Applied Water Science (2022) 12:69

https://doi.org/10.1007/s13201-022-01592-9

ORIGINAL ARTICLE

Suitable dam site identification using GIS‑based MCDA: a case study

of Chemoga watershed, Ethiopia
Yonas Gebresilasie Hagos1 · Tesfa Gebrie Andualem2,3 · Mequanent Abathun Mengie1 · Workineh Tadesse Ayele2 ·
Demelash Ademe Malede4,5

Received: 6 January 2022 / Accepted: 10 February 2022

© The Author(s) 2022

Abstract
Ethiopia is Africa's second-most populous country, after Nigeria, and is primarily a farming community with low produc-
tivity that is heavily reliant on rain-fed agriculture. Water scarcity, global warming, and rising population all necessitate
more effective water conservation methods. As a result, the demand for dams is increasing dramatically in order to provide
the community with safe drinking water, electricity, and irrigation to ensure food security. The goal of this study was to use
remote sensing and geographic information system (GIS) techniques in conjunction with the dam suitability stream model
and multi-criteria decision analysis to identify potential sites for multi-purpose dam construction. The study used six influ-
encing factors to find suitable dam sites, with the model's suitability stream and overall suitability output maps proposed
and evaluated as a result. Based on the topography and land use, the results showed that three proposed dam sites in the
upper part of the watershed are likely preferable for irrigation, fishery, and clean drinking water supply. The three proposed
dam sites in the watershed's lower reaches, however, are better suited to hydropower generation. In addition, remote sensing
and GIS are useful in dam/reservoir site selection because they allow decision-makers to create, manipulate, and manage
relevant thematic layers.

Keywords Chemoga watershed · Dams sites selection · Dam suitability · MCDA · Stream model

Introduction and is the fundamental element for sustainable development

(Guppy & Anderson 2017). Freshwater availability is likely
Water scarcity has been significantly increasing with urban to be one of the major societal challenges of the twenty-first
development and its associated growing population with its century, according to new global goals and commitments
ever-increasing demand because water is the basis of life such as the United Nations Millennium Development Goals
and livelihoods since it supports the health of ecosystems Water Mandate (Gleick 2014). Due to unplanned urbaniza-
tion, limited water resources, and ineffective regulations for
managing water supply and distribution, developing coun-
* Tesfa Gebrie Andualem tries are more vulnerable to water scarcity than developed
tesfag23@gmail.com
countries. To address the scarcity of water resources, it is
1
Irrigation, Drainage and Flood Control Sector, Ethiopian necessary to regulate runoff by building a dam and reservoir
Construction Design and Supervision Works Corporation, (Yuan & Su 1988).
Addis Ababa, Ethiopia Dams are built to store and safely retain large amounts
2
Department of Hydraulic and Water Resources Engineering, of water, which is then released for a variety of purposes
Debre Tabor University, Debre Tabor, Ethiopia including irrigation, hydropower, recreation, water sup-
3
Mawson Lakes Campus, UniSA STEM, University of South ply, flood protection, inland navigation, and so on (Yasser
Australia, Adelaide, Australia et al. 2013). Irrigation, hydropower, recreation, water sup-
4
Hydrology and Water Resources Management, Africa ply, and other projects all require the selection of a dam
Center of Excellence for Water Management, Addis Ababa site (Li 2019). The proper selection of dam sites is benefi-
University, Addis Ababa, Ethiopia
cial to ensuring project safety, reducing construction time,
5
Natural Resources Management Department, Debre Markos and lowering construction costs. As a result, early in the
University, Debre Markos, Ethiopia

13
Vol.:(0123456789)
 69 Page 2 of 26 Applied Water Science (2022) 12:69

Fig. 1  Location of the study area

Table 1  Information of the data Data type Original format Spatial resolution Source of data
sources used in this study sources

Rainfall data Point – Ethiopian meteorological agency

DEM Raster 12.5 m https://​search.​asf.​alaska.​edu/
LULC data Raster 10 m https://​livin​gatlas.​arcgis.​com/​landc​over/
Soil map Vector 1:250,000 MoWR
Geological map Vector 1:250,000 Geological survey of Ethiopia
Road network Vector – http://​geono​de.​wfp.​org

construction process, selecting and evaluating various suit- Sub-Saharan African countries. While there is plenty of
able dam sites are critical (Pan & Zhang 2021). water in the country, only 3% of water resources are acces-
Ethiopia has many rivers, with a likely average of 1575 sible (Keredin & Prasada 2016). The dam suitability stream
cubic meters of available water per capita per year among model (DSSM) and GIS-based MCDA techniques were used

Table 2  USDA-Soil Hydrologic soil group Soil textures Runoff potential Infiltration rates
conservation service (SCS) (HSG)
classification system, (1974);
ERA (2013) A Sand, loamy sand or sandy loam Low High
B Silt loam, or loam Moderately low Moderate
C Sandy clay loam Moderately high Slow
D Clay loam, silty clay loam, sandy clay, High Very slow
silty clay, or clay

13
Applied Water Science (2022) 12:69 Page 3 of 26 69

Table 3  Mean annual rainfall data for surrounding stations agricultural productivity in the community through irriga-
Station Location Mean annual rainfall
tion, clean drinking water, and other means. Many research-
(mm) (1986–2018) ers use GIS-based MCDA to identify a suitable area to build
Longitude Latitude
a dam/reservoir in their areas of interest, such as topography,
Debre Markos 37′ 43′ 44'' 10′ 19′ 32'' 1329.3 climate, hydrology, soils, agronomy, and socioeconomic fac-
Dembech’a 37′ 28′ 47'' 10′ 33′ 44'' 1364.1 tors (Luís & Cabra 2021; Nyirenda et al. 2021; Mohamed
Debre Elias 37′ 27′ 54' 10′ 17′ 34'' 1884.7 et al. 2021; Khudhair et al. 2020; Vikas & Tallavajhala 2020;
Yejube 37′ 44′ 37'' 10′ 9′ 14'' 1370.4 Zhenfeng et al. 2020; Omid et al. 2019; Rami et al. 2019;
Njiru & Siriba 2018; Faez &Abdul 2015; FAO 2003).
This study aimed to find suitable dam sites in the
to generate the suitable dam/reservoir site map in this study. Chemoga watershed, in order to assist decision-makers in
Chemoga watershed was one of the chosen watersheds, selecting the best location for a dam (s). This study assists
and it aids in the management and use of available water water resource planners, designers, and decision-makers
resources in order to alleviate water scarcity and increase involved in dam building planning to reinforce their study or

Fig. 2  Workflow of this study

13
 69 Page 4 of 26 Applied Water Science (2022) 12:69

emphasize their decision. Furthermore, it may prompt them varied between 1329.3 mm in Debre Markos and 1884.7 mm
to update their work in light of the proposed prospective dam in Debre Elias.
locations in this study. Then, by using the methodology out-
lined below, suitable dam site maps were generated based on Data sources
the selected influencing factors. Aside from that, this study
proposed several dam sites with high suitability and com- For this study, data such as digital elevation model (DEM),
puted their cross sections as well as some other parameters, hydrologic soil groups, rainfall, land use/cover, and geologi-
such as dam height and width, reservoir volume, 2D surface cal data were collected. The digital elevation model (DEM) of
area, 3D surface area, and catchment area. 12.5 m resolution downloaded from https://​search.​asf.​alaska.​
edu/ website is used for delineating the study watershed, gen-
erating slope and stream order maps. The land use/cover map
Materials and methods was downloaded from https://​livin​gatlas.​arcgis.​com/​landc​over/
with a 10 m resolution for the year 2020. The rainfall data for
Description of the study area the surrounding four weather stations (Debre Markos, Dem-
becha, Debre Elias, and Yejube) for a period extending from
The Chemoga watershed shown in Fig. 1 below covers an 1986 to 2018 was collected from Ethiopian Meteorological
area of 1162.02 ­km2 and found in the Amhara region, north- Agency, whereas, the soil map and geological map of the study
west Ethiopia, and geographically located between 10° 0′ area were obtained from the Ethiopian Ministry of Water, Irri-
30′′–10° 38′ 45′′ N latitude and 37° 26′ 30′′–37° 53′ 30′′ E gation and Energy (MoWIE) and the Geological Survey of
longitude. The river originates around Robe Gebeya town at Ethiopia respectively.
an altitude of 3951 m above mean sea level with slope varies
from almost flat to very steep and steadily decreases south- Selection criteria and methodology
westwards to its confluence with the Blue Nile River at an
altitude of 877 m a.m.s.l. The 33 years period (1986 to 2018) According to the Food and Agriculture Organization (FAO),
point rainfall data of four meteorological stations (Debre six principal factors for selecting potential dam and reservoir
Markos, Dembech’a, Debre Elias, and Yejube) that are found zone were topography; climate; hydrology; soils; agronomy;
within and around the watershed was collected from Ethio- and socioeconomics (FAO 2003). Therefore, six influenc-
pian Metrology agency. As the data received were monthly ing thematic layers were used, such as stream order/drainage
rainfall, annual rainfall for each year at each station, and density, slope; Runoff potential; land use/cover, geology, and
mean of 33 years for each station were calculated and then distance to the road. The method used in this study was GIS-
interpolated to inverse distance weight (IDW) in ArcGIS based MCDA and dam suitability stream model (DSSM) is
(Tables 1, 2, 3), the mean annual rainfall in the study region based on the stream order for identifying potential sites for

Fig. 3  Stream order map of the

study area

13
Applied Water Science (2022) 12:69 Page 5 of 26 69

Fig. 4  Drainage density map of

the study area

Fig. 5  Slope map of the study

area

constructing a dam. The model follows five main steps as platform; fifth, and finally, selecting and evaluating of dam site
shown in (Fig. 2) below. First, selection and preparation of based on the dam evaluation parameters. The workflow steps
affecting factors in raster layer format; second, reclassifica- of the study are discussed briefly below in Fig. 2.
tion of the influencing factors; third, assigning weight for
each influencing factor; fourth, overlay analysis in the ArcGIS

13
 69 Page 6 of 26 Applied Water Science (2022) 12:69

Fig. 6  a Land use/cover, b Hydrologic soil group, and c Curve number (CN) maps of the study area

Fig. 7  Rainfall map of the study

area

13
Applied Water Science (2022) 12:69 Page 7 of 26 69

Fig. 8  Runoff depth map of the

study area

Fig. 9  Geological formation

map of the study area

data using hydrology tools under the ArcGIS spatial ana-

Data analysis
lyst tools. Then generate the drainage density map for
this study area using line density tool in spatial analyst
Preparation and reclassification of factors
tools. Likewise, the slope map was prepared from DEM
in the ArcGIS platform spatial analyst tools using slope
A watershed modeling approach was applied to delineate
sub tool under surface tools, and the slope’s percentage
the Chemoga River watershed in the ArcGIS 10.3 plat-
was classified into five classes. In this study, the SCS-CN
form by using the 12.5 m resolution DEM. The drainage
method, a commonly used method for estimating direct
density map of the study area was obtained from DEM
runoff from a watershed, was used to generate the runoff

13
 69 Page 8 of 26 Applied Water Science (2022) 12:69

depth map of the watershed (USDA 1972). The land use/ In both of these two methods, stream order is idealized as a
cover map for the study area is classified into six classes, tree with a strong root and slenderer branches.
including water bodies, grassland, Shrubland, Cropland, According to the “top-down” system devised by Strahler,
Forestland, and settlements. The distance to access roads rivers of the first order are the outermost tributaries. If two
was created by the Euclidean distance tool in ArcGIS. rivers with different stream order merge, the resulting stream
Each criterion’s classification in a different class is based is given the higher of the two numbers (Strahler 1957). The
on expert opinions in the reclassification tool in ArcGIS Shreve method also gives the outermost tributaries the num-
(Engineers, Geologists, Hydrologists, and GIS experts) ber “1”. Unlike the Strahler method, at a confluence, the two
and literature (Zhenfeng et al. 2020; Ali et al. 2018; Njiru numbers are added together (Shreve 1966). In this study, the
& Siriba 2018). Strahler method of quantifying stream order is used (Fig. 3).
Then as the suitable dam site selection viewpoint, stream
Stream order and drainage density first order is highly suitable for constructing a dam and vice
versa.
Different methods for quantitative stream order are sug- Drainage density is defined as the total length of the chan-
gested by many researchers. The most used ones are Strahler nel in a drainage basin divided by the watershed area (Hor-
method (Strahler 1957) and Shreve method (Shreve 1966). ton 1932) and represented by the following equation:

Fig. 10  Distance from road

network map of the study area

Table 4  Saaty’s scale of relative importance (Saaty, 1980; Saaty and Vargas 1991)
Intensity of importance Definition Explanation

1 Equal importance Two activities contribute equally to the objective

3 Moderately importance Experience and judgment slightly favor one activity over
another
5 Strongly more important/much more important Experience and judgment strongly favor one activity over
another
7 Very strongly/far more important Demonstrated importance an activity is strongly favored and
its dominance demonstrated in practice
9 Extremely more important The evidence favoring one activity over another is of the
highest possible order of affirmation
2,4,6,8 Intermediate values between the two adjacent judgments When compromise is needed
Reciprocals Values for inverse comparison

13
Applied Water Science (2022) 12:69 Page 9 of 26 69

Table 5  Pairwise comparison Factor Stream order Slope Runoff poten- Land use Geology Dis-
of seven criterion matrix (Saaty, tial (mm) tance to
1980; Saaty and Vargas 1991) road

Stream Order 1
Slope 1/2 1
Runoff potential (mm) 1/3 1/2 1
Land use 1/5 1/4 1/3 1
Geology 1/7 1/6 1/5 1/3 1
Distance to road 1/9 1/8 1/5 1/5 1/3 1

Table 6  Pairwise comparison of Factor Stream order Slope Runoff Land use Geology Distance to road
seven criterion decimal matrix potential
(Saaty, 1980; Saaty and Vargas (mm)
1991)
Stream order 1.00 2.00 3.00 5.00 7.00 9.00
Slope 0.50 1.00 2.00 4.00 6.00 8.00
Runoff potential (mm) 0.33 0.50 1.00 3.00 5.00 7.00
Land use 0.20 0.25 0.33 1.00 3.00 5.00
Geology 0.14 0.17 0.20 0.33 1.00 3.00
Distance to road 0.11 0.13 0.14 0.20 0.33 1.00
Sum 2.29 4.04 6.68 13.53 22.33 33.00

Table 7  Normalized pairwise Factor Stream order Slope Runoff Land use Geology Distance to road
matrix calculated potential
(mm)

Stream order 0.44 0.49 0.45 0.37 0.31 0.27

Slope 0.22 0.25 0.30 0.30 0.27 0.24
Runoff potential (mm) 0.15 0.12 0.15 0.22 0.22 0.21
Land use 0.09 0.06 0.05 0.07 0.13 0.15
Geology 0.06 0.04 0.03 0.02 0.04 0.09
Distance to road 0.05 0.03 0.02 0.01 0.01 0.03

Table 8  Determined relative criterion weights

Factor Stream order Slope Runoff poten- Land use Geology Distance to road Criteria weight
tial (mm)

Stream order 0.44 0.49 0.45 0.37 0.31 0.27 0.39

Slope 0.22 0.25 0.30 0.30 0.27 0.24 0.26
Runoff potential (mm) 0.15 0.12 0.15 0.22 0.22 0.21 0.18
Land use 0.09 0.06 0.05 0.07 0.13 0.15 0.09
Geology 0.06 0.04 0.03 0.02 0.04 0.09 0.05
Distance to road 0.05 0.03 0.02 0.01 0.01 0.03 0.03

∑n where Dd is the drainage density, n is the number of streams,

L
(1)
i=1
Dd =
Abasin L is the stream length (km), and A is the drainage basin (km).
The drainage density map is generated from digital eleva-
tion model (DEM). The range of drainage density map in the

13
 69 Page 10 of 26 Applied Water Science (2022) 12:69

Table 9  The Eigenvector weights of each flood factor obtained after map was generated using the slope tool under the surface
the pairwise comparison group in the ArcGIS platform using 12.5 m resolution DEM.
Factor Normalized weight Influence (%) As shown in Fig. 5 below, the range of slope in the
Chemoga watershed varies from 0 to 513.4%. It is classi-
Stream order 0.39 39.0
fied into five intervals: almost flat (S < 3) and it covers 4.0%
Slope 0.26 26.2
area of the study area; gentle (3 < S ≤ 8) and it covers 24.8%
Runoff potential (mm) 0.18 17.9
area of the study area; moderate (8 < S ≤ 16) and it covers
Land use 0.09 9.3
29.9% area of the study area; steep (16 < S ≤ 28) and it cov-
Geology 0.05 4.9
ers 22.5% area of the study area; very steep (S > 28) and it
Distance to road 0.03 2.7
covers 18.8% area of the study area (Mohamed et al. 2021).
Sum 1.00 100

Rainfall‑runoff modeling (SCS‑CN)

Chemoga watershed varies from 0 to 4.95 ­km−1 as shown The SCS-CN method has been established in 1954 by
in Fig. 4. The developed drainage density map was grouped USDA SCS (Rallison 1980), defined in the Soil Conserva-
into five intervals according to their importance for dam site tion Service (SCS) by the National Engineering Handbook
suitability. These intervals are as follows: extremely suitable (NEH-4) Section of Hydrology to estimate surface runoff
(Dd > 4) and it covers 0.3% area of the study area; very suit- (Ponce & Hawkins 1996).
able (3 < Dd ≤ 4) and it covers 4.5% area of the study area; The Soil Conservation Service Curve Number approach
Adapted (2 < Dd ≤ 3) and it covers 18.0% area of the study is frequently used empirical methods to estimate the direct
area; Less suitable (1 < Dd ≤ 2) and it covers 34.3% area of runoff from a watershed (USDA 1972) in the study area.
the study area; Unsuitable (Dd < 1) and it covers 42.9% area The infiltration losses are combined with surface storage
of the study area. by the relation of:

Slope (S) P − Ia2

Q= ( ) (2)
P − Ia +S
The slope of an area is the main factor to select a suitable
dam site, and it significantly affects the runoff, recharge, and where Q = accumulated direct runoff (mm), P = accumu-
movement of surface water as well as the amount of sedi- lated rainfall (potential maximum runoff) (mm), Ia = ini-
mentation (Adham et al. 2016; Kadam et al. 2012). A slope tial abstraction including surface storage, interception, and

Table 10  Random inconsistency n 1 2 3 4 5 6 7 8 9 10

indices
RI 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.46 1.49

Table 11  Determined consistency ratios (CR)

Factor Stream order Slope Runoff potential Land use Geology Distance to road Weighted Criteria weight Consist-
sum value ency
measure

Stream order 0.44 0.49 0.45 0.37 0.31 0.27 2.34 0.39 6.42
Slope 0.22 0.25 0.30 0.30 0.27 0.24 1.57 0.26 6.48
Runoff potential 0.15 0.12 0.15 0.22 0.22 0.21 1.08 0.18 6.42
Land use 0.09 0.06 0.05 0.07 0.13 0.15 0.56 0.09 6.20
Geology 0.06 0.04 0.03 0.02 0.04 0.09 0.29 0.05 6.04
Distance to road 0.05 0.03 0.02 0.01 0.01 0.03 0.16 0.03 6.09
Total 1.00 1.00 1.00 1.00 1.00 1.00 CI 0.06
RI 1.24
CR 0.04
CR < 0.1
Consistency is acceptable

13
Applied Water Science (2022) 12:69 Page 11 of 26 69

Table 12  Weights of the S. No Factor Features Local weight Average weights

thematic layers and runoff map
with their assigned weights 1 Stream order 1st order 3.5 39.0
2nd order 6.9
3rd order 13.6
4th order 28.6
5th order 47.4
Drainage density (km/km2) Dd < 1 3.5
1 < Dd ≤ 2 6.8
2 < Dd ≤ 3 13.4
3 < Dd ≤ 4 26.0
Dd > 4 50.3
2 Slope (%) S<3 50.3 26.2
3<S≤8 26.0
8 < S ≤ 16 13.4
16 < S ≤ 28 6.8
S > 28 3.5
3 Runoff potential (mm) R > 1400 41.8 17.9
1400 ≥ R > 1300 26.4
1300 ≥ R > 1200 16.5
1200 ≥ R > 1100 10.4
R < 1100 4.9
4 Land use/landcover Barren land 35.1 9.3
Grassland 23.9
Shrubland 16.2
Cropland 11.0
water bodies 7.6
Forestland 4.0
Settlement 2.3
5 Geology Termaber basalts 31.6 4.9
Blue Nile basalts 26.4
Post-tectonic granites 14.6
Undifferentiated lower complex 10.1
Adigrate sandstone 7.0
Infra-adigrate clastics 4.7
Colluvium 3.2
Alluvium 2.3
6 Distant to roads (Km) <5 41.6 2.7
5 < S ≤ 10 26.2
10 < S ≤ 15 16.1
15 < S ≤ 20 9.9
> 20 6.2

infiltration prior to runoff (mm), and S = potential maximum Substituting 0.2S for Ia in Eq. 5, the SCS rainfall-runoff
retention (mm). equation becomes:
The relationship between Ia and S was developed from
experimental catchment area data. It removes the necessity (P − 0.2S)2
Q= (4)
for estimating Ia for common usage. The empirical relation- (P − 0.8S)
ship used in the SCS runoff equation is as follows:
S is related to the soil and cover conditions of the catch-
Ia = 0.2S (3) ment area through the CN. CN has a range of 0–100, and S
is related to CN by:

13
 69 Page 12 of 26 Applied Water Science (2022) 12:69

Table 13  Evaluation and essential parameters of the proposed dam sites

Parameters/proposed dams Dam-1 Dam-2 Dam-3 Dam-4 Dam-5 Dam-6

Location Longitude 363,221.05 349,405.30 348,241.29 333,927.34 343,980.52 337,751.36

Latitude 1,154,248.30 1,136,062.56 1,121,416.40 1,117,505.25 1,114,987.87 1,108,565.20
Stream order 3rd order 3rd order 4th order 4th order 4th order 4th order
Dam evaluation parameters Base elevation (m) 2432 2186 1243 1156 1147 992
Surface elevation (m) 2440 2200 1280 1220 1160 1040
Dam height (m) 8 14 37 64 13 48
Dam width (m) 245 186 875 344 173 187
2D Surface area(m2) 519,910 610,038 2,318,042 1,304,522 319,000 2,228,951
3D Surface area ­(m2) 522,441 612,292 2,328,583 1,359,061 321,124 2,266,706
Max volume ­(m3) 2,051,451 2,153,478 28,598,758 31,476,566 1,677,221 29,213,232
Catchment area (ha) 12,885 4384 63,620 28,562 71,788 80,394
Land Use coverage at full Barren land 0.0 0.0 0.0 0.0 1.9 0.0
reservoir condition (ha) Grassland 5.0 13.7 0.0 0.0 0.0 0.0
Shrubland 0.0 0.0 40.8 99.3 98.1 99.0
Cropland 95.0 82.6 56.5 0.0 0.0 0.0
water bodies 0.0 0.0 0.0 0.0 0.0 0.1
Forestland 0.0 0.4 2.7 0.7 0.0 0.9
Settlement 0.0 3.3 0.0 0.0 0.0 0.0

Fig. 11  Proposed dam sites (left) dam sites with suitability on stream (right) dam sites with overall suitability

13
Applied Water Science (2022) 12:69 Page 13 of 26 69

( ) Curve number (CN)

25400
S= − 254 (5)
CN
The CN (a dimensionless number ranging from 0 to 100)
is determined based on the hydrologic soil group and land
Hydrological soil groups (HSG) cover of the specific watershed. In the present research work
of Chemoga watershed, the CN values have been estimated
The USDA-Soil Conservation Service has developed four based on LULC classes of hydrologic soil groups A, B, C,
types of hydrologic soil groups based on infiltration rates and D, and the CN ranges vary between 35 and 100 (Fig. 6c).
namely A, B, C, and D in the soil classification system. The
detailed description of the four hydrologic soil groups based Rainfall analysis
on infiltration rate, texture, and runoff potential is discussed
as follows in (Table 2). For a rainfall-runoff factor, point rainfall data for 33 years
The Hydrological Soil Groups (HSG) map of the study (1986–2018) were collected from four stations (Debre
area was obtained from the Ethiopian Ministry of water Markos, Dembech’a, Debre Elias, and Yejube) within and
resources (MoWR) and processed by ArcGIS to convert the around the watershed were received from the Ethiopian
vector data to raster. Three hydrological soil groups (Fig. 6b) Metrology agency. As the data received were monthly rain-
have been found in the present study of Chemoga water- fall, annual rainfall for each year at each station, and mean
shed HSG—A and it covers 20.3% area of the study area; of 33 years for each station were calculated and then inter-
HSG—B and it covers 72.5% area of the study area and polated to inverse distance weight (IDW) in ArcGIS.
HSG—D and it covers 7.3% area of the study area.

Fig. 12  Evaluation parameters for proposed Dam site 1: a Catchment area b DEM c slope d triangulated irregular network (TIN), e contours, f
LULC

13
 69 Page 14 of 26 Applied Water Science (2022) 12:69

Fig. 13  3D view and cross sec-

tion of proposed dam site-1

Estimation of rainfall‑runoff (Debre Elias station). By using Eqs. 6 of the SCS-CN

method, it is estimated that the runoff rate varies between
The average annual surface runoff depth was determined 927 and 1,400 mm (Fig. 7). The runoff depth map of the
by using the SCS-CN method. It has revealed from the last study area was classified into five classes: these were R >
thirty-three years of rainfall data of four stations within 1400 mm, which covers 9.4% of the study area, 1400 mm ≥
and around the watershed. The mean annual rainfall var- R > 1300 mm, which covers 42.2% of the study area, 1300
ies between 1329.3 (Debre Markos station) and 1783.8 mm mm ≥ R > 1200 mm that covers 34.2% of the study area,

13
Applied Water Science (2022) 12:69 Page 15 of 26 69

Fig. 14  Evaluation parameters for proposed Dam site 2: a Catchment area, b DEM, c slope, d triangulated irregular network (TIN), e contours, f
LULC

1200 mm ≥ R > 1100 mm covering 8.3% of the study area, land use/ land cover type for dam and reservoir sites were
and R < 1100 covering 5.9% of the study area (Fig. 8). barren lands, grassland, and shrubland covers 41,004.6 ha
(34.6%) of the total study area.
Land use/land cover
Geology
Land use is an essential feature in surface runoff generation
(Jha et al. 2014). Land use/cover of an area has an extreme The geological character and rock type within a specific
impact on runoff velocity, infiltration process, evapotran- region affect the permeability of the dam which includes
spiration, and these processes have been played a vital role the capability of holding water for the dam (Marinos et al.
in the delineation of a suitable zone for the dam site. The 1997). Foundations are better on igneous rocks and hard
land use land cover map of the Chemoga watershed shown metamorphic rocks like granite, gneiss, quartzite, etc. The
in (Fig. 6a) expresses that there are seven major types of geological features of the Chemoga watershed were col-
land use/cover namely, Barren land (0.04%); Grassland lected from the Geological Survey of Ethiopia.
(2.65%); Shrubland (31.94%); Cropland (50.01%); water In this study, geological features have been identified into
body (0.05%); Forestland (4.55%) and Settlement (10.76%) eight units, such as Termaber Basalts, Blue Nile Basalts,
with their area coverage. Consequently, the most promising Adigrate Sandstone, Alluvium, Infra-Adigrate Classics,

13
 69 Page 16 of 26 Applied Water Science (2022) 12:69

Fig. 15  3D view and cross sec-

tion of proposed dam site-2

Colluvium, post-tectonic granites, Undifferentiated Lower Distance to the roads

Complex, and it was shown in Fig. 9. In the study area, we
have been found that Termaber Basalts spreads maximum The distance to the road network and rail is an essential
area and it was about 34.7% of the total area. Other features factor for selecting a suitable dam site as it is one of the
such as Blue Nile Basalts covers 23.4%; Adigrate Sandstone influencing socioeconomic criteria for site selection. It is
covers 20.2%; Alluvium covers 6.4%; Infra-Adigrate Clas- assumed that sites located far from the road networks were
sics covers 6.1%; Colluvium covers 5.0%; post-tectonic unsuitable for dam construction because it costs a large
granites cover 3.0%, and Undifferentiated Lower Complex amount of money to construct access roads and influence
covers 1.3% of the total area.

13
Applied Water Science (2022) 12:69 Page 17 of 26 69

Fig. 16  Evaluation parameters for proposed Dam site 3: a Catchment area, b DEM, c slope, d triangulated irregular network (TIN), e contours, f
LULC

the management process and vice versa (Njiru & Siriba judgment matrices to allocate weights (Tables 4, 5, 6, 7,
2018; Dorfeshan et al. 2014). A road network map for 8, 9, 10, 11) and the thematic layers of each level/criterion
the study area was collected from http://​geono​de.​wfp.​org class, and measure their relative importance using Saaty's
prepared for Humanitarian access by the United Nations s 1–9 scale. Weights were assigned to each factor class to
Office for Humanitarian Affairs (OCHA). The distance express the importance or preference of each factor class
from the road map of the study area has been classified relative to the other factor classes in generating suitable
into five classes: these were < 5 km and it covers 38.8% dam sites. This was done using related review literature,
area of the study area, 5 to 10 km and it covers 22.7% area field observation, and expert judgment to fill out a pair-
of the study area, 10 to 15 km and it covers 11.8% area of wise comparison matrix from which Eigenvectors and
the study area, 15 to 20 km and it covers 8.2% area of the consistency ratios were generated for each of the criteria
study area, > 20 km covers 18.5% area of the study area as being considered. The factors for choosing a suitable dam
shown in (Fig. 10). site are rated on a scale of 1 to 9, with 1 indicating equal
importance and 9 indicating one factor is more important
Analytic hierarchy process (AHP) methodology than the other. One is less important than the other when
the reciprocal of 1 to 9 (1/1 and 1/9) is used (Saaty 1980;
AHP is the most popular MCDM method for determin- Saaty & Vargas 1991). The basic steps to determine the
ing the weight of criteria or factors. In AHP, construct

13
 69 Page 18 of 26 Applied Water Science (2022) 12:69

Fig. 17  3D view and cross sec-

tion of proposed dam site-3

indicator's weight and consistency ratio (CR) were final- where n denotes the nth row and m denotes the mth column
ized on the assignment of weights to different criteria. elements of the judgment matrix.
Step 1. Establishment of judgment matrices (P) by pair- Step 2. Calculation of normalized weight.
wise comparison. This step is to normalize the matrix by totaling the
numbers in each column. Each entry in the column is then
⎛ P11 P12 ⋯ P1n ⎞
divided by the column sum to yield its normalized score.
⎜P P22 ⋯ P2n ⎟
P = ⎜ 21 (6) The sum of each column is 1.
⎜ ⋮ ⋯ ⋱ ⋮ ⎟⎟
( /∑ni )
⎝ Pn1 Pn2 ⋯ Pmn ⎠ Wn = GMn
n=1
GMn (7)

13
Applied Water Science (2022) 12:69 Page 19 of 26 69

Fig. 18  Evaluation parameters for proposed Dam site 4: a Catchment area, b DEM, c slope, d triangulated irregular network (TIN), e contours, f
LULC

where the geometric mean of the ith row of the judgment ∑

ni
(PW)n
matrices is calculated as: 𝜆max = (11)
ni wn
√ n=1

GMn = ni P1n P2n LPmni (8) where W is the weight vector (column). Random index (RI)
can be obtained from standard tables (Saaty 1980). In prac-
Step 3. Calculates a consistency ratio (CR) to verify the tice, a CR of 0.1 or below is considered acceptable. Any
coherence of the judgments. Now, calculate the consist- higher value at any level indicates that the judgments war-
ency ratio and check its value. The purpose of doing this rant re-examination.
is to make sure that the original preference ratings were Weight values were assigned for each factor and their
consistent. future classes based on their influence in selecting suitable
CI dam sites, with the most important factor receiving the high-
CR = (9) est weight and vice versa (Tables 12, 13). Order/drainage
RI
density is the most important factor in this study stream,
Consistency index (CI) is denoted as follows: with a weight of 39%, followed by slope, runoff potential,
land use, geology, and distance to the road, with weights of
𝜆max − ni
CI = (10) 26.2, 17.9, 9.3, 4.9, and 2.7%, respectively.
ni − 1

Max is the eigenvalue of the judgment matrix and it is

calculated as:

13
 69 Page 20 of 26 Applied Water Science (2022) 12:69

Fig. 19  3D view and cross sec-

tion of proposed dam site-4

Weighted overlay analysis stream order layer in favor of the drainage density layer
in order to obtain the study area's overall suitability map.
The potentially suitable dam site map was created using a
weighted index overlay analysis for the Chemoga water-
shed by adding the weight values of each thematic layer Result and discussion
and taking into account six influencing factors. The data
for each influencing factor were gathered from various Dam site suitability map
sources and analyzed using the ArcGIS Arc Map 10.3.1
platform as a geo-rectified thematic layer. Then, to gener- Two different dam suitability maps are finally obtained as
ate the suitability on the stream map, the first of two final "suitability on stream" and "overall suitability" for construct-
output maps is obtained by overlaying the influencing fac- ing a dam (Fig. 11). Both maps are divided into five levels
tors, including the stream order. Second, we ignore the of suitability: very high, high, moderate, less, and least suit-
able sites. As previously stated, the suitability of the stream

13
Applied Water Science (2022) 12:69 Page 21 of 26 69

Fig. 20  Evaluation parameters for proposed Dam site 5: a Catchment area, b DEM, c slope, d triangulated irregular network (TIN), e contours, f
LULC

map was obtained by overlaying the factors as a reclassified classed as unsatisfactory because there was no need to
raster layer and assigning the highest weight to it using the build a new dam there. The fourth location, located near
stream order factor. As a result, the analysis selected suitable the watershed's exit, is regarded as unacceptable since it
pixels on the streams, but other layers, such as slope, runoff is so close to the Blue Nile River, and constructing a dam
potential, LULC, geology, and distance to roads, were also in that region may affect future dams and reservoirs across
considered. the Blue Nile. However, if no dam can be built on the Blue
Fig. 11 shows that the first and third sites (counting Nile's downstream section, it may be considered as a suit-
from top to bottom) are undesirable because their cross- able dam location.
sectional widths are quite lengthy and span vast reservoir
regions, including agriculture and rural villages. As a Evaluation of proposed dam sites
result, the environmental and socioeconomic standards are
not met. The second site (counting from top to bottom) is The proposed dam sites are evaluated using nine parameters:
similarly located in a very favorable area, both in terms 3D surface area, 2D surface area, the maximum volume of
of stream compatibility and overall suitability; however, the reservoir, dam base elevation, dam surface elevation,
it is categorized as an unacceptable site due to the pres- dam height, dam width, catchment area, and contour close-
ence of an existing dam in that location. As a result, it was ness (Figs. 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23). The

13
 69 Page 22 of 26 Applied Water Science (2022) 12:69

Fig. 21  3D view and cross sec-

tion of proposed dam site-5

clipped reservoir coverage DEM is used to create contour There are large suitable areas in the northeast (middle)
maps with 1 m and 5 m contour intervals. Then, using the and around the outlet of the study area, as shown above
contours, a triangulated irregular network (TIN) was cre- in the overall suitability map in Fig. 11 (right). However,
ated to determine the reservoir's 2D, 3D surface area, and because of the existing dam/reservoir in the middle of the
volume. The surface volume sub tool in the 3D analyst tool watershed, which is mostly covered by rural settlements,
in the ArcMap platform is used to determine the reservoir's there is no need to build a new dam mean that” there is
2D, 3D surface area and volume, as well as the cross section an existing dam in that location and the model grouped as
(height and width) of the proposed dam, as shown in the fig- unsatisfactory. Areas with existing dams are considered
ures below. Finally, hydrological tools in the Spatial Analy- unsatisfactory because the model categorises locations as
sis Tools are used to generate the watershed using the dam good or unsatisfactory based on the likelihood of future dam
location as an outlet point to measure the catchment area. construction. The suitable area near the watershed's outlet

13
Applied Water Science (2022) 12:69 Page 23 of 26 69

Fig. 22  Evaluation parameters for proposed Dam site 6: a Catchment area b DEM c slope d triangulated irregular network (TIN), e contours, f
LULC

is also unsatisfactory because it is so close to the Blue Nile Validation of the dam site suitability map
River, and building a dam at the Chemoga watershed's out-
let may be influenced by future dams and reservoirs on the In various parts of the world, researchers are using vari-
Blue Nile. Furthermore, there are suitable areas at low-order ous models to find suitable dam sites, but it is critical to
streams that are otherwise unsuitable. Then, in this study, six adequately validate the output of the models with real-world
suitable dam sites in the Chemoga watershed were identified ground conditions or recorded observations. When deter-
by evaluating both stream map and overall suitability, as well mining the suitability of a dam site, it may be preferable to
as checking the listed dam evaluation parameters, with four validate the model output with existing dams in the water-
unsatisfactory sites that may be preferable for ponds and shed (Odiji et al. 2021). As a result, in this dam site suit-
small rainwater harvesting locations. ability study, the results obtained from the used model are
compared to the area with an existing dam and reservoir,

13
 69 Page 24 of 26 Applied Water Science (2022) 12:69

Fig. 23  3D view and cross sec-

tion of proposed dam site-6

which is selected and built with a detailed design. As shown multi-criteria decision analysis. The output suitability map
in Fig. 24, an existing dam in the watershed was built in the was created using six influencing thematic layers, including
model's generated output's very high suitable dam site loca- stream order/drainage density, slope, Runoff potential, land
tion. The model is then found to be appropriate for identify- use/cover, geology, and distance to the road. According to
ing dam site locations at the desk study level. related review literature, field observation, and expert judg-
ment, the most important features in identifying a suitable
dam site were stream order/drainage density (39%), slope
Conclusions (26.2%), and runoff potential (17.9%). Stream order is
important among the other features because it ensures that
The Chemoga watershed was investigated using remote enough water is available to be stored in the dam. Finally,
sensing and geographic information system (GIS) tech- based on the two output suitability maps, six potential dam
niques, as well as the dam suitability stream model and sites were proposed (Suitability on streams and Overall
suitability).

13
Applied Water Science (2022) 12:69 Page 25 of 26 69

Fig. 24  Existing dam site location comparison with model output for validation

The dam height ranges from 8 to 64 metres depending on Declarations

the cross section of the dam axis; the dam width ranges from
173 to 875 metres; the reservoir maximum storage capacity Conflict of interest There is no conflict of interest between authors.
ranges from 1.68 to 31.48 million cubic metres depending on
the distribution of topographic conditions in the surround- Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
ing area; and the reservoir 2D surface area ranges from 3.19 tion, distribution and reproduction in any medium or format, as long
to 231.8 ha among the proposed six dams. Settlements and as you give appropriate credit to the original author(s) and the source,
most of the land used for rain-fed agricultural purposes sur- provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
round the proposed dam sites in the upper part of the study
included in the article's Creative Commons licence, unless indicated
area, such as dam sites 1, 2, and 3. As a result, building a otherwise in a credit line to the material. If material is not included in
dam provides numerous benefits to the local community. the article's Creative Commons licence and your intended use is not
Proposed dam sites 4, 5, and 6 in the lower part of the water- permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
shed, based on topographic location and availability of irri-
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
gable command area, are not preferred for irrigation but are
most suitable for generating hydropower electric, fishery,
and recreation.
References
Author contributions YG Hagos, TG Andualem, and MA Mengie
were involved in the conceptualisation; YG Hagos and TG Andualem Adham A, Riksen M, Ouessar M, Ritsema CJ (2016) A methodology to
were involved in the methodology; YG Hagos and MA Mengie were assess and evaluate rainwater harvesting techniques in (semi-) arid
involved in the software; YG Hagos and TG Andualem were involved regions. Water 8:198. https://​doi.​org/​10.​3390/​w8050​198
in the validation; YG Hagos and TG Andualem were involved in the Ali J, Babak A, Mohsen H, Ian F, Mohammad K, Nastaran K, Erfan GT
resources; DA Malede and WT Ayele were involved in reviewing and (2018) A comparative study of the AHP and TOPSIS techniques
correcting the draft manuscript. for dam site selection using GIS: a case study of Sistan and Bal-
uchestan province, Iran. Geosciences 8:494
Dorfeshan F, Heidarnejad M, Bo A, (2014) Locating suitable sites
Funding The author(s) received no specific funding for this work.
for construction of underground dams through analytic hierarchy
process. In: international conference on earth, environment and
life sciences (EELS-2014). Dubai (UAE), pp 23–24

13
 69 Page 26 of 26 Applied Water Science (2022) 12:69

Ethiopian Road Authority (ERA) (2013) Drainage design manual. Nyirenda AM, Gumindoga W, Shumba A (2021) A GIS-based
Addis Ababa, Ethiopia approach for identifying suitable sites for rainwater harvesting
Faez HB, Abdul RMS (2015) Selection of rainwater harvesting sites by technologies in Kasungu District Malawi. Water SA 47(3):347–
using remote sensing and gis techniques: a case study of Kirkuk, 355. https://​doi.​org/​10.​17159/​wsa/​2021.​v47.​i3.​11863
Iraq. J Teknol 76(15):75–81 Odiji C, Adepoju M, Ibrahim I et al (2021) Small hydropower dam site
FAO (2003) Land and water digital media series, 26. Training course suitability modeling in upper Benue river watershed. Nigeria Appl
on RWH (CD-ROM). Planning of water harvesting schemes. Unit Water Sci 11:136. https://​doi.​org/​10.​1007/​s13201-​021-​01466-6
22 food and agriculture organization Rome, Italy Omid R, Zahra K, Mahmood S, Evelyn U, Davoud DM, Omid AN, Geor-
Gleick (2014) The worlds water: the bienniel report on freshwater. gia D, Dieu TB (2019) GIS-based site selection for check dams in
Island Press, Washington, p 475 watersheds: considering geomorphometric and topo-hydrological
Guppy L, Anderson K (2017) Water crisis report—the facts united factors. Sustainability 11:5639. https://​doi.​org/​10.​3390/​su112​05639
nations university institute for water, environment and health. Pan S and Zhang H. (2021). Comparative study on dam site selection in
Hamilton, Canada the pre-feasibility stage of Shitouzhai hydropower station
Horton RE (1932) Drainage-basin characteristics. Trans Am Geophys Ponce VM, Hawkins RH (1996) Runoff curve number: Has is reached
Union 13(1):350–361. https://d​ oi.o​ rg/1​ 0.1​ 029/T
​ R013i​ 001p0​ 0350 maturity? J Hydrol Eng 1(1):11–19
Jha MK, Chowdary VM, Kulkarni Y, Mal BC (2014) Rainwater har- Rallison RE (1980) Origin and evolution of the SCS Runoff equation. In:
vesting planning using geospatial techniques and multicriteria Proceeding of the symposium on water management 80 American
decision analysis. Resour Conserv Recycl 83:96–111. https://​doi.​ society of civil engineering boise ID
org/​10.​1016/j.​resco​nrec.​2013.​12.​003 Rami A, Abdallah S, Abdullah GY, AlaEldin I, Sunanda M, Mohamad
Kadam AK, Kale SS, Pande NN, Pawar NJ, Sankhua RN (2012) Iden- AK, Mohamed BAG (2019) Dam site suitability mapping and analy-
tifying potential rainwater harvesting sites of a semi-arid, basaltic sis using an integrated GIS and machine learning approach. Water
region of Western India using SCS-CN method. Water Resour 11:1880. https://​doi.​org/​10.​3390/​w1109​1880
Manag 26:2537–2554 Saaty TL (1980) The analytic hierarchy process planning, priority setting,
Keredin TS, Prasada PR (2016) Review on water resources and sources resource allocation. McGraw Hill, New York
for safe drinking and improved sanitation in Ethiopia. Int J Appl Saaty TL, Vargas LG (1991) Prediction projection and forecasting. Klu-
Res 2(3):78–82 wer Academic Publishers, Dordrecht, p 251
Khudhair, K N Sayl, Y Darama (2020) Locating site selection for rain- Shreve RL (1966) Statical law of stream numbers. J Geol 74(00221376,
water harvesting structure using remote sensing and GIS. In: 3rd 15375269):17–37
international conference on sustainable engineering techniques Strahler AN (1957) Quantitative analysis of watershed geomorphology.
(ICSET 2020), IOP conference series: materials science and engi- Trans Am Geophys Union 38(6):913–920
neering. Vol 881, p 012170. https://​doi.​org/​10.​1088/​1757-​899X/​ USDA (1972) Soil conservation service national engineering handbook
881/1/​012170 hydrology section 4. USDA, Washington, pp 4–10
Li XD (2019) Selection of dam site and dam type for hydropower sta- USDA-SCS (1974) Soil survey of travis county texas agricultural experi-
tions. Shaanxi Province Water Conserv 8:164–168 ment station. USDA soil conservation service, Washington
Luís A, Cabra P (2021) Small dams/reservoirs site location analysis in Vikas KR, Tallavajhala MVS (2020) GIS-based multi criteria decision
a semi-arid region of Mozambique. Int Soil Water Conserv Res. making method to identify potential runoff storage zones within
https://​doi.​org/​10.​1016/j.​iswcr.​2021.​02.​002 watershed. Ann GIS 26(2):149–168. https://​doi.​org/​10.​1080/​19475​
Marinos PG, Koukis G, Tsiambaos G, Stournaras G (1997) Engineer- 683.​2020.​17330​83
ing geology and the environment, vol 2. A.A. Balkema, Rotterdam Yasser M, Jahangir K, Mohmmad A (2013) Earth dam site selection
Mohamed A, Mohamed M, Mohamed S, Mohcine B, Jamal A (2021) using the analytic hierarchy process (AHP): a case study in the west
Identifying suitable sites for rainwater harvesting using runoff of Iran. Arab J Geosci 2013(6):3417–3426
model (SCS-CN), remote sensing and GIS based fuzzy analyti- Yuan J, Su R (1988) Fenhe reservoir siltation prediction and its preventa-
cal hierarchy process (FAHP). Geographia Technica 16:111–127 tion. Resour Sci 2:55–59
Njiru FM, Siriba DN (2018) Site selection for an Earth dam in Mbeere Zhenfeng S, Zahid J, Qazi MY, Atta-ur-Rahman MS (2020) Identification
North, Embu County—Kenya. J Geosci Environ Prot 6:113–133. of potential sites for a multi-purpose dam using a dam suitability
https://​doi.​org/​10.​4236/​gep.​2018.​67009 stream model. Water 12:3249. https://​doi.​org/​10.​3390/​w1211​3249

13