Applied Water Science (2024) 14:124

https://doi.org/10.1007/s13201-024-02190-7

ORIGINAL ARTICLE

Automated GIS‑based decision tool for rehabilitation of a drinking

water infrastructure in Tlemcen, Algeria
Yacine Abdelbaset Berrezel1,2 · Chérifa Abdelbaki1,2 · Bouchrit Rouissat1,3 · Tarik Boumaaza4 · Mohamed Saber5 ·
Matheus Goosen6 · Navneet Kumar7,8

Received: 30 July 2023 / Accepted: 22 April 2024 / Published online: 11 May 2024
© The Author(s) 2024

Abstract
Water resource management relies heavily on the utilization of decision-making systems to guide the strategic rehabilitation
and renovation of infrastructure within water distribution networks (WDNs). This study aims to develop and apply a scripting
tool in ArcGIS for decision-making in WDN. It combines a collection of decision support systems, including Geographic
Information Systems (GIS) and multi-criteria analysis, for infrastructure renovation and rehabilitation. The research is divided
into two main sections. The first focuses on creating and implementing the analytic hierarchy process tool in ArcGIS, while
the second discusses its application in the study area, which is the drinking water distribution network of an urban area in
Tlemcen, located in the Northwest of Algeria. The study’s results reveal that 25% of the network is in critical condition and
require immediate rehabilitation. Sixty-two percentage of the network is classified as moderately urgent, indicating a pressing
but less immediate need for intervention. Finally, the remaining 13% of the network requires attention over a longer timeframe
for rehabilitation. The main criteria influencing pipe prioritization are material (26% of total weight), laying date (26% of
total weight), and pressure (20% of total weight). This developed tool can be easily adapted and applied by engineers and
water management officers. It serves as a guide for decision-makers and planners in urban water management in Tlemcen
and can be replicated and applied to other areas worldwide.

Keywords Python · Scripting · Multi-criteria decision analysis (MCDA) · Geographic Information Systems (GIS) ·
Drinking water management

* Navneet Kumar Introduction

[email protected]
1
Department of Hydraulics, Faculty of Technology, Water shortage is a growing problem that impacts many
University of Tlemcen, P.B. 230, 13000 Tlemcen, Algeria communities around the world. It can be caused by different
2
Laboratoire EOLE, University of Tlemcen, P.B. 230, factors such as insufficient water infrastructure, institutional
13000 Tlemcen, Algeria failures, and demand that surpass supply. Molden (2020)
3
RISAM Laboratory, University of Tlemcen, P.B. 230, provides examples from different regions like California,
13000 Tlemcen, Algeria Egypt, Lesotho, and the Himalayas to illustrate how mis-
4
Oran Water and Sanitation Company, Administrative City management exacerbates water scarcity issues, even in areas
31027, 31000 Oran, Algeria with adequate water supplies. To minimize water scarcity
5
Water Resources Research Center, DPRI, Kyoto University, and energy consumption, Pardo et al. (2020) consider recov-
Goka‑sho, Uji City, Kyoto 611‑0011, Japan ering part of the energy by finding an efficient operating
6
Office of Research and Graduate Studies, Alfaisal University, strategy for the optimal location of pressure-reducing valves
P.O. Box 50927, 11533 Riyadh, Saudi Arabia in water distribution systems. The management of water dis-
7
Department of Ecology and Natural Resources Management, tribution systems (WDSs) can become complex due to issues
Center for Development Research (ZEF), University of Bonn, like leakage and pipe aging, which are commonly associ-
Genscherallee 3, 53113 Bonn, Germany ated with water scarcity (Brentan et al. 2022). One of the
8
Global Mountain Safeguard Research (GLOMOS), United paramount concerns in water management revolves around
Nations University, UN Campus, Platz der Vereinten addressing the challenges posed by aging and deteriorating
Nationen 1, 53113 Bonn, Germany

Vol.:(0123456789)
124 Page 2 of 17 Applied Water Science (2024) 14:124

water distribution networks (Aşchilean et al. 2017a). Specifi- multi-influencing factor (MIF) to create a perspective map-
cally, water pipe corrosion is a critical issue that necessitates ping of groundwater potential zone for the research region.
a thorough understanding for effective decision-making on The research region was categorized into five zones based
solutions and advancements (Amorocho-Daza et al. 2019). on the output map: very poor, poor, moderate, good, and
This may involve the exploration of novel materials or tech- excellent. Other studies compared several MCDA methods
nologies to prevent or mitigate corrosion. Reliable decision and analyzed their suitability for use in integrated asset man-
support procedures are essential in navigating this complex agement of water systems. Tscheikner-Gratl et al. (2017)
process (Abdelbaki et al. 2017). compared five different MCDM methods for use in an inte-
In this context, Hassan et al. (2018) introduced a model grated rehabilitation management scheme for small and
integrating two optimization techniques, the genetic algo- medium-sized municipalities. Yoo et al. (2014) proposed a
rithm (GA) and heuristic programming (HP), known as more reliable way to prioritize rehabilitation in water pipes
GA-HP, for designing water distribution infrastructure. In an through the use of a multi-criteria decision-making process
another study, Hassan et al. (2020) utilized a combination of that considers the hydraulic significance of the pipes.
a genetic algorithm and tree growing algorithm (GA-TGA) Likewise, the application of GIS technology in managing
to determine optimal pipe diameter and slope for the sewer water infrastructure has proven to be effective in determin-
system. Several studies have explored the use of decision ing the condition of the water distribution network spatially
support systems (DSSs) such as the multi-criteria decision and assists in predicting future situations. Abdelbaki et al.
analysis (MCDA) and Geographic Information Systems (2017) conducted a case study in Chetouane, Algeria where
(GIS) for effective drinking water network management. the areas of the water distribution network that require
Decision-making in emergencies is complex, requiring improvement were identified effectively using GIS analysis.
the integration of various criteria, models, and data sources Abdessamed et al. (2023) used GIS to study water quality in
(Pagano et al. 2018). Salehi et al. (2018) found that the use the Ain Sefra watershed in the Western part of Algeria. Their
of multi-criteria decision models can address the complex- research revealed that integrating the water quality index
ity of decision-making for the rehabilitation planning of (WQI) with GIS is an invaluable tool. By creating maps that
any town network. Așchilean and Giurca (2018) examined display WQI variations throughout the watershed, policy-
various rehabilitation technologies for water supply systems makers can make informed decisions and allocate resources
using multi-criteria analysis and chose pipeline techniques to more effectively manage water resources. In the same
among other alternatives to rehabilitate the distribution context of water supply management, (Kendouci et al. 2019)
networks in Cluj-Napoca. Tanyimboh and Kalungi (2009) employed GIS to spatially analyze the WDN and infrastruc-
applied the analytic hierarchy process (AHP) to determine ture in Bechar, an arid Algerian city. Their GIS analysis,
the optimal solution for designing and improving a water combined with a survey of 20% of residents, revealed 74%
distribution network in Wobulenzi, Uganda. dos Santos were unsatisfied with water quality/quantity, and 59% had
Amorim et al. (2020) used the AHP method to compare and inadequate pressure requiring pump usage.
prioritize three potential solutions for low-income housing GIS is needed in order to effectively arrange and analyze
in Brazil and found water-saving equipment to be the most a vast array of spatial information for use in decision-making
effective alternative. Al-Zahrani et al. (2016) used a fuzzy- models. The combination of GIS and the AHP method has
based decision support system and fuzzy AHP methods to been applied in various studies to assess and predict urban
identify the areas in the water distribution network that are water demands, prioritize rehabilitation plans in water dis-
most susceptible to vulnerability in Al-Khobar, Saudi Ara- tribution networks, and rank water mains for maintenance,
bia. (Amorocho-Daza et al. 2019) developed a methodology rehabilitation, or replacement. Panagopoulos et al. (2012)
using MCDA for water supply planning and evaluated alter- used the AHP method in conjunction with GIS to assess
natives systematically to determine the views of different and predict current and future urban water demands in the
stakeholders in relation to non-economic factors for alter- Greek city of Mytilene. Tabesh and Saber (2012) developed
nate water supply infrastructure in the City of Santa Marta, a decision support tool for prioritizing rehabilitation plans
Colombia. Gül and Firat (2021) utilized MCDA techniques in the water distribution network by utilizing GIS in the city
like the Elimination and Choice Expressing Reality known of Mashhad, located in the Northeast of Iran. Mesalie et al.
by (ELECTRE) to identify the most critical regions in the (2021) conducted a risk assessment of the drinking water
water distribution systems that need rehabilitation in order distribution system at Bahir Dar Institute of Technology in
to reduce water losses in Maltaya, Turkey. Ethiopia. To evaluate and prioritize risks across the network,
Another study conducted by Senapati and Das (2022) they developed an approach that combined GIS and AHP.
focuses on assessing groundwater potential in the agricul- The combination allowed to assess both structural and cus-
ture-dominated Cooch Behar district of West Bengal, India. tomer point risks across the distribution system. The findings
GIS-based two MCDM methods have been using, AHP and revealed that the components of the campus water supply
Applied Water Science (2024) 14:124 Page 3 of 17 124

distribution system should be replaced before they become improved upon the previous work by creating a Python tool
obsolete and suitable preventive measures should be taken. to integrate the analytical hierarchy process into ArcGIS.
Kabir et al. (2015) have developed a model that can rank Although the Python programing language is actively
water mains within a distribution network and identify sensi- used, both in industry and academia, for a wide variety of
tive and fragile pipes to support the need for maintenance, purposes, no previous work developed the tool for network
rehabilitation, or replacement combined with a GIS model. rehabilitation using AHP methods scripted in GIS using
The integration of MCDA and GIS forms a powerful tool Python in the urban environment, which holds significant
for making effective decisions by being able to display the importance for efficient utilization and management of
result of a multi-criteria decision on a spatial map for bet- urban water resources. One tool already integrated in Arc-
ter visualization and understanding. It allows spatial data GIS that can be combined with the AHP method is overlay
and analysts’ criteria to be combined to solve spatial prob- analysis, which identifies the best or most preferred loca-
lems (Rajabi et al. 2011). In studies conducted in Batu tions by applying weights to several layers and combining
Pahat, Malaysia (Ali Ahmad et al. 2015), and the city of them into a single output. However, the overlay analysis only
Malatya, Turkey (Kilinç et al. 2018), MCDA using AHP was accepts raster data as an input, which can result in a piece of
applied to identify suitable sites for constructing reservoirs lost information by converting vector data into raster data.
and evaluate options for inadequate pipeline functionality, Therefore, a tool that uses vector inputs such as the pipeline
respectively. layer to identify the sections that must be rehabilitated or
The use of GIS technology in analyzing spatial data renewed by weighting all the alternatives using the AHP
and criteria can aid in solving spatial problems and mak- method is highly needed.
ing informed decisions. There are various decision-making Some methodologies and strategies have been proposed
systems and models that have been created for different to rehabilitate intermittently supplied networks using multi-
purposes. For instance, Curry et al. (2020) developed a criteria procedures. These methods address complex plan-
scientific approach to aid in the decision making process ning issues such as identifying the inefficiencies within the
regarding the renovation or removal of significant dams. water distribution network system. Brentan et al. (2022)
Kessili and Benmamar (2016) proposed a methodology for developed a multistage optimization procedure for reha-
prioritizing sewer rehabilitation networks in Algeria’s capi- bilitating a water distribution network, consisting of three
tal city by ranking different criteria using AHP. Lima et al. different operations: pipe replacements, leakage fixing, and
(2021) conducted a multifaceted multi-criteria evaluation pump optimization, however, their study is limited because
that incorporated the perspectives of both local stakehold- the pipe replacements are based only on aging pipes factor,
ers and decision makers to identify priority watersheds for and other criteria such as different materials of pipes were
implementing groundwater management strategies. The not considered. Juan et al. (2022) proposed a comprehensive
authors emphasized that the model should be utilized as methodology that consists of three stages aimed at rehabili-
a supplementary tool in the decision-making process and tating deteriorated water systems. The first stage, termed Ini-
not solely relied upon as the decisive approach. In another tial system assessment, involves identifying vulnerable areas
related study, Zolfaghary et al. (2021) utilized GIS and a and critical supply hours. The second stage, known as sec-
MCDA to assess the feasibility of using treated wastewater torization, entails defining optimal district metered areas to
from urban areas for irrigation purposes in Golestan prov- reduce water losses and increase the supplied water through
ince located in Northern Iran. improved control of flows and pressure. Finally, the third
Although many programmers have worked on devel- stage determines the optimum investments for asset rehabili-
oping such a tool, only a few have implemented it in the tation, including upgrading storage tanks, replacing pipes,
water sector, especially on renovation and rehabilitation of repairing leaks, and replacing pumps. They used EPANET
water distribution networks. Initially, decision application through a Python integrated development environment and
tools were developed in ArcGIS for land use assessment. the NSGA-II algorithm (Non-dominated Sorting Genetic
Marinoni (2004) described the implementation of an ana- Algorithm) to optimize the rehabilitation of a water distri-
lytical hierarchy process in ArcGIS using the visual basic bution system with intermittent service, including replac-
programming language to facilitate land use assessment. ing pipes and repairing leaks. This approach prioritizes pipe
Boroushaki and Malczewski (2008) further extended the replacements based on performance and cost, within budget
work of Marinoni and concentrated on incorporating GIS constraints, to improve system performance. It is important
with an expanded version of the analytical hierarchy process to mention, though, that their detection of vulnerable regions
for spatial MCDA. However, the method used a combina- depended solely on the supply ration (SR), which is the sup-
tion of weighted summation GIS procedures in the evalu- plied water over the expected water demand. Brentan et al.
ation process to calculate the weights of each alternative. (2022) utilized graph theory in their methodology to identify
The output was presented in a raster format. Elhaj (2018) the most important pipes in the WDN. This enables targeted
124 Page 4 of 17 Applied Water Science (2024) 14:124

repair or rehabilitation of these pipes to improve the perfor- inadequate construction or repair of pipes, and a failure to
mance of the WDN in providing intermittent water supply adequately maintain and replace the pipelines. The water
(IWS). supply system is poorly structured as a result of extensions
One of the main innovations of our work is the develop- made without basic design (Abdelbaki et al. 2012). There are
ment of a tool that allows integration of the AHP method several deficiencies in the network, including frequent inter-
with GIS for vector data. This integration enables the use of ruptions in the water supply, high levels of water loss due to
spatial data to aid in the decision-making process for rehabil- leaks (60%), and a lack of attention to the aging and outdated
itating water distribution networks, which can be particularly piping infrastructure. (Abdelbaki et al. 2014). So far, the
useful for identifying priority areas for intervention. The tool water networks are facing serious management issues.
also generates tables containing the criteria and indicators
used in the decision-making process. Additionally, the tool Data used
calculates weights for each criterion and indicator by com-
paring them pairwise. This allows for prioritization of the The database of the WDN used in this research was obtained
network sections that need rehabilitation or renewal. from various sources (Abdelbaki 2014; Berrezel et al. 2022).
The tool is user-friendly, catering to scientists, engineers, The main data were water pipes network. The length of the
and water management officers. It serves as a guide for analyzed network (main pipes) spans 65 km, featuring pipes
decision-makers and planners in urban water management with a diameter that ranges from 40 to 600 mm, and made
and holds the potential for replication in diverse global con- of different materials such as steel, galvanized steel, and
texts. The integration of AHP with GIS offers an innovative high-density polyethylene (PEHD). The water distribution
approach to support decision-making in water distribution system relied on gravity and using 11 tanks with capacities
network rehabilitation, with broad applicability in similar ranging from 1500 to 2000 cubic meters, located at altitudes
contexts. between 800 and 1000 m. The study area’s background map
was obtained from Google Earth Pro and calibrated using
the georeferencing tool of ArcGIS. This process involved
Materials and methods importing the Google Earth Pro image into ArcGIS and
identifying common features shared between the image and
Study area other georeferenced data layers. Control points were strategi-
cally placed on these common features within both datasets.
The developed tool was applied to the water distribution
network of the Southern part of the urban agglomeration of Methodology
Tlemcen in Algeria, which encompasses the towns of Tlem-
cen, Chetouane, and Mansourah. This region, located in the The research methodology of this paper consists of three
Western part of Algeria, covers around 112 ­Km2 forming phases: algorithm development, AHP tool development,
the Tlemcen inner basin. The boundaries of this basin are and tool application (Fig. 2). The main goal of this tool is
defined by Lalla Setti cliff to the South, Ain El Houtz hills to to extract the attribute data from the water network layer
the North, Oum El Allou to the East, and the small volcanic in ArcGIS and prioritize pipes based on their laying date,
cone mountains of Beni Mester to the West (Abdelbaki et al. diameter, material, etc. The weight calculation of the pipes
2019). The research site located in the Southern region of is carried out using the AHP method, allowing the user to
the urban aggregation of Tlemcen encompasses an area of determine their emergency level in the WDN. The algorithm
3.4 ­Km2 as depicted in Fig. 1. The topography of the study development phase outlines the main instructions and opera-
area is rough and the soil is of a semi-rocky nature, with tions utilized to run the tool. The AHP tool development
the presence of a significant horst formation. This horst’s phase illustrates how to combine the tool with the AHP
surface is primarily composed of karstic limestones originat- method using Python modules and packages within ArcGIS,
ing from the Upper Jurassic and Lower Cretaceous periods. also demonstrating the script or the code of the program.
These carbonate formations constitute a notable karst land- The final phase involves applying the AHP tool in the study
scape, overlaid by substantial Neogene sedimentation layers. area to validate the generated findings. A testing process was
Groundwater within this karst region is extensively extracted undertaken to accomplish this. The system was designed
via boreholes, serving as a crucial source of potable water to work with a vector layer as vector files are used when
(Bensaoula 2007; Fellah et al. 2016).The elevation is char- descriptive information needs to be stored. Vector data are
acterized by steep inclines with elevation drops ranging from preferred in this case because it allows to represent the geo-
800 to 1100 m (Berrezel et al. 2022). graphical features of water distribution network in a more
The WDN of the study area is characterized by a tremen- precise and accurate way. Vector data consists of points,
dous rate of water loss that exceeds 50% due to aging pipes, lines, and polygons, each with its attribute data, including
Applied Water Science (2024) 14:124 Page 5 of 17 124

Fig. 1  a Algeria’s location within the African continent; b Tlemcen The water distribution network of the study area; f The digital eleva-
city’s position within Algeria; c Tlemcen town’s placement within tion model (DEM) representing the study area; g Topographic pro-
Tlemcen city; d The study area’s location within Tlemcen town; e files of the main directions

Fig. 2  Phases in the develop-

ment and application of a multi- Algorithm AHP tool
criteria decision-making tool development

Creating the Execution and

scripting tool validation of the tool

Tool Scripting process Application of the

parameter tool in a study area
124 Page 6 of 17 Applied Water Science (2024) 14:124

information about the laying date, diameter, and material Algorithm development
of the pipes. This information is critical for calculating the
pipe’s emergency level using the AHP method. Prior to scripting the program, the proposed algorithm was
In this study, the water distribution network and its attrib- utilized to define a step-by-step set of instructions. The main
ute data were primarily represented using raster data. How- goal of this tool was to use data from the pipe layer and
ever, it is important to note that raster data, being composed calculate the emergency level by employing the Analytic
of pixels, lacks the capability to store different attributes Hierarchy Process (AHP) method. The AHP method relies
directly. As a result, using vector data would be a more suit- on pairwise comparisons of criteria, which are presented
able approach for accurately representing the water distribu- as weight matrices represented as tables within the ArcGIS
tion network and extracting the necessary attribute data for environment and accessible via the table of contents. The
prioritizing the pipes. Despite the limitations of raster data, algorithm illustrated in Fig. 3 outlines the main instructions
the study proceeded with some tests using raster data. The used to run the tool.
approach involved converting the network layer into raster The initial steps involve importing the file containing the
layers for conducting overlay analysis and classifying the pipe network layer as an input (Step 1), and specifying which
network. Nonetheless, this process led to the loss of valu- attribute fields from the pipe layer will be used as criteria for
able information due to the inherent nature of raster data the AHP analysis, such as the diameter and material of the
representation. To enhance decision-making accuracy and pipes (Step 2). The program then creates a table containing
align more closely with reality, the program incorporates the user-defined criteria to facilitate pairwise comparisons
binary comparisons, initially employed by the AHP method between them based on the Saaty scale (Saaty 1980), which
to establish the relative importance or priority of different allows the user to indicate the relative importance of one
criteria. This functionality enables researchers to refine anal- criterion over another (Step 3).
yses by making necessary adjustments or redoing certain To ensure the consistency of the judgments made during
evaluations. When assessing attributes of the water distri- the pairwise comparisons, the program performs a consist-
bution network such as pipe material or diameter, binary ency check (Step 4). If the inconsistency exceeds a certain
comparisons are instrumental in identifying critical factors threshold, the user may need to revise their comparisons.
and prioritizing them over others. However, while binary Once the pairwise comparisons are completed, the pro-
comparisons significantly contribute to the analysis by incor- gram automatically calculates the weight or priority of each
porating vector data and uncertainty modeling, assessments criterion by generating a pairwise comparison matrix and
can be greatly enhanced, particularly through sensitivity performing the necessary computations (Step 5). The same
analysis. This type of analysis determines how alterations method is used to calculate the weights of any sub-criteria,
to input variables affect the model’s output, enabling identi- if applicable, by creating additional pairwise comparison
fication of the sensitivity of decision outcomes to changes in tables.
physical patterns or hydraulic parameters. Such analysis can The program then generates fields or columns in the main
pinpoint the key drivers of system performance and evaluate pipe layer to store the calculated weights for each criterion
the viability of proposed interventions, ultimately leading to and sub-criterion (Step 6). As part of the summation process,
more informed and effective decision-making in the manage- separate fields or columns are created to store the weighted
ment of drinking water networks. values for each individual criterion or sub-criterion (Step 7).
The calculated weights for each criterion and sub-criterion

Fig. 3  The algorithm utilized to

run the AHP tool in ArcGIS Input Creating tables of Pairwise comparison & Test of
layer assign weight of tables consistency
criteria & sub - created
network (matrix) created
it i

Is
No CR<0.1?

YES

Output weighting of Summing all Creating Assign each weight in

criteria in layer weight in the fields of each corresponding field
network fields indicator
Applied Water Science (2024) 14:124 Page 7 of 17 124

are assigned to their respective fields or columns in the pipe A square matrix was constructed to arrange the pairwise
layer’s attribute table (Step 8). comparisons of the different criteria according to Eq. 1.
Finally, to obtain the overall emergency level or prior-
ity score for each pipe section, the program calculates the ⎡ 1 a12 ⋯ a1n ⎤
� � ⎢ 1∕a12 1 ⋯ a2n ⎥
sum of the weighted values for all criteria and sub-criteria. A = aij = ⎢ (1)
⎢ ⋮ ⋮ ⋱ ⋮ ⎥⎥
This summation is performed by dividing the sum of the
⎣ 1∕a1n 1∕a2n ⋯ 1 ⎦
weighted values by the maximum value of the sum, resulting
in the absolute weight or emergency level for each pipe sec- where A is the decision matrix, ­aij are comparisons between
tion. The calculated emergency levels can then be used for elements I and j for all i, j ∈ {1, 2,,, n}.
decision-making purposes, such as identifying high-priority The values in the resulting pairwise comparison matrix rep-
areas for maintenance or emergency response within the pipe resent the relative importance of each criterion, and the matrix
network. should be normalized by dividing each value by the sum of
the values in its corresponding row. The weight is derived by
AHP tool development determining the mean of every row (Boukhari et al. 2018).
To assess the accuracy of the pairwise comparisons and
Analytic hierarchy process (AHP): Saaty (1980) proposed to determine the level of consistency in the judgments, the
the AHP method in the 1980s; it is a way for comparing, Consistency Index (CI), Consistency Ratio (CR) and Random
weighting, and ranking the options and criteria. In order to Index (RI) can be used. These are calculated using Eq. (2):
make a decision in an instructed manner, Saaty has devel-
oped 4 steps (Saaty 1988, 2001, 2008): (i) identifying and CR = CI∕RI (2)
clarifying the problem’s context; (ii) dividing the decision-
where:
making problem into levels, starting at the top with the
RI: Random Consistency Index, (RI) can be determined
decision’s objective, moving down to specify criteria and
from Table 2
options, and ending with a collection of alternatives; (iii)
making pairwise comparison matrices and determining the CI = (𝜆_ max −n)∕(n − 1) (3)
relative weights; (iv) assessing the accuracy of the pairwise
comparisons can be done by computing the consistency where:
index (CI) and the consistency ratio (CR). λ_max: the dominant eigenvalue of the pairwise compari-
To compare criteria, Saaty (1980) proposed a scale that son matrix.
can be used to identify which one is more important than the n: the order of the matrix.
other (refer to Table 1). In this particular case study, the Saaty The consistency ratio (CR) determines the validity of pair-
scale was used to compare different attributes of pipes in the wise comparisons. If the CR exceeds 10%, the evaluation cri-
water network, such as laying date, diameter, material, etc. By teria must be reevaluated. The final weight is determined by
assigning a numerical value to the relative importance of each multiplying the weights assigned to individual indicators at the
attributes, we can use the AHP method to calculate the weights lowest level by the weights assigned to higher levels of criteria
of each pipe and prioritize them according to their emergency and dimensions.
level in the water distribution network. The AHP tool for ArcGIS was implemented using the
Python programming language and leveraged ArcGIS mod-
ules for various stages of the process. The tool calculates the
emergency level using data from a pipe layer through four
Table 1  Saaty numerical scale (Saaty 1980) steps: data input, matrix creation, calculation, and output.
Intensity of importance Definition To add a script tool in ArcGIS, the first step is to create a
tool by entering its name, label, and optionally a description.
1 The same importance The tool is used to perform a specific task and requires cer-
3 Moderate importance tain input parameters to be set in order to function correctly.
5 Strong importance These parameters are defined in the wizard panel and can be
7 Extreme importance accessed and used in the source code as input. Each param-
9 Absolute importance eter has a unique name, a data type that specifies whether it

Table 2  Saaty (Saaty 1980) n 2 3 4 5 6 7 8 9 10 11

values of random index
RI 0.00 0.52 0.89 1.11 1.25 1.35 1.40 1.45 1.49 1.52
124 Page 8 of 17 Applied Water Science (2024) 14:124

is a table, layer, field, or something else, and certain proper- laying date of each section will be taken into account, with
ties such as whether it is required or optional. Some param- older sections being given higher priority; (3) the diam-
eters may also have a filter to restrict the values that can be eter, pressure, and velocity sub-criteria may be evaluated
entered or may allow for the input of multiple values instead based on the size of the pipe, the pressure level at which it
of just one. is situated, and the water velocity within it, respectively.
The Python program for ArcGIS consists of four phases: Overall, the process of defining the hierarchical structure
in the first phase, the program prompts the user to input the and characteristics of the area’s pipe network is important
name of a network layer and select fields to use as criteria. because it allows the user to identify the main structure of
A table is created with the same columns and rows as the the network in an organized way, and to determine which
selected fields, with a domain for ranking criteria in a scale sections of the network should be rehabilitated or renewed.
proposed by (Saaty 1980). This allows the user to compare This information is then used in the second step of the pro-
the importance of each criterion in the table. In the second cess, which involves applying the AHP tool to analyze and
phase, the program calculates the weight of each criterion in prioritize the different elements of the pipe network.
the table. In the third phase, the program adds the fields of To illustrate the calculation process for the comparison of
indicators to the main layer and updates the rows by adding various criteria through pairwise comparisons, the follow-
the weights of each criterion and sub-criteria. In the fourth ing steps exemplify the assessment of sub-criteria (mate-
and final phase, the program calculates the sum of all fields rial, diameter of the pipe, and laying date). The elements are
added by the user and divides it by the maximum value of arranged in matrix A.
the sum to obtain the absolute weight of the indicators.
⎡ 1 1∕4 1∕3 ⎤
Application of the tool A = ⎢4 1 2 ⎥
⎢ ⎥
⎣ 3 1∕2 1 ⎦
To apply this tool in the study area, there are two main
The importance of these elements is assessed by totaling
steps. The first step is to define the hierarchical structure
the values in the columns of matrix A, and then dividing
and characteristics of the area’s pipe network, while the
each value by the sum of the column. The importance is
second step involves the application of the AHP tool.
derived by computing the average of each row. The sum of
Defining the hierarchical structure & characteristics of
each column: [8, 1.75, 3.33].
the area’s pipe network involves dividing pipe networks
Dividing each value in the column by the total of the
into specific sections based on their specific character-
column:
istics such as material, diameter, and year of construc-
tion (Tscheikner-Gratl et al. 2017). These characteristics ⎡ 1∕8 (1∕4)∕1.75 (1∕3)∕3.33 ⎤ ⎡ 0.13 0.14 0.10 ⎤
are the influencing criteria which represent different ele- ⎢ 4∕8 1∕1.75 2∕3.33 ⎥ = ⎢ 0.50 0.57 0.60 ⎥
ments of the rehabilitation decision (Tscheikner-Gratl ⎢ ⎥ ⎢ ⎥
⎣ 3∕8 (1∕2)∕1.75 1∕3.33 ⎦ ⎣ 0.38 0.29 0.30 ⎦
et al. 2015). In the present study, two main criteria were
taken into consideration: the physical characteristics and The weight is determined by finding the mean of each
the hydraulic performance of the network. The physical row:
structure of the network criterion is divided into three
sub-criteria, which are materials used, laying date, and ⎡ 0.122 ⎤
diameter of pipe (D). The hydraulic behavior of the net- ⎢ 0.558 ⎥
⎢ ⎥
work criterion was divided into two sub-criteria which ⎣ 0.320 ⎦
are water pressure (P), and water velocity (V). The user
can define additional criteria based on data availability to Det |A − 𝜆I| = 0
increase the reliability of the decision-making. Figure 4
illustrates the hierarchical structure to define which part This allows to find the eigenvalues of the matrix A.Where
of the network should be rehabilitated by classifying the I: the identity matrix. λ: the eigenvalues of the Matrix.
sections in emergency levels; each criterion is subdivided �
into sub-criteria with 3 indicators. The ratings approach �⎡ 1 1∕4 1∕3 ⎤ ⎡ 1 0 0 ⎤��
�⎢ �
is used when there are many alternatives to be evaluated. Det � 4 1 2 ⎥ − 𝜆.⎢ 0 1 0 ⎥� = 0
�⎢ ⎥ ⎢ ⎥ �
�⎣ 3 1∕2 1 ⎦ ⎣ 0 0 1 ⎦��
Instead of comparing the options pairwise, the appraisal �
is carried out using indicators (categories) to evaluate or
λ is calculated as 3.018, n: the order of the matrix, n = 3, RI:
assess the performance or effectiveness of an alternative
Random Consistency Index, RI = 0.52, λ_max: the principal
as following: (1) materials: comparison between the least
eigenvalue of the matrix, λ_max = 3.018, Consistency Index:
desirable and the most desirable; (2) pipe laying date: the
Applied Water Science (2024) 14:124 Page 9 of 17 124

Fig. 4  The hierarchical structure

and characteristics of the area Goal: Defining the emergency levels of pipes to rehabilitate water

Goal
pipe network supply network

Criteria
Physical structure Hydraulic structure

Diameter
Materials Pressure

Sub-Criteria
Levels

Pipe laying Velocity

date

P<2bar

Steel
Indicators

Cast Iron 100mm<D

PEHD

V<0.5 m/s
10 years
0.5<V<1.5 m/s
15 years
20 years 1.5<V<3 m/s

CI = (λ_max−n)/(n  −  1) = (3.018–3)/(3–1) = 0.009, Consist- is calculated automatically using the weighting tool in the
ency ratio: CR = CI/RI = 0.009/0.52 = 1.73% second script.
CR is less than 10%. Thus, the comparison matrix is The third script is used to create a table of indicators that
consistent. can be used to evaluate the different alternatives (elements)
To use the AHP tool in the process of analyzing and of the pipe network. To use this script, the user must input
prioritizing elements of a pipe network, the user must first the table of criteria calculated in the previous step and speci-
specify the criteria and options for each level of the hierar- fies the indicators that will be compared and weighted. The
chy. The tool will then create tables with the same column resulting table of indicators is then used in the next step of
and row names for binary comparison and calculation of the process which involves creating fields in the main layer
the weights of the criteria and indicators. The user must
run 5 scripts to get the result. The steps below demonstrate
how to use the AHP tool to get the best results.
The first script involves creating a table that combines the
criteria to be weighted, with the selected criteria belonging
to the same group (e.g., main criteria) as shown in Fig. 5.
For example, the laying date, materials and the diameter of
the pipe are three sub-criteria that belong to the main cri-
terion named hydraulic structure. The user then determines
the level of importance of elements in the matrix according
to the Saaty scale (Table 1) and the weight of each criterion
Fig. 5  Weight of each criterion obtained by the AHP Tool
124 Page 10 of 17 Applied Water Science (2024) 14:124

and using a summing tool to calculate the final weight of and technical assessments provided a comprehensive evalu-
each alternative. In the fourth script, fields are created in ation of the water supply network, enabling the identifica-
the main layer, each field representing the weight of one tion of areas for improvement and development of a plan to
sub-criterion. In our case, we have 5 sub-criteria and 15 ensure the continuous and sustainable operation of the water
indicators (Fig. 5). supply system.
The summing tool is used in the fifth script to calculate
the final weight of each pipe section by adding the values
in the fields created in the previous step. After calculating Results and discussion
the final weights of each pipe section, the user can clas-
sify them by order and prioritize the maintenance actions or Criteria weighting and influence on emergency
other decision-making processes related to the pipe network. classification
In this study, three classes were proposed to categorize
the sections by their weights, with green indicating low pri- In order to classify and group WDN maintenance into three
ority (level 3), yellow indicating medium priority (level 2), levels of emergency, the AHP tool was used to determine
and red indicating high priority (level 1). This classification the weight of each section based on pairwise comparisons of
system allows for the identification of the priority portions criteria and indicators. The results indicate that the criteria
of the maintenance action and the classification of pipelines “Material” and “Laying Date” have more influence than the
into immediate, intermediate, and long-range plan-repair other criteria as they were given more importance by the
categories in the water distribution network. user (Fig. 5). Figure 6 displays the weights of the indica-
tors, with “Cast Iron” and the laying date “20 years” having
Validation of the tool the highest weight compared to the other indicators. This
implies that pipes made of cast iron and with the laying date
This study builds up on the works of previous researchers of 20 years are the most likely to undergo renewal or rehabil-
(Abdelbaki et al. 2012, 2014, 2019), who have examined itation. Tscheikner-Gratl et al. (2017) found that the majority
various aspects of the drinking water network operations, of pipeline failures in the water distribution network occur
including conducting surveys and diagnosing issues arise in cast iron and steel pipes. These materials are no longer
from water supply disruptions. By incorporating the insights used in modern water networks due to their tendency to fail
and findings from these studies, a more comprehensive and and cause disruptions. Aşchilean et al. (2017b) stated that
nuanced understanding of the emergency levels within the sections of steel pipes experienced significant water losses
study area was developed. The methodology used in this when compared to polyethylene pipes, which have fewer fail-
study relied on the Analytic Hierarchy Process (AHP), ures in the water distribution network. The pressure criteria
which enabled us to classify the area into low, medium, and the indicator “4 < P < 7 bar” can also impact the results
and high emergency levels based on a detailed assessment because it is widely accepted that water pressure for domes-
of multiple factors. The use of field surveys conducted by tic use should not exceed 4 bars, water leaks can appeared
(Abdelbaki 2014; Abdelbaki et al. 2014) allowed us to if this value is exceeded (Abdelbaki et al. 2014). In other
validate the obtained results and ensured their alignment words, it is critical to consider the user’s judgment while
with the actual ground conditions, thereby enhancing the selecting the criteria and indicators.
credibility and robustness of the present study. The survey The tool displayed the calculation results on the map,
was designed to evaluate the quality of service provided to making it easier to detect deficient parts of the network
subscribers, including the quality of repair work and cus- which are highlighted in red color (Fig. 7). The sections of
tomer complaints. Both customer and manager surveys were the network that are characterized by indicators with a high
conducted to evaluate the satisfaction levels of customers
with the current water supply service and to identify areas
of improvement. Following the survey, various technical Table 3  Technical indicators for the Tlemcen urban group
indicators were assessed as shown in Table 3, such as net-
Technical indicators Tlemcen urban group
work yield, linear losses index (LLI), linear index of repair,
annual average rate of network renewal, and operating pres- Network yield (%) 93%
sure. These technical indicators were used to evaluate the Linear losses index (LLI) 23–46 ­m3/day/km
efficiency and effectiveness of the water supply network, Linear index of repair 2–45 repairs/yr/km
with the aim of identifying any issues or weaknesses in the Annual avg. network renewal rate (%) 22% (150 km
technical aspects of the network and recommending cor- renewed out of
680 km)
rective measures to improve the overall performance of the
Operating pressure (bars) 0.8–10
water supply system. The combination of customer surveys
Applied Water Science (2024) 14:124 Page 11 of 17 124

Fig. 6  Weight of each indicator obtained by the AHP Tool

weight will typically tend to be in the red zone. During the Figure 7) aids in easy identification of various parts
pairwise comparison of the indicators that have influenced and to prioritize them. The length of the primary pipeline
the state of the network, such as the laying date (26%), the being examined is 65 km, with a diameter ranges from 40
pressure (20%), or network materials (26%), it was found to 600 mm, made of various materials such as cast iron,
that these indicators had higher weights based on the user’s PEHD, and steel. The classification of these pipes (which
judgment of their importance (72% of the overall weight). refers to grouping the pipes based on their weights) has
Results are displayed on the map. allowed the prioritization of the parts of the network based
on their level of emergency, determined by order, category,

Fig. 7  The emergency level of the network

124 Page 12 of 17 Applied Water Science (2024) 14:124

and color assigned to each pipe. For instance, pipes with classified sections according to their weight, and proposed
higher weights (and therefore, higher levels of emergency) the chronological order of the renewal interventions.
are assigned a higher order, a more urgent category, and a
more distinctive color (such as red) to clearly indicate their Spatial indicator analysis for WDN prioritization
higher priority status. Similarly, pipes with lower weights
and lower levels of emergency are assigned lower orders, One way to analyze the relative weight and location of spe-
less urgent categories, and less distinctive colors (such as cific indicators on the network is to use ArcGIS to display
green) to reflect their lower priority status. Figure 8 shows their distribution on the map. For example, the “material”
that 25% of the network is in the critical phase (level 1) and criterion and its sub-criterion, the “PEHD” indicator, can
should be rehabilitated with first priority, 62% is in medium be examined to understand their contribution to the overall
emergency (level 2), and the rest 13% covers the low level prioritization of the pipes. By displaying the distribution
of emergency (level 3). Several researchers classified the of this material on the network, it’s clear that its presence
network in 3 level (low—medium—high) in terms of reha- in areas of high priority such as the red zone, with 3.5 km
bilitation zones such as (Kabir et al. 2015; Tscheikner-Gratl of PEHD pipes is 3.5 km is low, as shown in Fig. 9. This
et al. 2016). (Salehi et al. 2018) classified the network into information can help decision-makers to make informed
4 zones; each zone has a specific characteristic, so the prior- decisions about rehabilitation and maintenance actions
itization of the pipes has been done separately. Prioritizing based on the specific indicator’s relative weight and loca-
various pipes with high weights leads to renewal measures. tion on the network.
The link between prioritizing and renewing is used in many This analysis method enables the decision maker to
results in similar research topics. (Boutebba et al. 2014), follow each indicator and observe its influence on the
network, then take action for rehabilitation. (Salehi et al.

Fig. 8  Weights of each emer-

gency level obtained by the
AHP Tool

Fig. 9  Linear network of PEHD

pipes
Applied Water Science (2024) 14:124 Page 13 of 17 124

2018) classified network zones by selecting the pipes with MCDA methods that offer more robust mechanisms for han-
higher rehabilitation priorities considering technical and dling subjectivity in weighting. Future research endeavors in
non-technical criteria such as the climate and any special this regard will aim to compare different MCDA methods to
criteria. It’s essential to choose the effective criteria when better understand their strengths and limitations to improve
sufficient network information is available. decision-making processes in water distribution network
There are many criteria to consider for WDS rehabilita- rehabilitation. This will contribute to a better understand-
tion. Mesalie et al. (2021) used multiple criteria, such as ing of the feasibility and effectiveness of different MCDA
pipe age, pipe material, water quality parameters, water methods in handling subjectivity in weighting to achieve
availability from supply sources, adequacy of water pres- more accurate and reliable prioritization results.
sure and flow rate for customers, and the condition of stor-
age facilities and pumps. (Tscheikner-Gratl et al. 2016) Field surveys and technical indicators for WDN
found that significant factors such as pipe age, pipe length, emergency level validation
pipe diameter, pipe slope, and pipe shape, as well as low
network material heterogeneity and vulnerability, greatly Abdelbaki (2014), conducted field surveys to investigate the
influenced the priority model’s result for pipe deteriora- status of WDN in Tlemcen area, the surveyed regions are
tion. However, the importance of certain indicators, such shown in Fig. 10. This allowed us to compare and validate
as the “Diameter” criterion, may be perceived differently our findings. The study conducted included a total of 100
by different users. respondents, comprising of 30 employees of the Algérienne
In this particular study, the “Diameter” criterion had a Des Eaux (ADE), the water company, and 70 subscribers in
low weight, but other researchers like (Blindu 2004) found the urban area of Tlemcen. The satisfaction levels reported
that it is directly related to the importance of the pipe. by respondents in the surveyed areas vary across the four
This highlights the subjectivity of prioritization and the regions: Birouana, Boudghene, El Kalaa Inferieure, and Sidi
importance of carefully considering the relative weights Tahar. The satisfaction levels are divided into three cate-
of different indicators. Structural deterioration of water gories: Good, Moderate, and Poor. The satisfaction levels
mains is affected by physical factors (Kabir et al. 2015). with network management vary significantly as illustrated
(Brentan et al. 2022) cited that replacing old pipes that in Fig. 11. Birouana has a majority of respondents reporting
are prone to deterioration and leakages, may improve the moderate satisfaction (50%), followed by good satisfaction
reliability and efficiency of the WDS and better meet the (27%) and poor satisfaction (23%). Boudghene, on the other
needs of users, yet their analysis did not consider other hand, has a higher percentage of moderate satisfaction (45%)
important factors such as the type of materials used for with good satisfaction (25%) and poor satisfaction (30%). El
the pipes, pressure levels, or pipe diameters, which can Kalaa Inferieure experiences a majority of poor satisfaction
also significantly influence the performance and longevity (61%), lower levels of good satisfaction (24%), and moder-
of the WDS. Although the age of installation criterion is ate satisfaction (15%). Sidi Tahar stands out with the highest
stated as the most important parameter in many literatures percentage of good satisfaction (40%), followed by moder-
(Rajani and Tesfamariam 2007; Kabir et al. 2015), still ate satisfaction (45%), and the lowest percentage of poor
Kabir et al. (2015) showed that these factors’ relationships satisfaction (15%). Interestingly, the AHP rankings closely
are not linear, thus calculating the pipe failure weights mirror the survey satisfaction levels in most cases, indicating
requires a complex representation of the relationships. a consistent prioritization pattern. The AHP values for each
While previous studies have primarily focused on the satisfaction level further support this alignment.
prioritization of pipes and the classification of network Our classification of emergency levels was consistent
zones by selecting pipes with higher rehabilitation pri- with 90% of the findings from the survey, confirming the
orities, our study extends beyond this by implementing a validity of our results.
novel tool to facilitate prioritization and conduct a more Furthermore, the technical indicators presented in Abdel-
comprehensive assessment. Additionally, our approach baki et al.’s study were instrumental in validating our find-
enables the prioritization process to be more dynamic and ings. For instance, the non-revenue water indicator in the
adaptable, allowing for multiple prioritization scenarios survey indicated a high percentage of water losses, which
to be explored and compared. aligned with our high emergency level category areas. Sim-
It is important to acknowledge that despite the advantages ilarly, the customer complaints indicator revealed a high
of our implemented tool, there are certain limitations that number of complaints in certain areas, which was also con-
need to be addressed. One such limitation is the subjective sistent with our classification of high emergency level areas
nature of weighting criteria and indicators, which can intro- (Abdelbaki et al. 2014).
duce bias into the decision-making process. To overcome
this limitation, it is necessary to explore and integrate other
124 Page 14 of 17 Applied Water Science (2024) 14:124

Fig. 10  Spatial distribution of surveyed regions

Fig. 11  Comparison between

the satisfaction levels of the
surveys and the AHP raking
Applied Water Science (2024) 14:124 Page 15 of 17 124

Conclusions N.K., M.S., and M.G.; approval of the version of the manuscript to be
published by Y.B., C.A., B.R., T.B., N.K., M.S., and M.G. All authors
have read and agreed to the published version of the manuscript.
A new tool has been developed to aid in managing the water
distribution network system in Tlemcen. This tool utilizes Funding Open Access funding enabled and organized by Projekt
ArcGIS and multiple criteria to identify and prioritize the DEAL.
maintenance of faulty network components based on their
level of emergency. The study’s results revealed that 25% Declarations
of the network is in a critical state and should be rehabili- Conflict of interest The authors declare no conflict of interest.
tated first, while 62% is in a medium state of emergency,
and 13% is in a low state. The main criteria influencing the Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
prioritization of pipes are material (26% of total weight),
tion, distribution and reproduction in any medium or format, as long
laying date (26% of total weight), and pressure (20% of total as you give appropriate credit to the original author(s) and the source,
weight). Among the indicators, cast iron pipes (26% of total provide a link to the Creative Commons licence, and indicate if changes
weight) and pipes with a laying date of 20 years (26% of were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
total weight) had the highest weights.
otherwise in a credit line to the material. If material is not included in
Additionally, the validation phase of the study was cru- the article's Creative Commons licence and your intended use is not
cial in ensuring the accuracy and reliability of the obtained permitted by statutory regulation or exceeds the permitted use, you will
results. Field surveys confirmed that the GIS-based approach need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
is effective in identifying potential water infrastructure fail-
ures in the study area, thereby strengthening the credibility
of this study and reinforcing the tool’s potential to guide
water management decision-making processes in Tlemcen References
and other regions. The user-friendly nature of the tool allows
Abdelbaki C, Touaibia B, Amine AM, Fadila KS (2012) Applied sys-
scientists, engineers, and water management officers to uti- temic approach to water supply network the case of an urban clus-
lize it as a guide for urban water management in Tlemcen ter of Tlemcen—Algeria. Procedia Eng 33:30–37. https://​doi.​org/​
and potentially other regions. 10.​1016/j.​proeng.​2012.​01.​1173
Abdelbaki C, Touaibia B, Mahmoudi H et al (2014) Efficiency and per-
This study represents a novel approach to the manage-
formance of a drinking water supply network for an urban cluster
ment of water distribution networks, combining GIS with at Tlemcen Algeria. Desalin Water Treat 52:2165–2173. https://​
multi-criteria decision analysis to prioritize rehabilitation doi.​org/​10.​1080/​19443​994.​2013.​870497
and maintenance actions. The integration of these tools Abdelbaki C, Benchaib MM, Benziada S et al (2017) Management
of a water distribution network by coupling GIS and hydraulic
allows for a more efficient and informed decision-making modeling: a case study of Chetouane in Algeria. Appl Water Sci
process, leading to improved water management and reduced 7:1561–1567. https://​doi.​org/​10.​1007/​s13201-​016-​0416-1
water losses. Furthermore, the tool’s applicability extends Abdelbaki C, Touaibia B, Ammari A et al (2019) Contribution of GIS
beyond Tlemcen, making it suitable for implementation in and hydraulic modeling to the management of water distribution
network. In: Koutsopoulos K, de Miguel GR, Donert K (eds)
various parts of the world. Geospatial challenges in the 21st century. Springer International
Moving forward, to improve the management of water Publishing, Cham, pp 125–150
distribution networks, it is recommended to enhance and Abdelbaki C (2014) Modélisation d’un réseau d’AEP et contribution
customize the developed tool. This could involve including à sa gestion à l’aide d’un SIG - Cas du Groupement Urbain de
Tlemcen. Thesis, Abou Bekr Belkaid
other criteria or indicators, such as environmental and social Abdessamed D, Jodar-Abellan A, Ghoneim SSM et al (2023) Ground-
factors related to water distribution management. Addition- water quality assessment for sustainable human consumption in
ally, integrating real-time monitoring abilities into the tool arid areas based on GIS and water quality index in the watershed
could provide ongoing observations into network perfor- of Ain Sefra (SW of Algeria). Environ Earth Sci 82:510. https://​
doi.​org/​10.​1007/​s12665-​023-​11183-9
mance, allowing for proactive maintenance and response Ali Ahmad B, Shahabi H, Bin Ahmad B (2015) Application of GIS
to changes in the system. This would ultimately improve based multi-criteria analysis in site selection of water reservoirs
operational efficiency and resilience. (Case study: Batu Pahat Malaysia). RJASET 9:995–1005. https://​
doi.​org/​10.​19026/​rjaset.​9.​2593
Al-Zahrani M, Abo-Monasar A, Sadiq R (2016) Risk-based prioritiza-
Author contributions Conception and design of study were performed tion of water main failure using fuzzy synthetic evaluation tech-
by Y.B., C.A., B.R., T.B., and N.K.; acquisition of data was provided nique. J Water Supply Res Technol AQUA 65:145–161. https://​
by Y.B., C.A., B.R., and T.B.; analysis and/or interpretation of data was doi.​org/​10.​2166/​aqua.​2015.​051
done by Y.B., C.A., B.R., and M.G.; drafting the manuscript was pre- Amorocho-Daza H, Cabrales S, Santos R, Saldarriaga J (2019) A new
sented by Y.B., C.A., B.R., and T.B.; revising the manuscript critically multi-criteria decision analysis methodology for the selection of
for important intellectual content was analyzed by Y.B., C.A., B.R.,
124 Page 16 of 17 Applied Water Science (2024) 14:124

new water supply infrastructure. Water 11:805. https://​doi.​org/​ Juan S, Bohorquez J, Salcedo C, et al (2022) A phased methodology for
10.​3390/​w1104​0805 the optimal rehabilitation of a network with intermittent supply of
Aşchilean I, Badea G, Giurca I et al (2017a) Choosing the optimal water based on hydraulic criteria. Valencia (Spain)
technology to rehabilitate the pipes in water distribution systems Kabir G, Tesfamariam S, Francisque A, Sadiq R (2015) Evaluating
using the AHP method. Energy Procedia 112:19–26. https://​doi.​ risk of water mains failure using a Bayesian belief network model.
org/​10.​1016/j.​egypro.​2017.​03.​1109 Eur J Oper Res 240:220–234. https://​doi.​org/​10.​1016/j.​ejor.​2014.​
Aşchilean I, Badea G, Giurca I et al (2017b) Determining priorities 06.​033
concerning water distribution network rehabilitation. Energy Pro- Kendouci MA, Bendida A, Mebarki S, Kharroubi B (2019) Study of
cedia 112:27–34. https://​doi.​org/​10.​1016/j.​egypro.​2017.​03.​1055 the management efficiency of the drinking water supply in arid
Așchilean I, Giurca I (2018) Choosing a water distribution pipe reha- areas: case of Bechar city (southwest of Algeria). Appl Water Sci
bilitation solution using the analytical network process method. 9:1–7. https://​doi.​org/​10.​1007/​s13201-​019-​1081-y
Water 10:484. https://​doi.​org/​10.​3390/​w1004​0484 Kessili A, Benmamar S (2016) Prioritizing sewer rehabilitation pro-
Bensaoula F (2007) Etude de la karstification à partir des données jects using AHP-PROMETHEE II ranking method. Water Sci
de forages : le cas des Monts de Tlemcen (Algérie). Karstologia Technol 73:283–291. https://​doi.​org/​10.​2166/​wst.​2015.​490
49:15–24. https://​doi.​org/​10.​3406/​karst.​2007.​2596 Kilinç Y, Özdemir Ö, Orhan C, Firat M (2018) Evaluation of technical
Berrezel YA, Abdelbaki C, Benabdelkrim MEA (2022) Coupling of performance of pipes in water distribution systems by analytic
GIS and Hydraulic modeling in management of an urban water hierarchy process. Sustain Cities Soc 42:13–21. https://​doi.​org/​
distribution network—a case study of Tlemcen (Algeria). In: 10.​1016/j.​scs.​2018.​06.​035
Heggy E, Bermudez V, Vermeersch M (eds) Sustainable energy- Lima ML, Barilari A, Massone HE, Pascual M (2021) Incorporating
water-environment nexus in deserts. Springer International Pub- local researchers’ and decision makers’ preferences for ground-
lishing, Cham, pp 253–258 water resources management in a spatial multi-voiced decision
Blindu I (2004) Outil d’aide au diagnostic du réseau d’eau potable model. J Environ Manage 302:113954. https://​doi.​org/​10.​1016/j.​
pour la ville de Chisinau par analyse spatiale et temporelle des jenvm​an.​2021.​113954
dysfonctionnements hydrauliques. Phdthesis, Ecole Nation- Marinoni O (2004) Implementation of the analytical hierarchy process
ale Supérieure des Mines de Saint-Etienne ; Université Jean with VBA in ArcGIS. Comput Geosci 30:637–646. https://​doi.​
Monnet—Saint-Etienne org/​10.​1016/j.​cageo.​2004.​03.​010
Boroushaki S, Malczewski J (2008) Implementing an extension of Mesalie RA, Aklog D, Kifelew MS (2021) Failure assessment for
the analytical hierarchy process using ordered weighted averag- drinking water distribution system in the case of Bahir Dar insti-
ing operators with fuzzy quantifiers in ArcGIS. Comput Geosci tute of technology, Ethiopia. Appl Water Sci 11:1–24. https://​doi.​
34:399–410. https://​doi.​org/​10.​1016/j.​cageo.​2007.​04.​003 org/​10.​1007/​s13201-​021-​01465-7
Boukhari S, Djebbar Y, Amarchi H, Sohani A (2018) Application of Molden D (2020) Scarcity of water or scarcity of management? Int
the analytic hierarchy process to sustainability of water supply and J Water Resour Dev 36:258–268. https://​doi.​org/​10.​1080/​07900​
sanitation services: the case of Algeria. Water Supply 18:1282– 627.​2019.​16762​04
1293. https://​doi.​org/​10.​2166/​ws.​2017.​194 Pagano A, Pluchinotta I, Giordano R, Fratino U (2018) Integrating
Boutebba K, Bouziane M, Bouamrane A (2014) Aide à la décision “Hard” and “Soft” infrastructural resilience assessment for water
pour l’optimisation de la gestion des réseaux d’alimentation en distribution systems. Complexity 2018:e3074791. https://d​ oi.​org/​
eau potable. LARHYSS 11:279–296 10.​1155/​2018/​30747​91
Brentan B, Zanfei A, Souza RG, et al (2022) Optimal rehabilitation Panagopoulos GP, Bathrellos GD, Skilodimou HD, Martsouka FA
procedure for intermittent water supply systems. Valencia (Spain), (2012) Mapping urban water demands using multi-criteria analy-
p 15 sis and GIS. Water Resour Manage 26:1347–1363. https://d​ oi.o​ rg/​
Curry RA, Yamazaki G, Linnansaari T et al (2020) Large dam renewals 10.​1007/​s11269-​011-​9962-3
and removals—Part 1: building a science framework to support a Pardo MÁ, Riquelme AJ, Jodar-Abellan A, Melgarejo J (2020) Water
decision-making process. River Res Appl 36:1460–1471. https://​ and energy demand management in pressurized irrigation net-
doi.​org/​10.​1002/​r ra.​3680 works. Water 12:1878. https://​doi.​org/​10.​3390/​w1207​1878
dos Santos Amorim JMB, de Bezerra S, TM, Silva MM, de Sousa Rajabi M, Mansourian A, Talei M (2011) A comparing study between
LCO, (2020) Multicriteria decision support for selection of Ahp, Ahp-Owa and fuzzy Ahp-Owa multi-criteria decision mak-
alternatives directed to integrated urban water management. ing methods for site selection of residential complexes in Tabriz-
Water Resour Manage 34:4253–4269. https://​doi.​org/​10.​1007/​ Iran. J Env Stud 37:77–92
s11269-​020-​02671-9 Rajani B, Tesfamariam S (2007) Estimating time to failure of cast-iron
Elhaj K (2018) Creating a Python toolbox to implement the analytical water mains. In: Proceedings of the institution of civil engineers—
hierarchy process method in ARCGIS water management vol 160, pp 83–88. https://​doi.​org/​10.​1680/​
Fellah HS, Adjim M, Bensaoula F (2016) Identification and cartog- wama.​2007.​160.2.​83
raphy of potential groundwater polluters in Tlemcen urban area Saaty TL (1980) The analytic hierarchy process: planning, priority
(northwest of Algeria). Desalin Water Treat 57:6007–6016. setting, resources allocation. McGraw, New York
https://​doi.​org/​10.​1080/​19443​994.​2014.​992971 Saaty TL (2001) Fundamentals of the analytic hierarchy process. In:
Gül Ş, Firat M (2021) Determination of priority regions for rehabili- Schmoldt DL, Kangas J, Mendoza GA, Pesonen M (eds) The
tation in water networks by multiple criteria decision-making analytic hierarchy process in natural resource and environmental
methods. Sigma J Eng Nat Sci 38:1481–1494 decision making. Springer, Netherlands, pp 15–35
Hassan WH, Jassem MH, Mohammed SS (2018) A GA-HP model Saaty TL (2008) Decision making with the analytic hierarchy process.
for the optimal design of sewer networks. Water Resour Manage IJSSCI 1:83. https://​doi.​org/​10.​1504/​IJSSCI.​2008.​017590
32:865–879. https://​doi.​org/​10.​1007/​s11269-​017-​1843-y Saaty TL (1988) What is the analytic hierarchy process? In: Mitra G,
Hassan WH, Attea ZH, Mohammed SS (2020) Optimum layout Greenberg HJ, Lootsma FA et al (eds) Mathematical models for
design of sewer networks by hybrid genetic algorithm. J Appl decision support. Springer, Berlin, Heidelberg, pp 109–121
Water Eng Res 8:108–124. https://​doi.​org/​10.​1080/​23249​676.​ Salehi S, Jalili Ghazizadeh M, Tabesh M (2018) A comprehen-
2020.​17618​97 sive criteria-based multi-attribute decision-making model for
Applied Water Science (2024) 14:124 Page 17 of 17 124

rehabilitation of water distribution systems. Struct Infrastruct Eng using a street section priority model. Urban Water Journal 13:28–
14:743–765. https://​doi.​org/​10.​1080/​15732​479.​2017.​13596​33 40. https://​doi.​org/​10.​1080/​15730​62X.​2015.​10571​74
Senapati U, Das TK (2022) GIS-based comparative assessment of Tscheikner-Gratl F, Egger P, Rauch W, Kleidorfer M (2017) Com-
groundwater potential zone using MIF and AHP techniques in parison of multi-criteria decision support methods for integrated
Cooch Behar district. West Bengal Appl Water Sci 12:43. https://​ rehabilitation prioritization. Water 9:68. https://​doi.​org/​10.​3390/​
doi.​org/​10.​1007/​s13201-​021-​01509-y w9020​068
Tabesh M, Saber H (2012) A prioritization model for rehabilitation Yoo DG, Kang D, Jun H, Kim JH (2014) Rehabilitation priority deter-
of water distribution networks using GIS. Water Resour Manage mination of water pipes based on hydraulic importance. Water
26:225–241. https://​doi.​org/​10.​1007/​s11269-​011-​9914-y 6:3864–3887. https://​doi.​org/​10.​3390/​w6123​864
Tanyimboh TT, Kalungi P (2009) Multicriteria assessment of optimal Zolfaghary P, Zakerinia M, Kazemi H (2021) A model for the use
design, rehabilitation and upgrading schemes for water distribu- of urban treated wastewater in agriculture using multiple criteria
tion networks. Civ Eng Environ Syst 26:117–140. https://​doi.​org/​ decision making (MCDM) and geographic information system
10.​1080/​10286​60070​18386​26 (GIS). Agric Water Manag 243:106490. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​
Tscheikner-Gratl F, Sitzenfrei R, Stibernitz C et al (2015) Integrated agwat.​2020.​106490
rehabilitation management by prioritization of rehabilitation areas
for small and medium sized municipalities. World environmental Publisher's Note Springer Nature remains neutral with regard to
and water resources congress 2015. American Society of Civil jurisdictional claims in published maps and institutional affiliations.
Engineers, Austin, pp 2045–2057
Tscheikner-Gratl F, Sitzenfrei R, Rauch W, Kleidorfer M (2016) Inte-
grated rehabilitation planning of urban infrastructure systems