Potential landfill site selection for solid waste disposal using GIS-based multi-criteria decision analysis (MCDA) in Yirgalem Town, Ethiopia

Abstract

Proper solid waste disposal is an important socioeconomic and environmental concern for all developing countries like Ethiopia. Due to a lack of proper disposal sites, infrastructure, and knowledge on suitable site selection, most cities/towns wrongly prefer water bodies, roadsides, and open areas as the best suitable for solid waste disposal. Such improper site selection will create morphological changes that lead to environmental hazards in the urban and surrounding areas. This study aims to identify suitable sites for solid waste disposal in Yirgalem town using the existing GIS-based multi-criteria decision analysis (MCDA) techniques. Thematic layers of road and stream network, settlement, land cover, slope, view shade, dumping site, river, and soil type and soil depth were considered as primary criteria and weights for criteria, and sub-criteria were assigned by MCDA analysis. Then, each criterion is mapped using the GIS technique, and a suitability map is prepared by overlay analyses. The consistency ratio of 0.049405 which is <0.10 endorsed the accuracy of the pairwise comparison matrix. The results indicate that 29.06% of the study area is unsuitable for the solid waste disposal site. However, 35.47% of the study area was marginally suitable, 27.84% moderately suitable, and 7.63% of the area was highly suitable is preferable for the solid waste disposal site.

Keywords:

• Municipal solid waste
• landfill site
• multi-criteria decision analysis
• remote sensing
• AHP
• GIS
• Yirgalem

Reviewing Editor:

• Sanjay Kumar Shukla, Edith Cowan University, Australia

Subjects:

• Earth Sciences
• Environmental Studies
• Environmental Management
• Environment & Health
• Environment & the City
• Databases

1. Introduction

Industrialization and economic development of the world country intensify avoidable waste, particularly in Asia and the Pacific region have led to increases in the quantity and complexity of generated waste avoidable human activities are called waste they may either be in solid or in a liquid state (Ejaz et al., 2010). Solid waste dump institutions solid waste has been discharged to closer or distant locations. In urban areas, solid waste is deposited on the outskirts and in an open area. After due time this area becomes part of the center of the given urban area. Plastic and other indecomposable solid waste long lasting at the location it is dumped. Solid waste management systems cover all actions that seek to reduce the negative impacts on health, the environment, and the economy. Developing countries are seriously facing associated problems in the collection, transportation, and disposal of communal solid waste (McDougall et al., 2008; Udomporn et al., 2009). In most developing countries, waste management is done by local authorities but different research has shown that these countries have insufficient and inefficient waste management systems (Barton et al., 2014).

Similarly, in Yirgalem town, the solid waste problem is increasing at an alarming rate due to rapid population growth and urbanization. This produced an increasing solid waste generation from different sources and most of the solid wastes that are generated remain uncollected and simply dumped in inappropriate places, such as in drainage lines, open spaces, and street sides. There were also no solid waste storage containers and dust bins on the roadside, and fence protecting the entrances of different animals. In addition, the problem is even becoming more severe since the solid waste is simply disposed of in one open dump site that has not been selected following any scientific way for suitability and it’s very close to the nearby water bodies like streams and rivers, education centres, and residential areas.

Generally, in Ethiopia, many researchers have conducted their studies on solid waste management by using Geographic Information System (GIS) based Multi-criteria Decision Making Analysis (MCDA) techniques (Assegid, 2014; Gedefaw, 2015; Kassie, 2016; Regassa et al., 2011). Specifically, in the Sidama Region, some research has been conducted on current solid waste management (SWM) challenges and practices, and assessment of the extent and coverage of urban SWM service delivery in urban centres (Belachew, 2019; Kayamo, 2022; Mefekir, 2009; Solomon, 2018). Those previous studies reported that solid waste management is a challenging issue in many towns and cities in the region. However, there is no/limited research in the identification of suitable areas for landfill site selection in the region, and town and city planners didn’t consider proper landfill site selection in their urban planning process. This very poor solid waste management system can cause severe environmental and human health problems in the study area. Thus this study is typically different from previous research studies by addressing such pressing site planning problems in the region as well as in Yirgalem town.

Currently, many technologies have emerged for environmental protection and sanitary, but Geographic Information Systems (GIS) and Remote Sensing (RS) are powerful integrated techniques to select the best possible solid waste dumping in recent essential technology (Ibrahim Mohammed et al., 2018). RS can provide information about various spatial criteria, such as land use and land cover, drainage density, slope, etc. (Bedasa & Wondwossen, 2019). Whereas, GIS aided in utilizing and creating the digital geo-database as a spatial clustering process and easily understood way for the solid waste dumping site selection process. In multi-criteria evaluation many data layers are to be handled by GIS and remote sensing to arrive at the suitable site, this can be achieved conventionally using GIS. Therefore, the study was aimed at providing suitable solid waste disposal sites by using GIS and remote sensing techniques to minimize the risk of ecological and human health problems, it is also helpful to set appropriate selection criteria for the identification of new solid waste dumping sites through scientific methods (Ibrahim Mohammed et al., 2018).

Traditional site selection for solid waste dumping approach resulted in many environmental and sanitary problems. To overcome these problems many analysis tools and techniques have been developed. Studies discussed the integration of GIS and multi-criteria decision analysis and its efficiency in solving site selection problems. The landfill site selection process involves multiple criteria and conflicting objectives in the decision-making, therefore, it considers not only scientific analysis and data mining (Ekmekçioğlu et al., 2010). The multi-criteria Decision-Making method supports decision-making that involves conflicting objectives and multiple criteria (Voogd, 2009). The MCDA methods assist the decision-makers in coming out with the best solutions or alternatives in any implementation of planning or any investment in a particular area (Doumpos & Zopounidis, 2014). The MCDM can deal with a large number of criteria, which makes it to be considered an efficient method for decision support tool (Korucu et al., 2013).

There are so many suitability analysis techniques that have been used with the advancement of programming and visualization technologies the capability and efficiency of spatial analysis become very close to the real solution problem of the practical world. Fuzzy logic is a very previous suitability analysis technique that dates back to 1965 by Zadeh. Logic, criteria standardization is very accurate with the use of fuzzy membership function (Khan & Samadder, 2014). Weighted Linear Combination is assumed as the second most important suitability analysis technique. It is used for Multi-Attribute Decision Making (MADM) and it has been widely used by many researchers in different countries to achieve suitable landfill sites for waste disposal (Khan & Samadder, 2014; Korucu et al., 2013). Thirdly Analytic Hierarchy Process (AHP) is the most popular suitability analysis technique. Association with GIS and related supportive technologies AHP is preferable for many researchers of different principles. It was introduced and established by Saaty in 1988 (Saaty, 1988). Finally ordered weighted average (OWA) is a recent MCDA integration technique that is used to rank criteria (OWA). OWA can identify specific uncertainties related to criteria ranking. There is some trade-off among the criteria which are controlled by the ordered weights thus, giving rise to the best output in the planning process (Dodgson et al., 2009; El Alfy et al., 2010; Majid & Mir, 2021; Siddiqui et al., 1996). In addition to saving time and money during the landfill siting process, Multi-Criteria Decision Analysis (MCDA) combined with the GIS approach and digital spatial data format can be used for long-term site monitoring (Abushandi & Alatawi, 2015; Kabite et al., 2012; Tadese et al., 2022; Zulu & Jerrie, 2017).

In general, many models of multi-criteria analysis have been separated or integrated two more models at a time. However, for the case study, multi-criteria decision-making analyses were used. Due to the analysis of multi-lateral attributes of multivariable almost all logical reasoning and technical supported analytical tools had been used. Standards were derived from literature and urban and environmental standards including policies and directives. In the present study, GIS-based multi-criteria decision analysis (AHP) methods were used to select a suitable landfill location by minimizing conflicting interests. Environmental and socio-economic factors including slope, view shade, elevation, soil type, and soil depth, land use, and land cover, distance to the road network, resident/settlement, dumping site, and river were weighted to create a landfill site because it gives a logical strategy for selecting a scientific site for an urban solid waste dump (Asefa et al., 2021). Hence, the study about suitable landfill site selection for waste management in Yirgalem Town is done by following the already existing GIS-based MCDA technique.

2. Methodology

2.1. Study site description

The study was conducted in Yirgalem Town, Sidama Regional State, Ethiopia. It is located at 6°44′–6°46′ N latitude and 38°24′–38°26′ E longitude (Figure 1). The study area has an elevation of 1600–1960 meters. Furthermore, it is the largest settlement in Dale Woreda (Yusuf et al., 2018). It is located 47 km from Hawassa, which is the capital city of the SNNPR and Sidama Regional State, and 311 km south of Addis Ababa, the capital of the country (Tekleyohannes, 2019). The total population of Yirgalem town is 64,507, of whom 31,737 are male and 32,770 are female (CSA, 2013). Yirgalem Town has a moderate climate with a minimum and maximum annual temperature of 14 and 30 °C, respectively. The study area obtained a bimodal rainfall with peaks in April, June, and August, with an annual rainfall of 1138–1690 mm (Yusuf et al., 2018). Yirgalem town has two sub-cities, namely Filwuha and Arada, with three kebeles within each sub-city.

Figure 1. Location map study area (Yirgalem, Town Ethiopia).

2.2. Materials and methods

This study focuses on solid waste disposal site selection using GIS-based MCDA techniques. The methodological strategy for this study includes thematic map generation, assigning weight factors, weighted overlay analysis by multi-criteria decision analysis, and site prioritization. For this study both primary and secondary were used, the primary data included field surveying (collecting ground control points and key informant interviews) and field observation (recording of current solid waste dumping areas via cameras). The secondary data were collected from USGS websites, existing archives, literature, governmental policies, urban development directives, reports, and articles (Table 1).

Table 1. Thematic layers used for potentially suitable landfill site selection, their sources, and details.

Parameters	Data types	Data details	Data sources
Land use	Landsat8OLI/TIRS	USGS, 30 × 30-m	https://earthexplorer.usgs.gov
Road network	Road network shapefile	Open Street Map Data	https://www.openstreetmap.org/
Resident areas	Boundary of resident areas	Google maps	https://www.google.com/maps/
Slope	ASTER GDEM	Version 3, 30 × 30 m	https://search.earthdata.nasa.gov
View shade	ASTER GDEM	Version 3, 30 × 30 m	https://search.earthdata.nasa.gov
Dumping site	Point location of dumping site	Google maps	https://www.google.com/maps/
River	ESRI shapefile	Space Science and Geospatial Institute: (SSGI) data	http://www.ssgi.gov.et/
Elevation	ASTER GDEM	Version 3, 30 × 30 m	https://search.earthdata.nasa.gov
Soil type	ESRI shapefile	Digital Soil Map of the World, FAO	http://www.fao.org/geonetwork/srv/en/
Soil depth	ESRI shapefile	Digital Soil Map of the World, FAO	http://www.fao.org/geonetwork/srv/en/

Source. Compiled by Authors.

2.3. Thematic data

In this study, ten thematic layers namely Sentinel 2 which has 10 m spatial resolution data were used to prepare a 2022 thematic map of land use/land cover that is classified into five classes. These classes include a built-up area, an agricultural area, vegetation, an open area, and agroforestry (the mixture of built-up and home garden vegetation). Soil drainage and soil depth with 150 m resolution were acquired from www.data.isic.org. Main road data were digitized from Google Earth Pro, stream networks were also derived from watershed analyses that generate the stream order, and the stream order was extracted from 30 m resolution DEM data which was acquired from www.earthexplorer.usgs.gov. The schematic structure of the methodological framework is presented in Figure 2.

Figure 2. General procedures followed in selecting a solid waste dumping suitable site.

To create a conducive/compatible working environment, all the collected spatial data (10 m Sentinel, 30 m DEM, 150 m resolution of Soil drainage and soil depth) were converted into similar cell size, scale, and accuracy. Before preparing the output maps, all the spatial data collected from different sources and resolutions and thematic layers were converted into the same cell size, scale, datum, and projection system (WGS 84 UTM zone 37N). The general work follows the study presented in bellow the following chart. First, all required data were prepared and clipped by study area extent (Yirgalem Town). Then all Suitability class standards, suitability criteria, and value of influencing weight of each variable and weight of each class were derived from related research, UNDP standard and environmental feasibility, and urban solid waste management standard of the FDRE.

2.4. Criteria determining factors

Solid waste should be dumped at a location of a certain distance from the geographical feature, public services, and settlements. Each piece of location should be evaluated for its feasibility for dumping solid waste. The multi-criteria decision-making analysis requires well-organized and normalized input data and a standardized degree of suitability classes of variables.

The scale of suitability ranged from class 1 to class 5, with 5 being the most suitable for solid waste dumping and 1 being very low suitable (Table 2). The suitability levels for each factor will be ranked as highly suitable-S1, Moderately suitable-S2, Marginally suitable-S3, Not suitable-N2, and permanently not suitable-N1 based on the structure of literature and expertise view land suitability classification. Then finally based on the degree of influence of the variable weighting overlay analysis will be done and lastly, the map of the current suitable site solid waste disposal will be made. Table 1 depicts the criteria and degree of suitability of each class of variables.

Table 2. Suitability criteria (standard).

No	Variables	S3	S2	S1	N2	N1
1	Road (m)	901–1392	751–900	626–750	625–4001	<400
2	Settlement (m)	>700	401–500	301–400	201–300	<200
3	Land cover/use/	Open area	Vegetation	Agroforestry	Agriculture	Built-up
4	Slope (%)	0–10	11–40	41–60	61–70	>70
5	View shade	Not visible			Suitable
6	Dumping site (m)	1501–2161	1201–1500	751–1200	501–750	0–500
7	River (m)	651–850	651–850	351–650	201–350	0–200
8	Elevation	>1854	1729–1853	1696–1728	851–1695	0–850
9	Soil type	Clay	Silt	Loam	Sandy loam	Sandy
10	Soil depth (cm)	80–280	56–80	50–65	35–50	19–33

Highly suitable-S1, Moderately suitable-S2, Marginally suitable-S3, Not suitable-N2 and Permanently not suitable-N1.

Sources. Korucu et al. (2013), Majid and Mir (2021), and Mekuria et al. (2019).

2.5. Suitability model selection

Spatial Multi-Criteria Decision Analysis (MCDA) supports decision-making and site selection in the sense that it sets the framework for evaluating alternatives based on multiple evaluation criteria spatial MCDA is also known as ‘Overlay Analysis’, implying that a map overlay process is taking place and several layers are combined in order a decision map to be produced. Such a decision map may then be used as a guide, indicating the most suitable location for the development of a specific activity (Sisay et al., 2021).

2.6. AHP method as a multi-criteria decision-making tool

The analytical Hierarchical process (AHP) is a multi-criteria decision making strategy. AHP is a reliable method that examines several variables based on expert viewpoints in a GIS environment with practical knowledge to establish decision-making (Şener et al., 2011). AHP was utilized by Saaty (1980, 1990) as an effective technique for making decisions that were centred on a group of factors and created a hierarchical configuration by giving each factor a weight to reduce the complexity of the decision-making process. Since the final result in AHP depends entirely on these processes, the only way to acquire accurate results is to normalize each parameter and apply the proper weights to it (Muralitharan & Palanivel, 2015; Roy et al., 2022). The Saaty scale of importance is shown in Table 3, with 1 indicating ‘equal importance’ and 9 indicating ‘extreme importance’. In this study, the weight of different factors is assigned using the AHP technique. There are other several weight determination methods, but AHP is regarded as the most appropriate and compatible method in solid waste site suitability analysis. Based on a range of expert opinions, field experience, and numerous literature reviews, the weight of all ten criteria is therefore assigned using the AHP method (Islam & Raja, 2021; Tayyab et al., 2021; Zeng & Huang, 2017). Nonetheless, the subsequent procedures are employed to compute the Pairwise comparison matrix of diverse theme strata and evaluate their coherence (Agarwal & Garg, 2016; Saaty, 1980).

Table 3. Scale of relative importance for the pairwise comparison.

Importance	Definition	Explanation
1	Equal importance	Two variables are required equally to the intended objective
3	Moderate importance	One variable is slightly more important than the other variable being compared
5	Strong importance	One variable is strongly important over the other variable being compared
7	Very strong importance	Activity is strongly favored and its dominance is demonstrated in practice.
9	Extreme importance	The evidence favouring one activity over another is of the highest possible order of affirmation
2, 4, 6, 8	Intermediate values	Intermediate values between the two adjacent judgments when compromise is needed.
1/2, 1/3, 1/4, 1/5, 1/6, 1/7, 1/8, 1/9		Reciprocal values of the previous appreciation

Source. Adapted from Alkaradaghi et al. (2019) and Saaty (1980).

As indicated in Table 3, the process judgment pairwise comparison was done through priority importance of intensity using numeric scale ranges 1 to 9 (Alkaradaghi et al., 2019). Thus, every alternative can be evaluated in terms of the decision criteria and each criterion can be estimated by its weight to implement a paired comparison in the matrix (relative scale of importance). As indicated in Table 2, the values of (a_ij) when (i = 1, 2, 3, … m) and (j = 1, 2, 3, … n) are used to indicate the performance values in the rows and columns of the matrix. The upper diagonal triangle of the matrix is filled with the values of comparison criteria, while the lower triangle of the matrix represents the reciprocal values of the upper diagonal, using Equation (1)(1) $$a_{\mathit{\text{ji}}}=\frac{1}{a_{\mathit{\text{ji}}} }$$(1) (Saat, 2006; Şener et al., 2011; Uyan, 2014). (1) $$a_{\mathit{\text{ji}}}=\frac{1}{a_{\mathit{\text{ji}}} }$$(1)

Where (a_ij) is the element of the row (i) and column (j) of the matrix. The comparison matrix for the relative importance of the criteria can be represented as follows in Equation (2)(2) $$\left[\begin{array}{cccc}a11\hfill & a12\hfill & \dots \hfill & a1n\hfill \\ a21\hfill & a22\hfill & \dots \hfill & a2n\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ a1m\hfill & a2m\hfill & \dots \hfill & \mathit{\text{anm}}\hfill \end{array}\right]\left[\begin{array}{c}\mathrm{W}1\hfill \\ \mathrm{W}2\hfill \\ \hfill \\ \hfill \\ \text{Wn}\hfill \end{array}\right]$$(2) : (2) $$\left[\begin{array}{cccc}a11\hfill & a12\hfill & \dots \hfill & a1n\hfill \\ a21\hfill & a22\hfill & \dots \hfill & a2n\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ a1m\hfill & a2m\hfill & \dots \hfill & \mathit{\text{anm}}\hfill \end{array}\right]\left[\begin{array}{c}\mathrm{W}1\hfill \\ \mathrm{W}2\hfill \\ \hfill \\ \hfill \\ \text{Wn}\hfill \end{array}\right]$$(2)

The eigenvectors for each row are calculated using geometric principles (Equation 3(3) $$\mathrm{\text{Egi}}={(a}_{11}\mathrm{*}{a}_{12\mathrm{ }}\mathrm{*}{a}_{13}\mathrm{*}{a}_{14}\dots \mathrm{*}{a}_{1n})\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.$$(3) ), multiplying the value for each criterion in each column in the same row of the original pair-wise comparison matrix, and then applying this to each row. Using Equation (2)(2) $$\left[\begin{array}{cccc}a11\hfill & a12\hfill & \dots \hfill & a1n\hfill \\ a21\hfill & a22\hfill & \dots \hfill & a2n\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ :\hfill & :\hfill & \hfill & :\hfill \\ a1m\hfill & a2m\hfill & \dots \hfill & \mathit{\text{anm}}\hfill \end{array}\right]\left[\begin{array}{c}\mathrm{W}1\hfill \\ \mathrm{W}2\hfill \\ \hfill \\ \hfill \\ \text{Wn}\hfill \end{array}\right]$$(2) above (Saaty, 1980). (3) $$\mathrm{\text{Egi}}={(a}_{11}\mathrm{*}{a}_{12\mathrm{ }}\mathrm{*}{a}_{13}\mathrm{*}{a}_{14}\dots \mathrm{*}{a}_{1n})\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.$$(3) where, Egi = eigenvalue for the row (i); n = the number of elements in a row (i) The priority vector or AHP weight is determined by normalizing the eigenvalue to 1 (divided by their sum), as follows in Equation (4)(4) $$\mathrm{\text{Pri}}=\frac{\mathrm{\text{Egi}}}{\left({\sum }_{i=1}^n\mathrm{\text{Egi}}\right)}$$(4) (Saaty, 1980). (4) $$\mathrm{\text{Pri}}=\frac{\mathrm{\text{Egi}}}{\left({\sum }_{i=1}^n\mathrm{\text{Egi}}\right)}$$(4)

The maximum eigenvalue or maximum lambda (λ_max) is obtained from the summation of products between each component of the priority vector and the sum of columns of the reciprocal matrix, as shown in the following Equation (5)(5) $${\lambda }_{\mathrm{\text{max}}}={\sum }_{j=1}^{\mathrm{n}}⌈W_j{\sum }_{i=1}^m{a}_{\mathit{\text{ij}}}⌉$$(5) (Belton & Stewart, 2002). (5) $${\lambda }_{\mathrm{\text{max}}}={\sum }_{j=1}^{\mathrm{n}}⌈W_j{\sum }_{i=1}^m{a}_{\mathit{\text{ij}}}⌉$$(5)

Where a_ij is the criteria in each column in the matrix, and W_j is the value of weight for each criterion which corresponds to the priority vector in the decision matrix. The consistency index represents the equivalent of the mean deviation of each comparative element and the standard deviation of the evaluation error from the true ones (Saaty, 1980), which will often turn out to be larger than the value describing a fully consistent matrix to provide a measure of severity of this deviation (Belton & Stewart, 2002; Saaty, 1980). Similarly, the consistency index value (CI) has been performed as follows in Equation 6(6) $$\mathrm{\text{CI}}=\frac{({\lambda }_{\mathrm{\text{max}}}\mathrm{ }‐\mathrm{ }n)}{(n\mathrm{ }‐\mathrm{ }1)}$$(6) (Saaty, 1980). (6) $$\mathrm{\text{CI}}=\frac{({\lambda }_{\mathrm{\text{max}}}\mathrm{ }‐\mathrm{ }n)}{(n\mathrm{ }‐\mathrm{ }1)}$$(6)

Where CI is the consistency index and n is the size or order of the matrix.

Finally, the consistency ratio (CR) is obtained by dividing the consistency index value (CI) by the random index value (Equation 7(7) $$\mathit{\text{CR}}=\raisebox{1ex}{$\mathit{\text{CI}}$}\!\left/ \!\raisebox{-1ex}{$\mathit{\text{RI}}$}\right.$$(7) ), here (RI = 1.56), and n is the size of the matrix (n = 13). Table 4 shows the mean random index value RI for a matrix with different sizes (Saaty, 1980). (7) $$\mathit{\text{CR}}=\raisebox{1ex}{$\mathit{\text{CI}}$}\!\left/ \!\raisebox{-1ex}{$\mathit{\text{RI}}$}\right.$$(7)

Table 4. The mean random index value (RI) (adapted from Saaty, 1980).

n	1	2	3	4	5	6	7	8	9	10	11	12	13
RI	0	0	0.58	0.9	1.12	1.24	1.32	1.41	1.45	1.49	1.51	1.48	1.56

Based on the standard stated by Alkaradaghi et al. (2019), each importance value and weight of each parameter were compared with each parameter. This was done based on flown surveying and an expert view. As indicated in Table 5, each parameter specifies a value for the weight that it merits by assuming the method of straightforward additive weighting. These weights are then used to organize the comparison matrix to revise the correct weight for each parameter (Alkaradaghi et al., 2019).

Table 5. Pair-wise comparison matrix, factor weights, and consistency ratio of factors.

	A	B	C	D	E	F	G	H	I	J
A	1	0.2	0.3	0.02	0.2	0.1	2	2	3	3
B	2	1	3	0.5	0.2	5	6	6	7	7
C	3	0.333333	1	3	3	0.3	5	6	7	7
D	5	2	0.333333	1	4	4	0.3	5	6	6
E	5	5	0.333333	0.25	1	3	0.5	5	5	6
F	10	0.2	3.333333	0.25	0.333333	1	4	4	5	5
G	0.5	0.166667	0.2	3.333333	2	0.25	1	3	4	4
H	0.5	0.166667	0.166667	0.2	0.2	0.25	0.333333	1	4	4
I	0.333333	0.142857	0.142857	0.166667	0.2	0.2	0.25	0.25	1	2
J	0.333333	0.142857	0.142857	0.166667	0.166667	0.2	0.25	0.25	0.5	1
CI = 0.073613, RI = 1.49, CR = 0.049405

2.7. Weighted overlay analysis (WOA) and identifying suitable areas

Weighted overlay analysis is an important simulation technique designed to create a composite map by integrating the geometry and features of each theme layer in a GIS environment (Mussa et al., 2020; Roy et al., 2022). Based on the MCDA concept, it is a multi-parametric method that enables users to integrate different raster layers for the final output (Roy et al., 2022). The goal of the WOA is to produce a cumulative possibility model that, as a function of the chosen parameters and their matching groups, depicts the possible groundwater zones in a particular location (Abijith et al., 2020). Therefore, WOA is used in this work to produce the final solid waste disposal site suitability map.

Since several studies have recently been conducted employing the WOA technique in GIS platforms to determine suitable landfill sites, WOA is considered as an effective method for Identifying suitable landfill sites (Andualem & Demeke, 2019; Ibrahim-Bathis & Ahmed, 2016; Rahmati et al., 2014). The WOA can be expressed based on the following Eq. 8(8) $$\mathrm{\text{WOA}}={\sum }_{j=1}^n{W}_i\mathrm{*}{R}_i$$(8) (Sk et al., 2020). (8) $$\mathrm{\text{WOA}}={\sum }_{j=1}^n{W}_i\mathrm{*}{R}_i$$(8)

Where W_i is considered as a particular decision criterion, R_i is the raster layer of that particular criteria and n is the number of decision matrix.

To obtain the suitability index value for the potential suitability areas, the WLC (Weighted Linear Combination) method was used based on the following Equation (9)(9) $$A_i={\sum }_{j=1}^n\mathit{\text{WI}}\mathrm{*}{C}_{\mathit{\text{ij}}}$$(9) (Alkaradaghi et al., 2019). To determine the suitable areas for solid waste disposal sites, there are ten factors were considered. The factors are road, stream network, settlement, land cover, slope, view shade, dumping site, river, soil type, and soil depth. Each factor thematic map was prepared using ArcGIS spatial analyst tools, and prepared maps were converted into a raster format. The overlay analysis of the final weighted factor map was produced the final suitable solid waste disposal site map was prepared using ArcGIS software by the following principle: (9) $$A_i={\sum }_{j=1}^n\mathit{\text{WI}}\mathrm{*}{C}_{\mathit{\text{ij}}}$$(9)

Where A_i: is the suitability index, WI: is the relative importance weight of the criterion, C_ij: is the grading value (i) under criterion (j), and n: is the total number of criteria. Where: A is a road, B is a stream network, C is a settlement, D is a land cover, E is the slope, F is view shade, G is a dumping site, H is a river, I is soil type, and J is soil depth.

The WLC method was applied to all criteria using the spatial analysis tools to assess the suitability index map. This is achieved by summing up the products by multiplying the sub-criteria rating values for each criterion (based on expert opinion in this sector) by the corresponding relative importance weight. Multi-criteria spatial modelling is an advanced GIS analysis. Weighting overlay analysis summarizes the weight influence of each factor based on the given criteria. By considering the input parameters and their weight that determine their rank of influence, the weighting calculates the probability of the degree of suitability of a given location for particular land use (Majid & Mir, 2021). The result of the weighted factor map in the ArcGIS software reveals four suitability indices for the municipal solid waste dumping site of Yirgalem town.

3. Result and discussion

3.1. Thematic map preparation

3.1.1. Road network

The dumping site should be accessible to road networks that enable to dumping of solid waste. So it shouldn’t be too far from the existing road to avoid the extra cost to access it. On the other side, it shouldn’t be too close to the existing road to keep sanitary for the road user, i.e. to avoid visual impacts, bad smiles, and environmental damage. After the Euclidian distance had been calculated the distance was classified into suitability classes based on the criteria standard (Table 2 above) the distance was classified into five classes (Figure 3). Therefore, a distance below four hundred meters (400 m) on both sides of the road is permanently not suitable. The distance from 401 to 625 m is not suitable-N2 and the distance from 626 to 750 m is marginally suitable-S3 and finally, the distance from 752 to 900 m is moderately suitable for S2 and lastly the distance from 901 to 1392 m highly suitable-S1.

Figure 3. Suitability criteria based on reclassified distance from the road network.

3.1.2. Stream network

The distant area from the lake and river stream network is preferable to that of the closer one. The study area river stream was generated by carrying out watershed analysis that resulted in a stream network as shown in Figure 4. Solid waste can be carried by surface running water from the upper altitude to the lower part and moves the garbage from the dumped area to the other alongside the river way. There the dumping site should be far from the river network. Based on the criteria in Table 1 and as depicted in Figure 4, the area located at the distance of 851 m and above is very suitable-S3, 651–850 m is moderately suitable, 351–650 m is marginally suitable, 201–350 m which closer to the stream is not suitable and finally less or equal to 200 m is the closest are is permanently not suitable. Mean that at a distance of (<200 m) from the stream network is risky to dump the solid waste. In Yirgalem town the stream networks were surrounded by a very sloppy area which increases the probability of solid waste running into the waterway and polluting water leading to environmental and health problems.

Figure 4. Suitability criteria based on reclassified distance from stream network.

3.1.3. Residential

Solid waste dumping sites should be far away from residential areas in order not to affect land values for future developments and expansion of the town and to protect the residential environment from pollution released from the settlement, commercial area, and other anthropogenic activities. Proximity from the residential area is categorized into five different degrees of suitability classes. According to the suitability standard table (Table 1), the study considered the reclassified distances shown in Figure 5 as permanently not suitable (<200 m), not suitable between 201 and 300 m, marginally suitable from 301 to 400 m, modestly suitable from 401 to 500 m and the proximity from 700 m to 2 km is very suitable for solid dumping.

Figure 5. Suitability criteria based on reclassified distance from the built up of the residential area.

3.1.4. Land use and land cover

The land use and land cover of the study area were analyzed using sentinel two with a ten-meter spatial resolution. The image was passed through all image analysis processes. This includes preprocessing and image processing. After the image was classified the result of the classification was evaluated in post-processing analysis. The land cover and land use of the town were some of the factors that were considered to select the best site for solid waste dumping. In the study area, agroforestry accounts for that area of (274.5 ha) located close to home that contains a combination of coffee (Arabica coffee) and banana (Musa acuminata) (Table 6). Vegetation class (465.48 ha) contains natural and plantation forest little bit combination of shrub and other fruit trees and class agriculture contains urban farming garden and outskirted farming land and it covers an area of 1663.92 ha. As indicated in Table 6, the open area contains bare soil, gourdes, and grassland of the study area, which covers 879.12 ha. Built-up 636.57 ha of the study area indicates that, all settlement areas and other urban utilities of the town like road churches, schools, banks, and related public services areas.

Table 6. Land use land cover area.

Land use class	Percentage	Area (m^2)	Area (ha)
Built-up	16	6,365,700	636.57
Agriculture	43	16,639,200	1663.92
Vegetation	12	4,654,800	465.48
Open area	22	8,791,200	879.12
Agroforestry	7	2,745,000	274.5
Total	100	39,195,900	3919.59

By reviewing different literature, journals, and other urban solid waste management guides, and documentation summary suitability criteria table (Table 2) was drawn. Several research studies showed that open area, agriculture, vegetation, agro-forester, and settlement areas were categorized into classes of, very suitable-S3, moderately suitable-S2, marginally suitable-S1, not suitable-N2, and permanently not suitable-N1, respectively (Choudhury & Das, 2012; Jamjan, 2009; Yahaya et al., 2010; Zain, 2009). These suitability classes of land cover and land use of the study area were analyzed after making field observations and the ground references were collected by using field surveying that was used for accuracy assessment (Figure 6).

Figure 6. Suitability criteria based on reclassified distance from the built-up of the residential area and land cover land use maps.

3.1.5. Accuracy assessment of the classifications

Based on the reference samples collected and the classification maps, the accuracy report of the analysis (Table 7 below) showed that the total accuracy of the maps in the study resulted from approximately 93.47% and kappa from 0.92. However, the total accuracy does not tell us the errors in each category in the classification maps. The amount of error made in each category was reported using the producer’s and user’s accuracy. The producer’s accuracy can be seen according to the interest of the map producer in how a certain area is classified or mapped, while the user’s accuracy can be seen in the point of view of the map user to indicate the probability that a pixel classified on the map/image represents that category on the ground (Congalton, 2001).

Table 7. Accuracy assessment (error matrix) reports of the classification maps for the study years.

	References
	Classes	Built-up	Agriculture	Open Area	Vegetation	Agroforestry	Total	UA%
Classification	Built-up	43	0	0	0	0	43	100
	Agriculture	1	39	1	0	1	42	92.86
	Open Area	1	1	13	0	0	15	86.67
	Vegetation	1	1	0	7	1	10	70
	Agroforestry	1	3	0	1	26	31	83.87
	Total	47	44	14	8	28	141
	PA%	91.49	88.64	92.86	87.5	92.86
	OAA%	93.47
	Kappa	0.92

3.1.6. Slope

The slope is one of the topographic factors that determine the potential suitability of a given area for a solid waste dumping site. It was derived from the ASTER Digital Elevation Model (DEM) 30*30 meter spatial resolution. In surface analysis, DEM was used as input, and slope was determined in percent. Accordingly, the slope in the percent map was prepared and reclassified into five slope suitability classes (Figure 7). The too steep slope would make it difficult to dig and construct solid waste well. So there the flat area is more suitable than that of the sharp slope. Accordingly, the degree of slope suitability was classified based on the suitability standard table (Table 2). The class is 0–10, 11–40, 40–60, 61–70, and >70% slope areas very suitable (S3), modernly suitable (S2), marginally suitable (S1), not suitable, and permanently not suitable, respectively. Drainage A combination of field data, topographic maps, aerial photographs, and satellite images are used for the analysis of the drainage network and all algorithms use digital elevation models (DEMs) as basic data (Chorowicz et al., 1992).

Figure 7. Suitability criteria based on reclassified slope and elevation.

3.1.7. Soil type and soil depth

According to Majid and Mir (2021), clay soil is highly suitable for solid waste dumping sites this is due to its low permeability and small texture. Silt soil is ideal second to clay soil. Loam, sandy loam soil is not suitable due to its large texture and high permeability, and sandy soil which permanently not suitable for a dumping site. The dumping site should be located at 80–280 cm soil depth. Accordingly, 20% of the study area is suitable by the criteria of soil depth (Figure 8). In contrast to this as shown in Figures 8(a,b) about 13% of the study is permanently unsuitable.

Figure 8. Suitability criteria based on reclassified soil texture and Soil depth maps.

3.1.8. Existing solid waste dumping site and viewshed

As depicted in Figure 9(a) some of the existing dumping sites of the study area were close one each other, particularly in the central and eastern parts of the town, two neighbour dumping sites were located within less than a 500 m radius which is categorized under permanently unsuitable site for solid waste dumping. This indicates only 14% of the dumping sites only placed in an appropriate location, there rest are not appropriate based on the criteria. According to Figure 9(b), about 76% of the study area is visible from the building up of the residential area. This means that only 24% of the area is suitable for a dumping site. The newly selected dumping site should be far from the existing one.

Figure 9. Suitability criteria based on reclassified distance in meters from the existing dumping site and view shade from built-up residential area maps.

3.1.9. Solid waste disposal site suitability analysis

The area coverage of each suitability index of the sites was calculated in the ArcGIS environment and showed that 549.085 ha (14.01%) and 589.78 ha (15.05%) of the study area are permanently unsuitable and unsuitable for solid waste disposal sites, respectively as the areas are environmentally unfavourable and economically (Table 8). This unsuitable (restricted) area includes close to surface water (river, streams) (area with a 350 m buffer zone), main roads (area with a <400 m buffer zone), areas with a steep slope (>70%), areas with close to road networks and far from road networks with a 350 m buffer zone. The main advantage of these areas' restrictions was to minimize their negative effects on the environment and public health as well as to minimize the cost of construction and maintenance of the solid waste disposal site. On the other hand, 70.94% of the study areas were suitable at different levels which means 299.21 ha (7.63%) of the study area is highly suitable, and 27.84% of the area is categorized in moderately suitable areas. The rest 1390.23 ha (35.47%) of the area were also categorized as marginally suitable areas (Figure 10).

Figure 10. Solid waste disposal site suitability Map of Yirgalem Town. NB: S3-highly suitable, S2-moderately suitable, S1-marginally suitable, N2-not suitable, and N1-permanently not suitable.

Table 8. Statistical analysis of the solid waste disposal site suitability.

No	Suitability	Area (km^2)	Area (ha)	Area (%)
1	Highly suitable	3,992,100	299.21	7.63
2	Moderately suitable	10,912,850	1091.285	27.84
3	Marginally suitable	13,902,300	1390.23	35.47
4	Not suitable	5,897,800	589.78	15.05
5	Permanently not suitable	4,490,850	549.085	14.01
	Total	39,195,900	3919.59	100

Therefore, with the population growth and high demand for agricultural land, the town planners should systematically and scientifically site and properly manage the proposed sites. It is also suggested that before landfilling, the solid wastes need to be processed by the window compost or aerobic composting and refuse-derived fuel methods.

3.1.10. Validation of suitable landfill sites model

Validation is one of the most important processes in assuring the validation of any model after it has been prepared is accuracy evaluation. In this study, the area under the curve (AUC) method was used to validate the landfill site selection suitability model. This validation method is the most trustworthy approach for verifying MCDA model performance has a simple design, inclusive, and equal forecasting character as compared to other methods (Darabi et al., 2021; Rahmati et al., 2020).

Land use, elevation, river, settlement, and dumping sites were used to validate the accuracy of the model through geospatial techniques and the AHP approach agreed with high-resolution Google Earth images. Besides, municipal reports and a review of the literature were also considered to properly validate the suitability model. Figure 11 shows the communicative percentage of area under landfill site suitability. The final AUC for the present study is 0.860 or 86.0%, which can be considered a good accuracy of the model (Roy et al., 2022). Similarly, high-resolution Google Earth images have also shown that the identified suitable areas were found in the open/bare land, which means the model has an excellent ability to predict the suitable areas (Figure 12).

Figure 11. Validation of model using the area under curve (AUC).

Figure 12. Validation of landfill suitable site with Google Earth.

4. Conclusion and recommendation

• The present study provides an optimal solution to characterize landfill-suitable sites in Yirgalem Town using GIS-based MCDA analysis. GIS and multi-criteria analysis can produce suitable sites for solid waste disposal which is a vital component in solid waste management.

• The use of GIS and MCDA considered environmental factors (protected areas, rivers, soil, slope, viewshed, streams networks, and soil depth), economic factors (roads), and social factors (land use land cover, dumping site, settlement, and urban centers) in arriving at the most suitable landfill sites.

• The most suitable landfill sites are located in the western and north-western parts of the study area.

• About 299.21 ha (7.63%) of the study area is highly suitable, 1091.285 ha (27.84%) moderately suitable areas, and 1390.23 ha (35.47%) of the area were also categorized as marginally suitable areas.

• While 589.78 ha (15.05%) of the study areas are not suitable, and 549.085 ha (14.01%) of the study area is categorized as permanently not suitable.

• This study is significantly important because it adds to our existing understanding of how to select appropriate landfill sites in Yirgalem town that are both scientifically credible and socially acceptable.

• This method of landfill selection could increase urban environmental and socioeconomic sustainability.

• The present study considered more factors to identify a suitable solid waste disposal site. However, due to the limitation of data source, groundwater table, drainage pipe networks, wind direction, and geology are not included as criteria for determination.

Acknowledgements

The author is grateful to Hawassa University, Wondo Genet College of Forestry and Natural Resources for the financial support and all the logistics. The author sincerely thanks Yirgalem Town administration for permitting us to carry out this study and their support during data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request.

Additional information

Funding

The author received funds from Hawassa University, Wondo Genet College of Forestry and Natural Resources.

Notes on contributors

Mikias Biazen Molla

Mikias Biazen Molla (PhD) is an Associate Professor (Senior Lecturer and Researcher) in GIS and Remote Sensing at Hawassa University, WGCRNRs, Ethiopia. He has been lecturing in different courses related to geospatial technology and environmental and/or natural resources-related studies. He conducted various research and community service activities and also advised so many MSc and PhD student research projects in his field of study. His current research interests include GIS, Remote Sensing, Environmental Planning and Management, urban green infrastructure and the environmental performance of Cities, Ecosystem services, Climate change, and geospatial technology applications in various fields of study.