论文信息

MOHAMED AHMED YAKUB / Ahmed Mohamed Ismail

Research Square

论文解读

AI 辅助整理
方法
Landsat 8、SRTM、CRU降雨与地质土壤数据,九图层AHP加权叠加,4
发现
地质与降雨权重合计52.4%,67.5%面积属良好区,但验证AUC仅0.486,仅宜区域规划。
机会点
验证精度偏低且优/差区无井检验,可引入实地电法或机器学习提升小样本区划可靠性。

论文摘要

摘要 地下水是索马里希尔谢贝利州中谢贝利地区乔哈尔区的主要供水来源,但钻孔选址工作一直缺乏对控制地下水赋存条件的系统评估。本研究通过整合遥感、地理信息系统和层次分析法(Analytical Hierarchy Process),识别并绘制了该区的地下水潜力分区图。研究从Landsat 8 OLI/TIRS影像、SRTM数字高程数据、2015至2025年CRU TS降雨记录以及辅助地质与土壤数据集中提取了九个专题图层,即地质、降雨、土壤、河网密度、土地利用与土地覆盖、坡度、线理密度、高程和归一化植被指数。各图层均按五点适宜性量表重新分类,并依据Saaty标度通过两两比较确定权重。地质(26.22%)和降雨(26.17%)成为主导控制因素,合计占总权重的52.39%,一致性比率为0.038,证实判断矩阵可接受。在ArcGIS中进行加权叠加分析,生成地下水潜力指数,并划分为四个等级:良好,面积3 023.99 km²,占该区67.54%;中等,1 306.52 km²,占29.18%;较差,135.22 km²,占3.02%;极好,11.66 km²,占0.26%。该图利用从希尔谢贝利州能源与水资源部获取的48个钻孔资料进行验证,并部分通过野外核查予以确认,总体一致率为47.9%,受试者工作特征曲线下面积为0.486。由于没有钻孔位于较差或极好等级区内,这两个等级未得到检验,因此该图建议用于区域规划,而非具体场地的钻孔选址。

全文

Research Square

Identification of Groundwater Potential Zones
Using Remote Sensing and GIS Techniques: A Case
Study in Jowhar, Somalia
MOHAMED AHMED YAKUB 
East Africa University
Ahmed Mohamed Ismail 
East Africa University
Research Article
Keywords: groundwater potential zones, remote sensing, geographic information systems, analytical
hierarchy process, weighted overlay, Jowhar, Somalia
Posted Date: November 13th, 2026
DOI: https://doi.org/10.21203/rs.3.rs-11340848/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License.  
Read Full License
Additional Declarations: No competing interests reported.

i

IDENTIFICATION OF GROUNDWATER POTENTIAL ZONES USING
REMOTE SENSING AND GIS TECHNIQUES: A CASE STUDY IN JOWHAR,
SOMALIA

Authors:
MOHAMED AHMED YAKUB, Email: Moasidow@gmail.com, East Africa University, Garowe
Somalia
AHMED MOHAMED ISMAIL, Email: a.moha1010@gmail.com, East Africa University,
Garowe Somalia

THESIS SUBMITTED IN FULFILMENT FOR THE DEGREE OF MSc IN WATER
RESOURCES MANAGEMENT

FACULTY OF ENGINEERING AND TECHNOLOGY

October, 2026

ii

ABSTRACT
Groundwater is the principal source of water supply in Jowhar District, Middle Shabelle Region,
Hirshabelle State, Somalia, yet borehole siting has proceeded with limited systematic assessment of the
conditions governing its occurrence. This study identified and mapped groundwater potential zones in
the district by integrating remote sensing, geographic information systems and the Analytical Hierarchy
Process. Nine thematic layers, namely geology, rainfall, soil, drainage density, land use and land cover,
slope, lineament density, elevation and the normalized difference vegetation index, were derived from
Landsat 8 OLI/TIRS imagery, SRTM digital elevation data, CRU TS rainfall records for 2015 to 2025
and ancillary geological and soil datasets. Each layer was reclassified onto a five-point suitability scale
and weighted by pairwise comparison on Saaty’s scale. Geology at 26.22 per cent and rainfall at 26.17
per cent emerged as the dominant controls, together accounting for 52.39 per cent of the total weight,
with a consistency ratio of 0.038 confirming the acceptability of the judgements. Weighted overlay
analysis in ArcGIS produced a groundwater potential index classified into four zones: Good, covering
3,023.99 km† or 67.54 per cent of the district; Moderate, 1,306.52 km† or 29.18 per cent; Poor, 135.22
km† or 3.02 per cent; and Excellent, 11.66 km† or 0.26 per cent. The map was validated against 48
boreholes obtained from the Ministry of Energy and Water Resources of Hirshabelle State and verified
in part by field inspection, giving an overall concordance of 47.9 per cent and an area under the receiver
operating characteristic curve of 0.486. Because no borehole lay within the Poor or Excellent zones, two
classes remained untested, and the map is accordingly recommended for regional planning rather than
for site-specific borehole siting.

Keywords: groundwater potential zones; remote sensing; geographic information systems; analytical
hierarchy process; weighted overlay; Jowhar; Somalia

iii

TABLE OF CONTENTS
DECLARATION .................................................................................................................................... ii
ACKNOWLEDGEMENT .................................................................................................................... iv
ABSTRACT............................................................................................................................................. v
LIST OF TABLES ............................................................................................................................... viii
LIST OF FIGURES ............................................................................................................................... ix
LIST OF ABBREVIATION .................................................................................................................. x

iv

CHAPTER ONE ..................................................................................................................................... 1
INTRODUCTION .................................................................................................................................. 1
1.1 Background of the Study .............................................................................................................. 1
1.2 Statement of the Problem ............................................................................................................. 5
1.3 Research Questions ....................................................................................................................... 7
1.4 Research Objectives ...................................................................................................................... 8
1.4.1 Specific Objectives ..................................................................................................................... 8
1.5 Scope and Limitations of the Study ............................................................................................. 8
CHAPTER TWO .................................................................................................................................. 10
LITERATURE REVIEW .................................................................................................................... 10
2.1 Groundwater Occurrence and Hydrogeology .......................................................................... 10
2.2 Remote Sensing for Groundwater Exploration........................................................................ 12
2.3 GIS-Based Methods for GWPZ Delineation ............................................................................ 16
2.4 Thematic Parameters and Their Hydrogeological Significance ............................................. 19
2.5 Review of Similar Studies in Africa and Comparable Regions .............................................. 24
2.6 Study Area Context ..................................................................................................................... 26
2.7 Research Gaps and Conceptual Framework ............................................................................ 28
CHAPTER THREE .............................................................................................................................. 31
RESEARCH METHODOLOGY ........................................................................................................ 31
3.1 Study Area ................................................................................................................................... 31

v

3.2 Data Sources .......................................................................................................................... 32
3.3 Research Method......................................................................................................................... 33
3.3.1 Thematic Layer Derivation ..................................................................................................... 36
3.3.3 Classification, Rating, and Scoring of Thematic Layers ...................................................... 36
3.3.4 Analytical Hierarchy Process (AHP) and Weight Assignment ............................................ 36
3.3.5 Weighted Overlay Analysis and GWPZ Map Generation ................................................... 38
3.3.6 Validation of the Groundwater Potential Zone Map ............................................................ 38
CHAPTER FOUR ................................................................................................................................ 40
RESULTS & DISCUSSION ................................................................................................................ 40
4.1 Geology ......................................................................................................................................... 40
4.2 Rainfall ......................................................................................................................................... 41
4.3 Soil ................................................................................................................................................ 42
4.4 Drainage Density ......................................................................................................................... 42
4.5 Land Use / Land Cover (LULC) ................................................................................................ 43
4.6 Slope ............................................................................................................................................. 44
4.7 Lineament Density ...................................................................................................................... 44
4.8 Elevation ...................................................................................................................................... 45
4.9 Normalized Difference Vegetation Index (NDVI) .................................................................... 46
4.10 Assignment of Weightages to the Thematic Layers Using AHP ........................................... 47
4.11 Mapping of the groundwater potential zones ......................................................................... 49

vi

CHAPTER FIVE .................................................................................................................................. 57
5.1 Conclusion.................................................................................................................................... 57
5.2 Limitations of the Study ............................................................................................................. 58
5.3 Recommendations ....................................................................................................................... 59
5.4 Suggestions for Further Research ............................................................................................. 60
REFERENCES ..................................................................................................................................... 62

1

LIST OF TABLES
Table 1. Summary of Data Sources and Specifications ......................................................................... 32
Table 2. The one-to-nine scale of parameters significance .................................................................... 37
Table 3: Reclassification and rating of layer suitability for the identification of groundwater potential
zones. ...................................................................................................................................................... 40
Table 4: Land-use/land-cover types and coverage of the study ............................................................. 44
Table 5: Pairwise comparison matrix of the selected thematic layers and their weights. ..................... 48
Table 6: Normalized weight of the matrix. ............................................................................................. 49
Table 7: Groundwater potential zones and corresponding areal distribution. ....................................... 50
Table 8 Summary of scores and weightage allocated to thematic layers influencing groundwater
potential. ................................................................................................................................................. 50
Table 9 Classification of the validation boreholes by measured yield. .................................................. 53
Table 10 Cross-tabulation of observed borehole productivity against predicted groundwater potential
class. ........................................................................................................................................................ 55

2

LIST OF FIGURES
Figure 1: location map of the study area ................................................................................................ 31
Figure 2. Flowchart for the identification of groundwater potential zones (GWPZs). ........................... 35
Figure 3: Rainfall map for the study area. .............................................................................................. 42
Figure 4: Geology (lithology) map of the study area. ............................................................................ 42
Figure 5: Map showing Drainage Density .............................................................................................. 43
Figure 6: Soil map of study area ............................................................................................................. 43
Figure 7: LULC Map of study area ........................................................................................................ 43
Figure 8: Slope Map of study area Figure 9: Lineament Density map of Study area ...................... 45
Figure 10: Showing Elevation of study area ........................................................................................... 46
Figure 11: NDVI (a) and reclassified NDVI (b). .................................................................................... 47
Figure 12: Mapping of groundwater potential zones (GWPZs). ............................................................ 52
Figure 13: Validation borehole locations overlaid upon the delineated groundwater potential zones. .. 54
Figure 14: Receiver operating characteristic (ROC) curve for the groundwater potential zone model. 56

3

CHAPTER ONE
INTRODUCTION
1.1 Background of the Study
Fresh water makes up approximately 2.5 per cent of the total water on Earth, with many of the portions
being locked up in glaciers and polar ice caps, and thus unavailable for instant consumption by humans
(Shiklomanov & Rodda, 2003; UNESCO, 2024). The need for freshwater has been steadily increasing
by nearly 1 per cent per year since the 1980s, according to the World Water Development Report 2024,
compiled by the United Nations Educational, Scientific and Cultural Organization (UNESCO), with
most of the demand stemming from population growth, expansion of agricultural land, urbanization, and
rapid industrialization. Groundwater plays a pivotal role in rural economies, supporting livelihoods in
nearly 90 per cent of these communities, and it is estimated that it provides about 25 per cent of the
irrigation water and almost half of the world’s drinking water supply (UNESCO, 2024; Famiglietti, 2021;
Taylor et al., 2022).
The World Water Development Report (WWDR) published by UNESCO (2023) also estimated the
water scarcity for at least one month a year in 2016 was around 930 million people, and by 2050, this
figure could rise between 1.7 and 2.4 billion. The projections illustrate the need for systematic
identification, mapping, monitoring and sustainable management of groundwater in order to ensure
future water security.
Groundwater has become the most strategically important freshwater resource, especially in the arid and
semi-arid countries where surface water resources are limited and unreliable, particularly in the face of
growing climate variability. More than one-third of the world’s largest groundwater systems are highly
stressed as a result of groundwater overexploitation, contamination, reduced recharge and non-
sustainable groundwater abstraction (UN-Water, 2024; Scanlon et al., 2023; Taylor et al., 2022). In
addition, Scanlon et al. (2023) observed that groundwater is responsible for ~50% of the total
groundwater withdrawn for domestic purposes globally and that systematic groundwater exploration and
sustainable management is crucial to improve water security, especially when faced with extended
droughts. Groundwater resources are even more important in low income and in post-conflict countries,
where surface water infrastructure is often found to be inadequate or destroyed, meaning that

4

groundwater is the only, or main, source of water for people’s livelihoods, agriculture and domestic water
needs.
In the last decade, there has been a great deal of research interest in the international scientific community
with regard to the assessment of Groundwater Potential Zone (GPZ) using Remote Sensing and
Geographic Information System (GIS). The employment of GIS-based Analytical Hierarchy Process
(AHP) in combination with remote sensing data from satellite imagery proved to be a successful
approach to accurately delineate groundwater recharge zones in semi-arid agricultural landscapes in
Europe, as shown by the study in the Uckermark district of Brandenburg, Germany (Rothe et al., 2022).
These hydrogeological and land use parameters were overlaid and zone maps were calculated using the
weighted overlay technique; the zone maps were further checked based on long-term groundwater level
monitoring data. In Asia, GWPZ has been categorized in the Jinan Karst Spring Basin of China using an
integration of hydrogeological datasets, GIS, remote sensing and AHP techniques to classify GWPZ in
the basin as good groundwater potential (67 per cent of the basin) (Shen et al., 2024). Similarly, Yousaf
et al. (2024) applied the remote sensing technique through GIS in Lahore, Pakistan and identified the
groundwater potential recharge zones in the study area with five classes ranging from extremely poor to
extremely good, and concluded that most of the study area is in poor to mediocre category because of
high impervious cover and groundwater depletion in the study area. In general, these global studies have
demonstrated that the RS GIS approach used for field surveys is an alternative that can be reproduced,
is economical and provides a comprehensive response to the spatial context that cannot be achieved with
conventional hydrogeological field surveys, which are often too costly in developing countries and post-
conflict situations.
The application of remote sensing and GIS in GWPZ mapping has been significantly increased at the
regional level in Africa, but there are critical data deficits particularly in data poor countries. A geospatial
assessment was carried out in the Igede area of Ekiti State in West Africa that used a combination of the
Analytical Hierarchy Process (AHP) and Geographic Information Systems (GIS) based Multi-criteria
analyses to derive the thematic layers such as geology, rainfall, Digital Elevation Model (DEM) data,
land use and land cover (LULC), Normalised Difference Water Index (NDWI), slope and lineament
density that were subsequently used in the creation of GWPZ maps, which were verified against the
borehole records (Odesola et al., 2024). The study supported the validity of the proposed method AHP
weighted overlay to delineate GWPZ in the geologically complex regions, which is an effective and

5

practical approach. In the study area in Central African Republic (CAR), groundwater potential zones
map was created using AHP method and GIS, groundwater was found to be highly controlled by the
lithology, drainage density, slope and land use in the areas of limited groundwater access in rural Central
African Republic. A study in the Buffalo Catchment in the Eastern Cape province of South Africa used
GIS, remote sensing and AHP to classify the GWPZ into five classes ranging from very poor to good
covering an area of about 1,237 km†, which revealed that the area with high GWPZ values corresponds
to an environment that is semi-arid, with low drainage density and favorable geological structures
(Moodley, 2022). These regional studies demonstrate the robustness and adaptability of the RS-GIS-
AHP framework in different geological conditions and climatic zones in sub-Saharan Africa.
In the East Africa context, there is significant research literature to guide GWPZ mapping through remote
sensing and GIS in countries like Ethiopia and Kenya. There is significant research literature in the East
Africa context that can guide GWPZ mapping using remote sensing and GIS in countries like Ethiopia
and Kenya. The applications of weighted overlay analysis of thematic layers, such as geology, lineament
density, soil type, drainage density, slope, rainfall and LULC, have been consistently shown to be able
to effectively delineate groundwater potential at the watershed scales in the Gorezen Watershed, the
Muger Watershed, Abay Basin and the Teji River Catchment of southwest Shewa Zone in Ethiopia
(Mulu et al., 2024; Duguma, 2025; Workie et al., 2024). In Ethiopia, Melese and Belay (2022) achieved
an accuracy of 82.9 per cent in the prediction of GWPZ using borehole yield data and in Baringo County,
Kenya, Ombasa et al. (2022) showed that GWPZ mapping was done with good reliability. Somalia is
situated in the Horn of Africa region which has an adverse climate and hydrogeological profile and
Somalia is an area with a pronounced and well-documented water insecurity, but has received
comparatively little systematic attention in the scholarly literature on geospatial GWPZ mapping.
Somalia is one of the water-stressed countries globally. The country experiences an arid to semi-arid
climate, very little perennial surface water and extreme rainfall variability (Ismail et al., 2026).
According to the Food and Agriculture Organization Somalia Water and Land Information Management
Program (FAO SWALIM, 2024), Jubba and Shabelle Rivers are the only perennial water sources in
Somalia, and the vast majority of the people (about 95 per cent) rely on groundwater resources for
domestic water supply, livestock watering and small-scale irrigation. The main equipment used for
abstraction of groundwater is borehole, shallow wells and springs, with borehole depths of 90 to 250
metres and the yield ranging from 5 to 20 cubic metres per hour based on the different aquifer systems.

6

Groundwater resources in Somalia are however largely restricted in availability due to poor water quality
in parts of the country, as well as a lack of hydrogeological information, and the almost complete
destruction of water sector infrastructure caused by decades of civil conflict (FAO SWALIM, 2024;
ReliefWeb, 2024).
Jowhar, the administrative capital of the Middle Shabelle Region and Hirshabelle State of Somalia,
presents significant potential for systematic groundwater potential zone mapping. The town and its
surrounding areas are situated along the middle reaches of the Shabelle River, which serves as the
principal source of water for domestic consumption, irrigation, and livestock production. However, the
river exhibits marked seasonal variability, characterized by recurrent flooding during the rainy seasons
and substantially reduced flows or complete drying during prolonged droughts, thereby affecting water
availability and agricultural productivity (Ministry of Energy and Water Resources [MoEWR], 2022;
Food and Agriculture Organization [FAO], 2023). According to the National Bureau of Statistics (NBS)
Somalia Population Estimation Survey 2022, Jowhar District continues to experience population growth,
increasing demand for reliable water resources and underscoring the importance of identifying
sustainable groundwater sources (National Bureau of Statistics [NBS], 2022).
The groundwater system underlying Jowhar District is associated with Quaternary alluvial deposits
bordering the Shabelle River together with deeper sedimentary aquifer systems. Nevertheless, the spatial
distribution, depth, recharge characteristics, storage potential, and productivity of these aquifers remain
inadequately documented in the available scientific literature. Consequently, groundwater potential
mapping using Remote Sensing (RS) and Geographic Information Systems (GIS) has become an
important scientific approach for identifying suitable groundwater development zones to support
sustainable water resource management in the district.
Groundwater recharge, occurrence, storage, and movement are controlled by several interrelated
hydrogeological and environmental factors. These include geology and lithology, which determine the
nature and permeability of aquifer materials; lineament density, which influences groundwater
infiltration through fractures and faults; geomorphology, which controls landforms and infiltration
characteristics; soil type, which affects infiltration and water retention capacity; land use and land cover,
which influence surface runoff and groundwater recharge; drainage density, which determines runoff
concentration and infiltration opportunities; slope, which regulates runoff velocity and infiltration rates;
and rainfall, which provides the primary source of groundwater recharge (Saaty, 1980; Freeze & Cherry,

7

1979; Fetter, 2018; Ismail et al., 2026). These factors collectively determine groundwater availability
and are therefore widely adopted in groundwater potential zone mapping studies.
The integration of thematic layers extracted from satellite imagery, Digital Elevation Models (DEMs),
geological maps, and ancillary spatial datasets constitutes the foundation of Remote Sensing and GI S-
based groundwater potential zone (GWPZ) mapping. In GWPZ studies, thematic parameters such as
geology and lithology, lineament density, geomorphology, soil type, land use and land cover, drainage
density, slope, and rainfall are integrated using the Analytical Hierarchy Process (AHP), a structured
multi-criteria decision-making technique developed by Saaty (1980). AHP assigns relative weights to
the groundwater conditioning factors through pairwise comparisons before integrating them using
Weighted Index Overlay (WIO) analysis within a GIS environment to produce a composite groundwater
potential map. The resulting maps are commonly classified into groundwater potential zones ranging
from very poor to very high and are subsequently validated using borehole yields or well productivity
data.
Although Remote Sensing, GIS, and AHP have been widely applied for groundwater potential mapping
in many parts of Africa and other regions of the world, their application in Somalia remains limited.
Existing groundwater investigations in Somalia have largely focused on hydrogeological surveys,
borehole exploration, and groundwater resource assessments with limited integration of spatial multi-
criteria decision analysis for groundwater potential delineation. Furthermore, no comprehensive GIS –
Remote Sensing –AHP-based groundwater potential zone mapping study has been undertaken
specifically for Jowhar District despite its increasing water demand, climate variability, and dependence
on groundwater resources. This study therefore seeks to fill this knowledge gap by integrating Remote
Sensing, GIS, and the Analytical Hierarchy Process to delineate groundwater potential zones in Jowhar
District, Middle Shabelle Region, Somalia. The study aims to generate a scientifically reliable
groundwater potential map that will support groundwater exploration, sustainable water resource
management, and evidence-based planning in the district.
1.2 Statement of the Problem
In the East African region there is a significant amount of literature of GWPZ mapping in Ethiopia and
Kenya with the help of Remote Sensing and GIS. A weighted overlay analysis of thematic layers like
geology, lineament density, soil type, drainage density, slope, rainfall, and LULC have been proven to

8

be effective in identifying groundwater potential at watershed level in various studies in Ethiopia, namely
in the Gorezen Watershed, the Muger Watershed in Abay Basin, and the Teji River Catchment of
southwest Shewa Zone (Mulu et al., 2024; Duguma, 2025; Workie et al., 2024). In Ethiopia, Melese and
Belay (2022) reported a mapping accuracy of 82.9 per cent for GWPZ prediction using borehole yield
data, and Ombasa et al. (2022) proved that mapping was reliable in Baringo County, Kenya. Given the
challenge of this region, with Somalia sharing this climate and hydrogeological condition, Somalia has
been a region of relative neglect in scholarly literature on geospatial GWPZ mapping despite the well-
attested and acute water insecurity in the region.
Somalia is among the water stressed countries in the world. The country has an arid to semi-arid climate,
with a lack of perennial surface water and high rainfall variability. According to the Food and Agriculture
Organization Somalia Water and Land Information Management Programme (FAO SWALIM, 2024),
in the country, the only perennial water bodies are the Jubba River and Shabelle river with majority of
the population (estimated at around 95 per cent) relying on groundwater for drinking water, livestock
watering and small scale irrigation. The main groundwater abstraction facilities are boreholes, shallow
wells and springs, the borehole depths are generally in the range of 90–250 meters and the groundwater
abstraction rates of springs and shallow wells are between 5 and 20 cubic meters per hour depending on
the aquifer system. New groundwater resources development in Somalia, however, is significantly
hampered by low water quality in certain areas, lack of hydrogeological data, almost complete
destruction of infrastructure in the water sector over many years of civil conflict, and lack of systematic
groundwater mapping programmes at the district or regional level (FAO SWALIM, 2024; ReliefWeb,
2024).
Jowhar is the administrative capital of the Hirshabelle Federal Member State and the Middle Shabelle
region, and appears to be a good candidate for systematic groundwater potential zone mapping. The town
and the surrounding district are located along the middle course of the Shabelle River, which is the main
source of water for agricultural, domestic and livestock activities in the area. The Shabelle River,
however, is very episodically affected by flooding, with one of the most extreme floods on record
occurring in May 2023, and occasionally affected by droughts when it dries up (Ministry of Energy and
Water Resources, Somalia, 2022). The groundwater system permeating the Jowhar District is linked to
alluvial deposits of the Quaternary age along the margins of the Shabelle River and the deeper geologic
aquifer deposits, but the distribution, depth and yield of these aquifer systems are not sufficiently

9

understood in the published scientific literature. Given the estimated population of the district of 622,700
and the anticipated urbanization in the Jowhar district, which is likely to increase its water demand, the
systematic identification of groundwater potential zones in the Jowhar District using remote sensing and
GIS techniques is a scientific and humanitarian imperative.
The integration of various thematic layers obtained from satellite images, digital elevation models,
geological maps and ancillary spatial data is the basis of the GWPZ mapping through remote sensing
and GIS. The themes normally used in the GWPZ studies are generally geology, lithology, lineament
density, geomorphology, soil type, land use and land cover, drainage density, slope, and rainfall. These
parameters have been assigned relative weights using methods like Analytical Hierarchy Process (AHP)
(Saaty, 1980), a structured multi-criteria decision-making method that quantifies the relative importance
of each parameter by comparing one by one, and is integrated in a weighted index overlay (WIO) analysis
within a GIS environment to create composite groundwater potential index maps. These maps are then
subdivided into areas with different groundwater potentials, usually in the range of very poor to very
good, and checked with existing borehole or well yield data, if available. Although this method has been
used and tested in many geological and climatic environments, it has not been used systematically in the
Jowhar District and has not been applied to the Middle Shabelle area of Somalia as a whole, which is a
gap in knowledge that is being addressed by this study.
1.3 Research Questions

  1. What is the spatial distribution pattern of groundwater potential zones in the Jowhar District of
    the Middle Shabelle region, Somalia?
  2. Which remote sensing and GIS-derived thematic layers including geology, lineament density,
    soil type, drainage density, slope, land use and land cover, geomorphology, and rainfall most
    significantly influence the delineation of groundwater potential zones in the study area?
  3. How does the integrated weighted index overlay analysis informed by the Analytical Hierarchy
    Process perform in generating a reliable and internally consistent groundwater potential zone
    map for the Jowhar District?
  4. To what extent does the groundwater potential zone map produced by this study align with
    observed data from existing boreholes and wells in the district, and what does this validation
    indicate about the reliability of the model output?

10

1.4 Research Objectives
The study is to identify and map groundwater potential zones in the Jowhar District, Middle Shabelle
region, Somalia, using remote sensing and GIS-based techniques.
1.4.1 Specific Objectives

  1. To assess the relative influence variables on the occurrence and distribution of groundwater in
    the study area.
  2. To integrate the derived thematic layers using weighted overlay analysis informed by the
    Analytical Hierarchy Process (AHP) to produce a composite groundwater potential index map
    for the Jowhar District.
  3. To validate the derived groundwater potential zone map against available borehole and well yield
    data in the study area to determine the accuracy and reliability of the RS-GIS-AHP model output.
    1.5 Scope and Limitations of the Study
    This study will be conducted in Jowhar District, Middle Shabelle Region, Hirshabelle State, Somalia,
    focusing on delineating groundwater potential zones using Remote Sensing (RS), Geographic
    Information Systems (GIS), and the Analytical Hierarchy Process (AHP). The analysis will cover the
    district’s administrative boundaries and the hydrogeomorphic characteristics of the Middle Shabelle
    River Basin within the study area. Spatial datasets, including Landsat 8 and Landsat 9 imagery, Sentinel-
    2 imagery, the 30 m Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM), and
    published geological and soil maps, will be used to derive thematic layers influencing groundwater
    occurrence. The study will primarily rely on geospatial analysis and will not include groundwater quality
    analysis, aquifer pumping tests, or continuous groundwater level monitoring.
    The study may be constrained by limited availability of borehole logs and groundwater yield data for
    validating the groundwater potential map. The spatial resolution of freely available satellite imagery and
    DEM data may also limit the identification of fine-scale geological and structural features. In addition,
    security challenges in some parts of Jowhar District may restrict field verification of remotely generated
    outputs. Finally, although the AHP is a widely accepted multi-criteria decision-making technique, the
    assignment of factor weights involves expert judgment, which may introduce some degree of
    uncertainty. These limitations will be minimized through the use of validated geospatial datasets,

11

established AHP procedures, and comparison of the results with available secondary data and published
literature.

12

CHAPTER TWO
LITERATURE REVIEW
2.1 Groundwater Occurrence and Hydrogeology
2.1.1 Principles of Groundwater Occurrence and Movement
Groundwater is the water that fills the voids, fractures and pores in subsurface geologic formations and
flows from areas of recharge to areas of discharge because of gravity and hydraulic gradients (Freeze &
Cherry, 1979). Groundwater occurrence in any locality is controlled by a combination of geological,
climatic/weather, geomorphological and land-use factors, all of which help to define the capacity of the
subsurface to receive, store, and transmit water. These parameters are fundamental in determining the
potential zones of groundwater systematics because each one of them is a physically meaningful proxy
of one or more hydrogeological processes.
Groundwater is found in two major subsurface zones; the unsaturated zone (vadose zone) where pores
are partially filled with air and water, and the saturated zone where all the available pore spaces a re
occupied by water. The top of the saturated zone is known as the water table in unconfined aquifer
systems and piezometric surface in confined systems. The vertical infiltration of precipitation to the
water table is the main source of aquifer recharge and is facilitated by the intensity of precipitation,
duration of precipitation, soil texture and structure, slope gradient, land cover type, and depth and
permeability of the overlying geologic units (Scanlon et al., 2023). Recharge is typically episodic,
spatially variable and is strongly controlled by localized geological and geomorphological features and
structures such as permeable alluvial channels, fractured rock outcrops and pediment surfaces in arid
and semi-arid regions with low and highly variable annual rainfall.
The capacity to yield water to wells depends on the two most important characteristics of an aquifer, the
transmissivity and the storability, which are dependent upon the porosity, permeability, and thickness of
the water-bearing formation. Primary porosity is the porosity of the original fabric of the rock or
sediment, and is generally high in alluvial sediments, sandstones and gravels, these being productive
aquifer in many semi-arid parts Africa. Post-depositional processes like fracturing, faulting, and
chemical dissolution create secondary porosity, which is the primary source of groundwater storage and
transmission in hard rock areas including crystalline basement rocks and carbonates, and is the primary
source of groundwater in hard rock areas (Mulu et al., 2024). The groundwater potential in a basement

13

rock environment is thus very much dependent on the orientation, density and connectivity of fracture
systems that are commonly mapped as lineaments from remote sensing data.
2.1.2 Hydrogeological Settings of Arid and Semi-Arid Regions in East Africa
The East African region comprises very complex hydrogeological settings, ranging from the deeply
weathered crystalline basement rocks of the East African Rift System and the uplands, through to the
sedimentary aquifer systems of the Somali-Ethiopian plateau and the alluvial systems in the main river
valleys. The annual rainfall in most parts of Kenya, Ethiopia, Djibouti, Somalia and Eritrea ranges from
less than 100 millimeters in the driest areas to about 600 millimeters at the sub-humid margins. In such
settings, contamination of the water resources in the form of high salinity and fluoride concentrations is
often observed as a result of the physical water scarcity, which makes it challenging to identify and
develop practical freshwater aquifer systems hydrogeologically (Kebede, 2022).
In Kenya and in northern Ethiopia, groundwater is closely related to the rift valley geological framework,
with fault-bounded sedimentary basins and volcanic features containing productive aquifers, reached by
deep boreholes. In lowland areas of Ethiopia and Somalia, the alluvial aquifers are the main source of
accessible groundwater, while groundwater is present in the Quaternary floodplain sediments at
relatively shallow depths. The studies were carried out in the Ewaso Ng’iro–Lagh Dera Basin (EGB) of
Kenya that spanned arid hard rock terrain and showed that integrated RS, GIS and AHP techniques could
be used to effectively delineate groundwater recharge zones, where it was found that the most influential
parameter in determining spatial groundwater potential was geological lineaments followed by drainage
network characteristics (Ombasa et al., 2022). This discovery supports the broad conclusion that the
density and connectivity of fracture systems in the basement rock are the most important factors
determining groundwater occurrence and accessibility in this type of environment, as has been shown in
many other studies.
The general East African literature also supports the view that the potential of groundwater resources in
alluvial and floodplain environments, as in the Shabelle River valley of Somalia, depends mainly on the
thickness, lateral extent, and hydraulic connectivity of permeable alluvial sediments, and on the depth
of the water table below the land surface. The World Bank (2022) groundwater assessment for Somalia
included a description of shallow (Quaternary) aquifers connected to the Shabelle and Jubba Rivers,

14

characterized by groundwater at depths below 15 metres in some areas along the river, and the most
strategically available in southern Somalia.
2.1.3 Aquifer Types Relevant to Somalia and the Jowhar Geology
The geology of southern Somalia (i.e. Middle Shabelle, the area where Jowhar is located), involves a
layer of sedimentary rocks that overlie older basement rocks. The subsurface geology in the southern
part of Somalia consists of Quaternary alluvial and aeolian deposits, Neogene and Palaeogene
sedimentary rocks including the sandstones of the Jesomma Formation, and the limestones and
evaporites of the Mustaxiil Formation, and deeper Mesozoic sedimentary rocks (World Bank, 2022).
Hydrogeological attributes and characteristics of each of these formations are unique which control their
potential as a water-bearing aquifer system.
The most readily available aquifer system in the study area is the Quaternary alluvial deposits that are
adjacent to the Shabelle River in the Jowhar District. These deposits are composed of unconsolidated
sands, silts and clays deposited by the river during cycles of flooding over many thousands of years, and
are generally more than 10 metres thick in the most favourable areas (World Bank 2022). These
sediments normally have adequate permeability and specific yield to allow groundwater abstraction by
shallow dug wells or hand dug boreholes and in the district are generally the most targetted for
community water supply development. Hydraulic connectivity to the river system is known for
moderately productive aquifers with high storage capacity that under high water table conditions
maintain low flows in the Shabelle river, in deeper sedimentary aquifer systems, especially the Jesomma
sandstone formation as identified in the larger Somali-Ethiopian basin (Awulachew et al., 2012).
The aquifer system of the Jowhar District can be classified into three major types: unconfined alluvial
aquifers in the floodplain sediments of the Quaternary, semi-confined sediments of the Neogene and
fractured or karstic limestone where it is close to the surface (Mustaxiil). Existing literature lacks
information about the spatial distribution and hydrogeological characteristics of these aquifer systems,
especially at the district scale, which makes GWPZ mapping a systematic study using GIS and RS a
necessity.
2.2 Remote Sensing for Groundwater Exploration
2.2.1 Overview of Satellite Data Used in Hydrogeological Studies

15

In the last 30 years, the use of satellite remote sensing for hydrogeological studies has grown rapidly due
to the progress in spatial image processing and GIS related analysis coupled with the progressive
improvement of data quality and availability of multi-spectral and radar satellite data. A number of
remote sensing methods have significant advantages over traditional hydrogeological field survey
methods, especially in the post-conflict Somalia where field operations are highly restricted, such as:
synoptic spatial coverage, temporal repeatability, cost-effectiveness and the ability to produ ce spatially
continuous data over large and physically inaccessible areas (Ashraf et al., 2024).
The Landsat satellite series co-owned by NASA and the United States Geological Survey (USGS) has
been one of the most widely used multispectral data sources in GWPZ mapping studies around the world.
The Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS) are multispectral
instruments on both Landsat 8 (launched in 2013) and Landsat 9 (scheduled for launch in 2021) with 30-
metre spatial resolution, covering the visible wavelengths to shortwave infrared and thermal infrared,
respectively. These bands are especially useful for geological and structural mapping, LULC
classification, the extraction of spectral indices (e.g. Normalized Difference Vegetation Index, NDVI
and Normalized Difference Water Index, NDWI) and the identification of soil and surface lithol ogy.
Ethiopia (Workie et al., 2025) and Nigeria (Odesola et al., 2024) studies have established the use of
Landsat 8 imagery for effective derivation of GWPZ thematic layers for these semi-arid environments.
The European Space Agency (ESA) Copernicus Earth Observation programme operates the Sentinel -2
satellite, which offers multi-spectral data at a spatial resolution of 10 meters in the visible and near-
infrared spectral bands and 20 meters in the shortwave infrared spectral bands, and a revisit time of about
five days at the equator. The superior spatial and spectral resolution of Sentinel-2 compared to Landsat
renders it very useful for detailed applications in geological and land cover mapping. In addition,
Sentinel-1 SAR data was found to be more effective than optical data for extracting structural lineaments,
with higher lineament densities, in Egypt, where Ashraf et al. (2024) conducted their research.
The Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) is a topographic
dataset, with a spatial resolution of approximately 30 meters globally, and is the most commonly used
one in the mapping studies within GWPZ. It is the main data source from which morphometric
parameters such as slope, drainage density, drainage pattern, stream order and watershed boundaries are
calculated. The ASTER Global Digital Elevation Model (ASTER GDEM), which has a similar
resolution of 30 m, is an alternative or complement to the topographic data set, and has better vertical

16

accuracy in some types of terrain. Mulu et al. (2024) and Duguma (2025) conducted studies in Ethiopia,
which found that SRTM-derived DEM data can serve as a valuable foundation for morphometric analysis
in semi-arid watersheds and produce drainage network maps and slope maps that are useful for capturing
the spatial controls on groundwater recharge and flow.
2.2.2 Lineament Extraction from Satellite Imagery
One of the most important uses of remote sensing in groundwater exploration is the delineation of
lineaments, which are zones of increased secondary permeability in the subsurface where groundwater
infiltrates, migrates, and accumulates, such as faults, fractures, joints and shear zones. Lineament density
is a widely used parameter that is a good indicator of rock fracturing and, therefore, the potential for
productive groundwater in an area. In crystalline basement rock areas, high lineament density is always
correlated with high groundwater potential, but in sedimentary formations, the lineaments can point to
higher groundwater potential zones where fracture zones cut through low permeability stratigraphy
(Ashraf et al., 2024).
Lineaments are generally generated from satellite imagery by a combination of multispectral or radar
information directional filtering and automatic or semi-automated line detection algorithms that can be
implemented within a GIS environment. Studies have shown that optical sensors (e.g. Sentinel-2 and
Landsat 8) are very effective at mapping longer lineaments with clear tone variation, whereas radar
sensors (e.g. Sentinel-1 SAR) are more effective at detecting shorter lineaments because they are less
sensitive to solar illumination conditions and more to surface roughness and topographic shadowing
(Ashraf et al., 2024). Hill shade and slope images derived from the DEM are also commonly employed
in lineament mapping to help highlight subtle lineations that can be seen with SRTM at different
azimuths of illumination, related to fault zones and fracture sets. This has been confirmed by studies in
Ethiopia which revealed that the geology and lineament density and geomorphology are always the most
important variables in the AHP-weighted GWPZ analysis, and thus, the primary step in the remote
sensing workflow is lineament mapping (Bekele et al., 2025).
2.2.3 Drainage Pattern Analysis Using DEM
Analysis of drainage network patterns generated from the DEM data is a basic part of GWPZ mapping
because drainage patterns also give indirect information of the underlying geology, topographic controls
on water movement, and relative ability of different terrain units to produce surface runoff or

17

groundwater recharge. Groundwater recharge potential decreases with increasing drainage density in
most geological conditions because high drainage density results in high surface runoff and low
infiltration and vice versa for low drainage density (Mulu et al., 2024).
Drainage network extraction is done by DEM based method through hydrological processing tools
available in GIS software (i.e. ArcGIS or QGIS) by following standard methodology of Stream Ordering
(Strahler method). The hydraulic response of watersheds and the susceptibility of their watersheds to
surface runoff and/or groundwater recharge is described by the drainage network, the stream order
hierarchy and basin morphometric parameters such as bifurcation ratio, elongation ratio and basin form
factor. The theme of drainage density in terms of the spatial variability in the study area is an important
theme for GWPZ mapping in semi-arid environments as they are areas with low drainage density, which
are assigned the highest groundwater potential rating in the weighted overlay analysis, as indicated by
Workie et al., 2025.
2.2.4 Land Use and Land Cover Classification and Its Influence on Recharge
Land use and land cover (LULC) is an important thematic parameter in the mapping of GWPZ since it
directly controls the amount of infiltrating precipitation for recharging the groundwater system and the
amount of surface runoff and evapotranspiration. Dense vegetation cover (forests, woodlands and
grasslands) slows surface water, reduces the energy of raindrops, increases soil organic matter and
structure, and promotes moderate infiltration rates; agricultural lands with seasonal crops and crop
residues also promote moderate infiltration. However, the urban and peri-urban regions with higher
impervious surfaces like roads, buildings and concrete structures result in a significant reduction in
infiltration capacity and increase in surface runoff, thereby causing groundwater depletion in rapidly
urbanizing regions (Sreerama Naik et al., 2024; Shaikh et al., 2024).
LULC maps are generated from a supervised or unsupervised classification of multispectral satellite data
(typically Landsat 8/9 or Sentinel-2) using classification algorithms like Support Vector Machine,
Random Forest or Maximum Likelihood, within GIS or remote sensing software. Confusion matrix is a
method that is commonly used to estimate the accuracy of LULC classification from ground truth points
collected independently, and an overall classification accuracy of > 85% is regarded as acceptable in the
literature for use in GWPZ analysis (Odesola et al., 2024). It was found that in the East African context,
LULC changes such as increasing the extent of agricultural land at the cost of forest have a significant

18

impact on groundwater recharge rates, and current and accurate LULC data is important to incorpo rate
into GWPZ assessments over a 20-year time period (Gashaw et al., 2022).
2.3 GIS-Based Methods for GWPZ Delineation
2.3.1 Weighted Index Overlay Analysis
he GIS-based technique most widely used in the spatial integration of multiple thematic layers in GWPZ
mapping studies is the weighted index overlay (WIO) analysis. The general concept of WIO is to
combine multiple spatial data layers representing different hydrogeological parameters together in a GIS
environment and perform arithmetic operation (sum, difference, product) with the individual layers to
produce a single composite groundwater potential index map by assigning a weight that reflects the
relative influence of the individual layers on groundwater occurrence (Adugna & Awoke, 2024). The
groundwater potential index (GPI) is then classified into a finite number of groundwater potential ordinal
classes (usually 4-5) ranging from very poor to very good, according to the statistical distribution of the
groundwater potential index over the study area.
The mathematical expression of WIO analysis can be written as the sum of the product of the weights of
the thematic layers and the standardized score for each thematic layer at each spatial unit (pixel). Ea ch
parameter should have a percentage weight that adds up to 100 percent or unity, meaning that each
parameter makes its share in the overall index. The weight assignment to each thematic layer is the most
conceptually and methodologically important step, because it is the least consistent with varying weights,
which creates significantly different GWPZ maps. Weights are usually established using the AHP
framework or through expert opinion, but other methods, such as the frequency ratio method, weight of
evidence, and machine learning algorithms have been used in the literature. The numerous studies that
have been conducted have consistently shown that the WIO analysis using weights derived from AHP
can generate GWPZ maps with a level of validation accuracy that is comparable to the more
computationally expensive machine learning techniques, yet with greater methodological transparency
and reproducibility (Mihret & Wuletaw, 2025).
2.3.2 The Analytic Hierarchy Process in Groundwater Studies
The Analytic Hierarchy Process (AHP) is a structured multicriteria decision making method, which was
introduced by Saaty in 1980, is used to systematically quantify the relative importance of multiple

19

decision criteria by using pairwise comparisons (Saaty, 1980). The AHP framework is used to determine
the weights of the various thematic layers (geology, lineament density, drainage density, slope, LULC,
soil type, rainfall) in the context of the GWPZ mapping, as they are deemed by expert judgement or by
empirical evidence from the literature to be of relative hydrological significance. The pairwise
comparison process consists of assigning a rating between 1 and 9 to the parameters of each pair on
Saaty’s nine-point comparison scale, which represents the importance of the parameters, ranging from
’equal importance’ (1) to ’extremely dominant’ (9).
The AHP framework generates a pairwise comparison matrix which is then normalized to produce a
priority vector (AHP weight) for each parameter using the eigenvalue analysis. The Consistency Ratio
(CR) is used to test the internal consistency of the pairwise comparison matrix to be less than 0.10 (10
per cent) for the weightings to be deemed to be acceptably consistent. If the CR is more than 0.10, it
means that the pairwise comparisons made are not consistent enough and the judgements should be
adjusted. The internal coherence of the weight assignment process has been confirmed by studies that
have used AHP in the mapping of GWPZ in Ethiopia, with consistently CRs below 0.10 even in well-
designed studies (Diriba et al., 2024; Adugna & Awoke, 2024). Rainfall and geology/lithology are the
two main factors that are most frequently given the highest weight in GWPZ studies in sub-Saharan
Africa, with geology/lithology being the primary factor in determining capacity of the geology to store
water and rainfall in determining groundwater recharge rates (Mulu et al., 2024; Bekele et al., 2025).
Many of the practical benefits of AHP seen in GWPZ studies are as follows: conceptually
understandable, relatively easy to calculate using readily available GIS software, flexible to incorporate
a range of qualitative and quantitative information, and the ability to aggregate multiple expert sets of
weights into a consistent set of weights. The groundwater potential is generally similar for all the models,
with AHP-based models appearing to be slightly more conservative in identifying high groundwater
potential areas when compared to other MCDA based methods such as Fuzzy AHP, the Multi-
Influencing Factor (MIF) method and machine learning methods like Random Forest (Workie et al.,
2025; Mihret & Wuletaw, 2025).
2.3.3 Multi-Criteria Decision Analysis Frameworks
Multi-criteria decision analysis (MCDA) is a family of structured decision-making methods that, in cases
of multiple criteria (often conflicting) and where a spatially differentiated evaluation of an attribute of

20

interest is required, can be used to reach a decision. Hydrogeological mapping can be used to integrate
the multi-dimensional nature of the groundwater occurrence in a systematic framework, such as a MCDA
framework, of which AHP-based WIO analysis is the most common. However, other MCDA
frameworks have been used in GWPZ mapping such as MIF method, Fuzzy AHP approach, logistic
regression, weight of evidence (WOE), frequency ratio analysis and, more recently, machine learning
classifiers like Random Forest, Boosted Regression Trees, and Artificial Neural Networks.
Unlike AHP, the MIF method is based on the number of factors each parameter is directly and indirectly
affected by to assign weights to thematic layers, thus reflecting the interconnection of each parameter
with the entire hydrogeological system in an effective manner rather than just based on expert opinion.
Mulu et al. (2024) conducted a study in the Gorezen Watershed, Ethiopia, which used both AHP and
MIF methods, concluding that the fusion of the two methods led to a better spatial agreement of GWPZ
classification when compared to single method use. The Fuzzy AHP method is an extension of the
classical AHP that overcomes the discrete nature of the GWPZ maps by assigning triangular fuzzy
numbers to the scores of pairwise comparisons of the experts’ judgements and eventually producing
smoother GWPZ maps that could better represent the inherent gradational nature of GWP distributions
across the landscape (Workie et al., 2025).
Thematic Layers can be integrated using the following means:
The thematic layers are integrated into a GIS-based GWPZ map with the help of raster-based overlay
analysis in ArcGIS or other GIS software, based on a standardized procedure which involves the
following steps: data collection, data processing and map preparation, reclassification of each thematic
layer into a uniform rating scale, application of weights using AHP or other MCDA approach, calculation
of the weighted overlay, classification of the GWPZ, and validation. The computation of the overlays
can only be done once each thematic layer is reclassified on a common rating scale on a ordinal sc ale
ranging from 1 to 5, with the higher rating having a greater groundwater potential (Diriba et al., 2024).
The most frequently used thematic layers used in GWPZ mapping studies in East Africa and similar
semi-arid areas are geology/lithology, geomorphology, lineament density, drainage density, slope,
rainfall, soil type and texture and LULC. These 8 parameters are consistently included in the studies
analyzed for this review and, in some cases, supplementary parameters like the Topographic Wetness
Index (TWI), Topographic Position Index (TPI) and Elevation are also added as indicators of surface

21

hydrological behavior. For example, studies undertaken in Ethiopia (Mulu et al., 2024; Workie et al.,
2025; Bekele et al., 2025), Nigeria (Odesola et al., 2024) and South Africa (Moodley, 2022) all include
these 8 parameters.
2.3.4 Integration of Thematic Layers
The integration of multiple thematic layers in the mapping of GWPZs is achieved by using a standardized
process of data collection, data processing, map preparation, reclassification of each thematic layer into
a uniform rating scale, weight assignment (either by AHP or other MCDA approach), weighted overlay
computation, GWPZ classification and validation, carried out in GIS (ArcGIS or equivalent), by raster-
based overlay analysis. To enable computation of overlay, each thematic layer has to be reclassified to
an ordinal rating scale, usually 1 to 5, with 5 being the most groundwater potential and 1 being the least
(Diriba et al., 2024).
The most frequently used thematic layers used for GWPZ mapping studies in East Africa and similar
semi-arid regions are geology, lithology, geomorphology, lineament density, drainage density, slope,
rainfall, soil type and soil texture, and LULC. All of the studies analyzed for this review also contain
these eight parameters, but the extra layers (Topographic Wetness Index (TWI), Topographic Position
Index (TPI) and elevation as a supplementary indicator of surface hydrological behavior) are sometimes
added to the study. The studies were conducted in Ethiopia (Mulu et al., 2024; Workie et al., 2025;
Bekele et al., 2025), Nigeria (Odesola et al., 2024) and South Africa (Moodley, 2022).
2.4 Thematic Parameters and Their Hydrogeological Significance
2.4.1 Geology
The geology of the aquifer are the most universally accepted thematic parameters used in the
construction of GWPZ, since the physical and chemical characteristics of the rock or sediment of which
an aquifer is composed will directly impact the porosity, permeability and water-storage capacity of the
aquifer. The hydrogeological properties of the different lithological units are vastly different:
unconsolidated sediments (sands and gravels) have high primary porosity and permeability; sandstones
or limestones have intermediate primary porosity and permeability and weathered or fractured zones are
also sources of secondary porosity in these rocks; crystalline basement rocks such as granite, gneiss and
schist have little primary porosity, but may contain significant secondary porosity through weathered
zones or fracture systems. Geology always has the highest or second highest weight in AHP-based

22

GWPZ analysis across East Africa indicating its first or second consideration in controlling the spatial
distribution of groundwater potential (Mulu et al., 2024; Eshetu et al., 2024).
2.4.2 Geomorphology and Landforms
The recognition of landform types with different occurrence characteristics of groundwater is a key
element of geomorphology in providing critical hydrogeological information. High groundwater
potential is associated with geomorphological units such as pediment surfaces, valley fills, alluvial fans,
floodplains and buried channels which are areas of active or historical sediment deposition where
permeable materials are deposited. Rocky hills, ridges and elevated plateaus, on the other hand, are
related to high surface runoff and low groundwater recharge. Commonly, geomorphological thematic
layers used for GWPZ mapping are generated using the analysis of DEM coupled with interpretation of
satellite imagery and in most studies from East Africa, the AHP weight range is between 15 and 22 per
cent (Eshetu et al., 2024; Bekele et al., 2025).
2.4.3 Lineament Density
Lineament density is the linear density of the structural lineaments and is the most prominent remote
sensing indicator of secondary porosity and permeability in the subsurface. In crystalline basement rock
areas, a high lineament density is correlated wit h a well‑connected fracture network that improves
groundwater infiltration, storage and transmissivity. However, low lineament density, especially in
impermeable rock types, are more likely to be associated with poor groundwater potential. The findings
from numerous studies conducted in Ethiopia, Kenya and South Africa have been consistent in that the
two or three most important groundwater parameters in the mapping exercise are lineament density and
that zones of high lineament density correspond to high borehole yield in the validation exercise
(Moodley et al., 2022; Mulu et al., 2024).
Quantitatively, lineament density is defined as the total length of lineaments per unit area, calculated
using the equation:
𝐿𝐷 = ∑ =1 𝐿

where LD is the lineament density (km/km†), Li is the length of the ith individual lineament segment, n
is the total number of lineament segments within the study area or cell, and A is the area of the designated

23

cell or the total region under consideration (in km†). In some hydrogeological contexts, a secondary
count‑based density (Dn = n / A) is also used to highlight areas with a high frequency of lineament
intersections, which are particularly favorable for groundwater flow due to enhanced fracture
connectivity.
For manual extraction of lineament density often conducted for validation purposes or in areas lacking
advanced GIS software a systematic grid‑overlay method is employed. A transparent grid of uniform
cell size (commonly 1 km × 1 km, though the dimensions depend on the scale of the satellite imagery
and the target aquifer scale) is overlaid onto the printed or digitized lineament map. Within each grid
cell, the cumulative length of all visible lineament traces is carefully measured using a calibrated ruler,
a cartographic wheel measure, or a digital scale. This measured total length is then divided by the actual
ground area represented by the grid cell to derive the local lineament density value. The computed value
is assigned to the centroid of each respective grid cell, and these point‑values are subsequently
interpolated and contoured to generate the final lineament density surface used in the GWPZ mapping
exercise.
2.4.4 Drainage Density
As described in Section 2.3.3, the drainage density is inversely proportional to the potential of
groundwater recharge; and low drainage density is considered as a permeable substrate which favors
infiltration, rather than runoff generation. For GWPZ mapping studies, the hydrologically conditioned
drainage network derived from a DEM is used to complete a kernel density estimation (KDE) in a GIS
environment and this is reclassified into the rating classes such that the lowest drainage density will
capture the highest groundwater potential score. The study in the Muger Watershed (Chala et al., 2025)
and Gorezen Watershed (Mulu et al., 2024) of Ethiopia showed that the areas of low drainage density in
permeable geological formations strongly have higher yields of boreholes than the areas of high drainage
density in impermeable crystalline rocks.
Fundamentally, drainage density is determined using the following equation:
𝐷𝐷 =
∑ =1 𝐿


24

where DD is the drainage density (km/km†), L sj is the length of the j th stream segment, m is the total
number of stream segments within the watershed or kernel, and A is the total area of the watershed or
the specific GIS raster cell (in km†). In modern GWPZ analysis, this conventional arithmetic calculation
is replaced or supplemented by kernel density estimation (KDE). The KDE applies a moving window of
a specified radius across the stream network, employing a smoothing kernel function (e.g., quartic or
Gaussian) to compute the density per unit area for each output raster cell. This approach produces a
continuous, spatially‑smoothed drainage density surface that effectively captures the spatial variability
of stream network concentration, which is then reclassified into the appropriate groundwater potential
scoring classes.
2.4.5 Slope and Topography
Slope gradient is one of the basic factors that determine the distribution of precipitation between surface
runoff and groundwater recharge. Steep slopes tend to cause surface water to flow rapidly and erode,
decreasing the time that surface water has to infiltrate, and thus the probability of infiltration into the soil
and eventual groundwater recharge; gentler slopes allow surface water to pond or to flow slowly,
increasing the likelihood of infiltration into soil and subsequent infiltration to groundwater. The slope is
generally calculated from the DEM and is derived from the rate of change of elevation in the horizontal
(x) and vertical (y) directions. The slope angle is mathematically expressed as:
θ = arctan(√()2 + ()2)  180"
where θ is the slope in degrees and z is the elevation. It is usually divided into categories of gentle to
very steep slope (less than 5 degrees to 90 degrees). Slopes less than 5 degrees will have the highest
groundwater potential rating in GWPZ mapping. The findings of this study in the Gidabo Watershed in
Ethiopia (Diriba et al., 2024) showed that most of the high groundwater potential zones were located at
gentle slopes which also coincided with fine‑textured soils and high-density lineament networks.
2.4.6 Rainfall and Precipitation
Groundwater recharge is primarily controlled by rainfall in most semi-arid environments and thus the
spatial and temporal variations in rainfall is the prime determinant of the spatial pattern in groundwater

25

potential. The depth and permeability of the vadose zone, slope, soil type and land cover will modulate
the impact of rain as a recharge driver, but it is better able to recharge areas with higher mean annua l
rainfall. The rainfall data for GWPZ mapping studies are usually derived from interpolated rainfall
surfaces created from rain gauge station readings or derived from satellite observations of precipitation
like CHIRPS (Climate Hazards Group InfraRed Precipitation with Stations) and PERSIANN, which are
especially useful in regions with limited ground-based rain gauges (Eshetu et al., 2024). The thematic
parameters of a GWPZ study in an AHP context is usually dominated by the weight of rainfall (20-25%),
which is the recharge source in Ethiopia (Eshetu et al., 2024; Bekele et al., 2025).
2.4.7 Soil Type and Texture
Soil physical properties significantly influence the speed with which infiltrating water moves through
the vadose zone to the water table, especially the texture (the relative amounts of sand, silt and clay); the
structure (the arrangement of soil particles); and the hydraulic conductivity (the ability of water to pass
through the soil). The hydraulic conductivity of the coarse textured sandy soils is high, allowing for high
infiltration rates, and thus high groundwater recharge rates, while the fine textured clay rich soil has a
low hydraulic conductivity and will allow for water to pond on the surface of the soil or to surface run
off. Soil texture data are usually obtained from national or international soil maps (such as the FAO
World Soil Database and the USDA Soil Taxonomy) and classified as to ground water permeability and
ground water recharge potential in GWPZ mapping. In the Muger Watershed (Duguma, 2025) it was
found that low matrix permeability of the soil on gently sloping terrain with high numbers of lineaments
showed high groundwater potential, even though the lineaments facilitated preferential pathways for
recharge.
2.4.8 Land Use and Land Cover
The relationship between LULC and groundwater recharge and infiltration have been outlined in Section
2.3.4. LULC usually has a moderate weight (8 to 15 per cent) in the AHP analysis, given the fact that it
is not a primary geologic control of groundwater occurrence, but rather a surface modifier of the recharge
process when considering the thematic parameter significance for groundwater mapping. Areas heavily
dominated by vegetation, such as forests, woodlands, and shrublands are always rated as having the
highest potential for groundwater whereas urban and bare land areas are always rated as having the
lowest potential for groundwater. Land Use and Land Cover (LULC) maps are created from recent

26

satellite data to reflect the temporal dynamism of LULC, which is a result of continuous process of land
degradation, agricultural expansion and urbanization within the study area, as indicated in the literature
(Gashaw et al., 2022; Kawara et al., 2024).
2.5 Review of Similar Studies in Africa and Comparable Regions
2.5.1 Case Studies from East and Northeast Africa
GIS and RS-based GWPZ mapping has become the most prolific research area in sub-Sa haran Africa
and data collection has been progressing rapidly in Ethiopia, with several peer-reviewed research studies
for a variety of geological, climatic and hydrological environments. The studies undertaken by Kindie
et al. (2019), Melese and Belay (2022), Godif and Manjunatha (2023) and Manderso et al. (2024) used
the GIS-based MCDA methods to delineate GWPZ using seven to nine thematic layers all found
lithology and rainfall to be the most influencing factor in the Lake Tana Basin. Good to excellent
agreement with yields obtained from borehole records was achieved in these studies, with the agreement
index being higher than 75 per cent in most of the cases.
Mulu et al. (2024) used the AHP and MIF methods in the Gorezen Watershed which is semi-arid and
dominated by basalt rocks, and the GWPZ was obtained by overlaying nine thematic layers, resulting in
five GWPZ categories. The fractured basalt zones that had high lineament density were identified as the
most critical groundwater potential zones and were supported by the results of similar studies in the
Ethiopian and Kenya Basalt terrains. Based on the GIS-AHP, the GWPZ delineation in the Teji River
Catchment of Southwest Shewa Zone was conducted, revealing that around 30.88 per cent of the
catchment was classified as good groundwater potential zone; the spatial distribution of GWPZ indicated
a good correlation with alluvial landform and low slope terrain. The validity of the RS-GIS-AHP
approach was validated by research conducted in the Gidabo Watershed in the Main Ethiopian Rift
(Diriba et al., 2024) which showed a GWPZ model accuracy of 76.8 per cent on the basis of spring and
well data using the Receiver Operating Characteristic (ROC) curve method, indicating the validity of the
approach in a structurally complex rift valley environment.
Studies in Baringo County, Kenya, by Ombasa et al. (2022) showed that there is a good spatial
correlation between high and very high groundwater potential areas and the location of productive
boreholes. The study highlighted the specific relevance of adding geological lineament data to alluvial
formation maps in semi-arid areas underlain by basement rocks, where secondary permeability

27

controlled by lineaments is the main option for groundwater access. Studies have shown the applicability
of AHP and WIO analysis for delineation of GWPZ in sedimentary and mixed geological environments
respectively with drainage density as the controlling parameter in alluvial formations and lithology in
sandstone formations in the Sudan and in the entire Nile basin region.
2.5.2 Methodological Comparisons and Lessons Learned
The methodological strategies used in the reviewed studies share some common features and can be
generalized to the present study. There are several commonalities and lessons to be learned from the
studies reviewed for this chapter when it comes to methodological approaches that were utilized. In the
first instance, the validation accuracy of GWPZ mapping produced by the RS-GIS-AHP fr amework
ranges from 75% to 88% in various geographical, geological and climatic conditions, indicating the
applicability and reliability of the RS-GIS-AHP framework as a GWPZ mapping methodology (Adugna
& Awoke, 2024; Diriba et al., 2024; Pial et al., 2025). Second, the spatial pattern and level of accuracy
of the generated GWPZ map are highly dependent on the number and nature of the thematic layers
included in the analysis: studies with a sufficient number (8 or more) of carefully selected thematic
layers, which cover the various geological, morphological, hydrological, and land use aspects of
groundwater occurrence tend to achieve higher validation accuracy than studies with fewer layers. Third,
the method of validation (based on borehole yield data, spring discharge, ROC curve analysis or
agreement index calculations) has a significant impact on the accuracy estimates reported for GWPZ
maps, and is not standardized across studies, which makes it difficult to make direct comparisons of
accuracy across studies.
From the East African literature, the critical methodological lesson is the need to consider the weight
assignment process in the specific study area geological and hydrogeological setting and conditions, not
necessarily to blindly use weights from studies carried out in different hydrogeological settings.
Lithological formations in alluvial areas like the Shabelle river valley in the Jowhar District may be
differently weighted in terms of geology and drainage density as compared to basement rock areas where
groundwater is more likely to occur in fracture controlled secondary porosity. A key strength of the AHP
is the ability to incorporate expert judgements which are relevant to the local setting, and where borehole
data is limited for calibration of parameter weights.

28

2.6 Study Area Context
2.6.1 Geology and Hydrogeology of the Jowhar and Middle Shabelle Region
Jowhar the administrative centre of the Middle Shabelle region lies in the southern lowlands of Somalia
at an altitude of about 100 metres above sea level. The region has an area of about 22,663 sq km, and is
bounded by the Shabelle River which flows for about 150 km through the region and then goes south
into the coastal swamps to reach its confluence with the Jubba River and final loss in coastal swamps
near Jilib. Quaternary alluvial deposits of the floodplain of the Shabelle River dominate the regional
geology of the Middle Shabelle, overlying Neogene and Palaeogene sedimentary rocks of the Somali
Basin, including the sandstones of the Jesomma Formation, and the limestones and evaporite lithologies
of the Mustaxiil Formation (World Bank, 2022; Awulachew et al., 2012).
The most readily available and most important aquifer system in the study area is the Quaternary alluvial
deposits in the Jowhar District along the corridor of the Shabelle River. The deposits are laterally
extensive beds of sand and gravel interbedded with fine-grained, silty and clayey horizons that formed
during successive flood events, the coarser-grained deposits comprising productive unconfined aquifers
thicker than 10 meters. Groundwater found in these alluvial deposits is relatively shallow (usually < 15
meters below ground surface) and is hydraulically linked to the Shabelle River and so has a dynamic and
variable relationship with river stage in the immediate floodplain surroundings. The World Bank (2022)
geologic assessment also identified alluvial Quaternary sands as the main shallow aquifer that could be
tapped for community water supply borehole drilling in the Shabelle River corridor in the south of
Somalia.
The core of the Middle Shabelle region has relatively deeper sedimentary sequences such as the
sandstones of the Jesomma Formation that are known to host moderately productive confined aquifers
with good storage capacity to support base flows in the Shabelle River during the dry season (Awulachew
et al., 2012). The Mustaxiil Formation made up of Limestones, Gypsums and related evaporite
lithologies is known to form karstic aquifer systems in the Somali-Ethiopian border area but may also
occur in the subsurface of the Middle Shabelle at depth. Spatial distribution, thickness and hydraulic
properties of these deeper formations are not well characterised in the Jowhar District and a systematic
borehole database of the hydraulic properties is not available.
2.6.2 Climate and Hydrology of the Study Area

29

The Middle Shabelle region has an arid to semi-arid climate with bimodal rainfall pattern characterized
by two rainy seasons: Gu (long rains) during April to June and Deyr (short rains) during October to
December. The rainfall amount in the region is about 150-500 mm per year, with 150 mm to the north
and interior part of the district and 500 mm towards the coastal fringe where it is influenced by the Indian
Ocean monsoon system. Precipitation levels are highly variable both geographically and from one year
to the next, with years of severe droughts and floods, such as the prolonged drought from 2016 –2022
(five consecutive below-average rainy seasons), followed by years of extreme floods due to above-
average precipitation in the highland areas of Ethiopia that feed the Shabelle River (FAO SWALIM,
2024; Ministry of Energy and Water Resources, Somalia, 2022).
The most significant surface water resource in the Middle Shabelle region is the Shabelle River in which
water is used for irrigated agriculture, livestock watering, recharge of the floodplain aquifers by flooding
and replenishment of groundwater. The Shabelle River has also a strong inter-annual and seasonal flow
variability, however: during the rainy seasons (Gu and Deyr), the discharge can increase rapidly with the
rainfall in the highland areas of Ethiopia, resulting in a very high discharge of the river in low-lying areas
of the district, which can lead to severe flooding; during the dry seasons (January-March and July-
September), the river discharge can drop to very low levels or even become zero in certain years, leaving
communities completely reliant on groundwater source. The Middle Shabelle and southern Somalia
regions received below-average seasonal rainfall in late 2024, and the FAO SWALIM Groundwater
Monitoring Bulletin (2024) revealed a chronic vulnerability of the water supply in the region to climatic
variability, as expected under the influence of La Niæa, with a 60 to 80 per cent below normal rainfa ll
forecast for the 2024 to 2025 dry season.
2.6.3 Existing Water Resources and Usage Patterns
Three main sources of water supply exist in the Jowhar District: Shabelle River, Dug wells and scoop
holes in the floodplain alluvium and deep borehole that taps into the confined sedimentary aquifers. The
Somali population, including residents of Jowhar and its nearby rural area, mainly depend on
groundwater for their year round drinking water supply while the Shabelle River is mainly an irrigation
water source during the flood season, according to the FAO SWALIM (2024) groundwater monitoring
programme. Generally, the boreholes in Middle Shabelle region produce 5-20 cubic meters per
hour/discharge from available aquifer formations while shallow dug wells produce 2.5-10 cubic meters
per hour/discharge but are prone to contamination and may go dry during extended dry season periods.

30

The International Crisis Group (2025) reported that water access and control over irrigation and borehole
facilities has been a major aspect of the political economy of conflict in the Middle Shabelle region, and
non-state actors in the Middle Shabelle have taken advantage of the region’s water scarcity for resource
extraction and population control. This conflict situation has greatly hampered systematic collection of
data on groundwater resources, and has made a significant contribution to the lack of reliable
hydrogeological databases for Jowhar District. Humanitarian organizations like UNICEF, OCHA, and
the Water, Sanitation and Hygiene (WASH) cluster have installed boreholes for emergency water supply
in Jowhar and the surrounding areas, but the location, depth and yield of these boreholes, along with the
geological log, are not systematically documented in any publicly accessible database, creating
significant difficulties for the validation of GWPZ maps in this study area through borehole-based
methods.
2.7 Research Gaps and Conceptual Framework
2.7.1 Identified Gaps in Existing Literature
Based on the above literature search, the present study is justifiable as it has identified some major gaps
and significance of the study. The lack of published peer-reviewed research using remote sensing and
the GIS-based GWPZ mapping methodologies in the Jowhar District or in the Middle Shabelle region,
Somalia, in general is the most basic disparity. The RS-GIS-AHP methodology has been validated in
many similar lowland river basin environments in the semi-arid regions of East Africa, such as Ethiopia,
Kenya and Sudan, but has not been documented for any application in the settings of river basins in the
South of Somalia. This is an important knowledge gap, both because of the critical water insecurity
situation for the area and because there is the potential to generate evidence from geospatial GWPZ
mapping that can be used to inform investment in groundwater infrastructure in a post conflict recovery
scenario.
A second gap is the lack of systematic hydrogeological data both for the Jowhar District and for the
whole of southern Somalia. In Somalia, borehole logs have been produced for decades and are becoming
increasingly scarce, with the majority of performance data for boreholes being lost and the country’s
hydrogeological maps almost nonexistent, in comparison to similar countries in East Africa, specifically
Ethiopia and Kenya, which have produced significant borehole logs through government geological
surveys and international development organizations. The current global and regional hydrogeological
databases, such as the World Bank (2022) groundwater assessment, and the FAO SWALIM monitoring

31

programme, only offer coarse scale data regarding aquifer systems at the national or regional level, which
is inadequate for mapping the GWPZ for siting boreholes and planning water supplies for the district
scale in Jowhar.
A third gap is related to the methodological literature about GWPZ mapping in data-poor conflict-
stressed contexts. The application and validation of the RS-GIS-AHP approach in areas that lack
comprehensive hydrogeological databases, e.g. Somalia, is challenging because of methodological
difficulties that have not been addressed in the literature. Further research is needed to address questions
related to the selection of the most relevant thematic parameters when detailed geological maps are not
available, as well as to the calibration of AHP weights when borehole yield data is not available in the
country to facilitate cross-reference, and to development of alternative validation methodologies using
proxy indicators of groundwater occurrence, which are frontier challenges to the applicability of
geospatial GWPZ methods in post-conflict humanitarian situations.
2.7.2 Theoretical and Analytical Framework
This study is based on three converging theoretical frameworks: hydrogeological systems theory,
geospatial multi-criteria decision analysis and integrated water resource management. The
hydrogeological systems theory is rooted in the assumption that the occurrence, storage and
transmissivity of groundwater are spatialized functions of the geology, morphology, hydrology and land
use of the landscape, and that these variables can be mapped, quantified, and combined to create a
spatialized model of groundwater potential (Freeze & Cherry, 1979; Scanlon et al., 2023). From
geospatial MCDA theory, the study follows AHP-WIO framework as the major analytical tool for
amalgamating the various thematic layers into a composite groundwater potential index based on the
theoretical arguments presented by Saaty (1980) and rigorously validated in the GWPZ mapping
literature discussed in this chapter.
The study is motivated by the theory of integrated water resource management, which outlines that a
spatially explicit evidence base, based on the principle of maximising groundwater development in areas
of highest hydrogeological potential, can be used to target infrastructure investments for the most
efficient and equitable groundwater access minimising the risk of dry borehole failures and resources
being depleted. Operational context is crucial in this study, as it is both a post-conflict and data-scarce
semi-arid district with predominantly groundwater as the most critical and vulnerable freshwater

32

resource, and reflects the need to achieve universal access to safely managed drinking water as stated in
the Sustainable Development Goals by 2030.
The analytical framework of the study, which is conceptually illustrated in this chapter and
operationalized in Chapter Three involves the sequential derivation of eight thematic layers from the
satellite remote sensing data and ancillary spatial data, namely geology, lineament density,
geomorphology, drainage density, slope, rainfall, soil type, and LULC; the assignment of weights to the
thematic layers through the AHP pairwise comparison process guided by the published literature and
hydrogeological expert knowledge; computation of a composite groundwater potential index using WIO
analysis in a GIS environment; classification of the groundwater potential index into five groundwater
potential zones; lastly, validation of the resulting groundwater potential zone map against available
borehole and well yield data from the Jowhar District. This framework follows the best practice in
GWPZ mapping literature, and is customized to the data availability limitations and hydrogeologic
conditions in the study area

33

CHAPTER THREE
RESEARCH METHODOLOGY
3.1 Study Area
Jowhar District is situated in the Middle Shabelle region of south-central Somalia, some 90 kilometers
north of Mogadishu. The district is located between latitude 2°00’, 3°00’ N and longitude 45°00’, 46°00’
E. This is estimated to cover an area of about 2,600 square kilometers and it includes a low-lying alluvial
plain that is crossed by the Shabelle River, one of the two permanent rivers in Somalia (SWALIM, 2012).
The Shabelle River corridor is historically important with regard to the agricultural economy of the
district, but the annual floods and droughts have caused a lack of stability in livelihoods.

Figure 1: location map of the study area

34

The climatic zone of Jowhar District is a semi-arid to arid climate having a bi-modal rainfall pattern.
The two main rainy seasons locally referred to as the Gu (April-June) and Deyr (October-December)
bring mean annual rainfall of 300 to 450 mm not enough to cater for the increasing domestic and
agricultural water needs of the populace (FAO-SWALIM, 2017). The temperatures are also high,
throughout the entire year, with annual averages between 25°C to 35°C, adding to the scarcity of surface
water due to high evapotranspiration rates.
The geology of the district is largely Quaternary alluvial deposits of Shabelle River flood plain deposits
underlain by the older sedimentary sequences, including limestone, of the Oligocene to Miocene age.
From a hydrogeological point of view, these geological units are regarded as very important because
they might contain unconfined and semi-confined aquifer systems. Even with this potential, there are
still limited systematic hydrogeological studies and borehole data to support the assessment of
groundwater resources and the development of groundwater plans, and remote sensing and GIS are being
used to support groundwater development planning.
1.2 Data Sources
The datasets employed in this study are summarized in Table 1. Landsat 8 OLI/TIRS imagery covering
path 162 and row 052 was retrieved from the United States Geological Survey (USGS) and served as the
basis for deriving lineament density (LD), the normalized difference vegetation index (NDVI), and land
use/land cover (LULC) maps of the study area. Existing geological information was revised using data
supplied by the Ministry of Water Resources and Energy (MOWRE), from which the lithological units
of the area were delineated. A Shuttle Radar Topography Mission Digital Elevation Model (SRTM
DEM) at 30 m spatial resolution was likewise acquired from the USGS website on the same date and
used to extract slope, drainage density, and elevation layers. Precipitation records spanning 2015 –2025
were sourced from the CRU TS v. 4.10 database, which supplies gridded estimates at a 0.5° × 0.5°
resolution. Finally, soil information obtained from the Ministry of Water and Energy was supplemented
with FAO soil data to characterize the soil types of the study area.
Table 1. Summary of Data Sources and Specifications
Data Types Data
Source
Data type and
acquisition data
The function of
the data
Reference

35

SRTM DEM
(30 m)
USGS Grid
2025-03-10
SL, DD, and
EL, maps
https://earthexplorer.usgs.gov/
Rainfall CRU TS
v. 4.10/
0.5° × 0.5°/2015–
2025
Rainfall map https://crudata.uea.ac.uk/cru/dat
a/hrg/cru_ts_4.10/cruts.260409
1129.v4.10/pre/
Soil data FAO Polygon/2023 Soil map Digital Soil Map of the World
v.3.6 - Datasets - "FAO
catalog"
Landsat 8
OLI/TIRS
USGS 2026/02/09 LULC, LD, and
NDVI
maps
https://earthexplorer.usgs.gov/
Geological data USGS Polygon
1:250,000/2020
Lithology map https://www.usgs.gov/centers/ce
ntral- energy-resources-science-
center/science/world-geologic-
maps
Note: USGS –United States Geological Survey; CRU TS vs. 4.10 –Climatic Research Unit gridded Time Series version 4 data; MOWRE
–Ministry of Water Resources and Energy; SL –Slope; DD –Drainage density; EL –Elevation; LULC –land use land cover; NDVI-
Normalized Difference vegetation Index; LD –Lineament density.
3.3 Research Method
Various criteria were employed to identify groundwater potential zones. Raster maps of the relevant
variables were generated using GIS (version 10.8) and Envi (version 5.6) from remote sensing data.
These thematic layers included land use/land cover, slope, elevation, NDVI, lineament density, drainage
density, soil texture, and geology. Atmospheric and radiometric corrections were applied with the Semi-
Automatic Classification Plugin. In GIS, processes such as conversion, interpolation, categorization,
reclassification, augmentation, and filtering were used to process the satellite data. Elevation data from
the Shuttle Radar Topography Mission Digital Elevation Model (SRTM DEM), with a 30 m ×30 m
resolution, improved accuracy in both vertical and horizontal measurements. This data was used to
produce an elevation map in ArcGIS. The elevation information was then used to generate slope raster
data for each grid cell by calculating the maximum rate of change between neighboring cells. Slope
classes were classified using the Slope Spatial Analyser tool. Drainage density, inferred from the slope
data, was calculated to measure the spacing of stream channels, defined as the total stream length per
unit area (Sar et al., 2015). The stream order was determined through flow accumulation and flow
direction operations in ArcGIS. The drainage lines were assessed by evaluating stream lengths gr eater
than 1000 units ("value > 1000"), and drainage density (km/km †) was calculated as shown in Eq. (1).
#$ = ∑ $%&'%=) (1)

36

where Di is the total stream length in stream order ith (km), and A is the basin area (km2). Utilizing PCI
Geomatics 2016, the Landsat 8 OLI/TIRS image (path 163, row 58), exhibiting 5% cloud cover, was
extracted to compute the lineament density of the study region. In recent years, high-resolution satellite
imagery has proven to be an effective technique for mapping groundwater in arid and semi-arid regions.
The thematic layer for lineament density (LD) is defined as the cumulative length of all identified
lineaments relative to the area of the catchment under consideration (Das, 2019a). The calculation of
lineament density was done using Eq. (2).
*$ = ∑ (,%&)'%=) (2)
where Li is the length of the ith lineament, ∑Li is the sum length of all lineaments in kilometers, and A
is the grid size in square kilometers. Lineament extraction was performed with random lines in various
regions based on hillshade analysis. Subsequently, lineaments were extracted using ArcMap 10.8. The
manual extraction of lineaments is more dependable than alternative automated techniques (Alshehri and
Abdelrahman, 2023). The geological, or lithological, thematic map of the study area was downloaded
from USGS and updated with the latest (2015) geological maps of the Garowe district from the Ministry
of Water Resources and Energy, Somalia (MOWRE, 2015). The ESRI land cover of Sentinel-2, at a 10-
meter resolution, was used to determine the recent (2024) LULC map of the study area. We obtained the
soil map of the study area from FAO (2003) data and MOWRE’s Somalia soil map (Table 1). A rainfall
map of this research was created using the inverse distance weighting (IDW) interpolation approach.
The NDVI assesses the greenness of vegetation and serves as a strong indication of temporal changes in
vegetation within a specific area. The NDVI layer was created using the Near Infrared and Red bands in
the Raster Calculator of ArcGIS 10.8 Eq. (3) .
.$/0 =
.01−13$
.01+13$ (3)
where NIR represents the near-infrared band, whereas RED signifies the red band in the reflectanc e
values of the electromagnetic spectrum. Finally, the delineation of the groundwater potential zone
involved converting all data layers into raster format, maintaining a uniform spatial resolution of 30 m
×30 m. The relative importance of each element influencing groundwater potential was assessed by a
pairwise comparison of all features. Each map was categorized using computed weights. The weighted
overlay module in ArcGIS 10.8 was used to create a groundwater potential map by integrating all

37

thematic layers. The groundwater potential zones identified through GIS and RS techniques were
assessed against existing ground data from boreholes, springs, and hand-dug wells to confirm the validity
of the qualitative results. This enhanced confidence in the application of qualitative results derived from
GIS and RS analysis. The overall framework for data processing and analysis is illustrated in Fig. 2

Figure 2. Flowchart for the identification of groundwater potential zones (GWPZs).

38

3.3.1 Thematic Layer Derivation
Nine thematic layers will be developed to represent the spatial factors used in groundwater potential
zone (GWPZ) mapping. These include geology, lineament density, drainage density, slope,
geomorphology, land use/land cover (LULC), soil type, and rainfall. Each thematic layer will be
prepared from the corresponding spatial dataset using ArcGIS 10.8, ENVI, and Google Earth Engine
where appropriate. All datasets will be projected to the Universal Transverse Mercator (UTM) coordinate
system, clipped to the study area boundary, converted to raster format with a common spatial resolution
of 30 m, and reclassified into groundwater suitability classe

论文原文与出版信息

Research Square
查看论文原文

继续探索