This study provides new evidence on the effectiveness of a combined remote sensing and hydrogeological investigation method for deep groundwater development in complex geologic and geomorphologic situations in Ethiopia. The study was undertaken in the Elidar district of the Afar region of Northern Ethiopia. Due to the water availability and vegetation in this district, the majority of the population (total population of 79,000 people and 647,000 head of cattle) are dependent on pastoralist livelihoods. The current ratio of available water supply facilities to number of beneficiaries is 1:2,323 people and 1:19,029 head of cattle. A joint UNICEF-UNESCO groundwater investigation pilot project with the objective of improving drilling success rates in the Elidar district considered a three-phase approach. In phase 1 and 2, an overlay method was developed that combined data from radar, optical remote sensing and ground measurement (geology, hydrogeology, hydrology, hydro-meteorology, and geophysics). The overlay model identified the most promising site to undertake well drilling, considering a probability of drilling wells with sufficient amount of water and permissible water quality (defined as Q = >2 L/s and EC < 2,000 μms/cm²). In phase 3, three production boreholes were drilled and the results showed a 92% accuracy against the overlay model.

INTRODUCTION

Ethiopia is renowned for being the ‘water tower of Africa’ due to the presence of three principal drainage systems feeding the Blue and White Nile, which originate in the Ethiopian central highlands (Arsano & Tamrat 2005; Teklu et al. 2015). Despite this acknowledgement, there are limited data on Ethiopian hydrogeology. Deyassa et al. (2014) state that a comprehensive model that explains the hydrogeological set-up and genesis of basement aquifers in Ethiopia does not exist in the international literature (Cherenet 1993; Deyassa et al. 2014). This is due to the diversity of aquifer systems, which are described in detail in Kebede (2012). The most widely used map is at a scale of 1:2,000,000 (EIGS 1993, 1996). The major reason for the limited data relates to the vast landmass, complex geological variations and inaccessibility of lowland areas due to political insecurity. Ayenew et al. (2008) characterized the physiology of Ethiopia as a ‘high-altitude volcanic plateau tapering into rift valley and peripheral lowlands’ (Ayenew et al. 2008, p. 97). This is depicted in Figure 1.
Figure 1

Location map showing the volcanic regions and major river basins (Ayenew et al. 2008).

Figure 1

Location map showing the volcanic regions and major river basins (Ayenew et al. 2008).

The most geologically complex zone of the country is the north-eastern Afar region, which is characterized by a highly variant hydrogeology and physiology. The topography of Afar ranges from the upper Rift Valley at an altitude of more than 2,000 m above sea level to the Danakil Depression, which is 120 m below sea level. In between, there are vast areas of the Melka–Sedi–Amibara zone in the Middle Awash valley at 750 m above sea level (Ayenew 2007). Salinization of the groundwater aquifers is a major limitation to effective groundwater development in the Afar region, with the most affected area being the arid southern and eastern Afar region (Ayenew 2007). The study area for this paper is the Elidar district (located in the east of the Afar region). Elidar is characterized by grabens (100–200 m above sea level) with interspersed undulating forests (400 m above sea level). These physiological features are interspersed with acidic volcanic centres. The region experiences low rainfall, high evapotranspiration and average temperatures >40°C. Due to the vegetation in this district, the majority of the population (79,000 people and 647,000 head of cattle) are dependent on pastoralist livelihoods. The current ratio of water supplies is 1:2323 people and 1:19029 head of cattle.

Literature on the hydrogeology of Elidar has been limited to conventional resistivity techniques for shallow groundwater identification similar to those used by Arvidsson et al. (2011). Limited published investigations have been carried out with regard to the groundwater resources of Elidar, and they are predominantly related to the Afar depression and its inference on the groundwater potential of the Djibouti basin (Jalludin & Razack 2004). These conventional shallow groundwater investigation techniques have proved ineffective in locating groundwater in Elidar due to the depth of the groundwater occurrence. To overcome this limitation, the reviewed literature suggests the use of GIS and remote sensing tools to delineate deep water bearing zones (Dar et al. 2010). Studies by Magesh et al. (2011) note that these tools have wide application in the management of various natural resources, but are highly underutilized for deep groundwater mapping. They further state that a combination of remote sensing and ‘ground trothing’ is an effective way of delineating potential groundwater zones. This is evident in studies by Murthy (2000), Leblanc et al. (2003) and Tweed et al. (2007), who used remote sensing to increase the accuracy at a lower cost and quicker timeframe. In a recent World Bank review of Earth Observation for Water Resources Management (World Bank 2016), it is recommended to use remote sensing to provide fast and reliable information for water-related decision-making. Therefore, this paper presents the methods used from a combined remote sensing and ground trothing overlay method from the Afar region of Ethiopia.

MATERIALS AND METHODS

During phase 1, groundwater potential estimates were calculated by overlaying data from radar and optical remote sensing in maps of 1:100,000 scale. In phase 2, all existing information (geology, hydrogeology, hydrology, hydro-meteorology, geophysics) were gathered and a base map of 1:250,000 was developed. The overlay model developed based on the findings of phase 1 and 2 studies identified the most promising site to undertake well drilling considering a probability of drilling wells with a sufficient amount of water and permissible water quality (defined as Q = >2 L/s and EC < 2,000 μms/cm²). The study was undertaken in the Elidar woreda (district). Elidar woreda (district) is located in the north-eastern region of Afar. It has an area of approximately 13,119 km2, is bounded in the north by Afdera woreda, in the west by Dubti, in the northeast by the international border with Eritrea and in the southeast by Djibouti. To reduce the probability of dry or negative boreholes, this study followed the three phases outlined in Figure 2.
Figure 2

Phases of implementation.

Figure 2

Phases of implementation.

Remote sensing data

In phase 1, the groundwater analysis (GwA) methodology developed by Acacia Water was used to develop the groundwater potential maps (Acacia 2014). The Acacia GwA is designed to give ‘an indication of the areas with the highest potential for fresh groundwater occurrence’ (Acacia 2014). First, data from a range of satellite data sources were compiled. These included from SRTM, MODIS, TRMM, LANDSAT, AMSR-E, SMOS, RADAR-SAT 2, Google Earth and Data Phase 1. The specificity of these data was used at varying depths depending on the strength of the satellite imagery. Table 1 compares the penetration data (resolution) from the different satellites, while definitions of remote sensing data sources and platforms are indicated in Table 2.

Table 1

Remote sensing source and resolution

SourceResolutionData
ASTER VNIR 15 m  
Landsat -5-7-8, ASTER SWIR 30 m Vegetation indices like MODIS, Landsat, can provide an indication of the water availability at the rooting depth, which may stretch to slightly larger depths, but still remain an indication of only the shallowest presence of soil moisture 
SRTM 90 m 
MODIS 250 m 
ENVISAT ASAR 1,000 m  
AMSR-E, SMOS >1,000 m Soil moisture information derived from satellite data (AMSR-E, SMOS and ASAT) provides information about the water availability in the first (tens of) centimetres of the soil 
TRMM, ARC-2 0.25 degree, 0.1 degree (TRMM and ARC-2) a full map and the temporal variance of the precipitation over the area is obtained 
SourceResolutionData
ASTER VNIR 15 m  
Landsat -5-7-8, ASTER SWIR 30 m Vegetation indices like MODIS, Landsat, can provide an indication of the water availability at the rooting depth, which may stretch to slightly larger depths, but still remain an indication of only the shallowest presence of soil moisture 
SRTM 90 m 
MODIS 250 m 
ENVISAT ASAR 1,000 m  
AMSR-E, SMOS >1,000 m Soil moisture information derived from satellite data (AMSR-E, SMOS and ASAT) provides information about the water availability in the first (tens of) centimetres of the soil 
TRMM, ARC-2 0.25 degree, 0.1 degree (TRMM and ARC-2) a full map and the temporal variance of the precipitation over the area is obtained 
Table 2

Definition of remote sensing data sources and platforms

SourceTranslation
ASTER Advanced Space-borne Thermal Emission and Reflection Radiometer 
VNIR Visible and near infrared 
Landsat -5-7-8 Continuous acquisition of satellite imagery of earth (versions 5, 7 and 8) 
SWIR Silicon wafer inspection with laser induced photoluminescence 
SRTM Shuttle Radar Topography Mission 
MODIS Moderate-Resolution Imaging Spectro-radiometer 
ENVISAT Environmental Satellite 
ASAR Advanced Synthetic Aperture Radar 
AMSR-E Advanced Microwave Scanning Radiometer – Earth Observing System 
SMOS Soil Moisture and Ocean Salinity 
TRMM Tropical Rainfall Measuring Mission 
ARC-2 Advanced Room Correction-System 2 
SourceTranslation
ASTER Advanced Space-borne Thermal Emission and Reflection Radiometer 
VNIR Visible and near infrared 
Landsat -5-7-8 Continuous acquisition of satellite imagery of earth (versions 5, 7 and 8) 
SWIR Silicon wafer inspection with laser induced photoluminescence 
SRTM Shuttle Radar Topography Mission 
MODIS Moderate-Resolution Imaging Spectro-radiometer 
ENVISAT Environmental Satellite 
ASAR Advanced Synthetic Aperture Radar 
AMSR-E Advanced Microwave Scanning Radiometer – Earth Observing System 
SMOS Soil Moisture and Ocean Salinity 
TRMM Tropical Rainfall Measuring Mission 
ARC-2 Advanced Room Correction-System 2 

Data from these sources were combined with precipitation, evapotranspiration, soil moisture, geology, geo-morphology, land use and existing well analysis to provide a S-GwA (shallow groundwater analysis) and D-GwA (deep groundwater analysis) conceptual model. These were presented as infiltration zones, accumulation zones and were based on the TRMM system. This system uses precipitation radar (PR), Microwave Imager (TMI) and Visible and Infrared Scanner (VIRS), Clouds and the Earth's Radiant Energy Sensor (CERES) and a Lightning Imaging Sensor (LIS) to estimate the precipitation. This estimate is merged with NOAA, GMS, GOES and Meteosat data to produce merged 3-hourly precipitation datasets in a 0.25 × 0.25 degree grid (dataset 3B42). The ARC-2 system uses inputs from two sources:

  • (1) 3-hourly geostationary infrared (IR) data centred over Africa from the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT);

  • (2) quality-controlled Global Telecommunication System (GTS) gauge observations reporting 24-hour rainfall accumulations over Africa.

Results of the remote sensing analysis are presented in 1:250,000 scale maps. A description of the district's characteristics is outlined in Figure 3. The principal points of population settlement are on the asphalt road from Serdo to Manda and onwards to Bure and the Eritrean border. Elevations vary within the district from 2,000 m above sea level to −272 m below sea level.
Figure 3

Physiographic map of Elidar woreda and location of main population centres.

Figure 3

Physiographic map of Elidar woreda and location of main population centres.

Hydrogeology and geology

In phase 2, UNICEF Ethiopia contracted a consortium led by UNESCO and other private consultancy firms to undertake a field hydrogeological study that included a water point inventory, water quality survey, groundwater recharge estimation and geophysical investigations at the ‘high potential sites’ identified by the remote sensing technology. Resistivity surveys were undertaken using the vertical electrical sounding technique. The WDJD-3 digital Earth resistivity meter was used to undertake the field work with a Schlumberger electrode configuration with a maximum half-current electrode spacing (AB/2) of between 330 m and 500 m. The results were plotted on a log–log plot of apparent resistivity versus half the current electrode spacing. Additionally, 50 water quality samples were collected and analysed to determine the hydrogeochemical characteristics of the district. Also, data were collected on the meteorology, rainfall, precipitation and evapotranspiration. The geology of the Elidar district was mapped using the existing data from the EIGS (1993, 1996) (Figure 4). Unconsolidated sediments were identified along the principal asphalt road where demand from the population was at its highest.
Figure 4

Geological map of Elidar woreda.

Figure 4

Geological map of Elidar woreda.

Water quality samples collected from the 50 sites were plotted using a box and whisker diagram (Figure 5). Results indicate TDS and conductivity levels of between 400 and 2,240 μms/cm² are of greatest concern.
Figure 5

Box and whisker plot of water samples from Elidar woreda.

Figure 5

Box and whisker plot of water samples from Elidar woreda.

The data were then combined using an expert judgement overlay method. Reference was made to Malezewski (1999) and Minor et al. (2007) to define the weightings. The study divided the overlay into four main parameters: geology (k), geomorphology, groundwater recharge potential and structural density. The assumptions for each of these factors included:

  • (1) Geology: General permeability, primary permeability (pore spaces) or regional permeability are associated with average jointing and fracturing. The storage capacity depends on pore spaces, and the movement depends on permeability resulting from interconnected weathering and cooling joints and rock mineral assemblage openings. Water storage or yield potentially increases with increasing general permeability.

  • (2) Geomorphology: Water storage or yield potentials increase with a decrease in slope steepness. Groundwater storage is poor over mountainous and steep slope areas.

  • (3) Groundwater recharge: Groundwater potential increases with the increase in recharge potential, and the rate of infiltration controls recharge potential.

  • (4) Structure: Secondary permeability along geological structures, generally considered as water bearing. The higher the structural density, the higher the water bearing capacity.

To assign weightings to each of these assumptions, an expert group consultation was held and the following expert judgement weightings were agreed (Table 3).

Table 3

Overlay method weightings

FactorClass and class weight
LayerParameterweight1234
Geomorphology 0.3 Flat low lying grabens Gentle to flat horst Gentle to steep horst Steep and mountainous 
50 30 15 
Drainage/drainage density 0.25 High Moderate   
70 30   
Hydrogeology/permeability 0.23 High Low to moderate Aquiclude  
60 35  
Structure/structural density 0.22 Major (regional) Moderate Local Poor 
55 25 15 
  Total 58.4 30.05 8.95 2.6 
  Classification   High Moderate Low to moderate Low 
FactorClass and class weight
LayerParameterweight1234
Geomorphology 0.3 Flat low lying grabens Gentle to flat horst Gentle to steep horst Steep and mountainous 
50 30 15 
Drainage/drainage density 0.25 High Moderate   
70 30   
Hydrogeology/permeability 0.23 High Low to moderate Aquiclude  
60 35  
Structure/structural density 0.22 Major (regional) Moderate Local Poor 
55 25 15 
  Total 58.4 30.05 8.95 2.6 
  Classification   High Moderate Low to moderate Low 

Topographical data and geomorphological maps were developed. Results of the hydrogeology and geology maps were presented as 1:100,000 scale maps.

Test drilling

In phase 3, groundwater suitability zones were identified from the remote sensing and the hydrogeology/geology maps. A probability model was then presented. To stress test the model, UNICEF contracted a private drilling company to drill four production boreholes in the high probability zones. An average success rate of 75% was applied to the contract. The contract was awarded to drill the four boreholes selected from the list of most feasible sites identified from a combined study of phase 1 and 2. In the contract, UNICEF agreed to pay for a maximum of three out of the four boreholes (assuming at least a 75% success rate) that are successful in striking water of acceptable quantity and quality as predefined in the contract agreement (Q > 2 L/s and with acceptable water quality). Any additional borehole drilled would be the liability of the contractor. Turnkey contracting was used following the FIDIC procurement rules of the National Competitive Bidding (NCB) procedures of the Government of Ethiopia Procurement Proclamation. Payment for positive boreholes was used as an incentive to reduce the potential number of abandoned or negative boreholes during the drilling process.

RESULTS

The overlay map produced as a result of this study is outlined in Figure 6, where weightage was given based on expert judgement for each of these categories.
Figure 6

Groundwater suitability map of Elidar woreda.

Figure 6

Groundwater suitability map of Elidar woreda.

Using the weightages, the district was divided into four groundwater probability zones (high, medium, low to moderate and low). Figure 7 indicates the location of the major asphalt road running from west to north east of the district and connecting the road to Eritrea and Djibouti with the region's capital (Semera). The sedentary majority of the population of the district live along this road. The model identified two zones of ‘high’ probability of groundwater occurrence along the road. These are marked in boxes in Figure 7, and are Dobi and Suula villages.
Figure 7

Zones of ‘high’ probability of groundwater occurrence along the road.

Figure 7

Zones of ‘high’ probability of groundwater occurrence along the road.

To further understand the potential required depths for improving drilling success rates in Dobi and Suula, the GWA conceptual model outlined in Figure 5 was produced. Figure 8 compares data from phase 1 and 2 of the investigation with existing borehole logs. The X axis is in km and demonstrates the distance from the district border to the location of the sites, and the Y axis indicates the mountain topography measured as metres above sea level. Using this conceptual model, estimations of both the probability of striking fresh groundwater from the lateral recharge from the highland region, combined with an approximation of borehole depths, can be ascertained.
Figure 8

Depth versus altitude conceptual model.

Figure 8

Depth versus altitude conceptual model.

The D-GWa and the S-GWa model estimates that boreholes will need to be drilled in Suula from 250 m above sea level to 88 m above sea level to ensure a borehole would strike water. The depth in Dobi will need to be from 120 m above sea level to 110 m above sea level to ensure access to a fresh water lens.

Table 4 compares the recommended drilling depths from the GWA model with the actual drilling depths achieved in phase 3 of the test drilling.

Table 4

Recommended drilling depths from GWA model

 Location
DescriptionDobe 1Dobe 2Suula 1
Recommended water strike depth (from GWA) 110 m a.s.l. 110 m a.s.l. 88 m a.s.l. 
Actual water level after drilling 113 m a.s.l. 116 m a.s.l. 98 m a.s.l. 
Recommended water quality (from GWA) 1,900 μS/cm 1,800 μS/cm 1,000 μS/cm 
Actual water quality (after drilling) 7,000 μS/cm 3,000 μS/cm 1,900 μS/cm 
SWL (m b.g.l) 2.5 207 
DD (m b.g.l)   112.4 
DWL (m b.g.l)   219.4 
Yield (L/s) 10 3.9 
Water quality Not acceptable Acceptable Acceptable 
Aquifer type Alluvial deposit Alluvial deposit Fractured basalt and scoria 
Remark Compressor test Compressor test Test pumping undertaken 
 Location
DescriptionDobe 1Dobe 2Suula 1
Recommended water strike depth (from GWA) 110 m a.s.l. 110 m a.s.l. 88 m a.s.l. 
Actual water level after drilling 113 m a.s.l. 116 m a.s.l. 98 m a.s.l. 
Recommended water quality (from GWA) 1,900 μS/cm 1,800 μS/cm 1,000 μS/cm 
Actual water quality (after drilling) 7,000 μS/cm 3,000 μS/cm 1,900 μS/cm 
SWL (m b.g.l) 2.5 207 
DD (m b.g.l)   112.4 
DWL (m b.g.l)   219.4 
Yield (L/s) 10 3.9 
Water quality Not acceptable Acceptable Acceptable 
Aquifer type Alluvial deposit Alluvial deposit Fractured basalt and scoria 
Remark Compressor test Compressor test Test pumping undertaken 

The drilling campaign was undertaken between September and December 2015. In phase 3, three production boreholes were drilled and the results showed a 92% accuracy against the overlay model.

CONCLUSION

In conclusion, this study provided a comprehensive understanding of the hydrogeology of the area that assured the drilling of productive wells. Each of the three phases of the study complemented each other. The remote sensing information helped to trace lineaments and provided a combination of high probability correlations for the identification of groundwater suitability zones. Field hydrogeological and geophysical surveys were used to verify the remote sensing study and identify suitable locations for drilling successful wells of acceptable groundwater quantity and quality (phase 2). As part of the field hydrogeological study, thorough investigation of existing groundwater outlets (springs, wells) was very useful to build the conceptual model and to verify groundwater occurrence at shallow and deep depths. The hydrogeological study was further enhanced from field checks of existing lithology and structures, analysis of information gathered from water wells and collection and verification of water quality data. Daily supervision of drilling works was required to ensure required drill depths and well construction were carried out as per the predefined design and locations. Analysis of drilling logs, test pumping results and water quality data helped to verify results of the first two phase studies and was used as an input to design the water supply system (phase 3).

However, it should be understood that, despite the positive implications of the three-phase approach, the availability of potential groundwater resources does not necessarily coincide with the settlement pattern of people. A phase (4) of the study to combine potential groundwater sites with socio-economic and demographic data should be considered in future studies. This would ensure that the groundwater is economically utilized to benefit people living within or in the surrounding areas of the identified potential sites.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the contribution of the collaborators in this research. The views expressed in this paper are solely those of the authors and do not necessarily reflect the views of all contributors.

REFERENCES

REFERENCES
Acacia
2014
Improving the drilling success rates in Afar and Tigray Regions – remote sensing component
.
UNICEF-UNESCO
,
Addis Ababa
,
Ethiopia
.
Arsano
Y.
Tamrat
I.
2005
Ethiopia and the Eastern Nile Basin
.
Journal of Aquatic Science
67
,
15
27
.
Arvidsson
K.
Stenberg
L.
Chirindja
F.
Dahlin
T.
Owen
R.
Steinbruch
F.
2011
A hydrogeological study of the Nhandugue River, Mozambique – a major groundwater recharge zone
.
Physics and Chemistry of the Earth
36
,
789
797
.
Ayenew
T.
Demlie
M.
Wohnlich
S.
2008
Hydrogeological framework and occurrence of groundwater in the Ethiopian aquifers
.
Journal of African Earth Sciences
52
,
97
113
.
Cherenet
T.
1993
Hydrogeology of Ethiopia and water resources development
.
Ethiopian Institute of Geological Surveys
,
Addis Ababa
,
Ethiopia
.
Dar
I. A.
Sankar
K.
Dar
M. A.
2010
Deciphering groundwater potential zones in hard rock terrain using geospatial technology
.
Environmental Monitoring and Assessment
173
,
597
610
.
Deyassa
G.
Kebede
S.
Ayenew
T.
Kidane
T.
2014
Crystalline basement aquifers of Ethiopia: their genesis, classification and aquifer properties
.
Journal of African Earth Sciences
100
,
191
202
.
EIGS
1993
Hydrogeological map of Ethiopia: 1:2,000,000 scale
.
Ethiopia Institute of Geological Survey
,
Addis Ababa
.
EIGS
1996
Explanation of the Geological Map of Ethiopia
.
Ethiopia Institute of Geological Survey
,
Addis Ababa
.
Kebede
S.
2012
Groundwater in Ethiopia – Features, Numbers and Opportunities
.
Springer-Verlag
,
Berlin, Heidelberg
.
Leblanc
M.
Leduc
C.
Razack
M.
Lemoalle
J.
Dagorne
D.
Mofor
L.
2003
Application of remote sensing and GIS for groundwater modeling of large semiarid areas: example of the Lake Chad Basin, Africa
. In:
Hydrology of Mediterranean and Semiarid Regions Conference
,
Montpellier
,
France
,
Red Books Series, vol. 278, IAHC, Wallingford, UK
, pp.
186
192
.
Malezewski
J.
1999
GIS and Multicriteria Decision Analysis
.
John Wiley and Sons
,
Chichester
,
UK
, p.
392
.
Teklu
B.
Paulien
A.
Adriaanse
I.
Mechteld
C.
Ter Horst
M. S.
Deneer
J. W.
Van den Brink
P. J.
2015
Surface water risk assessment of pesticides in Ethiopia
.
Science of the Total Environment
508
,
566
574
.
World Bank
2016
Earth Observation for Water Resources Management – Current use and Future Opportunities for the Water Sector
.
World Bank Group
,
Washington, DC
,
USA
.