hours ofsunlight per day was taken into account. The waterholding capacity of the soils was also entered in themodel—value of 16 cm (160 mm) was used, based onglobal soil mapping (Dunne and Willmott 1996). Amaximal estimated value for surface runoff of 8% wastaken into account. This was calculated by Clymans(2007) using the formula:

Runoff Coefficient RCð Þ ¼ 1000: Vtot=Atotð Þ=PAwhere Vtot = total volume of water leaving the catchment,Atot = total area of the catchment, and PA = precipitationfalling within the catchment.

Hydrological fieldworkA network of piezometers was installed in the upper partof the catchment to monitor the water-table level(Vandecasteele 2007). Five piezometers (A–E) wereinstalled using a hand bore, and have a maximal depthof 2 m. A further three piezometers (P1–P3) were installedwith a mechanical bore, up to a maximal depth of 4.7 m.Water levels were measured with a dip meter every 3 days.An inventory was made of hand-dug wells and springs.Discharges at the springs and the total catchmentdischarge, taken as the discharge at the cement damsituated at the lower limit of the catchment, weremeasured on a weekly basis. The positions of allpiezometers, wells and springs are given in Fig. 3.Measurements of the relative piezometer locations weremade with a laser theodolite total station. This was

Fig. 2 The eastern side of the May Zegzeg catchment, showing the lithological succession

643

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

necessary due to the low accuracy (an error of about 10 m)of the GPS measurements.

Groundwater flow modellingA representation of the groundwater flow for the wholecatchment was worked out using Visual MODFLOW(WHL 2005). The hydrological boundary (watershed) ofthe catchment used in the model was defined based on thetopographical map of the region (1:50,000, with contourline spacing of 20 m). The northern, eastern and westernlimits of the model lie on the watershed; the southern limitwas taken where the catchment borders the adjacentwaterway. The points registered were set out in ArcViewand could be overlaid on the model grid to define theactive cells within the catchment. The grid is comprised of100 cells in a N–S direction, and 87 cells in an E–Wdirection, with each cell having an area of 20×20 m (socovering a total area of 2,000 m×1,740 m). The model

was divided into three layers with varying hydrologicalproperties which correspond to the Trap Basalts aquifer(120 m thick), Amba Aradam Sandstone aquitard (50 m),and Antalo Limestone aquifer (190 m). The permanentsprings and main waterways in the catchment wereincluded as drains.

Results and discussion

Geologic and stratigraphic setting

Geological interpretation of the vertical electricalsounding (VES)Soundings VES 1 and VES 2 were taken in the upper partof the catchment, so reflect the properties of the basaltsand underlying sandstone. VES 3 was carried out in thelower part of the catchment on the Antalo Limestones, sothat variations in resistivity will be due to the varying

Fig. 3 Geological map of the May Zegzeg catchment. The sub-division of the Antalo Formation into units A3 and A4 is represented by ablack dashed line. The profiles along which lithologs were taken on both sides of the catchment are indicated

644

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

proportions of limestone and marl within the formation.For all interpretations it has been assumed that thetopmost soils (within the unsaturated zone) have thegreatest resistivity, that saturation with water will lowerthe resistivity, and that the sandstones have a greaterresistivity than both the limestone and basalt. Aninterpretation curve (Fig. 5) was fitted around themeasurements using VES interpretation software (DeBreuck et al. 1980). An example of the results obtainedis given in Table 1 for sounding 2. The vertisols are foundto have a thickness of 7.34 m, overlying 15 m of basaltsand an approximate 45 m of sandstone. The limestone canalso be differentiated here, though their thickness cannotbe defined.

FaultingA branch of the East African Rift Valley runs throughEthiopia; its major fault lines run in a N–S direction andform the escarpment at the edge of the Rift Valley and theDanakil Depression. Fault belts related to this largersystem are found in the region, and run on an echelon ina WNW direction (Chernet and Eshete 1982). In the MayZegzeg catchment, a difference in elevation of anestimated 19 m between the east and west sides of thecatchment (the western side lying higher), the location ofspring 3, as well as the presence of a large landslidesuggest there may be a fracture zone present. Lithologsmeasured and described in the eastern and western sidesof the catchment (located on Fig. 3) also showedsignificant lateral variation in lithology. Geo-electricalprofiling was carried out to investigate this hypothesis.Profile 1 gave much lower resistivity values (0.2–21.8ohm.m) than Profile 2 (4.6–134.6 ohm.m), due to thegreater thickness of the basalt higher up in the catchment.If fracture zones were present, they would be represented

by relatively lower resistivity values due to the relativelygreater proportion of water infiltrating along these joints.Figure 6 indicates low resistivity values encounteredin both profiles and gives a possible structural interpreta-tion—that of a multiple fracture zone. The fracture zonemay be made up of parallel, shallow fault lines with anapproximate NNW–SSE orientation.

LandslidesLandslides off the edge of the Amba Aradam Sandstonecliff into the valley are common in the catchment.Especially the lacustrine deposits are known to be proneto sliding (Nyssen et al. 2002). According to Moeyersonset al. (2006), these landslides are indicative of formerlywetter conditions (more developed perched water tables inthe area during the late Pleistocene to middle Holocene),and are mostly no longer active. The water table is nowpresumed to be too low for the presence of surplus waterto remain the driving force for slump movement. How-ever, the May Ntebteb debris flow has been found to creepdown slope at a rate of 3–6 cm/year (Nyssen et al. 2002),and was reactivated an estimated 70 years ago. This

Fig. 5 Graphical representation of the measured resistivity (p) andthe interpretation curve developed using VES interpretation soft-ware for sounding 2

Fig. 4 The hydrological cycle represented in a schematic cross-section of the May Zegzeg catchment. Abbreviations used are basalt (B),sandstone (Sst) and limestone (Lst) for the lithologies, and ΔS for the change in soil water storage. Not to scale

645

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

reactivation of movement is thought to be due to gullyincision over the last hundred years, and more recentreactivation may be due to the prohibition of grazing andwoodcutting in the catchment as of 8 years ago (Nyssen etal. 2010), which has increased infiltration rates on thelobe. The original trigger for the movement of the MayNtebteb debris flow may be linked to the underlyingfracture zone, which resulted in instability of the collu-vium. Smaller landslides have occurred in gullies on thecliffside of the catchment and are probably due to highsurface runoff during the rainy season.

Water balance of the unsaturated zoneThe total yearly values for potential evapotranspiration(PET), actual evapotranspiration (AET), precipitation,water deficit and surplus are given in Table 2; monthlyAET, precipitation and water surplus values are repre-sented in Fig. 7. Water surpluses occur mostly aroundAugust, their magnitude depending on the quantity anddistribution of rainfall and variations in temperature. Themaximal deficit usually occurs in February or March. Thevalues calculated for water surplus show a large variationfrom year to year, and reflect the unreliability of rainfall inthe region. The highest surplus (395 mm) as well as totalprecipitation (962 mm) was in 1998. 2002 and 2003 wereyears of widespread drought in northern Ethiopia, and

show low water surpluses of 42 and 27 mm respectively.The year 2004 had the lowest total precipitation(530 mm), but had a higher surplus value than both2002 and 2003 (79 mm). This can be related to the spreadof precipitation over the year; 2002 and 2004 have thesame PET values, but the lower surplus in 2002 can beattributed to higher total AET values reflecting the largerspread of precipitation over that year. 2004 shows a highersurplus than expected, as it has the lowest calculated AETdue to concentration of rainfall in a shorter period.

Piezometric water level and its evolutionThe piezometric measurements are represented in Fig. 8.Rising water levels can be related to the infiltration ofrainfall, adding to the groundwater store. Water levels fallsignificantly if there are consecutive days with little or norainfall. An accumulation of moisture in the ground overthe rainy season leads to a peak in water level towards theend of the rainy season, during the second week ofSeptember, in the higher parts of the plateau (P2 and P3).The water table is locally shallower in low-lying fieldswith relatively gentler gradients as was the case for thelocation of piezometers C and A. It is inferred that thewater table will show greater yearly fluctuations higher upin the basalts. In the lower part variations were oftenwithin 2 m, whereas the higher-lying piezometers showeda variation of 4.5 m (P3) during the rainy season.

The calculated discharges at the cement dam (seeFig. 3) are represented graphically in Fig. 9. The dischargeis seen to increase systematically from a minimum of12.33 m3/day at the beginning of the rainy season to amaximum of 276.48 m3/day at the end of measurements.As measurements were taken weekly, little variation isseen with individual rainfall events, though they influencethe rate of increase in discharge. The rising trend reflectsthe accumulation of water in the catchment, and anincrease in sub-surface flow over time. Average springdischarges measured were 0.13 m3/day for spring 3, and5.63 m3/day for spring 4. This value for spring 3 is merelya fraction of its actual total discharge, which occurs in amore diffuse way, and which is estimated to be in the

Fig. 6 A possible interpretation of the fracture zone, based onmeasured anomalies in the two profiles

Table 2 Yearly totals for potential evapotranspiration (PET), prec-ipitation (P), actual evapotranspiration (AET), deficit and surplus, allgiven in mm

Year Precipitation PET AET Deficit Surplus

1995 833 724 444 278 3421996 900 700 537 163 2781997 776 736 609 127 771998 962 783 521 261 3951999 821 746 458 289 2882000 678 761 514 247 1112001 679 756 483 273 1682002 577 783 481 300 422003 548 788 484 301 272004 530 787 414 371 792005 725 725 467 258 1802006 661 763 580 183 12Average 724 754 499 254 167

Table 1 VES sounding 2, giving resistivity, thickness and an int-erpretation of the distinguishable layers

Resistivity (Ωm) Thickness (m) Description

10,000 0.34 Compacted vertisol, dry3 1.00 Loose vertisol, saturated35 6.00 Vertisol, saturated30 15.00 Consolidated basalt500 45.00 Sandstone40 - Limestone

646

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

same range as that measured for spring 4. Both springs areimportant permanent sources of water in the catchment.

The water levels measured on a single day can beused to give, through interpolation, a map of theelevation of the water table, and by extrapolation thegroundwater flow within the Trap Basalts. Figure 10shows the map of the absolute level of the water tablefor the section of the Trap Basalts in which piezometerswere placed, based on piezometric measurements. Thelevel represented is actually a perched water table, andthe regional water table is assumed to lie some 200 mfurther down in the limestone sequence. The water flowis principally towards the permanent springs S3 and S4.The depth of the water table will decrease withdecreasing elevation of the ground surface in thedirection of groundwater flow.

Groundwater flow and hydrogeologicalcharacteristics of the formationsThe study area forms part of the Tekeze drainage basin.The groundwater regime in the area may be locallyperched above geological formations less permeable towater, as is the case in this catchment, but generally wateris found to percolate down to the lower Antalo Limestone

and Adigrat Sandstone. This conceptual model of ground-water flow is shown in Fig. 4. The exposed AntaloLimestone and the Trap Basalts form areas which allowdirect recharge. Streams are intermittent—the waterwayswithin the study area remain dry much of the year, andflashfloods are often experienced during the rainy seasondue to high surface runoff. Permeability of the grounddiffers greatly according to the underlying geology, and sowill determine to a great extent the amount of rainwaterinfiltrated as well as the amount of groundwater perma-nently or temporarily stored.

The basalt series forms an unconfined aquifer abovethe Amba Aradam Sandstone aquitard. Intercalated flu-vial/lacustrine sediments may also act as aquitards,forming locally perched water tables. Water is drained ina N–S direction towards the contact with the AmbaAradam Sandstones. Two permanent springs are present,and a multitude of temporary springs emerge at the base ofthis layer during the rainy season. The basalts themselveshave a low permeability, though the overlying vertisolshave a high water-holding capacity, and are morepermeable. However, this permeability can be greatlyreduced by compaction, which may lead to high surfacerunoff along pathways and roads.

Fig. 7 Monthly values for precipitation, actual evapotranspiration (AET) and surplus (in mm) over the period 1998–2006

Fig. 8 Water levels (depth to the water table in cm) measured for the period of fieldwork, 2006

647

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

The uppermost sediments of the Amba Aradam Sand-stone are highly compacted and have a low permeability.Water is most likely drained out of the layer along internalfractures and joints. Permeability varies within theformation, according to the grain size, weathering andjointing of the rocks. The rocks are cliff-forming, and thesteep gradient means that much of the water is transporteddownstream by surface runoff rather than percolation andgroundwater flow (Nyssen 2001). The Antalo Limestonebeds are cliff-forming and allow little or no primaryinfiltration. The highly jointed and weathered nature of therocks (Chernet and Eshete 1982) does, however, givethem a good secondary permeability, and the formation isan aquifer. Temporary local springs were found at thecontact between these massive limestone blocks and theintercalated finer marl beds, which act as aquitards.Groundwater flow is roughly N–S towards the southernlimit of the catchment.

The groundwater flow model considers the wholeArgako-Zenako catchment as a closed basin. The springsand main waterways were introduced as drains, and wereassigned an estimated hydraulic conductance of 10 m2/day, based on the measured discharges. The value used forthe recharge is 167 mm; this is the average value of thesurplus calculated over the period 1998–2006. Estimated

hydraulic conductivities for each layer are given inTable 3. The model is quasi-3D; in aquifer layers, onlyhorizontal flow components (Kh) are considered, while inaquitard layers, only vertical flow components (Kv) areconsidered. The hydraulic conductivity of the upperaquifer in the Trap Basalts has been estimated bycomparing calculated and observed hydraulic gradients.Sensitivity analysis has shown that the hydraulic resist-ance of the Amba Aradam Sandstone must be at least 2×106 days in order to keep water in the upper aquifer of theTrap Basalts, and for the lower aquifer in the AntaloLimestone, a low transmissivity value KhD (as the productof horizontal hydraulic conductivity and thickness) of10 m²/day was used, based on the practical absence ofprimary permeability. The groundwater flow model for theTrap Basalts and the Antalo Limestone is given in Fig. 11.Figure 12 compares measured piezometric levels to thosecalculated by the model for the vertisols/Trap Basalts.There is a reasonable relationship, which serves as avalidation of the model.

The groundwater balance of the model shows that 30%of the recharge is drained by the springs or discharges insurface water on the basalt plateau. The remaining 70%percolates through the sandstones into the limestonebasement and is discharged into the May ZegZeg water-way that forms the southern limit of the model area.

Conclusion

A geological map of the catchment was established(Fig. 3) based on a literature review and terrain observa-

Fig. 9 Total discharge measured at the cement dam for the period of fieldwork, 2006

Fig. 10 The interpolated level (in m a.s.l) of the water table in thevertisols/Trap Basalts based on measured piezometric water levelmeasurements for 11 and 14 August 2006

Table 3 Hydraulic parameters used in the groundwater flow model

Lithology KhD(m²/day)

Kh(m/day)

Kv(m/day)

Aquifer 1 Trap Basalts - 0.025 -Aquitard Amba Aradam

Sandstone- - 2.5×10–5

Aquifer 2 Antalo Limestone 10 - -

Kh horizontal hydraulic conductivity; Kv vertical hydraulicconductivity; D aquifer thickness (m)

648

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

tions. Landslides present in the catchment were alsomapped. Geo-electrical measurements allowed estimationof the thickness of the vertisols and consolidated basalts at7.34 and 15 m respectively, and identification of apossible multiple fracture zone in the northwestern partof the catchment.

The soil-water budget calculated for the period 1995 to2006 showed that water surpluses and groundwaterrecharge occurred mostly in August, and their values (onaverage 167 mm/year) were highly variable depending onthe temporal distribution of rainfall throughout the year. Ifrainfall is highly concentrated during a rainy season watersurpluses will in turn be much higher. Piezometersinstalled in the vertisols/Trap Basalts showed that meas-ured water levels could be closely related to the quantityand temporal distribution of rainfall in the catchment, withsome variation according to the specific location of thepiezometer. An accumulation of infiltrated water caused arise in all water levels measured during the course of therainy season to a peak in September. A map of this

Fig. 11 Calculated piezometric levels in the vertisols/Trap Basalts and Antalo Limestone. The area for which the water table was mappedbased on measured piezometric levels in the north of the catchment is indicated by the box

Fig. 12 A comparison of measured and calculated piezometricwater levels in the Trap Basalts

649

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy

perched water table was established, and the generaldirection of flow was seen to be towards the south.

Water flow in the catchment follows the topography,with the Trap Basalts forming an unconfined aquiferabove the much less permeable Amba Aradam Sand-stones. Water is retained in the basaltic series and theirweathered products to form a perched water table. Waterflow through the sandstone and limestone series is mostlikely along internal bedding planes and joints. There is areasonably good relationship between the measured waterlevels, and those modeled in the vertisols/Trap Basalts.

References

ABEM (2009) Instruction manual Terrameter SAS 1000/40000,ABEM printed matter no. 93109. www.abem.se/files/upload/Manual_Terrameter.pdf. Cited 15 June 2006

African Studies Center(2007) Country map of Ethiopia. Universityof Pennsylvania, Philadelphia, PA. www.africa.upenn.edu. Cited5 March 2007

Arkin Y, Beyth M, Dow DB, Levitte M, Haile T, Tsegaye H (1971)Geological map of Mekelle sheet ND 37-11 Tigre Province.Ministry of Mines, Energy & Water Resources, GeologicalSurvey of Ethiopia, Addis Ababa

Bosellini A, Russo A, Assefa G, Solomon T (1997) The Mesozoicsuccession of the Mekele Outlier, Tigre Province, Ethiopia.Mem Sci Geol 49:95–116

Chernet T, Eshete G (1982) Hydrogeology of the Mekelle area(ND37-11). Ministry of Mines and Energy Memoir 2, Ministryof Mines, Addis Ababa, 50 pp

Clymans W (2007) Impact of soil and water conservation measureson the hydrology and sediment budget of a subtropical catch-ment May Zegzeg (Northern Ethiopia). MSc Thesis, KULeuven, Belgium

Croley TE, Hartmann HC (1985) Resolving Thiessen polygons. JHydrol 76(3–4):363–379

De Breuck W, Beeuwsaert E, Van Camp M, Vandenheede J (1980)VES interpretation program manual. Department of AppliedGeology and Hydrogeology, Ghent University, Ghent, Belgium

Dunne KA, Willmott CJ (1996) Global distribution of plant-extractable water capacity of soil (Dunne). Int J Climatol 16

(8):841-859. Data set available online. http://www.daac.ornl.gov. Cited 20 October 2006

Ernstson K, Kirsch R (2006) Geoelectrical methods, basic princi-ples. In: Kirsch R (ed) Groundwater geophysics: a tool forhydrology. Springer, Heidelberg, pp 85–108

Kirsch R (2006) Groundwater geophysics: a tool for hydrogeology.Springer, Berlin

Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M (2006) Ageand backfill/overfill stratigraphy of two tufa dams, TigrayHighlands, Ethiopia: evidence for late Pleistocene and Holocenewet conditions. Paleogeography 230:165–181

Mohr PA (1962) The geology of Ethiopia. “Il Poligrafico” Asmara,University College of Addis Ababa Press, Addis Ababa

MU-IUC (2007) Digital database of climatological and stream flowdata of Geba catchment, obtained from National MeteorologicalServices Agency and Ministry of Water Resources. VLIR(Belgium), Mekelle University Institutional University Cooper-ation Programme, Mekelle, Ethiopia and Leuven, Belgium

Nyssen J (2001) Erosion processes and soil conservation in atropical mountain catchment under threat of anthropogenicdesertification: a case study from Northern Ethiopia. PhDThesis, KU Leuven, Belgium

Nyssen J, Moeyersons J, Poesen J, Deckers J, Haile M (2002) Theenvironmental significance of the remobilisation of ancient massmovements in the Atbara-Tekeze headwaters, northern Ethiopia.Geomorphology 49:303–322

Nyssen J, Clymans W, Descheemaeker K, Poesen J, VandecasteeleI, Vanmaercke M, Zenebe A, Van Camp M, Haile M,Haregeweyn N, Moeyersons J., Martens K, Gebreyohannes T,Deckers J, Walraevens K (2010) Impact of soil and waterconservation measures on catchment hydrological response: acase in north Ethiopia. Hydrol Proc 24(13):1880–1895

Thornthwaite CW, Mather JR (1955) The water balance. Publica-tions in Climatology, vol 8. Drexel Institute of Climatology,Centerton, NJ, pp 1–104

Vandecasteele I (2007) Hydrogeological investigation in the area ofMay Zegzeg, near Hagere Selam (Tigray, northern Ethiopia),MSc Thesis, Ghent University, Belgium

Wenner F (1916) A method of measuring earth resistivity. USBureau Standards Bull 12:469–478

WHL (2005) Visual MODFLOW, user’s manual, edn. 1. The industrystandard for 3D groundwater flow and contaminant transportmodeling. Waterloo Hydrogeologic Inc, Waterloo, ON, 611 pp

Wilmott CJ (1977) A Fortan IV Algorithm for calculating theClimatic Water Budget, Publications in Climatology, vol XXX,no. 2. Thorntwaite Assoc. Lab. for Climatology, Elmer, NJ

650

Hydrogeology Journal (2011) 19: 641–650 DOI 10.1007/s10040-010-0667-0

Author's personal copy