Flood Susceptibility Mapping by Integrating GIS‐Based Approach and Analytical Hierarchy Process (AHP): Wolaita Zone, Ethiopia

Abstract

ABSTRACT

Flooding is among the most devastating natural catastrophes affecting human life and property. Climate change and environmental degradation have exacerbated flooding disasters. Developing countries experience greater damage from flooding due to their low resilience, limited financial resources, weak early warning systems, and technological limitations. Accurate data, prediction, delineation of vulnerable areas, and formulation of local action plans can help minimize the extent of economic losses and fatalities due to flooding. In the current study area, limited data are available to lessen flood potential risks. Remote sensing and GIS approaches were adopted for mapping potential flood‐susceptible areas. Topographical, hydrological, and spectral indices conditioning factors were integrated, and a weighted overlay analysis was performed in ArcGIS. The results revealed that about 8.73%, 77.16%, and 14.08% of the study region are categorized as susceptible, moderately susceptible, and less susceptible to flooding, respectively. The findings would help government authorities and relevant bodies in developing early warning systems, advancing technology, creating local action plans, and formulating flood hazard mitigation and adaptation strategies.

Introduction

Floods can be described as excessive flows that surpass the capacity of river channels, reservoirs, lakes, dams, ponds, and other water bodies, causing water to overflow and inundate surrounding areas (Legese and Gumi 2020). Flooding is among the natural disasters and potentially devastating incidents that affect human life and property (Ullah et al. 2022). Globally, the frequency of flood hazards has considerably increased over the past three decades (Sajinkumar et al. 2022; Vishnu et al. 2019). Extreme rainfall, land use and land cover change, snowmelt, unplanned urbanization, dam failures, and rising groundwater are significant driving forces of flooding (Legese and Gumi 2020). Climate change and environmental degradation have further exacerbated flooding disasters (Mangini et al. 2018; Mostofi Zadeh et al. 2020). Globally, nearly 170 million people are affected by flooding yearly (Majeed et al. 2023). As the UN (2015) reported, floods alone affected 2.3 billion people globally between 1995 and 2015, causing nearly USD 386 billion in economic losses (Hussain et al. 2021). In 2021, the Global Natural Disaster Assessment Report indicated that floods alone caused the most significant number of fatalities, accounting for 4393 people and representing 41.87% of the total. Developing countries experience greater damage from flooding due to their low resilience, limited financial resources, weak early warning systems, and technological limitations (Tien Bui et al. 2020). In India, approximately 40 million hectares are identified as flood-prone areas, with nearly 8 million hectares affected by floods each year (Ray et al. 2019). Floods account for more than 90% of natural catastrophe damage in Niger (Adelekan and Asiyanbi 2016). Ethiopia is among the regions in Sub-Saharan Africa that are highly vulnerable to flooding, owing to its diverse topography, high rainfall variability, and increasing exposure to climate change impacts (Nigusse and Adhanom 2019). In Ethiopia, more than 80% of the annual precipitation falls during the primary rainy season (June–September), which is the primary cause of flooding (Nigusse and Adhanom 2019; Ogato et al. 2020; Getahun and Gebre 2015). In recent years, flash floods have significantly increased across the country (Billi et al. 2013; Desalegn and Mulu 2021). According to the OCHA (2024) report, heavy rain in April and early May 2024 caused flooding in several districts across the country, resulting in 590,000 people affected and 95,000 people displaced. Furthermore, it resulted in the death of approximately 2900 livestock and damaged 60,000 ha of crops.

Flood susceptibility mapping is a prerequisite for effective flood control and risk management, as it integrates numerous parameters (Mahato et al. 2021). Researchers have used GIS-based approaches and remote sensing to map flood susceptibility. They used various techniques like the analytical hierarchy process (AHP) (Ahmed et al. 2024; Almouctar et al. 2024; Dutta and Deka 2024), artificial neural networks (ANN) (Rudra and Sarkar 2023); frequency ratio (FR) (Addis 2023; Rahmati et al. 2016), machine learning (ML) (Diaconu et al. 2024; Maharjan et al. 2024; Zhu et al. 2024), weights of evidence (WoE) (Tehrany et al. 2014); random forest (RF) (Wahba et al. 2024; Wang et al. 2015); technique for order of preference by similarity to ideal solution (TOPSIS) (Mitra et al. 2023), deep learning (DL) (Li and Hong 2023; Ramayanti et al. 2022). All approaches have been successfully applied worldwide, yielding highly satisfactory results (Mahato et al. 2021). Nevertheless, it is impossible to determine the most effective model for flood susceptibility mapping (Tariq et al. 2023). Compared with purely statistical or machine learning methods, AHP is particularly advantageous in data-scarce environments, as it allows the integration of expert knowledge with available geospatial data while maintaining methodological transparency and logical consistency. It has been widely employed due to its robust practical applications (Ahmed et al. 2024; Msabi and Makonyo 2021; Souissi et al. 2020; Vojtek and Vojteková 2019).

Even though catastrophic floods cannot be entirely prevented, taking anticipatory measures and implementing effective management strategies can help reduce the damage (Das and Gupta 2021). Accurate data, public awareness and readiness, prediction, delineation of vulnerable areas, and formulation of local action plans can help minimize the extent of economic losses and fatalities due to flooding. In the Wolaita Zone, limited data and studies are available on flood-vulnerable areas, which hinders efforts to mitigate potential risks to human lives, infrastructure damage, and economic losses. Hence, this study focuses on mapping potential flood-susceptible regions using Multi-Criteria Decision Making (MCDM) analysis, remote sensing, and a GIS-based approach. It advances AHP-based flood susceptibility mapping by integrating a broader range of conditioning factors and refining their application to the Wolaita Zone context. Unlike many previous AHP–GIS studies that focus mainly on basic topographic and hydrological parameters, this work incorporates additional remote sensing–derived indices (SAVI, BSI, and MNDWI) to capture vegetation vigor, soil exposure, and surface water dynamics—critical drivers in landscapes undergoing rapid land-use change. The inclusion of terrain metrics, such as TWI, TRI, and SPI, further enhances the representation of hydrological and geomorphic processes beyond conventional approaches, thereby strengthening both contextual accuracy and methodological rigor. The results will help government agencies develop early warning systems, advance technological solutions, create local action plans, inform settlement planning, and formulate flood hazard mitigation and adaptation strategies.

Methodology

Study Area

Wolaita Zone, which covers an area of about 4509 km², with an elevation ranging from 649 to 2961 m.a.s.l., is located about 300 km from Addis Ababa (the country's capital). Geographically, it is situated between 37°13′23″ to 38°8′10″ E longitude and 6°30′23″ to 7°11′29″ N latitude (Figure 1). The zone experiences a bimodal rainfall pattern from March to October, with the mean annual rainfall extending from 1800 to 2218 mm/year. It experiences the mean maximum and minimum temperatures of about 29°C and 10.6°C, respectively. Major soil textures, including loam, clay loam, sandy clay loam, clay, and sandy clay, characterize the zone. According to the FAO Digital Soil Map of the World (1997), the following soil types are present: Eutric Nitosols, Haplic Xerosols, Plinthic Ferralsols, Eutric Cambisols, and Ochric Andosols. Based on the USGS World Geologic Maps (2000), the study zone comprises three different types of rocks: tertiary extrusive and intrusive rocks (Ti), quaternary extrusive and intrusive rocks (Qv), and quaternary (undivided) rocks (Q).

[IMAGE OMITTED. SEE PDF]

Data Acquisition

The input datasets were obtained from various sources for mapping flood-susceptible areas (Table 1). The Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) and Landsat 8–9 TIRS/OLI were accessed from USGS Earth Explorer. The rainfall data was retrieved from a satellite precipitation product from the Center for Hydrometeorology and Remote Sensing (CHRS) data portal, which offers real-time high-resolution (0.04° × 0.04°) data. The soil characteristics data were obtained from the FAO Digital Soil Map of the World (DSMW) and the International Soil Reference and Information Center (ISRIC) data hub. The Land Use Land Cover (LULC) dataset was retrieved from the European Space Agency (ESA) Worldcover_10m_2021_V200. The overall diagrammatic representation of the methodology flowchart is illustrated in Figure 2.

TABLE 1 Flood conditioning factors and data sources used for flood susceptibility mapping in the Wolaita Zone.

Flood conditioning factors	Data source	Source location
Elevation (m)	USGS SRTM 1 Arc-Second Global 30 m-v3
Slope (o)
Rainfall (mm)	CHRS 2012–2022
Distance from river networks (m)	USGS SRTM 1 Arc-Second Global 30 m-v3
Soil Adjusted Vegetation Index (SAVI)	USGS Earth Explorer Landsat 8–9 TIRS/OLI
Bare Soil Index (BSI)
Modified Normalized Difference Water Index (MNDWI)
Drainage density (km/km2)	USGS SRTM 1 Arc-Second Global 30 m-v3
Topographic Wetness Index (TWI)
Topographic Ruggedness Index (TRI)
Stream Power Index (SPI)
Soil texture	ISRIC Soil Data Hub—USDA
Soil type	FAO DSMW
LULC	ESA Worldcover_10m_2021_V200
Study area shapefile	HDX-OCHA

[IMAGE OMITTED. SEE PDF]

Flood Hazard Conditioning Factors Analysis

Topographical Conditioning Factors Thematic Maps Generation

The SRTM DEM raster datasets were imported into ArcMap, and a mosaic dataset was created. Elevation (m) and slope (%) were extracted from SRTM DEM through surface analysis. The Topographic Wetness Index (TWI) was derived from the SRTM DEM through hydrological analysis. It was calculated using Equation (1) (Moore et al. 1991).1$$$ \mathrm{TWI}=\mathit{\ln}\left(\frac{a}{\tan \beta}\right) $$$where $$$ \beta $$$ is the local slope angle in radians, $$$ a $$$ is the upslope contributing area per unit contour length in m²/m.

Hydrological Conditioning Factors Map Production

Drainage density was derived from the SRTM DEM through hydrological analysis (Equation 2). The mean annual rainfall from 2013 to 2022 was analyzed using IDW interpolation, considering the Z-value (grid code). The Stream Power Index (SPI) was analyzed using the SRTM DEM, incorporating hydrological and surface analysis. It was calculated using Equation (3).2$$$ \mathrm{Dd}=\frac{\sum \limits_{i=1}^n\left({S}_i\right)\ }{A} $$$3$$$ \mathrm{SPI}={a}^{\ast}\tan \left(\beta \right) $$$where S_i is the sum of the length of streams in km, and A is the area in km², Dd is the drainage density, SPI Stream Power Index, $$$ \beta $$$ is the local slope angle in radians, $$$ a $$$ is the upslope contributing area per unit contour length in m²/m.

Spectral Indices and LULC Map Preparation

Spectral indices were derived from Landsat 8–9 TIRS/OLI imagery. Equation (4) was used to calculate the Soil Adjusted Vegetation Index (SAVI) (Huete 1988). The green and shortwave infrared (SWIR) bands were used to analyze the MNDWI (Equation 5). Likewise, the Bare Soil Index (BSI) was calculated using near-infrared (NIR), red, blue, and shortwave infrared (SWIR) bands (Equation 6). The European Space Agency (ESA) Worldcover LULC legend and symbology were used and resampled to 30 m resolution employing the nearest resampling techniques.4$$$ \mathrm{SAVI}={\frac{\mathrm{NIR}-\mathrm{Red}}{\mathrm{NIR}+\mathrm{Red}+L}}^{\ast }1+L $$$5$$$ M\mathrm{NDWI}=\frac{\mathrm{Green}-\mathrm{SWIR}\ }{\mathrm{Green}+\mathrm{SWIR}\ } $$$6$$$ \mathrm{BSI}=\frac{\left(\mathrm{SWIR}+\mathrm{Red}\right)-\left(\mathrm{NIR}+\mathrm{Blue}\right)}{\left(\mathrm{SWIR}+\mathrm{Red}\right)+\left(\mathrm{NIR}+\mathrm{Blue}\right)} $$$where L is the correction factor, 0 for areas with high green vegetation cover, which is equal to NDVI; 0.5 for areas with moderate green vegetation cover; and 1 for areas with no green vegetation cover.

Soil Type and Texture Maps

The Ethiopian Digital Soil Map was exported from the DSMW and projected onto the WGS 1984 UTM Zone 37N. The study area's dominant soil type was clipped and converted to a raster data format. Soil texture was extracted from the African Soil Grid texture (defined by the USDA) and projected onto the WGS 1984 UTM Zone 37N. Then, the study area's soil texture was clipped, projected, and resampled to 30 m resolution.

AHP Model

Multi-criteria decision-making (MCDM) analysis is a widely adopted technique that enables users to determine the best option among multiple criteria. Thomas L. Saaty coined the concept of the AHP in 1980, and it has remained a widely used MCDM approach (Saaty 1980). Combining GIS and the AHP model improves decision-making by enabling the management and organization of vast amounts of geographical data (Majid and Mir 2021). After preparing the thematic layers, the ranks were assigned to the respective conditioning factors, taking into account expert opinion surveys, field survey experiences, stakeholder consultations, and existing literature. A purposive selection of experts in hydrology, disaster risk management, geomorphology, GIS, remote sensing, and agricultural officers was consulted using a structured questionnaire, to rate the relative importance of the conditioning factors in influencing flood susceptibility on a 1–9 Saaty scale. The inputs from experts and stakeholders were combined using the AHP method, where pairwise comparison matrices were constructed based on the responses, and average scores were calculated. Weight assignment, pair-wise (Table 2) and normalized pair-wise comparison matrices (Table 5), consistency index (Table 6; Equation 7), and principal eigenvector value were calculated based on (Saaty and Vargas 2012; Goepel 2018) preference scales. According to Saaty (2006), the random index for the fourteen contributing factors is 1.57. The consistency ratio (Equation 8) evaluates how reliable the judgments are compared to large samples of data. For the matrix to be consistent, it should be equal to or less than 0.1 (10%) (Majid and Mir 2021; Saaty 2006).7$$$ \mathrm{CI}=\frac{\lambda_{\mathrm{max}}-n}{n-1} $$$8$$$ \mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}} $$$where CI denotes consistency index, λ_max is the principal eigenvalue, n represents the number of constraints, CR denotes consistency ratio, and RI stands for random index.

TABLE 2 Pairwise comparison matrix.

Flood conditioning factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	Normalized weights	Criteria weights (%)
Rainfall (C1)	1														0.1547	15
River networks (C2)	1/2	1													0.1432	14
Elevation (C3)	1/2	1	1												0.1289	13
BSI (C4)	1/2	1/3	1/2	1											0.1097	11
SPI (C5)	1/2	1/2	1/3	1/2	1										0.0930	9
Slope (C6)	1/3	1/3	1/2	1/2	1/2	1									0.0819	8
Drainage density (C7)	1/3	1/3	1/3	1/3	1/2	1/2	1								0.0684	7
SAVI (C8)	1/3	1/3	1/2	1/2	1/3	1/2	1	1							0.0573	6
TWI (C9)	1/3	1/4	1/3	1/3	1/3	1/3	1/2	1/2	1						0.0474	5
MNDWI (C10)	1/3	1/4	1/3	1/4	1/4	1/3	1/3	1/2	1/2	1					0.0379	4
TRI (C11)	1/4	1/4	1/4	1/5	1/4	1/4	1/4	1/2	1/3	1/2	1				0.0260	3
LULC (C12)	1/4	1/5	1/5	1/5	1/4	1/4	1/5	1/3	1/3	1/3	1	1			0.0215	2
Soil type (C13)	1/5	1/6	1/6	1/6	1/4	1/6	1/6	/4	1/4	1/3	1/2	1	1		0.0169	2
Soil texture (C14)	1/7	1/7	1/7	1/7	1/5	1/7	1/7	1/6	1/5	1/4	1/3	1/2	1	1	0.0134	1
Sum	5.51	7.09	7.59	11.13	12.87	15.48	19.59	19.25	25.62	30.42	38.83	44.50	55.00	69.00	1	100

Potential Flood Susceptible Areas Mapping

After the thematic layers of the conditioning factors were prepared using GIS 10.8.2, weights and ranks were calculated using the AHP-MCDMA method. All flood predictor layers were transformed into raster format with uniform spatial resolution (30 m × 30 m) to ensure compatibility. Based on the AHP-derived weights, each raster layer was reclassified into suitability classes and assigned corresponding numeric scores on a standardized scale. The AHP weight of each factor was then applied as a multiplier to its respective derived weights (Equation 9; Table 3), ensuring that factors with higher relative importance were stored in a geodatabase. Weighted overlay analysis was applied to map flood-vulnerable areas by integrating all layers in the ArcGIS workspace. The final output was categorized into five classes: highly susceptible, susceptible, moderately susceptible, less susceptible, and not susceptible.9$$$ \mathrm{FSI}=\sum \limits_{j=1}^n\left({W_j}^{\ast }{R}_i\right) $$$where W_j denotes the normalized weight of the individual conditioning factor j, R_i represents the rank of the i class of the conditioning factor, and n denotes the number of conditioning factors.

TABLE 3 Flood conditioning factors, data sources, classes, class ranges, ratings, and criteria weights for flood susceptibility mapping in Wolaita Zone, Ethiopia.

Category	Flood conditioning factors	Data sources	Classes	Area	Susceptibility class ranges	Susceptibility class ratings	Criteria weight (%)
km2	%
Topographical factors	Elevation (m)	USGS SRTM 1 Arc-Second Global 30 m-v3	649–1123	275.96	6.12	Highly susceptible	1	13
1124–1474	1330.31	29.51	Susceptible	2
1475–1755	1081.75	23.99	Moderately susceptible	3
1756–2056	1517.95	33.67	Less susceptible	4
2057–2961	302.34	6.71	Not susceptible	5
Slope (o)	USGS SRTM 1 Arc-Second Global 30 m-v3	0–5	2135.73	47.48	Highly susceptible	1	8
6–11	1227.22	27.28	Susceptible	2
12–18	600.90	13.36	Moderately susceptible	3
19–28	368.24	8.19	Less susceptible	4
29–73	166.28	3.70	Not susceptible	5
Topographic Wetness Index (TWI)	USGS SRTM 1 Arc-Second Global 30 m-v3	−8.91–4.31	576.12	13.15	Not susceptible	5	5
4.32–5.98	1854.41	42.33	Less susceptible	4
5.99–8.03	1301.79	29.73	Moderately susceptible	3
8.04–11.4	496.52	11.33	Susceptible	2
11.5–23.8	151.73	3.46	Highly susceptible	1
Topographic Ruggedness Index (TRI)	USGS SRTM 1 Arc-Second Global 30 m-v3	0.11–0.32	112.04	2.49	Not susceptible	5	3
0.33–0.43	723.94	16.09	Less susceptible	4
0.44–0.51	1790.06	39.79	Moderately susceptible	3
0.52–0.61	1512.10	33.61	Susceptible	2
0.62–0.89	360.82	8.02	Highly susceptible	1
Precipitation	Mean annual Rainfall (2013–2022) (mm)	CHRS-UCI data portal	1102–1307	654.86	14.52	Less susceptible	5	15
1308–1504	1160.92	25.75	Less susceptible	4
1505–1723	1001.05	22.20	Moderately susceptible	3
1724–1933	1208.94	26.81	Susceptible	2
1934–2218	483.34	10.72	Highly susceptible	1
Surface Water	Distance from river networks (m)	USGS SRTM DEM	< 200	860.08	19.07	Highly susceptible	1	14
201–500	1142.48	25.34	Susceptible	2
501–900	1135.72	25.19	Moderately susceptible	3
901–2000	1261.12	27.97	Less susceptible	4
> 2001	109.68	2.43	Not susceptible	5
Drainage network	Drainage density (km/km2)	USGS SRTM 1 Arc-Second Global 30 m-v3	0–0.41	1314.62	29.15	Not susceptible	5	7
0.42–0.82	1308.80	29.03	Less susceptible	4
0.83–1.27	967.97	21.47	Moderately susceptible	3
1.28–1.79	651.81	14.46	Susceptible	2
1.8–2.99	265.93	5.90	Highly susceptible	1
Soil	Soil texture (30 cm depth)	ISRIC Soil Data Hub–USDA	Clay Loam	1157.2686	25.76	Susceptible	2	1
Loam	0.4698	0.01	Susceptible	2
Sandy Clay	0.2232	0.00	Moderately susceptible	3
Clay	3308.0319	73.63	Moderately susceptible	3
Sandy Clay Loam	27.0054	0.60	Less susceptible	4
Soil type	FAO DSMW	Ne—Eutric Nitosols	907.6176	20.13	Moderately susceptible	3	2
Xh—Haplic Xerosols	1191.7674	26.43	Less susceptible	4
Fp—Plinthic Ferralsols	45.4410	1.01	Highly susceptible	1
Be—Eutric Cambisols	1432.2438	31.76	Moderately susceptible	3
To—Ochric Andosols	904.2453	20.05	Less susceptible	4
WR—Water	27.7983	0.62	Susceptible	2
Spectral indices	Soil Adjusted Vegetation Index (SAVI)	USGS Earth Explorer Landsat 8–9 TIRS/OLI	−0.25 to −0.1	21.9609	0.49	Highly susceptible	1	6
−0.09–0.09	67.707	1.5	Susceptible	2
0.1–0.23	1651.664	36.63	Moderately susceptible	3
0.24–0.32	1771.813	39.29	Less susceptible	4
0.33–0.83	995.9121	22.09	Not susceptible	5
MNDWI	USGS Earth Explorer Landsat 8–9 TIRS/OLI	−0.68 to −0.23	1416.38	31.41	Not susceptible	5	4
−0.22 to −0.19	2008.06	44.53	Less susceptible	4
−0.18 to −0.066	994.74	22.06	Moderately susceptible	3
−0.065–0.098	67.34	1.50	Susceptible	2
0.099–0.18	22.54	0.50	Highly susceptible	1
SPI	USGS SRTM 1 Arc-Second Global 30 m-v3	−13.82 to −9.6	740.21	16.45	Not susceptible	5	9
−9.59 to −5.48	566.65	12.60	Less susceptible	4
−5.47 to −0.86	1452.52	32.29	Moderately susceptible	3
−0.85–2.26	1470.05	32.68	Susceptible	2
2.25–11.8	268.95	5.98	Highly susceptible	1
Bare Soil Index (BSI)	USGS Earth Explorer Landsat 8–9 TIRS/OLI	0.15–0.34	371.67	8.24	Not susceptible	5	11
0.35–0.38	1115.42	24.74	Less susceptible	4
0.39–0.41	1402.65	31.11	Moderately susceptible	3
0.42–0.46	1520.00	33.71	Susceptible	2
0.47–0.62	99.31	2.20	Highly susceptible	1
Land Use Land Cover	LULC	ESA Worldcover_10m_2021_V200	Herbaceous wetland	15.62	0.35	Highly susceptible	1	2
Permanent waterbodies	93.87	2.08	Highly susceptible	1
Bare/sparse vegetation	6.77	0.15	Susceptible	2
Built-up	34.14	0.76	Highly susceptible	1
Cropland	1681.13	37.28	Less susceptible	4
Grassland	727.24	16.13	Moderately susceptible	3
Shrubland	1387.99	30.78	Less susceptible	4
Tree cover	562.36	12.47	Not susceptible	5

Results and Discussion

Topographical Flood Predictors

The topographical flood predictors are depicted in Figure 3. Elevation is a significant parameter included in mapping flood inundation, considering that flooding and waterlogging are more likely in low elevations due to the downhill flow from higher topography (Merga et al. 2023; Siam et al. 2021). The lowest elevation, ranging from 649 to 1123 m a.s.l., is more prone to flooding due to its reduced capacity for natural drainage and its tendency to act as a convergence zone for surface water. These zones often coincide with river floodplains, valley bottoms, and depressions where accumulated runoff exceeds infiltration capacity, leading to recurrent inundation, as seen in the western portion of the study region (Figure 3a). This finding is consistent with studies in other Ethiopian basins, like the Dega Damot district (Negese et al. 2022) and the Gidabo watershed (Diriba et al. 2024), where low-lying topographies were repeatedly identified as the most vulnerable areas to flooding. About 29.51% of the study area's landscape is vulnerable to flooding (Table 3).

[IMAGE OMITTED. SEE PDF]

The slope gradient is a vital predictor in delineating flood susceptibility. Steep slopes are more likely to overflow, but flat locations are highly vulnerable to flooding (Kader et al. 2024). More than half, covering an area of 3362.95 km² (74.76%) of the region, falls under a flat slope, ranging from 0° to 11° (Figure 3b). Areas with a flat slope are highly prone to flooding due to reduced velocity of surface runoff, allowing greater opportunities for water stagnation, ponding, and infiltration saturation (Hagos et al. 2022). The Topographic Wetness Index (TWI) is a key factor that spatially expresses the variation in moisture across the landscape. It quantifies the accumulation of flow at a particular place or moves downward while being affected by gravity (Lee and Rezaie 2022). Figure 3c shows that regions with higher TWI (8.04 to 23.8), accounting for 648.25 km² (14.79%), are typically characterized by low slopes and large upslope catchments, which favor water convergence, soil moisture accumulation, and saturation, resulting in regions that are highly susceptible to flooding. Likewise, areas with a higher Topographic Ruggedness Index (TRI) (0.52 to 0.89), encompassing an area of 1872.92 km² and accounting for 41.63%, are prone to flooding (Figure 3d) because rugged terrain increases surface runoff, which reduces the soil's ability to absorb water.

Hydrological Flood Predictors

Hydrological flood predictors are portrayed in Figure 4. Precipitation is the primary cause of surface runoff and consequent flooding, with a strong positive correlation, as intense precipitation can exceed infiltration capacity, trigger rapid surface runoff, and overwhelm both natural and engineered drainage systems. Heavy rainfall increases flooding risks and vice versa (Merga et al. 2023). Rainfall is heavier in the western portion of the study region (1934 to 2218 mm), contributing to a high susceptibility to flooding. These areas encompass 483.34 km², accounting for 10.72% (Figure 4a). This result aligns with earlier studies conducted in Ethiopian basins, including the Wabi Shebele River Sub-basin (Merga et al. 2023), Gidabo Watershed (Diriba et al. 2024), and the Fetam watershed (Desalegn and Mulu 2021), which emphasized rainfall intensity as a dominant trigger of flood events.

[IMAGE OMITTED. SEE PDF]

Proximity to riverine environments is a principal factor, having a reciprocal relationship with flood hazards. Figure 4b shows river networks and buffer zones. Areas closer to river channels (< 200 m), covering an area of 860.08 km² (19.07%), are highly susceptible to flooding (Yin et al. 2023) because they are directly exposed to overbank flow, lateral water spreading, and sediment deposition during flood events. The present spatial analysis indicates that high-susceptibility zones are predominantly concentrated along major river corridors (Figure 7).

According to Kalédjé et al. (2019), there is a positive association between flood occurrence and drainage density; regions characterized by a higher drainage density are highly prone to flooding, as a denser stream network accelerates surface runoff concentration, shortens the lag time between rainfall and peak discharge, and reduces opportunities for infiltration, which increases the likelihood of flash flooding in catchments with compact drainage patterns. Higher drainage density (1.8 to 2.99) contributes to a higher probability of flooding, accounting for 5.9% of the study region (Figure 4c). Our findings align with maintained insights from Majeed et al. (2023) in Jhelum District, Pakistan, and Kader et al. (2024) in Bangladesh, corroborating the critical influence of dense drainage networks on flood vulnerability.

The Stream Power Index (SPI) indicates the erosivity of rivers and the surface runoff rate, with a higher SPI indicating a higher surface runoff capacity and vice versa (Edamo et al. 2022). The study area's west, southwest, and northeast regions experience higher SPI, extending from −0.85 to 11.8, covering an area of 1739 km² (38.66%), and are more susceptible to flooding (Figure 4d). These areas are more prone to channel incision, sediment transport, and stream bank erosion. Such conditions not only elevate localized flood risks but also intensify downstream flooding by accelerating runoff concentration and increasing peak discharges.

Spectral Indices and Land Use Practices

The Soil Adjusted Vegetation Index (SAVI) suggests that soil conditions and vegetation growth influence the distribution of precipitation over the landscape, which has an inverse relationship with flooding. Areas with high SAVI values indicate dense and healthy vegetation, which enhances infiltration, stabilizes soils, reduces surface runoff, and consequently lowers flood risk. Figure 5a illustrates that the west and southeast margins of the study region are characterized by lower SAVI values, ranging from −0.25 to 0.09, covering an area of 89.68 km², which is more prone to flooding. The Bare Soil Index (BSI) is another crucial spectral index used to quantify the extent of exposed or non-vegetated land surfaces, where high BSI values indicate extensive bare or degraded soils, which are typically more susceptible to flooding due to their reduced capacity to absorb rainfall and their tendency to promote surface runoff. Higher Bare Soil Index (BSI) values ranging from 0.42 to 0.62 were observed at the study region's west, south, southeast, and northeast margins. These areas encompass approximately 1619.31 km², which are highly likely to flood (Figure 5b).

[IMAGE OMITTED. SEE PDF]

The Modified Normalized Difference Water Index (MNDWI) is a significant spectral index considered in the present study to differentiate open water features, which has a direct relationship with flood susceptibility. The higher MNDWI values correspond to low-lying and water-retaining areas, indicating areas with greater surface water presence, proximity to wetlands, rivers, or saturated soils, all of which are inherently more vulnerable to flooding. The study regions with higher MNDWI values (−0.065 to 0.18) are more prone to flooding (Figure 5c). This demonstrates that integrating MNDWI with terrain-based factors enhances the predictive power of GIS–AHP frameworks for flood susceptibility mapping. The flood magnitude has a strong negative association with vegetation cover (Zhu et al. 2022). The study area regions, which are covered with barren land, waterbodies, wetlands, built-up areas, and sparse vegetation, are more vulnerable to flooding (Figure 5d).

Soil Characteristics

The soil types and texture are critical factors in determining flood-prone areas, as they accelerate or delay the precipitation's water-holding capacity and infiltration. Flooding danger rises as soil water-holding capacity declines due to increased surface runoff. The study region is characterized by a clay soil texture, encompassing an area of 3308.032 km² (73.63%), which is moderately prone to flooding due to its higher infiltration rate. The remaining study area has a clay loam soil texture, covering an area of 1157.27 km² (25.76%), which is susceptible to flooding due to its lower infiltration rate (Figure 6a). The Eutric Cambisols (Be) and Eutric Nitisols (Ne) soil type occupy the west and northcentral regions of the study area, covering 1432.24 and 907.62 km², respectively (Table 4), which are moderately susceptible to flooding (Figure 6b). Because of their low water holding capacity and penetration, Plinthic Ferralsols (Fp) are highly susceptible to flooding, covering an area of 45.44 km².

[IMAGE OMITTED. SEE PDF]

TABLE 4 Contextualization of related studies within regional and international frameworks.

Study area	Scholars	Flood conditioning factors (high to low influence)
Wabi Shebele River Sub-basin, Ethiopia	Merga et al. 2023	Elevation > slope > LULC > rainfall > soil type; distance from rivers
Ambo Town and its watershed, Ethiopia	Ogato et al. 2020	Land cover > slope > soil type > rainfall > drainage density > elevation
Gidabo Watershed, Ethiopia	Diriba et al. 2024	Slope > elevation > rainfall > LULC > drainage density > soil type
Fetam watershed, upper Abbay basin, Ethiopia	Desalegn and Mulu 2021	Slope > elevation > rainfall > LULC > drainage density > soil type
Adigrat Town, Tigray region, Ethiopia	Nigusse and Adhanom 2019	Slope > elevation > flow accumulation > LULC > flow direction > annual precipitation > water table
Dega Damot district, Ethiopia	Negese et al. 2022	Elevation > slope > flow accumulation > rainfall >> distance to rivers > drainage density > LULC > TWI > soil type > NDVI > curvature
Upper Awash River basin, Ethiopia	Hagos et al. 2022	Slope > elevation > drainage density > proximity to river > rainfall > soil texture > land use
Dosso region, Niger	Almouctar et al. 2024	Elevation > TWI; NDVI > slope > soil type > rainfall; distance from road; drainage density > distance from river > LULC >
Bangladesh	Kader et al. 2024	Elevation > slope > drainage density > rainfall > distance from the river > flow accumulation > curve number > permeability
Dhalai River, Tripura, India	Ahmed et al. 2024	Flow accumulation > elevation > slope > distance from river > rainfall > TWI > Land use > soil > profile curvature
Slovakia	Vojtek and Vojteková 2019	Slope > river density > distance from rivers > flow accumulation > elevation > curve number > lithology
Jhelum, Pakistan	Majeed et al. 2023	Distance from the river > drainage density > slope > elevation > rainfall > soil > geology > LULC
Subarnarekha basin, India	Das and Gupta 2021	Elevation > slope > distance from river > drainage density > TWI > land use > rainfall > geomorphology > soil texture > TRI > curvature > geology
Wolaita Zone, Ethiopia	The present study	Rainfall > distance from river networks > elevation > BSI > SPI > slope > drainage density > SAVI > TWI > MNDWI > TRI > LULC > soil type > soil texture

Limitations of the Study and Recommendations

The absence of detailed hydrological records and flood inventory data in the Wolaita Zone posed a significant limitation to this study, restricting the possibility of rigorous quantitative validation. Instead, a contextual validation was carried out by comparing the generated susceptibility zones with secondary data sources, including government disaster reports and community-level flood observations. The strong correspondence between identified high-susceptibility areas and historically flood-affected zones lends credibility to the model outputs despite data constraints. Moreover, the acceptable consistency ratio (CR = 0.044) (Tables 5 and 6), together with alignment to regional and global methodologies and findings (Table 4), confirms the robustness of the results and adaptability of the GIS-AHP framework for flood susceptibility mapping in diverse Ethiopian terrains and comparable environments worldwide. To address limitations, future research should integrate machine learning or ensemble approaches to refine factor weighting, develop local flood inventory databases for validation, and incorporate time-series remote sensing data to capture spatiotemporal flood dynamics. Strengthening collaborations with local institutions and engaging communities in participatory validation will further enhance the reliability and policy relevance of flood susceptibility mapping in data-scarce regions.

TABLE 5 Normalized pair-wise comparison matrix.

Flood conditioning factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	Sum	Normalized weights	Criteria weights (%)
Rainfall (C1)	0.1815	0.2820	0.2634	0.1798	0.1554	0.1938	0.1531	0.1558	0.1171	0.0986	0.1030	0.0899	0.0909	0.1014	2.1659	0.1547	15
River networks (C2)	0.0908	0.1410	0.1317	0.2696	0.1554	0.1938	0.1531	0.1558	0.1561	0.1315	0.1030	0.1124	0.1091	0.1014	2.0049	0.1432	14
Elevation (C3)	0.0908	0.1410	0.1317	0.1798	0.2332	0.1292	0.1531	0.1039	0.1171	0.0986	0.1030	0.1124	0.1091	0.1014	1.8042	0.1289	13
BSI (C4)	0.0908	0.0470	0.0659	0.0899	0.1554	0.1292	0.1531	0.1039	0.1171	0.1315	0.1288	0.1124	0.1091	0.1014	1.5354	0.1097	11
SPI (C5)	0.0908	0.0705	0.0439	0.0449	0.0777	0.1292	0.1021	0.1558	0.1171	0.1315	0.1030	0.0899	0.0727	0.0725	1.3017	0.0930	9
Slope (C6)	0.0605	0.0470	0.0659	0.0449	0.0389	0.0646	0.1021	0.1039	0.1171	0.0986	0.1030	0.0899	0.1091	0.1014	1.1469	0.0819	8
Drainage density (C7)	0.0605	0.0470	0.0439	0.0300	0.0389	0.0323	0.0510	0.0519	0.0781	0.0986	0.1030	0.1124	0.1091	0.1014	0.9581	0.0684	7
SAVI (C8)	0.0605	0.0470	0.0659	0.0449	0.0259	0.0323	0.0510	0.0519	0.0781	0.0658	0.0515	0.0674	0.0727	0.0870	0.8019	0.0573	6
TWI (C9)	0.0605	0.0352	0.0439	0.0300	0.0259	0.0215	0.0255	0.0260	0.0390	0.0658	0.0773	0.0674	0.0727	0.0725	0.6632	0.0474	5
MNDWI (C10)	0.0605	0.0352	0.0439	0.0225	0.0194	0.0215	0.0170	0.0260	0.0195	0.0329	0.0515	0.0674	0.0545	0.0580	0.5299	0.0379	4
TRI (C11)	0.0454	0.0352	0.0329	0.0180	0.0194	0.0162	0.0128	0.0260	0.0130	0.0164	0.0258	0.0225	0.0364	0.0435	0.3634	0.0260	3
LULC (C12)	0.0454	0.0282	0.0263	0.0180	0.0194	0.0162	0.0102	0.0173	0.0130	0.0110	0.0258	0.0225	0.0182	0.0290	0.3004	0.0215	2
Soil type (C13)	0.0363	0.0235	0.0220	0.0150	0.0194	0.0108	0.0085	0.0130	0.0098	0.0110	0.0129	0.0225	0.0182	0.0145	0.2372	0.0169	2
Soil texture (C14)	0.0259	0.0201	0.0188	0.0128	0.0155	0.0092	0.0073	0.0087	0.0078	0.0082	0.0086	0.0112	0.0182	0.0145	0.1870	0.0134	1
Conditioning factors	14	1	100

TABLE 6 Calculating consistency.

Flood conditioning factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	Weighted sum value	Normalized weights	WSV/CW (PEV)
Rainfall (C1)	0.1547	0.2864	0.2577	0.2193	0.1860	0.2458	0.2053	0.1718	0.1421	0.1136	0.1038	0.0858	0.0847	0.0935	2.3506	0.1547	15.1938
River networks (C2)	0.0774	0.1432	0.1289	0.3290	0.1860	0.2458	0.2053	0.1718	0.1895	0.1514	0.1038	0.1073	0.1016	0.0935	2.2344	0.1432	15.6029
Elevation (C3)	0.0774	0.1432	0.1289	0.2193	0.2789	0.1638	0.2053	0.1146	0.1421	0.1136	0.1038	0.1073	0.1016	0.0935	1.9933	0.1289	15.4670
BSI (C4)	0.0774	0.0477	0.0644	0.1097	0.1860	0.1638	0.2053	0.1146	0.1421	0.1514	0.1298	0.1073	0.1016	0.0935	1.6945	0.1097	15.4508
SPI (C5)	0.0774	0.0716	0.0430	0.0548	0.0930	0.1638	0.1369	0.1718	0.1421	0.1514	0.1038	0.0858	0.0678	0.0668	1.4300	0.0930	15.3800
Slope (C6)	0.0516	0.0477	0.0644	0.0548	0.0465	0.0819	0.1369	0.1146	0.1421	0.1136	0.1038	0.0858	0.1016	0.0935	1.2388	0.0819	15.1222
Drainage density (C7)	0.0516	0.0477	0.0430	0.0366	0.0465	0.0410	0.0684	0.0573	0.0947	0.1136	0.1038	0.1073	0.1016	0.0935	1.0065	0.0684	14.7068
SAVI (C8)	0.0516	0.0477	0.0644	0.0548	0.0310	0.0410	0.0684	0.0573	0.0947	0.0757	0.0519	0.0644	0.0678	0.0801	0.8509	0.0573	14.8543
TWI (C9)	0.0516	0.0358	0.0430	0.0366	0.0310	0.0273	0.0342	0.0286	0.0474	0.0757	0.0779	0.0644	0.0678	0.0668	0.6211	0.0474	13.1114
MNDWI (C10)	0.0516	0.0358	0.0430	0.0274	0.0232	0.0273	0.0228	0.0286	0.0237	0.0379	0.0519	0.0644	0.0508	0.0534	0.4884	0.0379	12.9029
TRI (C11)	0.0387	0.0358	0.0322	0.0219	0.0232	0.0205	0.0171	0.0286	0.0158	0.0189	0.0260	0.0215	0.0339	0.0401	0.3341	0.0260	12.8731
LULC (C12)	0.0387	0.0286	0.0258	0.0219	0.0232	0.0205	0.0137	0.0191	0.0158	0.0126	0.0260	0.0215	0.0169	0.0267	0.2843	0.0215	13.2509
Soil type (C13)	0.0309	0.0239	0.0215	0.0183	0.0232	0.0137	0.0114	0.0143	0.0118	0.0126	0.0130	0.0215	0.0169	0.0134	0.2330	0.0169	13.7556
Soil texture (C14)	0.0221	0.0205	0.0184	0.0157	0.0186	0.0117	0.0098	0.0095	0.0095	0.0095	0.0087	0.0107	0.0169	0.0134	0.1815	0.0134	13.5915
Number of comparisons = 91	PEV = 14.900	lmax-n = 0.9		CI = 0.07	RI (n = 14) = 1.57		CR = 0.0441	1	208.6

Conclusion

The study was conducted in the Wolaita Zone of Ethiopia and aimed to map flood-vulnerable areas by adopting GIS approaches and Remote Sensing in combination with MCDMA. Topographical, hydrological, and spectral indices conditioning factors were integrated, and a weighted overlay analysis was applied in ArcGIS. By embedding LULC and soil properties into the model, the study also captured anthropogenic and edaphic influences, marking a substantive improvement over existing AHP-based flood mapping studies, which are commonly underrepresented. The study revealed that about 380.49 km² (8.73%), 3362.5 km² (77.16%), and 613.73 km² (14.08%) of the study area are susceptible, moderately susceptible, and less susceptible to flooding, respectively (Figure 7). The susceptible areas are mainly observed in the study area's west, northwest, southwest, and south margins. These regions are characterized by relatively low-lying areas (649–1123 m.a.s.l), receiving higher rainfall (1724–2218 mm/year), close to river networks (< 200 m), higher drainage density (1.28–2.99 km/km²), higher TWI (8.04–23.8), higher SPI (−0.85–11.8), higher BSI (0.42–0.62), and higher MNDWI (−0.065–0.18). Besides, these regions are covered by shrublands, grasslands, barren land, and sparse vegetation. The findings of this study provide significant insights for local stakeholders, including government agencies, urban planners, agricultural extension offices, and disaster risk management authorities, by identifying areas most susceptible to flooding. The flood susceptibility map can serve as an evidence-based tool to guide land-use planning, infrastructure development, and settlement expansion, ensuring that new investments should avoid high-risk zones. For local farmers and agricultural cooperatives, the results highlight vulnerable farmlands where soil and water conservation practices, such as terracing, afforestation, and improved drainage systems, should be prioritized. Moreover, the result provides practical guidance for disaster preparedness and early warning initiatives, including the development of local action plans, which enable authorities to allocate resources more efficiently, formulate flood hazard mitigation and adaptation strategies, and strengthen community resilience.

[IMAGE OMITTED. SEE PDF]

Acknowledgments

The authors thank Gambella University and Wolaita Sodo University for providing facilities and opportunities. Furthermore, the authors extend their appreciation to the USGS for providing access to satellite data.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

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

References

Addis, A. 2023. “GIS–Based Flood Susceptibility Mapping Using Frequency Ratio and Information Value Models in the Upper Abay River Basin, Ethiopia.” Natural Hazards Research 3, no. 2: 247–256. https://doi.org/10.1016/j.nhres.2023.02.003.

Adelekan, I. O., and A. P. Asiyanbi. 2016. “Flood Risk Perception in Flood‐Affected Communities in Lagos, Nigeria.” Natural Hazards 80, no. 1: 445–469. https://doi.org/10.1007/s11069‐015‐1977‐2.

Ahmed, I., N. Das (Pan), J. Debnath, M. Bhowmik, and S. Bhattacharjee. 2024. “Flood Hazard Zonation Using GIS‐Based Multi‐Parametric Analytical Hierarchy Process.” Geosystems and Geoenvironment 3, no. 2: 100250. https://doi.org/10.1016/j.geogeo.2023.100250.

Almouctar, M. A. S., Y. Wu, S. An, et al. 2024. “Flood Risk Assessment in Arid and Semi‐Arid Regions Using Multi‐Criteria Approaches and Remote Sensing in a Data‐Scarce Region.” Journal of Hydrology: Regional Studies 54: 101862. https://doi.org/10.1016/j.ejrh.2024.101862.

Billi, P., Y. Tadesse, and R. Ciampalini. 2013. “The Increased Frequency of Flash Floods in Ethiopia: Climate Change or Human Impact?”.

Das, S., and A. Gupta. 2021. “Multi‐Criteria Decision‐Based Geospatial Mapping of Flood Susceptibility and Temporal Hydro‐Geomorphic Changes in the Subarnarekha Basin, India.” Geoscience Frontiers 12, no. 5: 101206. https://doi.org/10.1016/j.gsf.2021.101206.

Desalegn, H., and A. Mulu. 2021. “Flood Vulnerability Assessment Using GIS at the Fetam Watershed, Upper Abbay Basin, Ethiopia.” Heliyon 7, no. 1: e05865. https://doi.org/10.1016/j.heliyon.2020.e05865.

Diaconu, D. C., R. Costache, A. R. M. Towfiqul Islam, et al. 2024. “Developing a Flood Mapping Procedure Through Optimized Machine Learning Techniques. Case Study: Prahova River Basin, Romania.” Journal of Hydrology: Regional Studies 54: 101892. https://doi.org/10.1016/j.ejrh.2024.101892.

Diriba, D., T. Takele, S. Karuppannan, and M. Husein. 2024. “Flood Hazard Analysis and Risk Assessment Using Remote Sensing, GIS, and AHP Techniques: A Case Study of the Gidabo Watershed, Main Ethiopian Rift, Ethiopia.” Geomatics, Natural Hazards and Risk 15, no. 1: 2361813. https://doi.org/10.1080/19475705.2024.2361813.

Dutta, P., and S. Deka. 2024. “A Novel Approach to Flood Risk Assessment: Synergizing With Geospatial‐Based MCDM‐AHP Model, Multicollinearity, and Sensitivity Analysis in the Lower Brahmaputra Floodplain, Assam.” Journal of Cleaner Production 467: 142985. https://doi.org/10.1016/j.jclepro.2024.142985.

Edamo, M. L., K. Bushira, and T. Y. Ukumo. 2022. “Flood Susceptibility Mapping in the Bilate Catchment, Ethiopia.” H2Open Journal 5, no. 4: 691–712. https://doi.org/10.2166/h2oj.2022.128.

Getahun, Y. S., and S. L. Gebre. 2015. “Flood Hazard Assessment and Mapping of Flood Inundation Area of the Awash River Basin in Ethiopia Using GIS and HEC‐GeoRAS/HEC‐RAS Model.” Journal of Civil & Environmental Engineering 5, no. 4: 1.

Goepel, K. D. 2018. “Implementation of an Online Software Tool for the Analytic Hierarchy Process (AHP‐OS).” International Journal of the Analytic Hierarchy Process 10, no. 3: 1–10. https://doi.org/10.13033/ijahp.v10i3.590.

Hagos, Y. G., T. G. Andualem, M. Yibeltal, and M. A. Mengie. 2022. “Flood Hazard Assessment and Mapping Using GIS Integrated With Multi‐Criteria Decision Analysis in the Upper Awash River Basin, Ethiopia.” Applied Water Science 12, no. 7: 148. https://doi.org/10.1007/s13201‐022‐01674‐8.

Huete, A. R. 1988. “A Soil‐Adjusted Vegetation Index (SAVI).” Remote Sensing of Environment 25, no. 3: 295–309. https://doi.org/10.1016/0034‐4257(88)90106‐X.

Hussain, M., M. Tayyab, J. Zhang, et al. 2021. “GIS‐Based Multi‐Criteria Approach for Flood Vulnerability Assessment and Mapping in District Shangla: Khyber Pakhtunkhwa, Pakistan.” Sustainability 13, no. 6: 3126. https://doi.org/10.3390/su13063126.

Kader, Z., M. R. Islam, M. T. Aziz, et al. 2024. “GIS and AHP‐Based Flood Susceptibility Mapping: A Case Study of Bangladesh.” Sustainable Water Resources Management 10, no. 5: 170. https://doi.org/10.1007/s40899‐024‐01150‐y.

Kalédjé, P. S. K., J.‐R. N. Ngoupayou, A. F. Takounjou, M. Zebsa, F. Rakotondrabe, and J. M. Ondoa. 2019. “Floods of 18 and 19 November 2016 in Batouri (East Cameroon): Interpretation of the Hydro‐Meteorological Parameters and Historical Context of the Post‐Event Survey Episode.” Scientific World Journal 2019: 1–7. https://doi.org/10.1155/2019/3814962.

Lee, S., and F. Rezaie. 2022. “Data Used for GIS‐Based Flood Susceptibility Mapping.” Geo Data 4, no. 1: 1–15. https://doi.org/10.22761/DJ2022.4.1.001.

Legese, B., and B. Gumi. 2020. “Flooding in Ethiopia; Causes, Impact and Coping Mechanisms. A Review.” International Journal of Research and Analytical Reviews 7, no. 3: 707–717.

Li, Y., and H. Hong. 2023. “Modelling Flood Susceptibility Based on Deep Learning Coupling With Ensemble Learning Models.” Journal of Environmental Management 325: 116450. https://doi.org/10.1016/j.jenvman.2022.116450.

Maharjan, M., S. Timilsina, S. Ayer, B. Singh, B. Manandhar, and A. Sedhain. 2024. “Flood Susceptibility Assessment Using a Machine Learning Approach in the Mohana‐Khutiya River of Nepal.” Natural Hazards Research 4, no. 1: 32–45. https://doi.org/10.1016/j.nhres.2024.01.001.

Mahato, S., S. Pal, S. Talukdar, T. K. Saha, and P. Mandal. 2021. “Field‐Based Index of Flood Vulnerability (IFV): A New Validation Technique for Flood Susceptible Models.” Geoscience Frontiers 12, no. 5: 101175. https://doi.org/10.1016/j.gsf.2021.101175.

Majeed, M., L. Lu, M. M. Anwar, et al. 2023. “Prediction of Flash Flood Susceptibility Using Integrating the Analytic Hierarchy Process (AHP) and Frequency Ratio (FR) Algorithms.” Frontiers in Environmental Science 10: 1037547. https://doi.org/10.3389/fenvs.2022.1037547.

Majid, M., and B. A. Mir. 2021. “Landfill Site Selection Using GIS‐Based Multi‐Criteria Evaluation Technique. A Case Study of Srinagar City, India.” Environmental Challenges 3: 100031. https://doi.org/10.1016/j.envc.2021.100031.

Mangini, W., A. Viglione, J. Hall, et al. 2018. “Detection of Trends in the Magnitude and Frequency of Flood Peaks Across Europe.” Hydrological Sciences Journal 63, no. 4: 493–512. https://doi.org/10.1080/02626667.2018.1444766.

Merga, B. B., F. L. Mamo, M. B. Moisa, F. S. Tiye, and D. O. Gemeda. 2023. “Assessment of Flood Risk by Using Geospatial Techniques in Wabi Shebele River Sub‐Basin, West Hararghe Zone, Southeastern Ethiopia.” Applied Water Science 13, no. 11: 214. https://doi.org/10.1007/s13201‐023‐02019‐9.

Mitra, R., J. Das, and M. Kamruzzaman. 2023. “Application of TOPSIS Method for Flood Susceptibility Mapping Using Excel and GIS.” MethodsX 11: 102263. https://doi.org/10.1016/j.mex.2023.102263.

Moore, I. D., R. B. Grayson, and A. R. Ladson. 1991. “Digital Terrain Modelling: A Review of Hydrological, Geomorphological, and Biological Applications.” Hydrological Processes 5, no. 1: 3–30. https://doi.org/10.1002/hyp.3360050103.

Mostofi Zadeh, S., D. H. Burn, and N. O'Brien. 2020. “Detection of Trends in Flood Magnitude and Frequency in Canada.” Journal of Hydrology: Regional Studies 28: 100673. https://doi.org/10.1016/j.ejrh.2020.100673.

Msabi, M. M., and M. Makonyo. 2021. “Flood Susceptibility Mapping Using GIS and Multi‐Criteria Decision Analysis: A Case of Dodoma Region, Central Tanzania.” Remote Sensing Applications: Society and Environment 21: 100445. https://doi.org/10.1016/j.rsase.2020.100445.

Negese, A., D. Worku, A. Shitaye, and H. Getnet. 2022. “Potential Flood‐Prone Area Identification and Mapping Using GIS‐Based Multi‐Criteria Decision‐Making and Analytical Hierarchy Process in Dega Damot District, Northwestern Ethiopia.” Applied Water Science 12, no. 12: 255. https://doi.org/10.1007/s13201‐022‐01772‐7.

Nigusse, A. G., and O. G. Adhanom. 2019. “Flood Hazard and Flood Risk Vulnerability Mapping Using Geo‐Spatial and MCDA Around Adigrat, Tigray Region, Northern Ethiopia.” Momona, Ethiopian Journal of Science 11, no. 1: 90. https://doi.org/10.4314/mejs.v11i1.6.

Ogato, G. S., A. Bantider, K. Abebe, and D. Geneletti. 2020. “Geographic Information System (GIS)‐Based Multicriteria Analysis of Flooding Hazard and Risk in Ambo Town and Its Watershed, West Shoa Zone, Oromia Regional State, Ethiopia.” Journal of Hydrology: Regional Studies 27: 100659. https://doi.org/10.1016/j.ejrh.2019.100659.

Rahmati, O., H. R. Pourghasemi, and H. Zeinivand. 2016. “Flood Susceptibility Mapping Using Frequency Ratio and Weights‐Of‐Evidence Models in the Golastan Province, Iran.” Geocarto International 31, no. 1: 42–70. https://doi.org/10.1080/10106049.2015.1041559.

Ramayanti, S., A. S. Nur, M. Syifa, et al. 2022. “Performance Comparison of Two Deep Learning Models for Flood Susceptibility Map in Beira Area, Mozambique.” Egyptian Journal of Remote Sensing and Space Science 25, no. 4: 1025–1036. https://doi.org/10.1016/j.ejrs.2022.11.003.

Ray, K., P. Pandey, C. Pandey, A. P. Dimri, and K. Kishore. 2019. “On the Recent Floods in India.” Current Science 117, no. 2: 204. https://doi.org/10.18520/cs/v117/i2/204‐218.

Rudra, R. R., and S. K. Sarkar. 2023. “Artificial Neural Network for Flood Susceptibility Mapping in Bangladesh.” Heliyon 9, no. 6: e16459. https://doi.org/10.1016/j.heliyon.2023.e16459.

Saaty, T. L. 1980. The Analytical Hierarchy Process: Planning, Priority Setting, Resource Allocation. McGraw‐Hill International Book Company ‐ Business & Economics.

Saaty, T. L. 2006. “The Analytic Network Process.” In Decision Making With the Analytic Network Process, vol. 95, 1–26. Springer US. https://doi.org/10.1007/0‐387‐33987‐6_1.

Saaty, T. L., and L. G. Vargas. 2012. Models, Methods, Concepts & Applications of the Analytic Hierarchy Process. Vol. 175. Springer US. https://doi.org/10.1007/978‐1‐4614‐3597‐6.

Sajinkumar, K. S., A. Arya, A. Rajaneesh, et al. 2022. “Migrating Rivers, Consequent Paleochannels: The Unlikely Partners and Hotspots of Flooding.” Science of the Total Environment 807: 150842. https://doi.org/10.1016/j.scitotenv.2021.150842.

Siam, Z. S., R. T. Hasan, S. S. Anik, F. Noor, M. S. G. Adnan, and R. M. Rahman. 2021. “Study of Hybridized Support Vector Regression Based Flood Susceptibility Mapping for Bangladesh.” In Advances and Trends in Artificial Intelligence. From Theory to Practice, edited by H. Fujita, A. Selamat, J. C.‐W. Lin, and M. Ali, vol. 12799, 59–71. Springer International Publishing. https://doi.org/10.1007/978‐3‐030‐79463‐7_6.

Souissi, D., L. Zouhri, S. Hammami, M. H. Msaddek, A. Zghibi, and M. Dlala. 2020. “GIS‐Based MCDM—AHP Modeling for Flood Susceptibility Mapping of Arid Areas, Southeastern Tunisia.” Geocarto International 35, no. 9: 991–1017. https://doi.org/10.1080/10106049.2019.1566405.

Tariq, A., L. Hashemi Beni, S. Ali, S. Adnan, and W. A. Hatamleh. 2023. “An Effective Geospatial‐Based Flash Flood Susceptibility Assessment With Hydrogeomorphic Responses on Groundwater Recharge.” Groundwater for Sustainable Development 23: 100998. https://doi.org/10.1016/j.gsd.2023.100998.

Tehrany, M. S., B. Pradhan, and M. N. Jebur. 2014. “Flood Susceptibility Mapping Using a Novel Ensemble Weights‐of‐Evidence and Support Vector Machine Models in GIS.” Journal of Hydrology 512: 332–343. https://doi.org/10.1016/j.jhydrol.2014.03.008.

Tien Bui, D., N.‐D. Hoang, F. Martínez‐Álvarez, et al. 2020. “A Novel Deep Learning Neural Network Approach for Predicting Flash Flood Susceptibility: A Case Study at a High Frequency Tropical Storm Area.” Science of the Total Environment 701: 134413. https://doi.org/10.1016/j.scitotenv.2019.134413.

Ullah, I., B. Aslam, S. H. I. A. Shah, et al. 2022. “An Integrated Approach of Machine Learning, Remote Sensing, and GIS Data for the Landslide Susceptibility Mapping.” Land 11, no. 8: 1265. https://doi.org/10.3390/land11081265.

Vishnu, C. L., K. S. Sajinkumar, T. Oommen, et al. 2019. “Satellite‐Based Assessment of the August 2018 Flood in Parts of Kerala, India.” Geomatics, Natural Hazards and Risk 10, no. 1: 758–767. https://doi.org/10.1080/19475705.2018.1543212.

Vojtek, M., and J. Vojteková. 2019. “Flood Susceptibility Mapping on a National Scale in Slovakia Using the Analytical Hierarchy Process.” Water 11, no. 2: 364. https://doi.org/10.3390/w11020364.

Wahba, M., R. Essam, M. El‐Rawy, N. Al‐Arifi, F. Abdalla, and W. M. Elsadek. 2024. “Forecasting of Flash Flood Susceptibility Mapping Using Random Forest Regression Model and Geographic Information Systems.” Heliyon 10, no. 13: e33982. https://doi.org/10.1016/j.heliyon.2024.e33982.

Wang, Z., C. Lai, X. Chen, B. Yang, S. Zhao, and X. Bai. 2015. “Flood Hazard Risk Assessment Model Based on Random Forest.” Journal of Hydrology 527: 1130–1141. https://doi.org/10.1016/j.jhydrol.2015.06.008.

Yin, H., Q. Wu, S. Yin, S. Dong, Z. Dai, and M. R. Soltanian. 2023. “Predicting Mine Water Inrush Accidents Based on Water Level Anomalies of Borehole Groups Using Long Short‐Term Memory and Isolation Forest.” Journal of Hydrology 616: 128813. https://doi.org/10.1016/j.jhydrol.2022.128813.

Zhu, K., C. Lai, Z. Wang, Z. Zeng, Z. Mao, and X. Chen. 2024. “A Novel Framework for Feature Simplification and Selection in Flood Susceptibility Assessment Based on Machine Learning.” Journal of Hydrology: Regional Studies 52: 101739. https://doi.org/10.1016/j.ejrh.2024.101739.

Zhu, W., J. Chen, Q. Sun, Z. Li, W. Tan, and Y. Wei. 2022. “Reconstructing of High‐Spatial‐Resolution Three‐Dimensional Electron Density by Ingesting SAR‐Derived VTEC Into IRI Model.” IEEE Geoscience and Remote Sensing Letters 19: 1–5. https://doi.org/10.1109/LGRS.2022.3178242.