Landfill Site Selection Using Multi-Criteria Decision Analysis, Remote Sensing Data, and Geographic Information System Tools in Najran City, Saudi Arabia

Abstract

Many practical issues arise when dealing with solid waste management, and there are also environmental effects to be considered. Selecting landfill sites requires extra care with respect to many factors such as the environment, health hazards for people, and the cost of transportation. Furthermore, cities have their own rules, methods, and practices for managing and selecting the best locations for collecting solid waste. In this research, multi-criteria decision analysis (MCDA) was presented and used to evaluate the appropriateness of and suggest the best locations for landfill sites in Najran, KSA. Some remote sensing data and the ArcGIS software were used to prepare nine thematic layers, including drainage density, groundwater depth, land use, soil type, road network, surface elevation, surface slope, distance from residential areas, and distance from protected areas. To evaluate the layer weightings, a questionnaire with pairwise comparisons was distributed among experts and analyzed using the analytical hierarchy process (AHP) and fuzzy set technique. The one-factor-at-a-time (OFAT) sensitivity test was conducted to test the sensitivity of the evaluated weightings. A landfill suitability index (LSI) map was created using raster calculator tools and divided into five classes: limited suitability (LSI value 1.39–2.49), least suitable (2.50–3.03), relatively suitable (3.04–3.48), suitable (3.49–3.91), and most suitable (3.92–4.66). According to the statistical analysis, 18.0% and 18.2% of the total area were within the most suitable and suitable landfill regions, while 21.2%, 14.9%, and 9.5% accounted for relatively suitable, least suitable, and limited suitability areas, respectively. The employed technique and its findings can provide an appropriate guideline to assist the municipality of Najran city, regional planners, and decision-makers in selecting an optimal landfill site in the future. This study also presented some recommendations to enhance the suitability map of landfill sites in Najran city.

1. Introduction

The amount of municipal solid waste (MSW) being produced worldwide has risen significantly. The World Bank has estimated that due to a fast-growing global population, the generation of municipal waste is likely to increase to 2.2 billion tons by the year 2025. The generated volumes of MSW in Saudi Arabia have been increasing at an alarming rate, with a currently estimated volume of over 15 million tons per year [1]. This trend is expected to continue with the increase in population growth and urbanization, highlighting the urgent need for effective waste management strategies to be implemented. The MSW generation rate varies across the Arab Gulf countries, with Kuwait having the highest rate of 1.4 to 1.8 kg/person/day and Oman having the lowest rate of 0.7 to 1 kg/person/day. However, regardless of the specific rate, all Arab Gulf countries face significant challenges when it comes to managing their municipal solid waste [2]. Figure 1 shows the MSW generation rate for different Arab Gulf countries.

In Saudi Arabia, the MSW generation rate is estimated to be approximately 1.2 kg/person/day. Some of the challenges associated with this include inadequate infrastructure, insufficient funding for waste management projects, and a lack of awareness among the public regarding waste reduction and proper disposal practices. The efforts being made to address these challenges include selecting the best locations for landfill sites, implementing waste-to-energy technologies, and increasing public awareness regarding waste reduction and proper disposal practices.

Waste collection, sorting, and recycling performed in a scientific manner through waste management and evaluation can result in economic benefits, a clean environment, and the construction of sustainable cities [3]. Disposing of MSW via landfill is considered a financially efficient and environmentally responsible method [4,5]. Unfortunately, due to the influence of various factors such as social, environmental, technological, economic, and legal considerations, the task of pinpointing the whereabouts of landfill sites is complex and demanding [6,7,8].

Various techniques have been suggested by previous researchers to address the issues associated with evaluating waste site suitability. These strategies include utilizing geographic information system (GIS) mapping, conducting environmental impact assessments, and engaging with the local community and obtaining their input, as well as carrying out site suitability analysis that takes into account several factors such as the soil type, hydrogeology, topography, and distance from transportation networks [9,10,11]. The combination of GIS tools and MCDA study topics can be mutually beneficial due to the significant value they bring to assessing the different factors.

Although there have been various strategies published in the literature, only a few of them are appropriate for site selection. The analytic network process (ANP) and the analytic hierarchical process (AHP) have the greatest application in solving the landfill site selection problem [12,13,14,15]. The AHP is beneficial due to its ability to handle complex and multi-criteria decision-making processes. Moreover, the AHP allows decision-makers to structure a problem into a hierarchy of criteria and sub-criteria, assign relative weightings to each of these criteria, and then evaluate alternatives based on how well they satisfy the criteria [16,17,18]. Using the AHP for landfill site selection helps to ensure a transparent and systematic decision-making process that takes into account various environmental, social, economic, and technical factors. In addition, the AHP provides a mathematical framework for decision-making that enables decision-makers to prioritize and compare different criteria, leading to a more informed and objective decision. When compared to simple weighted average procedures, this integration provides users with both qualitative and quantitative outcomes [19,20,21,22]. Overall, the use of the AHP for landfill site selection can help mitigate potential conflicts and ensure that the chosen site is the most suitable based on a comprehensive and fair evaluation of all relevant criteria. Furthermore, the use of the AHP can also help to increase public trust and confidence in the decision-making process by providing a clear and rational methodology that can be easily communicated to the public, stakeholders, and other interested parties, ultimately leading to greater acceptance and support for the chosen landfill site.

This paper aimed to provide a comprehensive overview and comparison of two widely used multi-criteria decision-making methods, namely the fuzzy analytic hierarchy process (fuzzy AHP) and the analytic hierarchy process (AHP), and present a structured approach for documenting AHP/fuzzy AHP studies. This paper includes a case study to demonstrate the proposed documentation approach and provides insights into the practical implementation of these methods in real-world decision-making scenarios. Overall, this paper serves as a valuable resource for researchers and practitioners seeking to better understand the similarities and differences between fuzzy AHP and AHP, as well as how to effectively document and implement these methods in their own decision-making processes. Furthermore, this paper highlighted the best landfill location map for Najran city, KSA, and it considered factors such as drainage density, groundwater depth, land use, soil type, hydrogeology, topography, and existing infrastructure and transportation networks. Based on the analysis of these factors, several potential landfill sites were identified and evaluated to determine the optimal location for a new landfill site in Najran city, KSA. The final recommended landfill site was selected based on a comprehensive evaluation of all the factors analyzed, taking into account environmental, social, and economic considerations.

2. Study Area and Dataset

2.1. Study Area

Najran is a city in southwestern Saudi Arabia near the border with Yemen. It is the capital of Najran province and has a population of over 246,000 people. It is situated in a valley at the foot of three mountains, which gives it a unique landscape and climate compared to other regions in Saudi Arabia. The city is located between latitudes 16 and 19 degrees north and longitudes 43 and 48 degrees west (Figure 2a). The climate is hot and semiarid, as it is across the Arabian Peninsula. To find the ideal landfill locations, this region was chosen as the study area. The ground surface height variations in the chosen area range from 1128 to 1932 m above the mean sea level, with moderate to high topography. They include a broad range of land use activities, and residential, urban, and agricultural areas, as well as transportation systems.

The city of Najran’s population expanded dramatically from 254,000 in 2005 to 435,000 in 2023 (https://www.stats.gov.sa/, accessed on 22 March 2023) (Figure 2b), and the highest population increase occurred between 1995 and 2005, where the population increased from 115,000 to 254,000; this has led to increased urbanization and MSW production [23,24].

The city is known for its rich history and cultural heritage, with numerous archaeological sites in the surrounding areas that date back to ancient times. In recent years, Najran has also experienced significant economic growth and development with the establishment of new industries such as cement production, agriculture, and tourism. Najran city has many residential areas, as well as commercial and industrial zones. Additionally, Najran is known for its agriculture, particularly for growing dates and other crops in the surrounding fertile land. The city is also home to several historic sites, such as the Najran Museum and the ruins of Al-Ukhdood.

The city contains 75 residential areas, several markets, and enormous landfill sites where household trash from places including restaurants, houses, and service centers is deposited. Furthermore, there are numerous farms growing a variety of crops in the south side of the city. Based on the density of the population and the urban activity of the area, the main components of Saudi Arabian MSW include organic waste (37%), paper (28.5%), wood (8%), textiles (6.4%), plastic (5.2%), glass (4.6%), and others (10.6%). As a significant proportion of Saudi Arabia’s MSW is composed of organic material, this offers an excellent opportunity for biological waste-to-energy technologies. The city has four main industrial sectors: non-metallic (building) materials, plastic materials, metallic materials, and food, according to the Najran Chamber of Commerce and Industry’s 2014 report. Goods for consumption are produced in Najran factories, including food products such as juice, milk, and drinking water, in addition to products made from chemicals and plastic, including plastic bags, soap, liquid detergents, and household items, and non-metallic building materials such as brick, pre-cast concrete, cement, and numerous ceramic types.

2.2. Dataset

The digital elevation model (DEM) ALOS PALSAR RTC DEM with 12.5 m resolution data was downloaded free of charge from https://asf.alaska.edu/ (accessed on 15 March 2023). Figure 3a shows the DEM satellite data used to analyze the topography of the area for landfill site selection. The surface slope, elevation, drainage density, and mainstream of Wadi Najran (valley) were determined using this data to ensure that the landfill site was suitable in terms of safety, engineering, and environmental aspects. A Sentinel-2 satellite image (cloud-free with 10 m spatial resolution) was acquired on 11 December 2021, downloaded free of charge from the archives of the United States Geological Survey website (https://earthexplorer.usgs.gov/, accessed on 20 April 2023). Access to real-time road map data allows for more accurate and precise analysis, especially in rapidly changing environments where traditional surveyed road data may quickly become outdated. Figure 3b shows the integration of Sentinel-2 satellite data with the residential area and water well locations, which were used as the criteria for landfill site selection through the MCDA and GIS tools.

Google Earth maps were used to verify the classification of the Sentinel-2 image to obtain an accurate land use layer. This process allowed for a more accurate identification of suitable landfill sites, which is critical in minimizing the environmental impacts of landfill operations and ensuring public health and safety. All satellite images used were processed using ArcGIS 10.7 software, re-sampled to a spatial resolution of 10 m, and georeferenced to The Universal Transverse Mercator (UTM) zone 38 north coordinate system. The national geological database portal website (https://ngp.sgs.org.sa/, accessed on 24 April 2023) was used to provide the geology data necessary for the study, including information regarding rock types and geological structures that can affect groundwater flow. In addition, this study also used field surveys to collect data regarding the location, depth, and water quality of water wells in order to assess the potential impact of landfill sites on groundwater resources.

3. Methodology

Nine thematic layers were used in this work to obtain landfill sites: drainage density, groundwater depth, land use/cover, soil type, road network, surface elevation, surface slope, distance from residential areas, and distance from protected areas, based on previous studies [7,25,26]. The use of these thematic layers in conjunction with MCD analysis, remote sensing data, and GIS tools resulted in the identification and selection of optimal landfill sites that met environmental, social, and economic criteria.

The final weightings of each criterion were calculated using AHP and FAHP methods based on expert opinions. This ensured that the decision-making process was objective and transparent, and input from various stakeholders was considered in the final recommendations.

Each layer was divided into five sub-classes and assigned local weightings from one to five depending on its importance in the overall decision-making process.

Each pixel was assigned an integer number from one to five based on its suitability for landfill site selection. A value of five represented a pixel that was most suitable and a value of one represented a pixel that was less suitable. The weightings of the main criteria and the overall and global ratings for each layer were calculated based on expert opinions. The final suitability landfill map was generated by overlaying all of the thematic layers and assigning a final suitability score to each pixel based on the weighted sum of the scores of all layers. Figure 4 shows the methodological framework’s schematic structure.

The five main components for determining a suitable landfill location are presented in the following sections. Defining the main criteria, constraints, and their classification based on the literature review and their weightings using the group fuzzy best–worst technique is the main topic of Section 3.1. The definition of the sub-classes and weighting values for each layer are defined in Section 3.2. Preparing the thematic layer for using the GIS technique is presented in Section 3.3. In Section 3.4 and Section 3.5, brief explanations of the AHP and fuzzy multi-criteria decision-making process are provided.

3.1. Main Criteria Selection

Restrictions or constraints should be taken into account in order to regulate the environment that the solution must meet or stop actions from occurring or developing in an unsuitable manner. Some potential locations will need to be removed when the limitations for landfill sites have been determined, such as airports and archaeological sites [27,28]. In general, a landfill site should not be located in a secured area or private location. Constraints and the privacy distance of landfill sites are determined by the legislation of the KSA’s general environmental regulations and rules for implementation [29]. Several organizations in Saudi Arabia are in charge of the laws governing environmental protection. All of these organizations are governed by an Environmental Council, whose mission is to make Saudi Arabia one of the world’s top 20 most environmentally friendly countries [30]. Additionally, the constraints and criteria that were applicable to this study were chosen based on questionnaires and international expert reviews, such as those documented in [31,32,33,34].

A group of experts familiar with the study area conditions and waste management rules, including municipalities, geologists, environmental health engineers, and professors, were interviewed to define the most effective criteria and their assigned weightings. They were questioned and interviewed regarding their preferences and criteria for site selection, such as drainage and water factors, residential areas, water wells, transportation routes, and environmental factors. After analyzing their opinions, nine criteria, linked to environmental, social, and economic considerations, were chosen for the MCDA process. The drainage density $\left(C_1\right)$, groundwater depth $\left(C_2\right)$, land use $\left(C_3\right)$, soil type $\left(C_4\right)$, road network $\left(C_5\right)$, surface elevation $\left(C_6\right)$, surface slope $\left(C_7\right)$, distance from residential areas $\left(C_8\right)$, and distance from protected areas $\left(C_9\right)$ were selected, as most similar studies had used and recommended these factors [35].

Table 1. The nine selected criteria linked to three categories.

Categories	Criteria Name	Criteria Code
Environmental criteria	Drainage density	$C_1$
	Groundwater depth	$C_2$
	Land use	$C_3$
	Surface slope	$C_7$
Economic criteria	Surface elevations	$C_6$
	Soil type	$C_4$
	Road network	$C_5$
Social criteria	Distance from residential areas	$C_8$
	Distance from protected areas	$C_9$

shows the selected criteria linked to their categories. $C_1$ and $C_2$ are crucial factors in landfill site selection as they can have significant environmental and economic impacts, such as contamination of water sources and increased costs for treating contaminated water.

The following sub-sections describe each criterion used in the analysis and how it was prepared and weighted to determine the most suitable landfill site location.

3.1.1. Drainage Density $\left(C_1\right)$

Surface water pollution is one of the most important factors to take into account when choosing a landfill location; hence, landfill sites should not be located close to sources of surface water (wells, lakes, and streams). Wadies and wetlands are affected by toxic leaks and emissions from landfill sites. Low weighting was attributed to high drainage density areas for suitable landfill sites, and vice versa for low drainage density areas. The drainage density is indicative of the infiltration and permeability of a drainage basin and it is calculated by dividing the total length of all streams by their entire area [36]. When identifying and choosing landfill sites, several studies have used and categorized the drainage density factor as one of the most vital factors that must be considered and assigned a greater weighting [35,37]. ArcGIS was used to create a layer map for the drainage density criteria. Figure 5a shows the range of drainage density from 0 to 4 ($\mathrm{k}\mathrm{m}\mathrm{p}\mathrm{e}\mathrm{r}{\mathrm{k}\mathrm{m}}^2)$ in the Najran city region.

3.1.2. Groundwater Depth $\left(C_2\right)$

The groundwater depth (GWD) ${(C}_1)$ must be given consideration when choosing a landfill site location in order to reduce the danger of water pollution [38]. Areas with deep groundwater are more suitable as landfill sites than those with shallow groundwater. We used water well information collected by the Ministry of Environment and Agriculture (https://www.mewa.gov.sa/ar/Pages/default.aspx, accessed 24 April 2023). The kriging interpolation algorithm was used to generate a GWD map. The kriging method, which is more versatile than other interpolation and spatial averaging techniques, such as inverse distance weighing (IDW) or trend surface and splines, offers the most accurate linear spatial interpolation [39,40,41]. Many studies compared the performance of kriging with other interpolation techniques and consistently found that kriging outperformed them, especially in large areas with fewer sampling points [42]. Figure 5b shows the generated GWD layer. The GWD varied from 21 to 39 m below the earth’s surface.

To reduce the risk of leaching from solid waste and protect water aquifers from pollution, the landfill location should be situated in a region with deep groundwater [43]. The GWD in the study area was divided into five classes. The areas with depths of less than 24 m were assigned a score of 1. A score of 2 was assigned to areas with depths ranging from 24 to 28 m, and those with depths from 28 to 32 m were assigned a score of 3. Areas with depths from 32 to 36 m were assigned a score of 4, and areas with depths greater than 36 m were assigned a score of 5.

3.1.3. Land Use $\left(C_3\right)$

The physical elements of the land’s surface that are able to be observed and measured are referred to as the land cover; the term land use refers to how people use the land. LULC describes the economic and cultural practices that are carried out in a given region, which may be agricultural, residential, commercial, mining, etc. [44]. Certain land uses, such as agricultural and forestry, may not be taken into account due to international requirements for landfill site selection. As a consequence, it is essential to produce a land use map before selecting landfill sites. The LULC map was produced using the supervised maximum likelihood classification (MLC) [45]. Five classes were extracted: roads, the Wadi Najran area, buildings, agriculture, and bare lands. Figure 5c shows the LULC map. To validate the accuracy of the obtained LULC map, the error matrix was calculated [46,47].

Table 2. Overall accuracy and kappa coefficient for the Landsat image.

Class Name	Roads	Wadi Najran	Buildings	Agriculture	Bare Land	Sum
Roads	124	3	4	6	7	144
Wadi Najran	6	135	8	3	5	157
Buildings	7	5	148	5	6	171
Agriculture	9	4	5	80	4	102
Bare Land	3	9	3	3	145	163
Sum	149	156	168	97	167	737
Overall accuracy was 85.75% and kappa coefficient was 0.82.

illustrates the classification design matrix containing the overall accuracy and the kappa coefficient (85.75% and 0.82, respectively).

The layer land use criterion was classified into seven categories after adding another two layers: airports and residential zones. The study area’s land use map is shown in Figure 5c. Based on the literature review and the KSA, the bare land areas were given a weighting of 5 since this was the most suitable location for a landfill site. The roads were assigned a weighting of 4, areas with buildings were assigned a weighting of 3, and agriculture areas were assigned a score of 2. The lowest score of one was assigned to the Wadi Najran area. Table 3 shows assigned weightings for the different land use layers.

3.1.4. Soil Type Classification $\left(C_4\right)$

Using the “soil type classification” criterion, the study area should be insulated from pollution brought on by leachate transportation across the unsaturated zone of the soil layers [48]. Site selection procedures should totally ignore regions that are subject to landslides, volcanic activity, degradation, earthquakes, subsidence, and particularly the formation of karst. The study area’s soil is composed of quaternary surficial alluvial and swamp deposits and different lithological units: sand of different varieties (silt, clay) and minor Eolian sand; terrace gravel, gravel-plain deposits, and silt containing minor sand and gravel; plutonic rocks; and Hornblende granite. The lithological classes for the study area were categorized into five types. A score of 1 was assigned to gravel, sand, silty rocks, and granite, and a score of 2 was assigned to sand. A score of 3 was assigned to silt, and a score of 4 was assigned to rocks. The highest score of 5 was assigned to silty granite soil. Table 3 shows the assigned weightings for the different soil types.

3.1.5. Road Network $\left(C_5\right)$

The road network should be isolated from the land use map and taken into account when choosing landfill locations. According to the KSA’s general environmental regulations and rules, a landfill site cannot be located in protected zones around road networks. The minimum buffer distance for the location of a landfill site is dependent on the type of road; therefore, for the purposes of this study, it was determined to be 100 m. Figure 5e shows the road network in the study area. Table 3 shows the assigned weightings for different buffer zones for the road layer.

3.1.6. Surface Elevations $\left(C_6\right)$

The slope and land altitude criterion are considered to be major elements when choosing a landfill location. This is because transporting waste to the landfill location would be challenging in the presence of steep slopes [49]. Soil pollution results from drainage water leaching into flat regions and water bodies from steep hills. From the ALOS DEM image, the surface elevation map was prepared using ArcGIS 10.7 software. The elevation values inside the study area range from 1228 to 1932 m above the mean sea level. The elevation values were divided into five categories and given weightings of 5, 4, 3, and 1 for each group: 1120–1135, 1145–1160, 1160–1175, and more than 1190 m above the mean sea level. The study area’s elevation categories are shown in Figure 5c. Table 3 shows the assigned weightings for the different elevations.

3.1.7. Surface Slope $\left(C_7\right)$

Surface water flow around the landfill site, waste transportation, and access to the landfill site all depend on the soil slope [50,51]. High earth elevation variations cause a high slope on the surface, which allows contaminated leachate to infiltrate the surrounding residential areas. A digital map for slope values was extracted using the DEM of Najran city. The slope values inside the study area range from 0 degrees to 68 degrees. Five categories were identified for the surface slope: less than 5, 5–10, 10–15, 15–20, and more than 20 degrees; these were given ratings of 5, 4, 3, 2, and 1, respectively, according to the literature. Table 3 shows the assigned weightings for the surface slope.

Table 3. Sub-criteria classes and associated ratings.

Item	Main Criteria	Sub-Criteria Class	Rating	References
1	$\mathrm{Drainage}\mathrm{density}\left(\mathrm{k}\mathrm{m}\mathrm{p}\mathrm{e}\mathrm{r}{\mathrm{k}\mathrm{m}}^2\right)$	<0.75	5	[14,35,52]
		0.75–1.5	4
		1.5–2.25	3
		2.25–3.0	2
		>3.0	1
2	Groundwater depth (m)	<24	1	[35,53]
		24–28	2
		28–32	3
		32–34	4
		>34	5
3	Land use	Wadi Najran	1	[37,54]
		Agriculture	2
		Buildings	3
		Roads	4
		Bare Land	5
4	Soil type	Gravel	1	Saudi Geological Survey map scale 1:250,000, [37]
		Sand	2
		Silt	3
		Rocks	4
		Granite	5
5	Road distance (m)	100–250	5	[37,53]
		250–500	4
		500–750	3
		750–1000	2
		>1000	1
6	Surface elevation (m)	1140–1180	5	[37]
		1180–1220	4
		1220–1260	3
		1260–1300	2
		>1300	1
7	Surface slope (degree)	<5	5	[37,54]
		05–10	4
		10–15	3
		15–20	2
		>20	1
8	Distance from residential areas (m)	0–1000	1	[37,54]
		1000–2000	2
		200–3000	3
		3000–4000	4
		>4000	5
9	Distance from protected areas (m)	3000–4000	1	[37,53,55]
		4000–5000	2
		5000–6000	3
		6000–7000	4
		>7000	5

Table 3. Sub-criteria classes and associated ratings.

Item	Main Criteria	Sub-Criteria Class	Rating	References
1	$\mathrm{Drainage}\mathrm{density}\left(\mathrm{k}\mathrm{m}\mathrm{p}\mathrm{e}\mathrm{r}{\mathrm{k}\mathrm{m}}^2\right)$	<0.75	5	[14,35,52]
		0.75–1.5	4
		1.5–2.25	3
		2.25–3.0	2
		>3.0	1
2	Groundwater depth (m)	<24	1	[35,53]
		24–28	2
		28–32	3
		32–34	4
		>34	5
3	Land use	Wadi Najran	1	[37,54]
		Agriculture	2
		Buildings	3
		Roads	4
		Bare Land	5
4	Soil type	Gravel	1	Saudi Geological Survey map scale 1:250,000, [37]
		Sand	2
		Silt	3
		Rocks	4
		Granite	5
5	Road distance (m)	100–250	5	[37,53]
		250–500	4
		500–750	3
		750–1000	2
		>1000	1
6	Surface elevation (m)	1140–1180	5	[37]
		1180–1220	4
		1220–1260	3
		1260–1300	2
		>1300	1
7	Surface slope (degree)	<5	5	[37,54]
		05–10	4
		10–15	3
		15–20	2
		>20	1
8	Distance from residential areas (m)	0–1000	1	[37,54]
		1000–2000	2
		200–3000	3
		3000–4000	4
		>4000	5
9	Distance from protected areas (m)	3000–4000	1	[37,53,55]
		4000–5000	2
		5000–6000	3
		6000–7000	4
		>7000	5

3.1.8. Distance from Residential Areas $\left(C_8\right)$

The locations of the landfill sites should be far from residential areas to avoid noise and environmental pollution. On the other hand, they should not be too far away in order to decrease transportation costs. The study area consists of 75 residential communities; their boundaries are shown in Figure 2b. Five buffered zones received five different scores: zones that were less than 1000 m from residential communities received a score of 1, those between 1000 and 2000 m received a score of 2, those between 3000 and 4000 m received a score of 4, and zones that were more than 4000 m from residential communities received the best score of 5. Table 3 shows the assigned weightings of the buffered distance from a residential area.

3.1.9. Distance from Protected Areas $\left(C_9\right)$

There are many protected feature layers inside the study area that should be considered and identified when selecting potential landfill sites. Generally, the landfill site should be situated a buffer distance away from specific geographic features such as airports, cultural heritage areas, military zones, and university campuses [37]. This is in order to preserve such important features, which are recognized as the most valuable source of national resources. Based on the potential hazards to the environment and human health, as well as the government, scores of 1, 2, 3, 4, and 5 were assigned based on five buffering distances of 0–1000 m, 1000–3000 m, 3000–5000 m, 5000–7000 m, and more than 7000 m away from these locations. Table 3 shows the membership value considered for the protected area criterion.

3.2. Sub-Class Rating Values

Rating sub-criteria classes in the analysis of landfill site selection requires a comprehensive and scientifically sound approach that takes into account the existing literature, the unique geographical conditions of the target region, and near similar study areas. By combining these three factors, a more accurate and effective rating system can be developed. This may involve analyzing the region’s hydrology, topography, geology, and other relevant factors. By understanding the unique characteristics of the region, the rating system can be tailored to the specific needs of the area. For example, in the case of Najran municipality, an analysis of the region’s geographical conditions revealed that high drainage density increases the likelihood of surface water pollution, while shallow groundwater depth increases the risk of groundwater contamination. Therefore, lower drainage density and deeper groundwater were deemed more suitable sub-criteria for landfill site selection in Najran municipality.

The nine GIS layers that represented the main criteria were reclassified and assigned integer numbers from one to five as weightings. These weightings were then used to determine the optimal landfill site locations in accordance with the KSA’s general environmental regulations and rules [37]. In cases where such rules were absent, worldwide experiences or previous studies carried out in places with equivalent environmental conditions compared with the current study site were used to define these weightings. The limitations for each sub-class and their corresponding references can be found in Table 3. The rating system ranged from 1 to 5, with each value representing a level of suitability; a value of 1 represented restricted areas and a value of 5 represented the most suitable areas.

The rating values for each GIS layer were determined based on the limitations outlined in Table 3. The nine resulting GIS layers were reclassified according to predefined criteria and environmental regulations, as depicted in Figure 5.

Using buffer distance data recommended in Table 3 from previous studies, all nine layers were utilized to generate different maps with pixel values (weightings) ranging from 1 to 5 using ArcGIS software. These maps are depicted in Figure 6. The integer values assigned to each sub-criteria represent their suitability for landfill site selection, with 1 indicating restricted places, 2 as the least appropriate, 3 as somewhat appropriate, 4 as appropriate, and 5 as the most appropriate.

The factors of drainage density, groundwater depth, and land use were identified as critical in determining suitable landfill sites in Najran municipality based on the unique environmental and geographical characteristics of the region.

In terms of drainage density, Najran has a high density of wadis (seasonal watercourses) and valleys, which can increase the likelihood of surface water pollution. These areas act as conduits for water flow during the rainy season, and pollutants from landfill sites can easily contaminate these watercourses. Therefore, low drainage density areas were considered more suitable for landfill sites.

Groundwater can be accessed from a shallow alluvial aquifer located along Wadi Najran, but this also means that it is at a higher risk of becoming contaminated. Landfill sites with impervious liners and other protective measures can help prevent pollutants from seeping into groundwater, but a shallow water table can increase the risk of contamination. Therefore, it was deemed important to locate landfill sites in areas with deeper groundwater.

In terms of land use, Najran has significant agricultural activity, and locating a landfill site in an agricultural area could pose a risk to the local food supply and the environment. Therefore, it was necessary to consider land use when identifying suitable landfill sites.

While similar factors may be important in other regions, the specific characteristics of each region will influence the relative importance of different criteria. For example, regions with different topography, hydrology, and land use patterns may require different criteria to be considered when identifying suitable landfill sites. Therefore, it is important to conduct a detailed analysis of the unique environmental and geographical characteristics of each region to determine which factors are most important in the landfill site selection process.

3.3. GIS Analysis for Landfill Site Selection

The ArcGIS version 10.7 software was used to combine multiple thematic data layers into a single map. Pre-analysis and main analysis represent the two-stage processing of an ArcGIS analysis. The satellite data and maps must be integrated and unified during the pre-analysis process, as the satellite data differ greatly in terms of format, size, coordinate system, and scale. Producing land use maps, a surface slope layer, and buffer distances for specific geographic features such as airports, highways, rivers, and other such items constitutes part of the main analysis process. The raster calculator tool was used to combine all of the layers and their weightings to produce the final landfill site location map. The processing methodology used in the current study to define an appropriate map for landfill sites using the FAHP and ArcGIS software is shown in Figure 4. A landfill suitability index ($LSI$) map was created by multiplying each layer by its weighting, adding the results, and excluding restricted areas Equation (1).

$${LSI}_i={\sum }_{j=1}^n{w}_j{x}_{ij}$$

(1)

Here, ${LSI}_i$ is the suitability index for pixel $i{,w}_j$ is the weighting of criterion j, $x_{ij}$ is the value of pixel $i$ in relation to criterion $j$, and $n$ is the total number of used criteria. The best places for landfill sites will be those with a high $LSI$ value.

3.4. Classical AHP

The analytic hierarchy process (AHP) is frequently employed as a multi-criteria decision-making method (MCDM) [56]. The AHP uses pairwise comparisons of numerous options in relation to various criteria and offers a decision-aid tool for problems that involve many criteria. Any AHP model starts with the objective, and the second and third levels are the criterion and sub-criteria, respectively. The fourth level contains the alternatives presented in the end [57]. The AHP assists in calculating the weighting for each factor using the decision matrix’s normalized standard vector $A$, as described in Equation (2).

$$A=\left[\begin{array}{ccc}1& \dots & {sc}_n\\ \cdots & 1& \dots \\ {sc}_n& \dots & 1\end{array}\right]$$

(2)

The typical preference scale for AHP is a 1–9 scale, which ranges from “equal significance” to “great importance”, while other evaluation scales, including 1–5, may also be used frequently. A is a reciprocal matrix. A nine-point scale of 1–9 and 1/2–1/9 was employed for the direct and inverse relationships, respectively.

Table 4. Nine-point scale pairwise comparisons for the Saaty and fuzzy triangular scales.

Saaty Scale	Importance	Fuzzy Triangular Scale
1	The same importance	(1,1,1)
2	Intermediate preference	(1,2,3)
3	Weakly important	(2,3,4)
4	Intermediate preference	(3,4,5)
5	Fairly important	(4,5,6)
6	Intermediate preference	(5,6,7)
7	Strongly important	(6,7,8)
8	Intermediate preference	(7,8,9)
9	Absolutely important	(9,9,9)

presents the nine-point scale pairwise comparisons for the Saaty and fuzzy triangular scales.

Matrix A is called the consistency matrix (CR) if all of the comparisons are perfect in the DM judgment matrix. The maximum eigenvalue or maximum lambda (${\lambda }_{max})$ of matrix A was used to determine the consistency of the judgments made. The ${\lambda }_{max}$ is one technique used by the AHP method to calculate the final weightings (w). The calculation of the weightings can be explained by Equation (3):

$$Aw\cong nw\Rightarrow \left(A-{\lambda }_{max}I\right)w=0\Rightarrow {\lambda }_{max}=\frac{1}{n}\sum _{i=1}^n\frac{{\left(Aw\right)}_i}{w_i}$$

(3)

where $A$ is the comparison matrix, $w$ is the weighting matrix, ${\lambda }_{max}$ represents the eigenvalues, and $\mathit{n}$ is the number of alternatives or criteria.

The performance evaluation consistency index (CI) is determined in Equation (4), which ensures precise and accurate judgments in reality.

$$CI=({\lambda }_{max}-n)/\left(n-1\right)\Rightarrow w=(w_1,w_1,\dots ,w_n)$$

(4)

The overall consistency ratio $\left(CR\right)$ of the criteria used should be less than 10%.

Table 5. The random consistency index.

$\mathit{n}$	1	2	3	4	5	6	7	8	9
$RI$	0.0	00	0.52	0.89	1.12	1.26	1.36	1.41	1.46

shows the random consistency index $\left(RI\right)$ dependent on the number of criteria used [16].

The CR values are calculated using Equation (5).

$$CR=CI/RI$$

(5)

If the $CR$ computed is equal to or greater than 0.10 in the pairwise comparison matrix’s judgment result, it is regarded as inconsistent. Therefore, it is essential to investigate and modify the judgments accordingly.

3.5. Fuzzy Set Theory

Fuzzy set theory is also used as a model technique for MCDM [58]. The fuzzy AHP technique can be thought of as an advanced analytical approach that evolved from the standard AHP. Numerous researchers who have studied fuzzy AHP, an extension of Saaty’s theory, have presented evidence that fuzzy AHP gives a better explanation of these types of decision-making processes than the standard AHP approaches [59,60,61,62]. Figure 7 illustrates a triangular fuzzy number (TFN), the triangular-donated $P_i=(l_i,m_i,u_i)$, where $l_i$, $m_i$, and $u_i$ are the lowest, most probable, and largest possible values, respectively. Equation (6) is the definition of the membership function of $M\left(\mathsf{\mu }\right)$.

$$\mathrm{\mu }(x|M)=\left\{\begin{array}{c}0\left(x<l\right)\\ \frac{x-l}{m-l}\left(l\le x<m\right)\\ \frac{\mathrm{\mu }-x}{\mathrm{\mu }-m}\left(m\le x<\mathrm{\mu }\right)\\ 0(x\ge \mathsf{\mu })\end{array}\right.$$

(6)

A judgment fuzzy matrix is created assuming we have three criteria, which include the ratio of the judgment matric $A_{n\times n}$ described in Equation (7):

$$\tilde{A}=\left(\begin{array}{ccc}(1,1,1)& (\mathrm{3,4},5)& (\mathrm{2,3},4)\\ {\tilde{a}}_{12}& (1,1,1)& (\mathrm{1,2},3)\\ {\tilde{a}}_{13}& {\tilde{a}}_{23}& (1,1,1)\end{array}\right)$$

(7)

where any ${\tilde{a}}_{12}$ means the left, center, and right values of the fuzzy number. The four steps detailed below can be used to carry out the FAHP calculation and obtain the weightings of the different criteria [63]:

Step 1: Calculate the geometric mean based on Buckley’s geometric [64], as in Equation (8):

$${\tilde{r}}_i={({\tilde{r}}_{i1}\times \otimes {\tilde{r}}_{i2}\dots {\tilde{r}}_{in})}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.}$$

(8)

where $n$ represents the rank of the matrix.

Step 2: Calculate fuzzy weighting values ${\tilde{w}}_i$, where “n” represents the rank of the matrix:

$${\tilde{w}}_i={\tilde{r}}_i\otimes {({\tilde{r}}_1\oplus {\tilde{r}}_2\oplus \dots \oplus {\tilde{r}}_n)}^{-1}$$

(9)

where ${\tilde{w}}_i=\left({lw}_i,{mw}_i,{uw}_i\right)$ represents the lower, middle, and upper values of the fuzzy weighting, respectively.

Step 3: De-fuzzification by calculating the center of the area.

$$w_i=\left(\frac{{lw}_i,{mw}_i,{uw}_i}{3}\right)$$

(10)

De-normalization, namely:

$$w_i=\frac{{\overline{w}}_i}{{\sum }_{j=1}^n{\overline{w}}_i}$$

(11)

and the eigenvectors are expressed as:

$$W={\left[w_1,w_2,\dots \dots ..w_n\right]}^T$$

(12)

Step 4: Calculate the greatest eigenvalue of the judgment matrix:

$${\mathit{\lambda }}_{\mathit{m}\mathit{a}\mathit{x}}={\sum }_{i=1}^n\frac{A{W}_i}{n{W}_i}$$

(13)

The mean random consistency index $\left(RI\right)$ is selected from Table 5 based on the number of used criteria $n$. Finally, the overall $CR$ (see Equation (5)) value should be calculated, and it should be less than 10% to verify the consistency of the FAHP. The consistency check is helpful for investigating the relationships between each index and evaluating the harmony of the experts’ opinions.

4. Results

A group of experts familiar with the study area conditions and waste management rules including municipalities, geologists, environmental health engineers, professors, stakeholders, and students was formed, and they were interviewed to determine the main and sub-criteria define their weightings. They were questioned and interviewed to identify the study’s constraints and criteria. After analyzing their opinions, the selected factors were classified into nine criteria based on three main categories: environmental, economic, and social.

4.1. Criteria Weighting Based on the AHP

Each thematic layer used nine layers and their sub-class received a weighting using the fuzzy AHP approach. Each criterion was given a relative weighting according to previous studies and expert opinion.

Table 6. The first assessment compared criteria using the AHP weighting method by an expert.

	${\mathit{C}}_1$	${\mathit{C}}_2$	${\mathit{C}}_3$	${\mathit{C}}_4$	${\mathit{C}}_5$	${\mathit{C}}_6$	${\mathit{C}}_7$	${\mathit{C}}_8$	${\mathit{C}}_9$	${\mathit{w}}_{\mathit{i}}$
$C_1$	1	2	3	5	6	7	8	8	9	0.331
$C_2$	0.5	1	1	4	5	6	6	6	8	0.219
$C_3$	0.33	1	1	2	2	4	5	5	7	0.158
$C_4$	0.2	0.25	0.5	1	1	3	3	4	6	0.089
$C_5$	0.16	0.2	0.5	1	1	2	2	2	5	0.071
$C_6$	0.14	0.16	0.25	0.33	0.5	1	1	2	4	0.043
$C_7$	0.12	0.16	0.2	0.33	0.5	1	1	2	3	0.040
$C_8$	0.12	0.16	0.2	0.25	0.5	0.5	0.5	1	3	0.031
$C_9$	0.11	0.12	0.14	0.16	0.2	0.25	0.33	0.33	1	0.017
${\lambda }_{max}$$=9.4,CI$$=0.05,RI$$=1.46,CR$ = 3.4%

shows the first expert opinion design matrix and obtained weighting based on the AHP algorithm, where $w_i$ is the weighting of criterion $i$.

The first expert opinions passed the consistency test, where the $CR$ value was less than 10%. All expert’s opinions were analyzed, and their final weightings were calculated.

Table 7. The AHP method utilized four experts’ opinions to assign different criterion weightings and $CR$ values.

id	${\mathit{C}}_1$	${\mathit{C}}_2$	${\mathit{C}}_3$	${\mathit{C}}_4$	${\mathit{C}}_5$	${\mathit{C}}_6$	${\mathit{C}}_7$	${\mathit{C}}_8$	${\mathit{C}}_9$	${\mathit{\lambda }}_{\mathit{m}\mathit{a}\mathit{x}}$	$\mathit{C}\mathit{R}\mathbf{\%}$
Exp.-1	0.331	0.219	0.158	0.089	0.071	0.043	0.040	0.031	0.017	9.4	3.4
Exp.-2	0.319	0.210	0.174	0.107	0.070	0.054	0.027	0.024	0.015	9.7	6.1
Exp.-3	0.253	0.242	0.173	0.111	0.089	0.065	0.026	0.022	0.018	9.4	3.7
Exp.-4	0.293	0.215	0.188	0.103	0.081	0.051	0.026	0.021	0.021	9.8	6.4
$w_{avg.}$	0.299	0.222	0.173	0.102	0.078	0.053	0.030	0.024	0.018

shows the obtained weightings and the final average weightings $w_{avg.}$ For each criterion based on all expert opinions.

4.2. Criteria Weighting Based on the FAHP

Table 8. The first assessment compared criteria using the FAHP weighting method by an expert.

	${\mathit{C}}_1$	${\mathit{C}}_2$	${\mathit{C}}_3$	${\mathit{C}}_4$	${\mathit{C}}_5$	${\mathit{C}}_6$	${\mathit{C}}_7$	${\mathit{C}}_8$	${\mathit{C}}_9$
$C_1$	(1,1,1)	(1,2,3)	(2,3,4)	(4,5,6)	(5,6,7)	(6,7,8)	(6,8,9)	(7,8,9)	(9,9,9)
$C_2$		(1,1,1)	(1,1,1)	(3,4,5)	(4,5,6)	(5,6,7)	(5,6,7)	(5,6,7)	(7,8,9)
$C_3$			(1,1,1)	(1,2,3)	(1,2,3)	(3,4,5)	(4,5,6)	(4,5,6)	(6,7,8)
$C_4$				(1,1,1)	(1,2,3)	(2,3,4)	(2,34)	(3,4,5)	(5,6,7)
$C_5$					(1,1,1)	(1,2,3)	(1,2,3)	(1,2,3)	(4,5,6)
$C_6$						(1,1,1)	(1,1,1)	(1,2,3)	(3,4,5)
$C_7$							(1,1,1)	(1,2,3)	(2,3,4)
$C_8$								(1,1,1)	(2,3,4)
$C_9$									(1,1,1)

shows the first expert opinion design matrix based on the FAHP algorithm.

Table 9. The nine criteria weightings used based on the FAHP method.

	$\mathbf{Geometric}\mathbf{Means}{\mathit{r}}_{\mathit{i}}$			$\mathbf{Fuzzy}\mathbf{Weighting}{\mathit{w}}_{\mathit{i}}$			Averaged Weight	Normalized Weight
$C_1$	3.617	4.479	5.232	0.218	0.331	0.491	0.347	0.325
$C_2$	2.340	2.957	3.563	0.141	0.219	0.335	0.231	0.217
$C_3$	1.661	2.138	2.706	0.100	0.158	0.254	0.171	0.160
$C_4$	0.956	1.206	1.533	0.058	0.089	0.144	0.097	0.091
$C_5$	0.682	0.956	1.262	0.041	0.071	0.118	0.077	0.072
$C_6$	0.449	0.584	0.735	0.027	0.043	0.069	0.046	0.044
$C_7$	0.424	0.544	0.705	0.026	0.040	0.066	0.044	0.041
$C_8$	0.324	0.418	0.596	0.019	0.031	0.056	0.035	0.033
$C_9$	0.198	0.231	0.281	0.012	0.017	0.026	0.018	0.017
Sum	10.649	13.513	16.613				1.067	1.000
Reverse	0.094	0.074	0.060
Inc. order	0.060	0.074	0.094

shows the results of the FAHP process based on the first expert opinion. Geometric means, fuzzy weighting, average eight, and normalized weightings were used for each criterion.

Table 10. Different criteria weightings and $\mathrm{C}\mathrm{R}$ values based on four expert’s opinions using the FAHP weighing method.

id	${\mathit{C}}_1$	${\mathit{C}}_2$	${\mathit{C}}_3$	${\mathit{C}}_4$	${\mathit{C}}_5$	${\mathit{C}}_6$	${\mathit{C}}_7$	${\mathit{C}}_8$	${\mathit{C}}_9$
Exp.-1	0.325	0.217	0.160	0.091	0.072	0.044	0.041	0.033	0.017
Exp.-2	0.315	0.209	0.173	0.109	0.072	0.055	0.028	0.025	0.015
Exp.-3	0.249	0.231	0.181	0.115	0.090	0.069	0.026	0.021	0.018
Exp.-4	0.309	0.226	0.171	0.094	0.075	0.057	0.027	0.020	0.021
$W_{avg.}$	0.299	0.221	0.171	0.102	0.077	0.056	0.030	0.025	0.018

shows the obtained weightings and the final average weightings $W_{avg.}$ for each criterion based on all expert opinions.

After applying both the AHP and FAHP methods to determine the criteria weightings $w_{avg.}$, it could be observed that they yielded very similar results. This is evident in Table 7 and Table 10, which show a close correlation between the obtained weightings from each method. Furthermore, when considering all layers representing environmental factors as a whole ($C_1,C_2,C_3,and{C}_7$), they accounted for 72.2% of the total weighting, while economic factors ($C_4,C_5,and{C}_6$) accounted for 23.5% and social criteria ($C_8,and{C}_9$) accounted for 4.3% of the total weighting.

4.3. Analysis of the LSI Classification Map

To determine the landfill suitability index, we used Equation (1) with tools in ArcGIS software version 10.7 and considered average weightings from both the AHP and FAHP methods. The resulting index values ranged between 1.39 and 4.66. These values were categorized into five suitability areas based on their level of suitability: limited suitability, least suitable, relatively suitable, suitable, and most suitable. These categories were overlaid with residential zone layers in Figure 8a, where the blue color indicates the limits of the most suitable areas and the red color indicates those that are suitable. Some of the most suitable and suitable landfill locations were located within residential areas; however, it is not appropriate to have landfill sites located within residential areas, as this can lead to various health hazards for the people living nearby. Landfill sites are known to emit toxic gases and foul odors that can cause respiratory problems, skin irritation, and even cancer in some cases. Additionally, these landfill sites attract scavengers, such as rats and other rodents, that carry diseases that can be harmful to humans. Therefore, it would be wise to locate landfill sites away from residential areas where possible so as not to put the health of innocent citizens at risk. Instead, alternative locations, such as remote areas, should be considered for landfill categorization to minimize the negative impact on nearby residents. Figure 8b displays the ultimate landfill classifications upon removal of the residential area after applying the erase tool in ArcGIS. The landfill suitability map can be made more practical and useable by overlaying it with the Sentinel-2 satellite image of the study area, as shown in Figure 8c. This will provide a clearer visualization for land managers to make informed decisions regarding waste disposal.

To sum up, after evaluating the most appropriate class for landfill sites, it was found that there are no viable locations in the western and central regions of the study area. Therefore, to minimize transportation expenses, we merged the suitable areas with the most favorable ones to produce a final landfill site suitability map. While Figure 8d may depict suitable landfill areas, merging two suitability classes could result in potential environmental hazards. The fact that the proposed areas are on bare land only adds to the concern as it suggests there is no existing infrastructure or ecosystem to mitigate any negative impact on the surrounding environment.

Table 11. LSI category statistics before and after the removal of suitability classes within residential zones.

Degree of Suitability	LSI Value	Area before	Area after
	(Range)	km2	km2
Limited suitability	1.39–2.49	26.13	05.86
Least suitable	2.50–3.03	41.03	05.67
Relatively suitable	3.04–3.48	58.33	15.53
Suitable	3.49–3.91	49.94	15.59
Most suitable	3.92–4.66	49.94	18.98
Total	-	49.39	61.63

provides a comparison of LSI categories before and after excluding areas within residential zones. The most suitable area, after the removal of residential areas, covered 18.98 square km, while the suitable area covered an area of 15.59 square km. Additionally, the relatively suitable, least suitable, and limited suitability areas covered an expanse of 15.53, 5.67, and 5.86 square kilometers, respectively, based on the statistics obtained from the LSI analysis in this study.

4.4. Sensitivity Analysis

Any multi-criteria input values and criteria arrays are susceptible to uncertainty and change. Consequently, a sensitivity analysis was performed. A sensitivity analysis demonstrates how sensitive the findings are to changes in factors and whether they are stable [65,66,67]. The final landfill suitability index coefficient at any pixel was certainly impacted by the estimated input weighting for each thematic layer at that pixel. The one-factor-at-a-time (OFAT) method was used to carry out a sensitivity analysis by changing the weightings of the criteria and quantifying changes in the outputs [68]. The weighting of each layer changed within a particular percentage interval, while all other layers were adjusted to their nominal values to meet the FAHP model, which required keeping the weighted sum equal to one. Nine sub-criteria with a range of ±20% weighting variation and 5% interval variation were exposed in order to conduct the sensitivity analysis. To show the spatial change in the results and complete the sensitivity test, 72 simulations were required to be conducted.

Table 12. Results of the simulation runs performed based on a ±20% variation in the weightings of the land use and slope criteria.

Change %	Cell Number in the Suitability Map					Cell Variation from the Base Weighting%					Sum of Cells
	Limited	Least	Relatively	Suitable	Most	Limited	Least	Relatively	Suitable	Most
$C_1$ (Drainage density)
20	251,873	413,257	571,822	509,907	501,662	3.6	0.7	2.0	2.1	1.6	2,248,521
15	245,534	420,099	567,307	513,358	501,923	6.1	2.4	2.8	2.8	1.6	2,248,221
10	246,510	422,462	582,612	496,701	500,236	5.7	2.9	0.1	−0.6	1.3	2,248,521
5	245,820	400,027	546,539	536,525	519,610	6.0	2.5	6.3	7.4	5.2	2,248,521
Base	261,380	410,381	583,370	499,464	493,926	0.0	0.0	0.0	0.0	0.0	2,248,521
−5	223,925	395,460	552,728	550,369	526,039	14.3	3.6	5.3	10.2	6.5	2,248,521
−10	275,409	464,263	578,378	454,521	475,950	5.4	13.1	0.9	9.0	3.6	2,248,521
−15	223,501	391,986	574,228	547,260	511,546	14.5	4.5	1.6	9.6	3.6	2,248,521
−20	261,604	446,618	561,073	393,576	485,650	0.1	8.8	3.8	21.2	1.7	2,148,521
$C_2$ (Groundwater depth)
20	255,333	3753,67	571,147	530,567	516,107	2.3	8.5	2.1	6.2	4.5	2,248,521
15	256,625	382,995	561,967	533,828	513,106	1.8	6.7	3.7	6.9	3.9	2,248,521
10	256,700	382,225	558,649	536,012	514,935	1.8	6.9	4.2	7.3	4.3	2,248,521
5	259,101	380,753	557,749	534,587	516,331	0.9	7.2	4.4	7.0	4.5	2,248,521
Base	261,380	410,381	583,370	499,464	493,926	0.0	0.0	0.0	0.0	0.0	2,248,521
−5	244,362	396,071	548,516	543,995	515,577	6.5	3.5	6.0	8.9	4.4	2,248,521
−10	247,616	408,247	574,078	519,552	499,028	5.3	0.5	1.6	4.0	1.0	2,248,521
−15	243,155	409,039	575,348	531,334	489,645	7.0	0.3	1.4	6.4	0.9	2,248,521
−20	243,840	414,095	572,340	520,502	497,744	6.7	0.9	1.9	4.2	0.8	2,248,521
$C_9$ (Distance from the protected area)
20	261,245	409,791	580,174	493,347	503,964	0.1	0.1	0.5	1.2	2.0	2,248,521
15	261,245	409,802	578,841	495,445	503,188	0.1	0.1	0.8	0.8	1.9	2,248,521
10	261,245	409,802	578,841	495,445	503,188	0.1	0.1	0.8	0.8	1.9	2,248,521
5	261,378	410,260	578,273	500,627	497,983	0.0	0.0	0.9	0.2	0.8	2,248,521
Base	261,380	410,381	583,370	499,464	493,926	0	0	0	0	0	2,248,521
−5	261,781	410,382	584,368	498,364	493,626	0.2	0.0	0.2	0.2	0.1	2,248,521
−10	261,995	410,172	584,742	499,668	491,944	0.2	0.1	0.2	0.0	0.4	2,248,521
−15	262,090	410,289	584,489	499,862	491,791	0.3	0.0	0.2	0.1	0.4	2,248,521
−20	262,092	410,331	583,110	501,760	491,228	0.3	0.0	0.0	0.5	0.5	2,248,521

presents the results of three examples of the sensitivity test conducted for the drainage density, groundwater depth, and distance from protected areas criteria.

Table 12 provides statistical data regarding the sensitivity tests conducted on $C_1,C_2,\mathrm{a}\mathrm{n}\mathrm{d}{C}_9$ criteria with a total of 24 scenarios. The three tested factors had the highest and lowest weighted criteria. Furthermore, determining the percentage change in pixel count for each category based on the modifications made to the layer weighting provided a reliable criterion to measure the layer sensitivity during testing. The average variation in pixel numbers was 5.1%, 4.2%, and 0.4% for the $C_1,C_2,\mathrm{a}\mathrm{n}\mathrm{d}{C}_9$ criteria, respectively.

Table 13. Sensitivity statistical results and pixel percentage change for three test criteria.

Factor	Min. Change %	Max. Change %	Avg. Change %
$C_1$	0.0	21.2	5.1
$C_2$	0.3	08.9	4.2
$C_9$	0.0	02.0	0.4

shows a summary of the statistical sensitively test for three tested factors. The LSI values appeared to be slightly sensitive to drainage density, groundwater depth, and the distance from protected areas (with an average variation of 5.1%, 4.2%, and 0.4%, respectively). This may be because of the reasonable weighting of all of these parameters.

It is important to note that these tests only focused on a limited number of factors. In order to accurately assess the effectiveness of any given solution or approach, a more comprehensive analysis must be conducted that takes into account additional variables such as topography, land use, and soil composition. Without considering all of the possible influences, the accuracy and usefulness of sensitivity tests become questionable. Figure 9 shows that the drainage density criterion was the most sensitive, with an average variation of 5.1%.

5. Discussion

With the growth in the population and living standards in Najran province, the production of municipal solid waste has risen significantly. The municipality is taking measures to manage both water supply and flooding by constructing the Najran Dam, located toward the west of Najran city. However, it is crucial to identify suitable landfill sites for disposing of such a substantial amount of solid waste while ensuring that people and the environment are protected from potential harm.

The current study successfully integrated the AHP and fuzzy AHP methods with various geospatial features to determine appropriate landfill site locations. Expert information was used to calculate criteria weightings, while satellite images and numerical data were collected to create nine thematic layers including drainage density, groundwater depth, land use, soil type, road network, surface elevation, and surface slope, as well as distance from residential areas and protected regions. Each layer was divided into five classes that were rated based on their suitability for a landfill site, ranging from one (restricted region) to five (most suitable location).

By combining the theory of fuzzy sets, Saaty’s AHP is improved by the FAHP [69]. Pairwise comparisons and a triangle with fuzzy numbers were used to determine the weightings of the importance of the decision-making criteria. The FAHP method produced nearly identical weighting values with negligible variation.

The sensitivity analysis involved varying the weights assigned to the different criteria in the AHP analysis and evaluating the impact of these variations on the overall rankings of the alternative solutions. This helped to identify any potential issues related to the interaction between the variables and provided insight into the robustness of the AHP results. The OFAT sensitivity method was utilized to evaluate the responsiveness of each criterion. This involved conducting 24 simulations that modified the weightings of three criteria by ±20% from their base weighting and comparing the resulting number of pixels for each landfill site category. The drainage density and groundwater depth criteria were found to be highly sensitive with a variation of between almost two and five percent in calculating the landfill suitability index. In order to incorporate consideration of local approval and environmental concerns, a post-suitability field investigation is recommended.

Several studies have been conducted using the AHP and geospatial analysis to determine suitable landfill sites. For example, a study conducted in Nigeria used the AHP and GIS to determine suitable landfill sites based on factors such as surface water bodies, land use, land slope, depth to groundwater, geology, lineament, soil media, built-up area, road network, and airport location [70]. Another study conducted in Iran used a similar approach to determine suitable landfill sites based on factors such as distance from residential areas, distance from rivers, and soil type [71]. In addition, a study conducted in Pakistan used almost the same factors and AHP to determine suitable landfill sites [72]. Ten thematic layers, drainage density, land use, surface slope, elevation, lineament density, normalized difference vegetation index, rainfall, distance from the airport, distance from the road, and soil type, were chosen and weighted for computation of the landfill suitability index [35].

In comparison to these studies, our study integrated the AHP and fuzzy AHP to determine suitable landfill sites based on nine geospatial features, including drainage density, groundwater depth, land use, soil type, road network, surface elevation, surface slope, distance from residential areas, and distance from protected regions. Our results showed that drainage density, groundwater depth, and land use were the most important factors in determining suitable landfill sites in Najran municipality’s environment.

In addition, our study found that the FAHP method produced almost identical weightings with a negligible variation compared to the AHP method. Furthermore, our study produced a landfill suitability index map with five categories of suitability ranging from limited suitability to the most suitable areas. This is similar to the approach used in a previous study conducted in Iran, which also produced a landfill suitability index map with five categories of suitability based on factors such as slope, distance from roads, and distance from urban areas [37]. Overall, our study provides a useful approach for determining suitable landfill sites that can be adapted to different contexts and can support decision-making in waste management. However, it is important to note that post-suitability field investigation is recommended to incorporate local approval and environmental concerns.

6. Conclusions

The current study included a variety of remote sensing datasets and decision-making approaches in order to provide an integrated framework for the landfill suitability site map for a portion of Najran city, Saudi Arabia. After the extensive literature review and expert opinion gathering, nine thematic layers were selected for this research, including drainage density, groundwater depth, land use, soil type, road network, surface elevation, surface slope, distance from residential areas, and distance from protected areas. These layers were developed using satellite and conventional data. Subsequently, the landfill suitability index was calculated and divided into five classes: limited suitability (1.39–2.49), least suitable (2.5–3.03), relatively suitable (3.04–3.48), suitable (3.49–3.91), and most suitable (3.92–4.66). As per the statistical analysis, after excluding residential areas, 30.8% and 25.3% of the total area were designated as the most suitable and suitable landfill site zones, respectively. Furthermore, 25.2%, 9.2%, and 9.5% of the total area accounted for the relatively suitable, least suitable, and limited suitability zones, respectively.

In terms of spatial variance, suitable and most suitable potential landfill sites were found in the northeastern, north central, and south central parts of the watershed. This study concluded that due to the Najran Dam, the high groundwater level, and high drainage density, there are no spaces appropriate for landfill sites on the western side of the study area. The sensitivity analysis was performed to determine the effectiveness of each parameter. The drainage density and groundwater depth play a significant role in the LSI values (5.2% and 4.2, respectively). This may be because of the high theoretical weighting assigned to these criteria during the calculation of the suitability zone of the landfill site.

The current study provides a scientific foundation for suitability analyses that consider multiple criteria. The FAHP with weighted overlay is the most effective and widely used method for choosing city landfill locations. It offers a useful method for assessing the suitability of landfill sites in the Najran watershed, and design engineers, regional managers, and decision-makers can use it as the starting point for discussing whether or not new urban landfill sites are required. Finally, it is recommended that the study’s results be compared with field studies in order to select the best landfill sites. The recommended landfill locations would require some additional fieldwork, such as field examinations of the geological and geotechnical characteristics, questionnaire studies to evaluate public acceptance and customer attitude, and evaluations of construction suitability.

The integration of the AHP and fuzzy AHP methods can lead to improved decision-making in complex and uncertain scenarios. The AHP is a widely used method for multi-criteria decision-making problems, while the fuzzy AHP can handle imprecise and uncertain information. By integrating these two methods, decision-makers can consider multiple criteria, incorporate fuzzy information, and improve the accuracy and transparency of the decision-making process. Additionally, the integration of these methods can lead to novel applications and insights in decision-making, which can drive innovation and new solutions. Therefore, the integration of the AHP and fuzzy AHP methods can enhance novelty in various ways and improve decision-making in real-world scenarios.

The findings and recommendations of this research will serve as a valuable resource for decision-makers and stakeholders involved in waste management in Najran city, KSA. This research also highlights the importance of conducting thorough site evaluations and considering various factors when selecting landfill locations to ensure the long-term sustainability and success of waste management systems. Furthermore, this report emphasizes the need for sustainable waste management practices that prioritize environmental protection and public health while also taking into account economic and social factors.

Overall, this study demonstrates the importance of a multidisciplinary approach in waste management planning and highlights the need for sustainable waste management practices that prioritize environmental protection, public health, and community involvement. The implementation of these sustainable waste management practices can serve as a model for other cities and regions facing similar challenges and contribute to the global effort toward achieving sustainable development goals. In conclusion, the proper management of waste is a critical issue that affects not only the environment but also public health and the overall well-being of communities.

7. Limitations and Future Works

One limitation of this study is the inability to conduct a multi-collinearity test due to the mix of categorical and numerical input layers. Future research can explore methods applicable to both types of variables. Another limitation is the reliance on human judgment for determining the final weights of the AHP, which could introduce errors or biases. Increasing the number of expert opinions can enhance accuracy. Additionally, a map including landfill sites can validate search results.

Future research could test the methodology under extreme weather conditions, contaminated sites, sensitive ecosystems, urban areas, and limited data availability. Integrating additional criteria, testing the methodology in different contexts and comparing it with other decision-making methodologies, and incorporating uncertainty and risk analysis are areas for future research.