Integrated GIS-Based MCDA and Machine Learning Techniques in Flood Susceptibility Mapping in Ala River Basin, Nigeria

Integrated GIS-Based MCDA and Machine Learning Techniques in Flood Susceptibility Mapping in Ala River Basin, Nigeria | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Integrated GIS-Based MCDA and Machine Learning Techniques in Flood Susceptibility Mapping in Ala River Basin, Nigeria Adedoyin Benson Adeyemi, Akinola Adesuji Komolafe, Catherine Lilian Nakalembe, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4863685/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Flooding is a recognized form of natural disaster that can lead to loss of life, destruction of critical infrastructure with consequences impacting critical sectors including agriculture and health. This study aims to map out flood susceptible areas within the Ala River basin of Ondo State, Nigeria by integrating the Analytical Hierarchy Process (AHP) Multi-Criteria Decision Analysis (MCDA) technique and Support Vector Machines (SVM) Machine Learning (ML) model. Nineteen factors including elevation, slope, aspect, curvature (profile and plan), roughness, flow direction, flow accumulation, drainage density, distance from the river, TWI, STI, SPI, soil, geology, NDVI, NDMI, LULC, and rainfall were considered as input parameters. Flood susceptibility maps generated from each of these approaches were combined to create a more comprehensive flood susceptibility map of the study area. The AHP analysis has a consistency ratio of 1.8%. Precision, recall, f1-score, accuracy score, and ROC-AUC curve were used in evaluating the AHP-MCDA and SVM-ML model. Based on the evaluation, the combined flood susceptibility map result showed the best performance with the AUC score 0.74, SVM-ML with a score 0.73, and the AHP-MCDA having the least score of 0.59. As these results demonstrate, multiple approaches are required to mitigate flooding. Flood prediction Geographic Information System Multi-Criteria Decision-Making Analysis Machine Learning Flood Susceptibility River Basin Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1.0 INTRODUCTION Floods are natural hazards whose impacts can be intensified by uncontrolled urbanization and land-use changes. When severe flooding occurs in developed areas, it can cause extensive damage to residential structures, industrial facilities, public infrastructure, agricultural land, and crops, resulting in substantial economic losses and potential losses of human life (Olatona et. al., 2018 ). According to the World Bank, over 1.81 billion people worldwide are directly exposed to flood depths exceeding 0.15 meters in a 1-in-100-year flood event as of 2022. Globally, flooding causes more human casualties and property damage than any other natural hazard, making it one of the most severe natural risks (Ologunorisa, 2001; Alcira and Martha, 1991; Ishaya et al., 2009 ). In Nigeria, numerous regions have experienced severe flooding, displacing millions, disrupting businesses, contaminating water supplies, and increasing waterborne disease risk (Etuonovbe, 2011). UNICEF reported in 2022, that due to severe flooding, over 2.5 million Nigerians, 60% of whom are children require humanitarian assistance, and are at heightened risk of malnutrition, death due to drowning, and waterborne illnesses. Several studies have shown that floods in urban centers in developing countries are a result of inadequate drainage systems, inefficient storm sewerage networks, and blockages in existing drainage infrastructure (Vanneuville et al. , 2011; Abhas et al. , 2012; Bakare et. al., 2019 ). These urban floods are among the most frequent and characteristic natural disasters in cities, with consequential effects including significant human social disruptions and extensive damage to infrastructure (Bakare et. al., 2019 ). Floods manifest in various forms, including river, coastal, urban, flash, and groundwater floods and an area's vulnerability to flood events is influenced by physical factors such as topography, land cover, climate change, rainfall duration and frequency, and soil permeability (Thakur et al. , 2011; Bakara et al. , 2019). Climate change is the primary driver increasing flood risk, altering the changing behavior of extreme rainfall events (Pederson et. al. , 2012; Olatona et. al., 2018 ). Increased precipitation, higher surface runoff, rising global temperatures, and rapid urbanization all contribute to flooding (Danumah et. al., 2016 ). Urban flooding is intensified by deforestation, inadequate urban planning, and rapid urbanization, worsened by buildings in wetlands and swamps that serve as critical flood buffers. Given these challenges, accurate flood risk assessment is crucial, especially in vulnerable communities, to inform effective planning, mitigation, adaptation, and response strategies. This assessment involves understanding of flood likelihood and potential consequences, considering factors such as hazard, exposure, and vulnerability (De Moel et. al., 2015 ). Flood impacts are particularly severe when extreme rainfall affect poor communities in flood-prone areas, such as areas along the Ala River basin in Akure, Ondo State, Nigeria. These areas often have poorly constructed buildings and have limited adaptive capacity (Olatona et. al., 2018 ). Komolafe et. al. ( 2015 ) found that high levels of vulnerability and a lack of coping mechanisms among the Nigerian population are the main factors exacerbating flood impacts and losses. With the potential increase in the intensity of floods and its associated risks due to climate change and urbanization, more importance will be placed on effective flood risk management (UNISDR, 2015 ). In Nigeria, traditional approaches to flood management at various levels are basically river channelization and construction of river embankment, which often yield less positive outcome whenever extreme events occur (Ibitoye et al., 2020 ). However, recent development in flood risk management proposes the use of effective flood risk management, which places premium on non-structural measures (Komolafe et al., 2018 ). One of the flood risk management elements is the mapping and identifying spatial location, distribution and severity of flood hazards; this is very crucial in the identification of element at risk and the quantification of the potential flood risk in the study area (Brémond et al. , 2013; Merz et al. , 2010; UNISDR, 2023). Effective flood risk management therefore requires identifying flood-prone areas and prioritizing interventions in these high-risk zones (Sayers et. al. , 2013). This underscores the importance of developing precise flood susceptibility maps, which delineate areas with varying degrees of flood risk based on a range of physical, environmental, and socioeconomic factors (Ighile et. al., 2022 ). These maps can serve as critical tools for decision-makers, planners, and emergency responders, enabling them to allocate resources, implement risk reduction measures, and enhance community resilience to flooding (Ighile et. al., 2022 ). Geographic Information Systems (GIS), Remote Sensing (RS), and Machine Learning (ML) offer powerful tools for addressing flood risk in large, diverse countries like Nigeria (Ighile et. al., 2022 ). With a land area of 923,768 square kilometers and varied topography ranging from coastal regions to mountainous areas, Nigeria presents complex challenges for flood prediction and management. GIS provides a robust platform for spatial analysis, enabling the integration of diverse physical, environmental, and socio-economic factors that contribute to flood risk across Nigeria's vast landscape (Efraimidou & Spiliotis, 2024 ). Remote sensing technologies allow for the continuous monitoring of large areas, providing up-to-date information on land use changes, rainfall patterns, and river dynamics that is crucial for flood risk assessment. ML models, including Artificial Neural Networks (ANN) (Rahman & Ramli, 2024 ), Support Vector Machines (SVM) (Zehra, 2020 ), and Random Forest (RF) (Alipour et. al., 2020 ), have shown effectiveness in predicting flood occurrences. These models can process the complex, large-scale datasets characteristic of a country like Nigeria, leveraging historical flood data and multiple input variables to generate accurate predictions (Mosavi et. al., 2018 ). The integration of these technologies is particularly valuable for Nigeria, where flooding challenges are increasing and vary significantly across regions. This research aims to establish a method combining GIS techniques with the Analytic Hierarchy Process (AHP)-based Multi-Criteria Decision Analysis (MCDA) and Machine Learning (ML) model to predict and map flood-susceptible areas in the Ala River Basin of Akure, Ondo State. While focused on a specific region, this approach has the potential for broader application across Nigeria's diverse landscapes. 2.0 STUDY AREA Ala River basin flows along the Akure metropolis in Ondo State, Nigeria (Fig. 1). The basin extends geographically between 5° 9’ 0” E, 7° 18’ 0” N and 5° 27’ 0” E, 7° 6’ 0” N, covering an area of approximately 348 km 2 at a mean elevation of about 302 m. The Ala River, which has the most geographical impact, originates in the northwest highlands at Ipinsa and flows southeast through densely populated parts of Akure city. Several tributaries, including Eleye, Ipalefa, Alagbaka, and Omiyeye, join Ala or Elegbin rivers at various points. According to the Köppen classification scheme, the area falls within the Aw climate type, characterized by distinct wet and dry seasons. The driest month typically has precipitation of less than 60mm. This seasonal variation in rainfall, combined with the basin's topography and urbanization, contributes significantly to its flood susceptibility. The Ala River basin features diverse geomorphological landscapes, spurs, saddles, valleys, and river channels. Geologically, it is part of Nigeria's southwest basement complex. The major rock types in the area are Chanockite, Migmatite gneiss, Quartzite, and Biotite gneiss (Olatona et. al., 2018 ). These Precambrian rocks have undergone tectonic activities, resulting in fracturing, jointing, and cracking. These geological characteristics can affect soil permeability and drainage patterns, potentially impacting flood behavior and risk in the region. The Ala River basin's susceptibility to floods is particularly high due to its topography, climate, and rapid urbanization. The densely populated areas of Akure city, situated along the river's path, are especially vulnerable. During the wet season, heavy rainfall combined with the basin's limited capacity to absorb and channel water efficiently leads to frequent flooding events. These floods often result in significant economic losses, displacement of residents, and damage to infrastructure. Figure 2 illustrates the severe impacts of flooding in the Ala River Basin. The images show inundated residential areas, damaged roads, and disrupted economic activities following a major flood event in 2023. This visual representation underscores the urgent need for effective flood management strategies in the region. 3.0 MATERIALS AND METHODS 3.1 Datasets The study employed a variety of geospatial and environmental data sets to assess flood susceptibility in the Ala River Basin. These data sets, their resolutions, and their sources are summarized in Table 1 . The selection of these specific data types was based on their relevance to flood risk factors and their availability for the study area. The integration of these diverse data sets allows for a comprehensive analysis of the physical, environmental, and historical factors contributing to flood susceptibility in the region. Table 1 Data Sets Used in Flood Susceptibility Analysis of the Ala River Basin S/n Data Resolution Source 1. ALOS PALSAR DEM Data 12.5m Alaska Satellite Facility Website 2. Landsat-9 OLI/TIRS Data 30m USGS Earth Explorer Website 3. Climate Data 30secs WorldClim Website 4. Soil Data - FAO Website 5. Geology Data - Nigeria Geological Survey Website 6. Historical Flood Inventory Data - Fieldwork and other Secondary sources Data processing involved selecting key factors contributing to flooding in the study area (Table 2 ). These factor’s values were classified and given ratings with respect to their importance (Table S.1 ). Using the Digital Elevation Model (DEM), we generated hydrologic and topographic factor layers including elevation, slope, aspect, curvature, topographic wetness index (TWI), distance from rivers, and drainage density. Soil and geology map layers were derived from their respective downloaded datasets. Landsat satellite imagery was used to produce environmental factor layers, including land use/land cover, normalized difference vegetation index (NDVI), and normalized difference moisture index (NDMI). A rainfall map layer was created using data from the WorldClim database. All thematic maps were produced and processed using ArcGIS tools. To standardize the analysis, each thematic map was reclassified into five categories based on the factor's contribution to flood susceptibility. This reclassification allowed for consistent comparison and integration of the various factors in the subsequent flood risk assessment. Table 2 Summary of Flood Susceptibility Factors Relevant to this Study Factor Description Range/Classes Significance for Flooding Elevation Height above a fixed reference point (Ighile et. al., 2022 ). 250m − 552m Lower elevations are more flood-prone. Slope Degree of inclination of the surface (Edamo et. al. 2022). 0° − 60° Lower slopes increase flood likelihood. Aspect Direction of the slope face (Cea and Costabile, 2022 ). -1° − 360° Affects the water flow direction. Curvature (Plan and Profile) Slope of the slope of a surface (Longley et. al. , 2011). Plan: -6 to 11; Profile: -7 to 6 Concave areas are more flood-prone. Roughness Variation in elevation between pixels. 0.111m – 0.889m Higher roughness may slow water flow. Flow Direction Direction of water flow for each cell (Edamo et. al. , 2022). 1, 2, 4, 8, 16, 32, 64, 128 Determines water movement. Flow Accumulation Number of cells draining into each cell. 0–2,190,524 Higher flow accumulation values indicate flood-prone areas. Drainage Density Density of stream network (Avand et. al., 2021 ). 0–3.16 km 2 Higher drainage density increases flood risk. Distance from River Proximity to water bodies. 0–1,374 m Closer areas are more flood-prone (Liu et. al. , 2021; Edamo et. al. , 2022). TWI Topographic Wetness Index indicates the amount of water present in an area. It calculates the probability of downward water flow due to gravity (Ighile et. al., 2022 ). 1.9–25 Higher values indicate higher flood risk. STI Sediment Transport Index describes the particles in water moving due to water flow. 0–22,366 Higher values may indicate higher flood risk (Edamo et. al. , 2022). SPI Stream Power Index calculates the erosive power of flowing water (Ighile et. al., 2022 ). 0–6,656,820.5 Higher values indicate higher erosive power, resulting in higher flood risk. Soil Soil types affect water absorption and runoff based on soil characteristics (Edamo et. al. , 2022). Lixisols, Acrisols, Fluvisols Soil types with lower water absorption result in to increase in runoff. Geology Rock types affect water permeability which determines the rate of water infiltration into the sub-zones. Granite gneiss, Porphyritic granite Low permeable rocks increase flooding. NDMI Normalized Difference Moisture Index determine moisture content of vegetation (Mahdizadeh & Perez, 2022). -0.19–0.2 Higher values indicate more moisture resulting to flood risk increase. NDVI Normalized Difference Vegetation Index determine the health strength of vegetation. 0.04–0.32 Higher values indicate more vegetation resulting to flood risk increase. Land Use/Cover Types of land use and land cover shows the relationship between human activities, natural environment and floods (Mojaddadi et. al. , 2017). Waterbody, Vegetation, Built-up, Bareland/Cropland, and Outcrop Vegetation reduces flood risk. Built-up areas increase flood risk. Bareland and impervious surfaces also increase flood risk because of increased runoff. Rainfall Annual precipitation 165.44mm − 189.09mm Higher rainfall increases flood risk 3.2 Methods 3.2.1 Multi-Criteria Decision Analysis (MCDA)- Analytic Hierarchy Process (AHP) A hierarchical structure based on the Analytic Hierarchy Process (AHP) model was developed for the selected criteria. This was followed by a pair-wise comparison matrix analysis of the selected influencing variables. Priority values were assigned, and relative weights for each variable were determined. The ensure robustness of the process, the consistency of evaluations and decisions was measured. Conclusions regarding priority variables were then synthesized to identify flood vulnerability zones, following established methodologies (Vignesh et al. , 2020). Flood influencing factors were compared using Multi-Criteria Decision Making (MCDM) to compute relative weights. An MCDM-based AHP model was employed to quantify the importance of the selected flood-inducing factors, determining their potential to cause flood hazards. Factors were categorized into sub-classes with ranks 1 to 5 based on Saaty's preference scale (Table S.3 ). The factors’ ranks for the AHP-MCDA analysis based on their priority weight include: rainfall (14.5%), elevation (12.8%), distance from river (11.4%), flow accumulation (9.8%), TWI (8.1%), drainage density (7.2%), slope (5.8%), NDVI (5.2%), NDMI (4.4%), LULC (4.0%), soil (3.6%), flow direction (2.6%), roughness (2.3%), SPI (1.9%), STI (1.6%), aspect (1.4%), plan curvature (1.2%), profile curvature (1.1%), and geology (1.0%) (Table S.3 ). Thematic layers for each factor, indicating vulnerability levels from high, to very low, were integrated using the Weighted Linear Combination (WLC) in ArcGIS to produce the AHP-MCDA flood susceptibility zone map. A Consistency ratio (CR) was calculated to assess pairwise comparison consistency defined as the consistency index (CI) to a random-like matrix (RI) (Table S.2 ), depending on the number of influencing factors considered (Saaty, 1990; Mahmoud and Gan, 2018; Komolafe et al., 2020 ). The Consistency Index (CI) is calculated using the formula: $$\:CI=\:\frac{{\lambda\:}_{-n}}{n-1}$$ Where λ is the average of the consistency vector and n is the number of factors. $$\:CR=\frac{CI}{RI}$$ Where CI is the Consistency Index and RI is the random-like matrix. The accepted limit of the CR is 0.1 or 10%. This threshold is considered sufficient to detect the impact of each criterion. In this study, the calculated CR for pairwise comparison analysis was 0.018 (1.8%), well within the accepted limit. 3.2.2 Support Vector Machines (SVM) Machine Learning Model Analysis The SVM model was used to predict the flood-prone zones in the study area. SVM, an effective supervised learning technique for regression and classification problems, operates using the structural risk minimization rule and statistical learning theory (Mosavi et.al., 2018 ). For flood inventory data was split 70:30 into training datasets and testing datasets. The model was evaluated using the precision, recall, f1-score, overall accuracy, and ROC-AUC curve. The SVM flood susceptibility map was produced, classifying the area into five flood risk levels (very high, high, moderate, low, and very low) using the standard deviation break method. 3.2.3 Multicollinearity Investigation Multicollinearity, the high correlation between two or more independent variables can lead to interpretation problems and reduce model accuracy. This study used Pearson's correlation coefficients to assess multicollinearity. The coefficients range from − 1 to 1, with 1 indicating perfect positive correlation, -1 perfect negative correlation, and 0 no correlation. The multicollinearity analysis for the SVM-ML revealed high correlations (> 0.8) between NDVI-NDMI (0.89) and Flow Accumulation-SPI (0.97). NDMI and Flow Accumulation were removed to improve prediction accuracy (Table S.4 ). 3.2.4 AHP-MCDA and SVM-ML Integration Analysis The AHP-MCDA and SVM-ML approaches were integrated using weighted overlay spatial analysis to a comprehensive and reliable flood susceptibility map. The area extent and percentage over each susceptibility zone were calculated. 4.0 RESULTS 4.1 Flood Causative Factors Nineteen potential flood-triggering factors were considered and categorized into 5 susceptibility classes (Table S.1 ), except for geology (2 classes: porphyritic granite and granite gneiss) and soil (3 classes: acrisol, fluvisols, and lixisols). Elevation, being one of the important factors that causes flood occurrence as shown in Fig. 4 a reveals the south eastern part of the study area to be at lower elevation (i.e. 250m) proving it to be more vulnerable to flooding compare to the northern part and a small section of the southern part, having higher elevation (i.e. 552m). Similar to elevation, the area with decreasing slope in the study area is more prone to flooding (Fig. 4 b). Aspect which affects soil humidity and flow direction of water make flood mapping more accurate. The aspect map of the study area is divided into 5 classes: (-1–71.05), (71.05–143.11), (143.11–215.16), (215.16–287.22), (287.22–359.27) indicating the very high, high, moderate, low, very low vulnerable to flood respectively. The two curvatures (plan and profile) of the study area (Fig. 4 d and 4 e) ranges between − 6 to 11 values and − 7 to 6 values respectively, with the least curvature value more vulnerable to flooding (Table S.1 ). Unlike the curvatures, the higher the roughness value in the study area, the more prone it is to the flooding. The flow accumulation map (Fig. 4 h) which is derived from the flow direction map (Fig. 4 g) is divided into 5 classes, with the highest flow accumulation being most vulnerable to flooding. The drainage density and distance from the river both reveal the measure of proximity to the river in the study area, though the drainage density shows the density levels of the river flow (Fig. 4 i and 4 j). The TWI, STI, and SPI are divided into 5 classes (Fig. 4 k, 4 l, and 4 m respectively), with their class with the highest values representing areas most vulnerable to flooding, based on the fact that increase in wetness and rate of particles transport in water flow in an area infers increase in vulnerability to flood. The study area’s soil classification indicated on the soil map (Fig. 4 n) are 3 types: lixisols, acrisols, and fluvisols, with the lixisols covering over 80% of the study area. The geology map shows 2 main rock types which are granite gneiss and porphyritic granite (Fig. 4 o). The NDMI (Fig. 4 p) and NDVI (Fig. 4 q) values typically ranges between − 1 to 1. From both the NDMI and NDVI maps, its observed that the moisture level is majorly high at the middle of the basin due to high presence of vegetation in the area, while there is less vegetation at the upper portion of the basin. The LULC classification of the study area was into 5 classes including vegetation, built-up, bare ground/cropland, outcrop, and water body (Fig. 4 r). The vegetation covering a larger area extent of 36.5%, the bare ground/cropland covering 26.1%, the built-up 20.4%, the outcrop 16.9% and the water body covering 0.0018%. The Ala River and its tributaries serve as the major water bodies across the study area. The rainfall data of the study area as shown on the map layer ranges between 165.44 to 189.09 rainfall values (Fig. 4 s). The spatial distribution of the annual rainfall was prepared using kriging interpolation method, and classified into 5 classes. The lowest values class to the highest values class waving from the northern part of the study area to the southern part. Area with the highest rainfall indicate the most prone area to flooding. 4.2 Flood Susceptibility Maps 4.2.1 AHP-MCDA Flood Susceptibility Map The result from the AHP-MCDA mapping reveal that majority of the Ala River basin falls within the moderate flood susceptibility zone (Fig. 5 a). While the northern part which is the upstream is identified to be less susceptible to flood, the southern part of the basin which is the downstream is majorly identified to be highly susceptible to flooding. The calculated area covered by each class: very high, high, moderate, low, and very low as represented on the map include 0.15 km² (0.04%), 70.10 km² (20.36%), 209.48 km² (60.84%), 63.74 km² (18.51%), and 0.83 km² (0.24%) respectively (Table 3 ). 4.2.2 SVM-ML Flood Susceptibility Map Figures 5 b show flood susceptibility map from the SVM-ML analysis. It can be deduced from the map that the SVM-ML model better uncovered the upstream of the river which is the northern part of the basin is also prone to flooding as much as the downstream. The very high and high susceptible zone covers an area extent of 8.10 km 2 (2.33%) and 33.01 km 2 (9.51%) respectively. The moderate, low, and very low susceptible zones cover an area extent of 50.89 km 2 (14.66%), 108.89 km 2 (31.36%), and 146.32 km 2 (42.14%) respectively as shown in Table 3 . 4.2.3 Combined Approach Flood Susceptibility Map The map from the integration of the AHP-MCDA and SVM-ML approaches expressed an improved map of the flood susceptibility zones of the study area (Fig. 6 ). The map identified majorly the southern part (downstream) of the basin as most flood-prone, even as the northern part (upstream) is not ignored. The settlements situated within the highly susceptible zones of the basin are Ago-Owo, Ago-Dada, Agogboro, Olokuta, Olobi, and Iluabo, at the downstream. While at the upstream, Oba-Ile, Alagbaka, Odo-Ijoka, Isolo, Tuyi street, Araromi, and Ayedun street etc. (Fig. 6 ). Table 3 compares the flood susceptibility analysis results using AHP-MCDA, SVM-ML, and Combined Approach methods in the Ala River Basin. The very high, high, moderate, low, and very low susceptible zones of the combined approaches map cover an area of 7.10 km 2 (2.06%), 49.61 km 2 (14.41%), 130.21 km 2 (37.82%), 156.80 km 2 (45.54%), and 0.57 km 2 (0.17%) respectively (Table 3 ). Table 3 Comparison of flood susceptibility analysis results using AHP-MCDA, SVM-ML, and Combined Approach methods in the Ala River Basin, showing area and percentage for each susceptibility level. Susceptibility Level AHP-MCDA SVM-ML Combined Approach Very high 0.15 km² (0.04%) 8.10 km² (2.33%) 7.10 km² (2.06%) High 70.10 km² (20.36%) 33.01 km² (9.51%) 49.61 km² (14.41%) Moderate 209.48 km² (60.84%) 50.89 km² (14.66%) 130.21 km² (37.82%) Low 63.74 km² (18.51%) 108.89 km² (31.36%) 156.80 km² (45.54%) Very Low 0.83 km² (0.24%) 146.32 km² (42.14%) 0.57 km² (0.17%) 4.3 Validation and Performance Assessment To evaluate the accuracy and performance of the flood susceptibility mapping, we employed several validation criteria; precision, recall, f1-score, overall accuracy and ROC-AUC. The performance metrics for each model presented in Table 4 . The AHP-MCDA approach appears to have the least overall accuracy of 67%, the SVM-ML model having an overall accuracy of 79%, while the combined approach demonstrated an overall accuracy of 81%, with notably higher performance in identifying non-flood areas compared to flood-prone areas. To validate the generated flood susceptibility maps, we applied the ROC-AUC curve analysis. Results showed that the combined SVM-ML and AHP-MCDA map marginally outperformed both the individual SVM-ML and AHP-MCDA maps, although the performance difference between the combined map and the SVM-ML map was minimal (Fig. 8 ). Table 4 Evaluation metric scores of the flood susceptibility models Model Flood Status Precision Recall F1-Score Accuracy AHP-MCDA No Flood (0) 0.70 0.85 0.77 67% Flood (1) 0.55 0.33 0.42 SVM-ML No Flood (0) 0.84 0.85 0.85 79% Flood (1) 0.69 0.67 0.68 Combined No Flood (0) 0.86 0.89 0.87 81% Flood (1) 0.71 0.68 0.70 5.0 DISCUSSION The integration of GIS with AHP-MCDA and SVM-ML for flood susceptibility mapping in the Ala River basin divulges promising results, revealing difference in the two approaches. The AHP-MCDA approach incorporated expert knowledge and preferences, revealed a larger percentage of the study area in the moderate flood susceptibility zone. It identified downstream areas as most vulnerable, primarily due to low elevation levels and high-water accumulation. This aligns with findings from Komolafe et al. ( 2020 ) and Edamo et al. (2022), who emphasized the significance of topography and rainfall in flood occurrence. The use of AHP-MCDA for flood susceptibility mapping is known to be appropriate for regional studies while been dependent on the characteristics of the region (Komolafe et. al., 2020 ). However, it’s limitation is in the level of experience and knowledge of the expert. In contrast, the SVM-ML approach, leveraging historical flood occurrence data alongside the flood causative factors, highlighted upstream areas with significant built-up activities as highly vulnerable to flooding. The model showed that the central part of the basin, characterized by less built-up area and more vegetation, has moderate to low flood susceptibility. Interestingly, the downstream area remained highly susceptible due to low elevation and high flow accumulation. This pattern suggests that rainfall also plays a crucial role in the study area's flooding frequency, with built-up areas experiencing an initial surge in water flow before it moves downstream. Both methods demonstrated their value, with the SVM-ML model proving to be more accurate in this particular study area. The AHP-MCDA method's strength lies in its ability to incorporate expert knowledge and systematically prioritize factors influencing flood susceptibility. On the other hand, the SVM-ML model excels in handling high-dimensional datasets and complex, non-linear relationships between input variables and flood susceptibility. The AHP-MCDA enables stakeholders to systematically prioritize and evaluate factors influencing flood susceptibility, while SVM-ML enhances the predictive accuracy of susceptibility maps. The integration of these approaches generated a more robust flood susceptibility map, capitalizing on the strengths of both methodologies. It is worth to mention that elements such as buildings, farmlands, and human population were identified in the flood prone area of the basin during a survey in the area, which were also identified in the high and very high flood susceptible zones of the combined flood susceptibility map. The effect of heavy rainfall coupled with built-up which causes increase in water runoff at the upstream is identified to be the leading cause of flood susceptibility in the upstream of the basin. Areas such as Ayedun, Araromi, Alagbaka, Oba-Ile, etc. at the upstream are revealed to be most exposed to the risk of flood in the study area and recently are mostly feeling the impact (Fig. 6 ). Ibitoye et. al., ( 2020 ) reported that the increase in the frequency of flood occurrence accompanied by its severe damages in the Ala River basin over the years is as a result of the presence of buildings, infrastructures and increasing human population in Akure metropolis. At the downstream area, farmlands and settlements such as Ago-Owo, Ago-Dada, Olokuta, Ala-Ajagusin, etc. are exposed to the risk of flood. The severity of flood in the downstream would continue to increase rapidly if prompt caution and measures are not taken. 6.0 CONCLUSION This study successfully mapped the flood susceptibility of the Ala River basin in Ondo State by employing AHP-MCDA and SVM-ML approached, considering 19 flood causative factors. The AHP-MCDA results indicated that approximately 20% of the study area exhibited very high and high susceptibility to flooding, concentrated mainly in the downstream areas. The SVM-ML model, however, showed that about 12% of the area had very high and high flood susceptibility, with a focus on upstream areas which are Ayedun, Araromi, Oke-Ijebu, Alagbaka, and Oba-Ile areas, with significant development and less-developed downstream areas. The central parts of the basin, characterized by less built-up area and more vegetation, exhibits moderate to low flood susceptibility. In contrast, the downstream area remains highly susceptible due to its low elevation and high flow accumulation. This pattern indicates that rainfall plays a significant role in the study area's flooding frequency. During periods of high rainfall, built-up areas experience an initial surge in water flow before it moves downstream. Consequently, while elevation is a crucial factor in flood susceptibility mapping, rainfall interacts complexly with topography to influence flood risk. The combined AHP-MCDA and SVM-ML flood susceptibility map revealed the complementary nature of the two approaches, providing a comprehensive view of the area's flood occurrence patterns. Built-up infrastructure in the upstream emerged as a major contributing factor, while high flow rates from upstream and reduced channel capacity likely contribute to downstream flooding. Based on these findings, we recommend prioritizing floodplain management in populated upstream areas through more stringent zoning laws and promotion of green infrastructure. Upgrading and maintaining drainage canals, particularly in downstream areas such as Ala-Ajagbusi, Kajola, Apafa-Aiyede, and Ago-Goke settlements is crucial for effective floodwater management. Additionally, public awareness campaigns should be conducted to inform locals about flood risks and preparedness strategies. While this integrated approach provides valuable insights for flood risk management in the Ala River Basin, it's important to note that the study did not cover flood depth and velocity. Future research should consider incorporating hydraulic modeling to provide a more comprehensive analysis of flood susceptibility. This could enhance the accuracy of predictions and offer more detailed information for flood management strategies. In conclusion, this study demonstrates the effectiveness of combining GIS, AHP-MCDA, and SVM-ML techniques in flood susceptibility mapping. The methodology developed here can be adapted for similar studies in other regions, contributing to improved flood risk management strategies globally. As climate change continues to alter precipitation patterns and increase flood risks worldwide, such integrated approaches will become increasingly valuable for urban planning and disaster preparedness. Declarations Competing Interests: There are no relevant financial or non-financial interests to disclose. Funding: No funds, grants, or other support were received during the preparation of this manuscript. Authors Contribution: All authors contributed to the study's conception and design. Material preparation, data collection, and analysis were performed by Adeyemi Adedoyin Benson, Ismaila Racheal Opeyemi, and Komolafe Akinola Adesuji. The first draft of the manuscript was written by Adedoyin Benson Adeyemi and Ismaila Racheal Opeyemi while Komolafe Akinola Adesuji, Catherine Lilian Nakalembe, Adebowale Daniel Adebayo, and Enoch Oluwaferanmi Babayemi helped improve the first draft of the manuscript. References Ajin, R.S., R.R. Krishnamurthy, M. Jayaprakash and P.G. Vinod, 2013. Flood hazard assessment of Vamanapuram River Basin, Kerala, India: An approach using Remote Sensing and GIS techniques. Adv. Applied Sci. Res., 4: 263-274 Ajjur, S. B., & Mogheir, Y. K. (2020). Flood hazard mapping using a multi-criteria decision analysis and GIS (case study Gaza Governorate, Palestine). Arabian Journal of Geosciences , 13 (2). https://doi.org/10.1007/s12517-019-5024-6 Akukwe, T. I., & Ogbodo, C. (2015). Spatial analysis of vulnerability to flooding in Port Harcourt Metropolis, Nigeria. SAGE Open , 5 (1), 215824401557555. doi:10.1177/2158244015575558 Albert, A. A., Samson, A. A., Peter, O. O. and Olufunmilayo, A. O. An Assessment of the Socio Economic Impacts of Soil Erosion in South-Eastern Nigeria, Shaping the Change, XXIII FIG Congress Munich, Germany, pp. 12, 2006. Alimi, S. A., Andongma, T. W., Ogungbade, O., Senbore, S. S., Alepa, V. C., Akinlabi, O. J., … Muhammed, Q. O. (2022). Flood vulnerable zones mapping using geospatial techniques: Case study of osogbo metropolis, Nigeria. The Egyptian Journal of Remote Sensing and Space Science , 25 (3), 841–850. doi:10.1016/j.ejrs.2022.08.003 Alipour, A., Ahmadalipour, A., Abbaszadeh, P., & Moradkhani, H. (2020). Leveraging machine learning for predicting flash flood damage in the Southeast US. Environmental Research Letters , 15 (2), 024011. https://doi.org/10.1088/1748-9326/ab6edd Avand, M., Moradi, H. R., & Ramazanzadeh Lasboyee, M. (2021). Spatial prediction of future flood risk: An approach to the effects of climate change. Geosciences , 11 (1), 25. doi:10.3390/geosciences11010025 Awosika, L.F and R. Folorunsho, 2000. Integrated monitoring of flood and information. Dissemination for Flood Early Warnings in Nigeria, Nigeria Institute for Oceanography and Maritime Research VI Lagos. Bakare, G. O., Omosulu, S. B., & Famutimi, J. T. (2019). Geo-Mapping of Areas Vulnerable to Ala-River Basin Flood Disaster Risk in Akure, Ondo State, Nigeria . doi:10.20944/preprints201911.0184.v1 Balasubramanian, A.. (2017). Soil Erosion- Causes and Effects. 10.13140/RG.2.2.26247.39841. Bettis III, E. A., Gully Erosion of Western Iowa. The Lowa Department of Natural Resources, Iowa Geological and Water Survey, 2005. Birmah, M. N., Kigun, P. A., Alfred, Y. B., Majidadi, ST., & Surajo, L. A. (2021). Flood Assessment in Suleja Local Government Area, Niger State, Nigeria. International Journal of Research Publication and Reviews , 2 (3), 219–239. Cabrera, Jonathan & Lee, Han Soo. (2019). Flood-Prone Area Assessment Using GIS-Based Multi-Criteria Analysis: A Case Study in Davao Oriental, Philippines. Water. 11 . 2203. 10.3390/w11112203. Cea, L., Costabile, P., 2022. Flood risk in urban areas: modelling, management and adaptation to climate change. A review. Hydrology 9 (3), 50. doi:10.3390/hydrology9030050. Chen, J., A.A. Hill and L.D. Urbano, 2009. A GISbased model for urban flood inundation. J. Hydrol., 373: 184-192. DOI: 10.1016/j.jhydrol.2009.04.021 Danumah, J.H.; Odai, S.N.; Saley, B.M.; Szarzynski, J.; Thiel, M.; Kwaku, A.; Kouame, F.K.; Akpa, L.Y. (2016). Flood risk assessment and mapping in Abidjan district using multi-criteria analysis (AHP) model and geoinformation techniques, (cote d’ivoire). Geoenviron. Disasters, 3 (10). De Moel, H., Jongman, B., Kreibich, H., Merz, B., Penning-Rowsell, E. C., & Ward, P. J. (2015). Flood risk assessments at different spatial scales. Mitigation and Adaptation Strategies for Global Change , 20 (6), 865–890. https://doi.org/10.1007/s11027-015-9654-z Efraimidou, E., & Spiliotis, M. (2024). A GIS-Based Flood Risk Assessment Using the Decision-Making Trial and Evaluation Laboratory Approach at a Regional Scale. Environmental Processes , 11 (1). https://doi.org/10.1007/s40710-024-00683-w El-Swaify, S. (2013). Impacts of Land Use Change on Soil Erosion and Water Quality-A Case Study from Hawaii Impacts of Land Use Change on Soil Erosion and Water Quality-A Case Study from Hawaii. Graf, W. L., The Arroyo problem-palaehydrology and palaeohydraulics in the short term, in K. J. Gregory (ed). Background to palaeohydrology: A perspective (Chihester, Wiley), 279-302, 2003 Hilborn, D., Gully Erosion Control, Ontario Ministry of Agriculture, Food and Rural Affairs, 2005 Hill, J., Megier, J., & Mehl, W. (1995). Land degradation, Soil Erosion and Desertification Monitoring in Mediterranean Ecosystems. Remote Sensing Review, 12, 107-130. http://dx.doi.org/10.1080/02757259509532278 Huong, H.T.L. and A. Pathirana, 2013. Urbanization and climate change impacts on future urban flooding in Can Tho city, Vietnam. Hydrol. Earth Syst. Sci., 17: 379-394. DOI: 10.5194/hess-17-379-2013 Ibitoye, M.O., Komolafe, A. A., Adegboyega, S. A., and Adebola, A. O, 2020. Analysis of vulnerable urban properties within river Ala floodplain in Akure, Southwestern Nigeria. Spatial Information Research, 28(4):431–445 DOI 10.1007/s41324-019-00298-6. Ighile, E.H. Shirakawa, H.; Tanikawa, H. (2022). Application of GIS and Machine Learning to Predict Flood Areas in Nigeria. Sustainability 14 , 5039. https://doi.org/10.3390/su14095039 Ishaya, S., Ifatimehin, O. O., & Abaje, I. B. (2009). Mapping Flood Vulnerable Areas in a Developing Urban Centre of Nigeria. Journal of Sustainable Development in Africa , 11 (4). Komolafe, Akinola & Adegboyega, Suleiman & Akinluyi, Francis. (2015). A Review of Flood Risk Analysis in Nigeria. American journal of environmental sciences. 11. 157 -166. 10.3844/ajessp.2015.157.166. Komolafe, A. A., Srikantha H., and Avtar R., 2018. Development of generalized loss functions for rapid estimation of flood damages: a case study in Kelani River basin. Applied Geomatics, 10 (1): 13-30, Komolafe, A. A., Awe, B. S., Olorunfemi, I. E., & Oguntunde, P. G. (2020). Modelling flood-prone area and vulnerability using integration of multi-criteria analysis and HAND model in the Ogun River Basin, Nigeria. Hydrological Sciences Journal, 65(10), 1766–1783. https://doi.org/10.1080/02626667.2020.1764960 Luino, F. (2016) ‘Floods’, Encyclopedia of Earth Sciences Series, pp. 1–6. doi:10.1007/978-3-319-12127-7_126-1. Mosavi, A., Ozturk, P., & Chau, K. W. (2018). Flood Prediction Using Machine Learning Models: Literature Review. Water , 10 (11), 1536. https://doi.org/10.3390/w10111536 Olatona, O. O., Obiora-Okeke O. A., and Adewumi J. R. (2018). Mapping of flood risk zones in Ala River Basin Akure, Nigeria. American Journal of Engineering and Applied Sciences , 11 (1), 210–217. doi:10.3844/ajeassp.2018.210.217 Oyinloye, M.A. and Olamiju, O.I. (2011) ‘Flood risk mapping and vulnerability analysis using GIS: empirical evidences from New Town area, Ondo, Ondo State, Nigeria’, Int. J. Society Systems Science, Vol. 3, No. 3, pp.291–304. Pedersen, A.N., P.S. Mikkelsen and K. ArnbjergNielsen, 2012. Climate change-induced impacts on urban flood risk influenced by concurrent hazards. J. Flood Risk Manage., 5: 203-214. DOI: 10.1111/j.1753-318X.2012.01139.x Poesen, J., Nachtergaele, J., Verstraeten, G. and. Valentin, C., Gully erosion and environmental change: importance and research needs, Catena, 50 (2-4), pp. 91-133, 2003. P. Sayers, Y. L.i, G. Galloway, E. Penning-Rowsell, F. Shen, K. Wen, Y. Chen, and T. Le Quesne. 2013. Flood Risk Management: A Strategic Approach. Paris, UNESCO. Rahman, A. J. A., & Ramli, N. A. (2024). Flood Prediction Using Artificial Neural Networks: A Case Study in Temerloh, Pahang. Qeios . https://doi.org/10.32388/tuz29y Schanze, J., 2006. Flood Risk Management- a Basic Framework. In: Flood Risk Management-Hazards, Vulnerability and Mitigation Measures, Schanze, J., E. Zeman and J. Marsalek (Eds.), Springer, pp: 149-167. Scheuer, S. and V. Meyer, 2007. FloodCalc. Software tool for the calculation of multicriteria flood damage and risk maps Sowmya, K., C.M. John and N.K. Shrivasthava, 2015. Urban flood vulnerability zoning of Cochin City, southwest coast of India, using remote sensing and GIS. Nat. Hazards, 75: 1271-1286. DOI: 10.1007/s11069-014-1372-4 Suleiman, Y.M, M.B. Matazu, A.A. Davids and M.C. Mozie, 2014. The application of geospatial techniques in flood risk and vulnerability mapping for disaster management at Lokoja, Kogi State, Nigeria. J. Environ. Earth Sci., 4: 54-61. Suresh, R. (2012) Soil and water conservation engineering. Delhi: Standard Publishers Distributors. UNISDR, 2015. Disaster Risk Reduction in the Post-2015 Development Agenda: Transforming Our World: The 2030 Agenda fpr Sustainable Development., in UNISDR, ed.: Geneva, Switzerland, UNISDR. Available at: https://www.unisdr.org/we/inform/publications/45417. UNOCHA. Nigeria: Humanitarian Response Plan; United Nations Office for the Coordination of Humanitarian Affairs: Abuja, Nigeria, 2018 Zehra, N. (2020). Prediction Analysis of Floods Using Machine Learning Algorithms (NARX & SVM). International Journal of Sciences: Basic and Applied Research , 49 (2), 24–34. Supplementary Files SUPPLEMENTARYRESULTS.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4863685","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":336793719,"identity":"f28f4ce0-9daa-4e69-8b53-d150f9d61233","order_by":0,"name":"Adedoyin Benson Adeyemi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8klEQVRIiWNgGAWjYDACCTBiYGBj7//4AEjz8BGvheeAsQFICxvRWhgkEswgegnp4J/d/PDGxzabaD6JhLTKrzl2MmwMzA8f3cBnyZ1jxpYz29Jy23geHLstuy0Z6DA2Y+McfNbcSDCT5m07nNvGnth2W3IbM1ALD5s0Pi3yN9K/Sf9t+5/bxpDMViy5rZ6wFoMbOWbSjG0Hcts40tgYP247TFiL4Z0zxZY955KBfjnDLM247TgPGzMBv8jdbt9440eZXe789h7Gjz+3Vdvzszc/fIzX+yDACI0LZh4wSUg5GPyBav1BlOpRMApGwSgYaQAAjdxIJLpS2kgAAAAASUVORK5CYII=","orcid":"","institution":"Federal University of Technology Akure","correspondingAuthor":true,"prefix":"","firstName":"Adedoyin","middleName":"Benson","lastName":"Adeyemi","suffix":""},{"id":336793720,"identity":"f0090e43-5dfa-4200-97a3-f36ef3ec2ed9","order_by":1,"name":"Akinola Adesuji Komolafe","email":"","orcid":"","institution":"Federal University of Technology Akure","correspondingAuthor":false,"prefix":"","firstName":"Akinola","middleName":"Adesuji","lastName":"Komolafe","suffix":""},{"id":336793721,"identity":"f45892fe-e930-42dd-83c2-286557223710","order_by":2,"name":"Catherine Lilian Nakalembe","email":"","orcid":"","institution":"University of Maryland at College Park: University of Maryland","correspondingAuthor":false,"prefix":"","firstName":"Catherine","middleName":"Lilian","lastName":"Nakalembe","suffix":""},{"id":336793722,"identity":"3dc24db3-f696-4e7e-94fe-643dd3a77f61","order_by":3,"name":"Racheal Opeyemi Ismaila","email":"","orcid":"","institution":"Federal University of Technology Akure","correspondingAuthor":false,"prefix":"","firstName":"Racheal","middleName":"Opeyemi","lastName":"Ismaila","suffix":""},{"id":336793723,"identity":"7ef1fb6b-921c-4d35-acd1-cb82752004f8","order_by":4,"name":"Adebowale Daniel Adebayo","email":"","orcid":"","institution":"University of Maryland at College Park: University of Maryland","correspondingAuthor":false,"prefix":"","firstName":"Adebowale","middleName":"Daniel","lastName":"Adebayo","suffix":""},{"id":336793724,"identity":"b44daa76-58ec-42a3-8e9c-9d52a417ebf2","order_by":5,"name":"Oluwaferanmi Enoch Babayemi","email":"","orcid":"","institution":"Federal University of Technology Akure","correspondingAuthor":false,"prefix":"","firstName":"Oluwaferanmi","middleName":"Enoch","lastName":"Babayemi","suffix":""}],"badges":[],"createdAt":"2024-08-05 17:32:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4863685/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4863685/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":63762597,"identity":"90b89ab8-0d98-426d-b2bc-60f9f0d95f98","added_by":"auto","created_at":"2024-09-02 06:36:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2715375,"visible":true,"origin":"","legend":"\u003cp\u003eMap of the Ala River Basin in Akure, Ondo State, Nigeria\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/8bba8cfbc87849b1c9901c7b.png"},{"id":63763212,"identity":"e8830238-2e8f-4ae9-8d66-b669181fd4f9","added_by":"auto","created_at":"2024-09-02 06:44:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3029587,"visible":true,"origin":"","legend":"\u003cp\u003eFlood Impacts in the Ala River Basin, (a) Flooded road infrastructure preventing movements (source: \u003ca href=\"https://www.premiumtimesng.com/news/more-news/606663-floods-ravage-ondo-communities-after-heavy-rains.html\"\u003ehttps://www.premiumtimesng.com/news/more-news/606663-floods-ravage-ondo-communities-after-heavy-rains.html\u003c/a\u003e), (b) Flooded residential area at Ayede-Ogbese, Akure (source: \u003ca href=\"https://dailypost.ng/2019/09/04/flood-ravages-communities-ondo-residents-rendered-homeless-photos/\"\u003ehttps://dailypost.ng/2019/09/04/flood-ravages-communities-ondo-residents-rendered-homeless-photos/\u003c/a\u003e), (c) Abandoned old St. Stephen Church on the Ala flood plain area, Akure (picture taken during dry season), (d) Submerged abandoned building on the Ala flood plain area, Akure (picture taken during dry season).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/97d75f324c863b3f145e72ac.png"},{"id":63762602,"identity":"63b177e3-c994-4fae-939a-f1af6e981d8c","added_by":"auto","created_at":"2024-09-02 06:36:52","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":204882,"visible":true,"origin":"","legend":"\u003cp\u003eMethodology Flowchart\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/33fe7c59c35d6abceb0e9394.png"},{"id":63762604,"identity":"b47b3a9f-ab47-4c9f-b455-3f47e93597f0","added_by":"auto","created_at":"2024-09-02 06:36:52","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":5321354,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of flood conditioning factors in the Ala River Basin, Akure, Nigeria: (a) Elevation, (b) Slope, (c) Aspect, (d) Plan Curvature, (e) Profile Curvature, (f) Roughness, (g) Flow Direction, (h) Flow Accumulation, (i) Drainage Density, (j) Distance from Rivers, (k) Topographic Wetness Index (TWI), (l) Sediment Transport Index (STI), (m) Stream Power Index (SPI), (n) Soil Types, (o) Geology, (p) Normalized Difference Moisture Index (NDMI), (q) Normalized Difference Vegetation Index (NDVI), (r) Land Use/Land Cover (LULC), and (s) Annual Rainfall.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/edfcbd51224f1d4327ffcb5e.png"},{"id":63762601,"identity":"c57b0214-aa5c-4a05-bc83-4492c2439a1e","added_by":"auto","created_at":"2024-09-02 06:36:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":675854,"visible":true,"origin":"","legend":"\u003cp\u003eFlood Susceptibility Map of the study area (a) AHP-MCDA (b) SVM-ML\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/e59a4a8ac2baf0dd2271bb89.png"},{"id":63763214,"identity":"692c41a8-cf1b-4b6f-a858-6fcce4220840","added_by":"auto","created_at":"2024-09-02 06:44:51","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":368113,"visible":true,"origin":"","legend":"\u003cp\u003eCombined AHP-MCDA and SVM-ML Flood Susceptibility Map of the study area\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/11166c2d60dd9a178c6d1a8c.png"},{"id":63763213,"identity":"7cb42697-b551-4df2-8b55-1da0c81770a7","added_by":"auto","created_at":"2024-09-02 06:44:51","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":33136,"visible":true,"origin":"","legend":"\u003cp\u003eChart showing the percentage of AHP-MCDA, SVM-ML and Combined Approach Flood Susceptible Zones in the study area\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/1cc0684440b426c33d46a01d.png"},{"id":63762596,"identity":"31dbd009-eed6-4b76-a897-79f23420f695","added_by":"auto","created_at":"2024-09-02 06:36:51","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":58272,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver Operating Characteristic (ROC) curves comparing the performance of AHP-MCDA, SVM-ML, and combined AHP-MCDA/SVM-ML flood susceptibility models for the Ala River Basin, Akure, Nigeria.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/09d43fdb52b7acb033201fbf.png"},{"id":65302597,"identity":"19c0581e-6fd4-4541-a588-71fcb597e780","added_by":"auto","created_at":"2024-09-25 23:56:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":15654762,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/a4714f08-5ca7-4738-8e9f-98437c2ef5b3.pdf"},{"id":63762599,"identity":"ccfe4952-cf2d-4c92-9331-19b1f7cd1c9c","added_by":"auto","created_at":"2024-09-02 06:36:51","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":43340,"visible":true,"origin":"","legend":"","description":"","filename":"SUPPLEMENTARYRESULTS.docx","url":"https://assets-eu.researchsquare.com/files/rs-4863685/v1/e57c2290754cfb0ca89619f5.docx"}],"financialInterests":"","formattedTitle":"Integrated GIS-Based MCDA and Machine Learning Techniques in Flood Susceptibility Mapping in Ala River Basin, Nigeria","fulltext":[{"header":"1.0 INTRODUCTION","content":"\u003cp\u003eFloods are natural hazards whose impacts can be intensified by uncontrolled urbanization and land-use changes. When severe flooding occurs in developed areas, it can cause extensive damage to residential structures, industrial facilities, public infrastructure, agricultural land, and crops, resulting in substantial economic losses and potential losses of human life (Olatona et. al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). According to the World Bank, over 1.81\u0026nbsp;billion people worldwide are directly exposed to flood depths exceeding 0.15 meters in a 1-in-100-year flood event as of 2022. Globally, flooding causes more human casualties and property damage than any other natural hazard, making it one of the most severe natural risks (Ologunorisa, 2001; Alcira and Martha, 1991; Ishaya et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Nigeria, numerous regions have experienced severe flooding, displacing millions, disrupting businesses, contaminating water supplies, and increasing waterborne disease risk (Etuonovbe, 2011). UNICEF reported in 2022, that due to severe flooding, over 2.5\u0026nbsp;million Nigerians, 60% of whom are children require humanitarian assistance, and are at heightened risk of malnutrition, death due to drowning, and waterborne illnesses. Several studies have shown that floods in urban centers in developing countries are a result of inadequate drainage systems, inefficient storm sewerage networks, and blockages in existing drainage infrastructure (Vanneuville \u003cem\u003eet al.\u003c/em\u003e, 2011; Abhas \u003cem\u003eet al.\u003c/em\u003e, 2012; Bakare et. al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). These urban floods are among the most frequent and characteristic natural disasters in cities, with consequential effects including significant human social disruptions and extensive damage to infrastructure (Bakare et. al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFloods manifest in various forms, including river, coastal, urban, flash, and groundwater floods and an area's vulnerability to flood events is influenced by physical factors such as topography, land cover, climate change, rainfall duration and frequency, and soil permeability (Thakur \u003cem\u003eet al.\u003c/em\u003e, 2011; Bakara \u003cem\u003eet al.\u003c/em\u003e, 2019). Climate change is the primary driver increasing flood risk, altering the changing behavior of extreme rainfall events (Pederson \u003cem\u003eet. al.\u003c/em\u003e, 2012; Olatona et. al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Increased precipitation, higher surface runoff, rising global temperatures, and rapid urbanization all contribute to flooding (Danumah et. al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Urban flooding is intensified by deforestation, inadequate urban planning, and rapid urbanization, worsened by buildings in wetlands and swamps that serve as critical flood buffers. Given these challenges, accurate flood risk assessment is crucial, especially in vulnerable communities, to inform effective planning, mitigation, adaptation, and response strategies. This assessment involves understanding of flood likelihood and potential consequences, considering factors such as hazard, exposure, and vulnerability (De Moel et. al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFlood impacts are particularly severe when extreme rainfall affect poor communities in flood-prone areas, such as areas along the Ala River basin in Akure, Ondo State, Nigeria. These areas often have poorly constructed buildings and have limited adaptive capacity (Olatona et. al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Komolafe et. al. (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) found that high levels of vulnerability and a lack of coping mechanisms among the Nigerian population are the main factors exacerbating flood impacts and losses. With the potential increase in the intensity of floods and its associated risks due to climate change and urbanization, more importance will be placed on effective flood risk management (UNISDR, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Nigeria, traditional approaches to flood management at various levels are basically river channelization and construction of river embankment, which often yield less positive outcome whenever extreme events occur (Ibitoye et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, recent development in flood risk management proposes the use of effective flood risk management, which places premium on non-structural measures (Komolafe et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). One of the flood risk management elements is the mapping and identifying spatial location, distribution and severity of flood hazards; this is very crucial in the identification of element at risk and the quantification of the potential flood risk in the study area (Br\u0026eacute;mond \u003cem\u003eet al.\u003c/em\u003e, 2013; Merz \u003cem\u003eet al.\u003c/em\u003e, 2010; UNISDR, 2023). Effective flood risk management therefore requires identifying flood-prone areas and prioritizing interventions in these high-risk zones (Sayers \u003cem\u003eet. al.\u003c/em\u003e, 2013). This underscores the importance of developing precise flood susceptibility maps, which delineate areas with varying degrees of flood risk based on a range of physical, environmental, and socioeconomic factors (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). These maps can serve as critical tools for decision-makers, planners, and emergency responders, enabling them to allocate resources, implement risk reduction measures, and enhance community resilience to flooding (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGeographic Information Systems (GIS), Remote Sensing (RS), and Machine Learning (ML) offer powerful tools for addressing flood risk in large, diverse countries like Nigeria (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). With a land area of 923,768 square kilometers and varied topography ranging from coastal regions to mountainous areas, Nigeria presents complex challenges for flood prediction and management. GIS provides a robust platform for spatial analysis, enabling the integration of diverse physical, environmental, and socio-economic factors that contribute to flood risk across Nigeria's vast landscape (Efraimidou \u0026amp; Spiliotis, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Remote sensing technologies allow for the continuous monitoring of large areas, providing up-to-date information on land use changes, rainfall patterns, and river dynamics that is crucial for flood risk assessment. ML models, including Artificial Neural Networks (ANN) (Rahman \u0026amp; Ramli, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), Support Vector Machines (SVM) (Zehra, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and Random Forest (RF) (Alipour et. al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), have shown effectiveness in predicting flood occurrences. These models can process the complex, large-scale datasets characteristic of a country like Nigeria, leveraging historical flood data and multiple input variables to generate accurate predictions (Mosavi et. al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe integration of these technologies is particularly valuable for Nigeria, where flooding challenges are increasing and vary significantly across regions. This research aims to establish a method combining GIS techniques with the Analytic Hierarchy Process (AHP)-based Multi-Criteria Decision Analysis (MCDA) and Machine Learning (ML) model to predict and map flood-susceptible areas in the Ala River Basin of Akure, Ondo State. While focused on a specific region, this approach has the potential for broader application across Nigeria's diverse landscapes.\u003c/p\u003e"},{"header":"2.0 STUDY AREA","content":"\u003cp\u003eAla River basin flows along the Akure metropolis in Ondo State, Nigeria (Fig.\u0026nbsp;1). The basin extends geographically between 5\u0026deg; 9\u0026rsquo; 0\u0026rdquo; E, 7\u0026deg; 18\u0026rsquo; 0\u0026rdquo; N and 5\u0026deg; 27\u0026rsquo; 0\u0026rdquo; E, 7\u0026deg; 6\u0026rsquo; 0\u0026rdquo; N, covering an area of approximately 348 km\u003csup\u003e2\u003c/sup\u003e at a mean elevation of about 302 m. The Ala River, which has the most geographical impact, originates in the northwest highlands at Ipinsa and flows southeast through densely populated parts of Akure city. Several tributaries, including Eleye, Ipalefa, Alagbaka, and Omiyeye, join Ala or Elegbin rivers at various points.\u003c/p\u003e \u003cp\u003eAccording to the K\u0026ouml;ppen classification scheme, the area falls within the Aw climate type, characterized by distinct wet and dry seasons. The driest month typically has precipitation of less than 60mm. This seasonal variation in rainfall, combined with the basin's topography and urbanization, contributes significantly to its flood susceptibility.\u003c/p\u003e \u003cp\u003eThe Ala River basin features diverse geomorphological landscapes, spurs, saddles, valleys, and river channels. Geologically, it is part of Nigeria's southwest basement complex. The major rock types in the area are Chanockite, Migmatite gneiss, Quartzite, and Biotite gneiss (Olatona et. al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). These Precambrian rocks have undergone tectonic activities, resulting in fracturing, jointing, and cracking. These geological characteristics can affect soil permeability and drainage patterns, potentially impacting flood behavior and risk in the region.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe Ala River basin's susceptibility to floods is particularly high due to its topography, climate, and rapid urbanization. The densely populated areas of Akure city, situated along the river's path, are especially vulnerable. During the wet season, heavy rainfall combined with the basin's limited capacity to absorb and channel water efficiently leads to frequent flooding events. These floods often result in significant economic losses, displacement of residents, and damage to infrastructure. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e illustrates the severe impacts of flooding in the Ala River Basin. The images show inundated residential areas, damaged roads, and disrupted economic activities following a major flood event in 2023. This visual representation underscores the urgent need for effective flood management strategies in the region.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"3.0 MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Datasets\u003c/h2\u003e \u003cp\u003eThe study employed a variety of geospatial and environmental data sets to assess flood susceptibility in the Ala River Basin. These data sets, their resolutions, and their sources are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The selection of these specific data types was based on their relevance to flood risk factors and their availability for the study area. The integration of these diverse data sets allows for a comprehensive analysis of the physical, environmental, and historical factors contributing to flood susceptibility in the region.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eData Sets Used in Flood Susceptibility Analysis of the Ala River Basin\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eS/n\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eData\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eResolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eALOS PALSAR DEM Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12.5m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAlaska Satellite Facility Website\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLandsat-9 OLI/TIRS Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSGS Earth Explorer Website\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClimate Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30secs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWorldClim Website\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSoil Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFAO Website\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGeology Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNigeria Geological Survey Website\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHistorical Flood Inventory Data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFieldwork and other Secondary sources\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eData processing involved selecting key factors contributing to flooding in the study area (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). These factor\u0026rsquo;s values were classified and given ratings with respect to their importance (Table \u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003eS.1\u003c/span\u003e). Using the Digital Elevation Model (DEM), we generated hydrologic and topographic factor layers including elevation, slope, aspect, curvature, topographic wetness index (TWI), distance from rivers, and drainage density. Soil and geology map layers were derived from their respective downloaded datasets. Landsat satellite imagery was used to produce environmental factor layers, including land use/land cover, normalized difference vegetation index (NDVI), and normalized difference moisture index (NDMI). A rainfall map layer was created using data from the WorldClim database.\u003c/p\u003e \u003cp\u003eAll thematic maps were produced and processed using ArcGIS tools. To standardize the analysis, each thematic map was reclassified into five categories based on the factor's contribution to flood susceptibility. This reclassification allowed for consistent comparison and integration of the various factors in the subsequent flood risk assessment.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSummary of Flood Susceptibility Factors Relevant to this Study\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFactor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRange/Classes\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSignificance for Flooding\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHeight above a fixed reference point (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e250m \u0026minus;\u0026thinsp;552m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLower elevations are more flood-prone.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDegree of inclination of the surface (Edamo \u003cem\u003eet. al.\u003c/em\u003e 2022).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026deg; \u0026minus;\u0026thinsp;60\u0026deg;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLower slopes increase flood likelihood.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDirection of the slope face (Cea and Costabile, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-1\u0026deg; \u0026minus;\u0026thinsp;360\u0026deg;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAffects the water flow direction.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCurvature (Plan and Profile)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSlope of the slope of a surface (Longley \u003cem\u003eet. al.\u003c/em\u003e, 2011).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePlan: -6 to 11;\u003c/p\u003e \u003cp\u003eProfile: -7 to 6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eConcave areas are more flood-prone.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRoughness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVariation in elevation between pixels.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.111m \u0026ndash; 0.889m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher roughness may slow water flow.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFlow Direction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDirection of water flow for each cell (Edamo \u003cem\u003eet. al.\u003c/em\u003e, 2022).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1, 2, 4, 8, 16, 32, 64, 128\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDetermines water movement.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFlow Accumulation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of cells draining into each cell.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;2,190,524\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher flow accumulation values indicate flood-prone areas.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrainage Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDensity of stream network (Avand et. al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;3.16 km\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher drainage density increases flood risk.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDistance from River\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eProximity to water bodies.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;1,374 m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCloser areas are more flood-prone (Liu \u003cem\u003eet. al.\u003c/em\u003e, 2021; Edamo \u003cem\u003eet. al.\u003c/em\u003e, 2022).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTWI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTopographic Wetness Index indicates the amount of water present in an area. It calculates the probability of downward water flow due to gravity (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.9\u0026ndash;25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher values indicate higher flood risk.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSTI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSediment Transport Index describes the particles in water moving due to water flow.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;22,366\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher values may indicate higher flood risk (Edamo \u003cem\u003eet. al.\u003c/em\u003e, 2022).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSPI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStream Power Index calculates the erosive power of flowing water (Ighile et. al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;6,656,820.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher values indicate higher erosive power, resulting in higher flood risk.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoil\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSoil types affect water absorption and runoff based on soil characteristics (Edamo \u003cem\u003eet. al.\u003c/em\u003e, 2022).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLixisols, Acrisols, Fluvisols\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSoil types with lower water absorption result in to increase in runoff.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGeology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRock types affect water permeability which determines the rate of water infiltration into the sub-zones.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGranite gneiss, Porphyritic granite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLow permeable rocks increase flooding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNDMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNormalized Difference Moisture Index determine moisture content of vegetation (Mahdizadeh \u0026amp; Perez, 2022).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.19\u0026ndash;0.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher values indicate more moisture resulting to flood risk increase.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNDVI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNormalized Difference Vegetation Index determine the health strength of vegetation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.04\u0026ndash;0.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher values indicate more vegetation resulting to flood risk increase.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand Use/Cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTypes of land use and land cover shows the relationship between human activities, natural environment and floods (Mojaddadi \u003cem\u003eet. al.\u003c/em\u003e, 2017).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWaterbody, Vegetation, Built-up, Bareland/Cropland, and Outcrop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVegetation reduces flood risk. Built-up areas increase flood risk. Bareland and impervious surfaces also increase flood risk because of increased runoff.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRainfall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAnnual precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e165.44mm \u0026minus;\u0026thinsp;189.09mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigher rainfall increases flood risk\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Methods\u003c/h2\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e3.2.1 Multi-Criteria Decision Analysis (MCDA)- Analytic Hierarchy Process (AHP)\u003c/h2\u003e \u003cp\u003eA hierarchical structure based on the Analytic Hierarchy Process (AHP) model was developed for the selected criteria. This was followed by a pair-wise comparison matrix analysis of the selected influencing variables. Priority values were assigned, and relative weights for each variable were determined. The ensure robustness of the process, the consistency of evaluations and decisions was measured. Conclusions regarding priority variables were then synthesized to identify flood vulnerability zones, following established methodologies (Vignesh \u003cem\u003eet al.\u003c/em\u003e, 2020).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFlood influencing factors were compared using Multi-Criteria Decision Making (MCDM) to compute relative weights. An MCDM-based AHP model was employed to quantify the importance of the selected flood-inducing factors, determining their potential to cause flood hazards. Factors were categorized into sub-classes with ranks 1 to 5 based on Saaty's preference scale (Table \u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003eS.3\u003c/span\u003e). The factors\u0026rsquo; ranks for the AHP-MCDA analysis based on their priority weight include: rainfall (14.5%), elevation (12.8%), distance from river (11.4%), flow accumulation (9.8%), TWI (8.1%), drainage density (7.2%), slope (5.8%), NDVI (5.2%), NDMI (4.4%), LULC (4.0%), soil (3.6%), flow direction (2.6%), roughness (2.3%), SPI (1.9%), STI (1.6%), aspect (1.4%), plan curvature (1.2%), profile curvature (1.1%), and geology (1.0%) (Table \u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003eS.3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThematic layers for each factor, indicating vulnerability levels from high, to very low, were integrated using the Weighted Linear Combination (WLC) in ArcGIS to produce the AHP-MCDA flood susceptibility zone map.\u003c/p\u003e \u003cp\u003eA Consistency ratio (CR) was calculated to assess pairwise comparison consistency defined as the consistency index (CI) to a random-like matrix (RI) (Table \u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003eS.2\u003c/span\u003e), depending on the number of influencing factors considered (Saaty, 1990; Mahmoud and Gan, 2018; Komolafe et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe Consistency Index (CI) is calculated using the formula:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:CI=\\:\\frac{{\\lambda\\:}_{-n}}{n-1}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere \u003cem\u003eλ\u003c/em\u003e is the average of the consistency vector and \u003cem\u003en\u003c/em\u003e is the number of factors.\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:CR=\\frac{CI}{RI}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere \u003cem\u003eCI\u003c/em\u003e is the Consistency Index and \u003cem\u003eRI\u003c/em\u003e is the random-like matrix.\u003c/p\u003e \u003cp\u003eThe accepted limit of the CR is 0.1 or 10%. This threshold is considered sufficient to detect the impact of each criterion. In this study, the calculated CR for pairwise comparison analysis was 0.018 (1.8%), well within the accepted limit.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e3.2.2 Support Vector Machines (SVM) Machine Learning Model Analysis\u003c/h2\u003e \u003cp\u003eThe SVM model was used to predict the flood-prone zones in the study area. SVM, an effective supervised learning technique for regression and classification problems, operates using the structural risk minimization rule and statistical learning theory (Mosavi et.al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor flood inventory data was split 70:30 into training datasets and testing datasets. The model was evaluated using the precision, recall, f1-score, overall accuracy, and ROC-AUC curve. The SVM flood susceptibility map was produced, classifying the area into five flood risk levels (very high, high, moderate, low, and very low) using the standard deviation break method.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e3.2.3 Multicollinearity Investigation\u003c/h2\u003e \u003cp\u003eMulticollinearity, the high correlation between two or more independent variables can lead to interpretation problems and reduce model accuracy. This study used Pearson's correlation coefficients to assess multicollinearity. The coefficients range from \u0026minus;\u0026thinsp;1 to 1, with 1 indicating perfect positive correlation, -1 perfect negative correlation, and 0 no correlation. The multicollinearity analysis for the SVM-ML revealed high correlations (\u0026gt;\u0026thinsp;0.8) between NDVI-NDMI (0.89) and Flow Accumulation-SPI (0.97). NDMI and Flow Accumulation were removed to improve prediction accuracy (Table \u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003eS.4\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e3.2.4 AHP-MCDA and SVM-ML Integration Analysis\u003c/h2\u003e \u003cp\u003eThe AHP-MCDA and SVM-ML approaches were integrated using weighted overlay spatial analysis to a comprehensive and reliable flood susceptibility map. The area extent and percentage over each susceptibility zone were calculated.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"4.0 RESULTS","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003e4.1 Flood Causative Factors\u003c/h2\u003e\n \u003cp\u003eNineteen potential flood-triggering factors were considered and categorized into 5 susceptibility classes (Table \u003cspan class=\"InternalRef\"\u003eS.1\u003c/span\u003e), except for geology (2 classes: porphyritic granite and granite gneiss) and soil (3 classes: acrisol, fluvisols, and lixisols). Elevation, being one of the important factors that causes flood occurrence as shown in Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ea reveals the south eastern part of the study area to be at lower elevation (i.e. 250m) proving it to be more vulnerable to flooding compare to the northern part and a small section of the southern part, having higher elevation (i.e. 552m). Similar to elevation, the area with decreasing slope in the study area is more prone to flooding (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb). Aspect which affects soil humidity and flow direction of water make flood mapping more accurate. The aspect map of the study area is divided into 5 classes: (-1\u0026ndash;71.05), (71.05\u0026ndash;143.11), (143.11\u0026ndash;215.16), (215.16\u0026ndash;287.22), (287.22\u0026ndash;359.27) indicating the very high, high, moderate, low, very low vulnerable to flood respectively. The two curvatures (plan and profile) of the study area (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ed and \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ee) ranges between \u0026minus;\u0026thinsp;6 to 11 values and \u0026minus;\u0026thinsp;7 to 6 values respectively, with the least curvature value more vulnerable to flooding (Table \u003cspan class=\"InternalRef\"\u003eS.1\u003c/span\u003e). Unlike the curvatures, the higher the roughness value in the study area, the more prone it is to the flooding. The flow accumulation map (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eh) which is derived from the flow direction map (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eg) is divided into 5 classes, with the highest flow accumulation being most vulnerable to flooding. The drainage density and distance from the river both reveal the measure of proximity to the river in the study area, though the drainage density shows the density levels of the river flow (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ei and \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ej). The TWI, STI, and SPI are divided into 5 classes (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ek, \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003el, and \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003em respectively), with their class with the highest values representing areas most vulnerable to flooding, based on the fact that increase in wetness and rate of particles transport in water flow in an area infers increase in vulnerability to flood. The study area\u0026rsquo;s soil classification indicated on the soil map (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003en) are 3 types: lixisols, acrisols, and fluvisols, with the lixisols covering over 80% of the study area. The geology map shows 2 main rock types which are granite gneiss and porphyritic granite (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eo). The NDMI (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ep) and NDVI (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eq) values typically ranges between \u0026minus;\u0026thinsp;1 to 1. From both the NDMI and NDVI maps, its observed that the moisture level is majorly high at the middle of the basin due to high presence of vegetation in the area, while there is less vegetation at the upper portion of the basin. The LULC classification of the study area was into 5 classes including vegetation, built-up, bare ground/cropland, outcrop, and water body (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003er). The vegetation covering a larger area extent of 36.5%, the bare ground/cropland covering 26.1%, the built-up 20.4%, the outcrop 16.9% and the water body covering 0.0018%. The Ala River and its tributaries serve as the major water bodies across the study area. The rainfall data of the study area as shown on the map layer ranges between 165.44 to 189.09 rainfall values (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003es). The spatial distribution of the annual rainfall was prepared using kriging interpolation method, and classified into 5 classes. The lowest values class to the highest values class waving from the northern part of the study area to the southern part. Area with the highest rainfall indicate the most prone area to flooding.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003e4.2 Flood Susceptibility Maps\u003c/h2\u003e\n \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.1 AHP-MCDA Flood Susceptibility Map\u003c/h2\u003e\n \u003cp\u003eThe result from the AHP-MCDA mapping reveal that majority of the Ala River basin falls within the moderate flood susceptibility zone (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea). While the northern part which is the upstream is identified to be less susceptible to flood, the southern part of the basin which is the downstream is majorly identified to be highly susceptible to flooding. The calculated area covered by each class: very high, high, moderate, low, and very low as represented on the map include 0.15 km\u0026sup2; (0.04%), 70.10 km\u0026sup2; (20.36%), 209.48 km\u0026sup2; (60.84%), 63.74 km\u0026sup2; (18.51%), and 0.83 km\u0026sup2; (0.24%) respectively (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.2 SVM-ML Flood Susceptibility Map\u003c/h2\u003e\n \u003cp\u003eFigures \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eb show flood susceptibility map from the SVM-ML analysis. It can be deduced from the map that the SVM-ML model better uncovered the upstream of the river which is the northern part of the basin is also prone to flooding as much as the downstream. The very high and high susceptible zone covers an area extent of 8.10 km\u003csup\u003e2\u003c/sup\u003e (2.33%) and 33.01 km\u003csup\u003e2\u003c/sup\u003e (9.51%) respectively. The moderate, low, and very low susceptible zones cover an area extent of 50.89 km\u003csup\u003e2\u003c/sup\u003e (14.66%), 108.89 km\u003csup\u003e2\u003c/sup\u003e (31.36%), and 146.32 km\u003csup\u003e2\u003c/sup\u003e (42.14%) respectively as shown in Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\n \u003ch2\u003e4.2.3 Combined Approach Flood Susceptibility Map\u003c/h2\u003e\n \u003cp\u003eThe map from the integration of the AHP-MCDA and SVM-ML approaches expressed an improved map of the flood susceptibility zones of the study area (Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e). The map identified majorly the southern part (downstream) of the basin as most flood-prone, even as the northern part (upstream) is not ignored. The settlements situated within the highly susceptible zones of the basin are Ago-Owo, Ago-Dada, Agogboro, Olokuta, Olobi, and Iluabo, at the downstream. While at the upstream, Oba-Ile, Alagbaka, Odo-Ijoka, Isolo, Tuyi street, Araromi, and Ayedun street etc. (Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e). Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e compares the flood susceptibility analysis results using AHP-MCDA, SVM-ML, and Combined Approach methods in the Ala River Basin. The very high, high, moderate, low, and very low susceptible zones of the combined approaches map cover an area of 7.10 km\u003csup\u003e2\u003c/sup\u003e (2.06%), 49.61 km\u003csup\u003e2\u003c/sup\u003e (14.41%), 130.21 km\u003csup\u003e2\u003c/sup\u003e (37.82%), 156.80 km\u003csup\u003e2\u003c/sup\u003e (45.54%), and 0.57 km\u003csup\u003e2\u003c/sup\u003e (0.17%) respectively (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eComparison of flood susceptibility analysis results using AHP-MCDA, SVM-ML, and Combined Approach methods in the Ala River Basin, showing area and percentage for each susceptibility level.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSusceptibility Level\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAHP-MCDA\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSVM-ML\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCombined Approach\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVery high\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.15 km\u0026sup2; (0.04%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8.10 km\u0026sup2; (2.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7.10 km\u0026sup2; (2.06%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHigh\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e70.10 km\u0026sup2; (20.36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e33.01 km\u0026sup2; (9.51%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e49.61 km\u0026sup2; (14.41%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eModerate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e209.48 km\u0026sup2; (60.84%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e50.89 km\u0026sup2; (14.66%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e130.21 km\u0026sup2; (37.82%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e63.74 km\u0026sup2; (18.51%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e108.89 km\u0026sup2; (31.36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e156.80 km\u0026sup2; (45.54%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVery Low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.83 km\u0026sup2; (0.24%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e146.32 km\u0026sup2; (42.14%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.57 km\u0026sup2; (0.17%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003e4.3 Validation and Performance Assessment\u003c/h2\u003e\n \u003cp\u003eTo evaluate the accuracy and performance of the flood susceptibility mapping, we employed several validation criteria; precision, recall, f1-score, overall accuracy and ROC-AUC. The performance metrics for each model presented in Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e. The AHP-MCDA approach appears to have the least overall accuracy of 67%, the SVM-ML model having an overall accuracy of 79%, while the combined approach demonstrated an overall accuracy of 81%, with notably higher performance in identifying non-flood areas compared to flood-prone areas.\u003c/p\u003e\n \u003cp\u003eTo validate the generated flood susceptibility maps, we applied the ROC-AUC curve analysis. Results showed that the combined SVM-ML and AHP-MCDA map marginally outperformed both the individual SVM-ML and AHP-MCDA maps, although the performance difference between the combined map and the SVM-ML map was minimal (Fig. \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab7\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eEvaluation metric scores of the flood susceptibility models\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModel\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eFlood Status\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePrecision\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRecall\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eF1-Score\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eAHP-MCDA\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eNo Flood (0)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.70\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.85\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.77\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cem\u003e67%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eFlood (1)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.55\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.33\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.42\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cem\u003eSVM-ML\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eNo Flood (0)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.84\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.85\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.85\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cem\u003e79%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eFlood (1)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.69\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.67\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.68\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eCombined\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eNo Flood (0)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.86\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.89\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cem\u003e0.87\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cem\u003e81%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFlood (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.70\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n\u003c/div\u003e"},{"header":"5.0 DISCUSSION","content":"\u003cp\u003eThe integration of GIS with AHP-MCDA and SVM-ML for flood susceptibility mapping in the Ala River basin divulges promising results, revealing difference in the two approaches. The AHP-MCDA approach incorporated expert knowledge and preferences, revealed a larger percentage of the study area in the moderate flood susceptibility zone. It identified downstream areas as most vulnerable, primarily due to low elevation levels and high-water accumulation. This aligns with findings from Komolafe et al. (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Edamo \u003cem\u003eet al.\u003c/em\u003e (2022), who emphasized the significance of topography and rainfall in flood occurrence. The use of AHP-MCDA for flood susceptibility mapping is known to be appropriate for regional studies while been dependent on the characteristics of the region (Komolafe et. al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, it\u0026rsquo;s limitation is in the level of experience and knowledge of the expert.\u003c/p\u003e \u003cp\u003eIn contrast, the SVM-ML approach, leveraging historical flood occurrence data alongside the flood causative factors, highlighted upstream areas with significant built-up activities as highly vulnerable to flooding. The model showed that the central part of the basin, characterized by less built-up area and more vegetation, has moderate to low flood susceptibility. Interestingly, the downstream area remained highly susceptible due to low elevation and high flow accumulation. This pattern suggests that rainfall also plays a crucial role in the study area's flooding frequency, with built-up areas experiencing an initial surge in water flow before it moves downstream.\u003c/p\u003e \u003cp\u003eBoth methods demonstrated their value, with the SVM-ML model proving to be more accurate in this particular study area. The AHP-MCDA method's strength lies in its ability to incorporate expert knowledge and systematically prioritize factors influencing flood susceptibility. On the other hand, the SVM-ML model excels in handling high-dimensional datasets and complex, non-linear relationships between input variables and flood susceptibility. The AHP-MCDA enables stakeholders to systematically prioritize and evaluate factors influencing flood susceptibility, while SVM-ML enhances the predictive accuracy of susceptibility maps. The integration of these approaches generated a more robust flood susceptibility map, capitalizing on the strengths of both methodologies.\u003c/p\u003e \u003cp\u003eIt is worth to mention that elements such as buildings, farmlands, and human population were identified in the flood prone area of the basin during a survey in the area, which were also identified in the high and very high flood susceptible zones of the combined flood susceptibility map. The effect of heavy rainfall coupled with built-up which causes increase in water runoff at the upstream is identified to be the leading cause of flood susceptibility in the upstream of the basin. Areas such as Ayedun, Araromi, Alagbaka, Oba-Ile, etc. at the upstream are revealed to be most exposed to the risk of flood in the study area and recently are mostly feeling the impact (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Ibitoye et. al., (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) reported that the increase in the frequency of flood occurrence accompanied by its severe damages in the Ala River basin over the years is as a result of the presence of buildings, infrastructures and increasing human population in Akure metropolis. At the downstream area, farmlands and settlements such as Ago-Owo, Ago-Dada, Olokuta, Ala-Ajagusin, etc. are exposed to the risk of flood. The severity of flood in the downstream would continue to increase rapidly if prompt caution and measures are not taken.\u003c/p\u003e"},{"header":"6.0 CONCLUSION","content":"\u003cp\u003eThis study successfully mapped the flood susceptibility of the Ala River basin in Ondo State by employing AHP-MCDA and SVM-ML approached, considering 19 flood causative factors. The AHP-MCDA results indicated that approximately 20% of the study area exhibited very high and high susceptibility to flooding, concentrated mainly in the downstream areas. The SVM-ML model, however, showed that about 12% of the area had very high and high flood susceptibility, with a focus on upstream areas which are Ayedun, Araromi, Oke-Ijebu, Alagbaka, and Oba-Ile areas, with significant development and less-developed downstream areas. The central parts of the basin, characterized by less built-up area and more vegetation, exhibits moderate to low flood susceptibility. In contrast, the downstream area remains highly susceptible due to its low elevation and high flow accumulation. This pattern indicates that rainfall plays a significant role in the study area's flooding frequency. During periods of high rainfall, built-up areas experience an initial surge in water flow before it moves downstream. Consequently, while elevation is a crucial factor in flood susceptibility mapping, rainfall interacts complexly with topography to influence flood risk.\u003c/p\u003e \u003cp\u003eThe combined AHP-MCDA and SVM-ML flood susceptibility map revealed the complementary nature of the two approaches, providing a comprehensive view of the area's flood occurrence patterns. Built-up infrastructure in the upstream emerged as a major contributing factor, while high flow rates from upstream and reduced channel capacity likely contribute to downstream flooding.\u003c/p\u003e \u003cp\u003eBased on these findings, we recommend prioritizing floodplain management in populated upstream areas through more stringent zoning laws and promotion of green infrastructure. Upgrading and maintaining drainage canals, particularly in downstream areas such as Ala-Ajagbusi, Kajola, Apafa-Aiyede, and Ago-Goke settlements is crucial for effective floodwater management. Additionally, public awareness campaigns should be conducted to inform locals about flood risks and preparedness strategies.\u003c/p\u003e \u003cp\u003eWhile this integrated approach provides valuable insights for flood risk management in the Ala River Basin, it's important to note that the study did not cover flood depth and velocity. Future research should consider incorporating hydraulic modeling to provide a more comprehensive analysis of flood susceptibility. This could enhance the accuracy of predictions and offer more detailed information for flood management strategies.\u003c/p\u003e \u003cp\u003eIn conclusion, this study demonstrates the effectiveness of combining GIS, AHP-MCDA, and SVM-ML techniques in flood susceptibility mapping. The methodology developed here can be adapted for similar studies in other regions, contributing to improved flood risk management strategies globally. As climate change continues to alter precipitation patterns and increase flood risks worldwide, such integrated approaches will become increasingly valuable for urban planning and disaster preparedness.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCompeting Interests:\u003c/h2\u003e \u003cp\u003eThere are no relevant financial or non-financial interests to disclose.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eNo funds, grants, or other support were received during the preparation of this manuscript.\u003c/p\u003e\u003ch2\u003eAuthors Contribution:\u003c/h2\u003e \u003cp\u003eAll authors contributed to the study's conception and design. Material preparation, data collection, and analysis were performed by Adeyemi Adedoyin Benson, Ismaila Racheal Opeyemi, and Komolafe Akinola Adesuji. The first draft of the manuscript was written by Adedoyin Benson Adeyemi and Ismaila Racheal Opeyemi while Komolafe Akinola Adesuji, Catherine Lilian Nakalembe, Adebowale Daniel Adebayo, and Enoch Oluwaferanmi Babayemi helped improve the first draft of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAjin, R.S., R.R. Krishnamurthy, M. Jayaprakash and P.G. Vinod, 2013. Flood hazard assessment of Vamanapuram River Basin, Kerala, India: An approach using Remote Sensing and GIS techniques. Adv. Applied Sci. Res., 4: 263-274 \u003c/li\u003e\n\u003cli\u003eAjjur, S. B., \u0026amp; Mogheir, Y. K. (2020). Flood hazard mapping using a multi-criteria decision analysis and GIS (case study Gaza Governorate, Palestine). \u003cem\u003eArabian Journal of Geosciences\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(2). https://doi.org/10.1007/s12517-019-5024-6\u003c/li\u003e\n\u003cli\u003eAkukwe, T. I., \u0026amp; Ogbodo, C. (2015). Spatial analysis of vulnerability to flooding in Port Harcourt Metropolis, Nigeria. \u003cem\u003eSAGE Open\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(1), 215824401557555. doi:10.1177/2158244015575558\u003c/li\u003e\n\u003cli\u003eAlbert, A. A., Samson, A. A., Peter, O. O. and Olufunmilayo, A. O. An Assessment of the Socio Economic Impacts of Soil Erosion in South-Eastern Nigeria, Shaping the Change, XXIII FIG Congress Munich, Germany, pp. 12, 2006. \u003c/li\u003e\n\u003cli\u003eAlimi, S. A., Andongma, T. W., Ogungbade, O., Senbore, S. S., Alepa, V. C., Akinlabi, O. J., \u0026hellip; Muhammed, Q. O. (2022). Flood vulnerable zones mapping using geospatial techniques: Case study of osogbo metropolis, Nigeria. \u003cem\u003eThe Egyptian Journal of Remote Sensing and Space Science\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(3), 841\u0026ndash;850. doi:10.1016/j.ejrs.2022.08.003\u003c/li\u003e\n\u003cli\u003eAlipour, A., Ahmadalipour, A., Abbaszadeh, P., \u0026amp; Moradkhani, H. (2020). Leveraging machine learning for predicting flash flood damage in the Southeast US. \u003cem\u003eEnvironmental Research Letters\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(2), 024011. https://doi.org/10.1088/1748-9326/ab6edd\u003c/li\u003e\n\u003cli\u003eAvand, M., Moradi, H. R., \u0026amp; Ramazanzadeh Lasboyee, M. (2021). Spatial prediction of future flood risk: An approach to the effects of climate change. \u003cem\u003eGeosciences\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), 25. doi:10.3390/geosciences11010025\u003c/li\u003e\n\u003cli\u003eAwosika, L.F and R. Folorunsho, 2000. Integrated monitoring of flood and information. Dissemination for Flood Early Warnings in Nigeria, Nigeria Institute for Oceanography and Maritime Research VI Lagos.\u003c/li\u003e\n\u003cli\u003eBakare, G. O., Omosulu, S. B., \u0026amp; Famutimi, J. T. (2019). \u003cem\u003eGeo-Mapping of Areas Vulnerable to Ala-River Basin Flood Disaster Risk in Akure, Ondo State, Nigeria\u003c/em\u003e. doi:10.20944/preprints201911.0184.v1\u003c/li\u003e\n\u003cli\u003eBalasubramanian, A.. (2017). Soil Erosion- Causes and Effects. 10.13140/RG.2.2.26247.39841.\u003c/li\u003e\n\u003cli\u003eBettis III, E. A., Gully Erosion of Western Iowa. The Lowa Department of Natural Resources, Iowa Geological and Water Survey, 2005. \u003c/li\u003e\n\u003cli\u003eBirmah, M. N., Kigun, P. A., Alfred, Y. B., Majidadi, ST., \u0026amp; Surajo, L. A. (2021). Flood Assessment in Suleja Local Government Area, Niger State, Nigeria. \u003cem\u003eInternational Journal of Research Publication and Reviews\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(3), 219\u0026ndash;239.\u003c/li\u003e\n\u003cli\u003eCabrera, Jonathan \u0026amp; Lee, Han Soo. (2019). Flood-Prone Area Assessment Using GIS-Based Multi-Criteria Analysis: A Case Study in Davao Oriental, Philippines. \u003cem\u003eWater. 11\u003c/em\u003e. 2203. 10.3390/w11112203.\u003c/li\u003e\n\u003cli\u003eCea, L., Costabile, P., 2022. Flood risk in urban areas: modelling, management and adaptation to climate change. A review. \u003cem\u003eHydrology 9 \u003c/em\u003e(3), 50. doi:10.3390/hydrology9030050.\u003c/li\u003e\n\u003cli\u003eChen, J., A.A. Hill and L.D. Urbano, 2009. A GISbased model for urban flood inundation. J. Hydrol., 373: 184-192. DOI: 10.1016/j.jhydrol.2009.04.021 \u003c/li\u003e\n\u003cli\u003eDanumah, J.H.; Odai, S.N.; Saley, B.M.; Szarzynski, J.; Thiel, M.; Kwaku, A.; Kouame, F.K.; Akpa, L.Y. (2016). Flood risk assessment and mapping in Abidjan district using multi-criteria analysis (AHP) model and geoinformation techniques, (cote d\u0026rsquo;ivoire). \u003cem\u003eGeoenviron. Disasters, 3\u003c/em\u003e(10).\u003c/li\u003e\n\u003cli\u003eDe Moel, H., Jongman, B., Kreibich, H., Merz, B., Penning-Rowsell, E. C., \u0026amp; Ward, P. J. (2015). Flood risk assessments at different spatial scales. \u003cem\u003eMitigation and Adaptation Strategies for Global Change\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(6), 865\u0026ndash;890. https://doi.org/10.1007/s11027-015-9654-z\u003c/li\u003e\n\u003cli\u003eEfraimidou, E., \u0026amp; Spiliotis, M. (2024). A GIS-Based Flood Risk Assessment Using the Decision-Making Trial and Evaluation Laboratory Approach at a Regional Scale. \u003cem\u003eEnvironmental Processes\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1). https://doi.org/10.1007/s40710-024-00683-w\u003c/li\u003e\n\u003cli\u003eEl-Swaify, S. (2013). Impacts of Land Use Change on Soil Erosion and Water Quality-A Case Study from Hawaii Impacts of Land Use Change on Soil Erosion and Water Quality-A Case Study from Hawaii.\u003c/li\u003e\n\u003cli\u003eGraf, W. L., The Arroyo problem-palaehydrology and palaeohydraulics in the short term, in K. J. Gregory (ed). Background to palaeohydrology: A perspective (Chihester, Wiley), 279-302, 2003 \u003c/li\u003e\n\u003cli\u003eHilborn, D., Gully Erosion Control, Ontario Ministry of Agriculture, Food and Rural Affairs, 2005 \u003c/li\u003e\n\u003cli\u003eHill, J., Megier, J., \u0026amp; Mehl, W. (1995). Land degradation, Soil Erosion and Desertification Monitoring in Mediterranean Ecosystems. Remote Sensing Review, 12, 107-130. http://dx.doi.org/10.1080/02757259509532278\u003c/li\u003e\n\u003cli\u003eHuong, H.T.L. and A. Pathirana, 2013. Urbanization and climate change impacts on future urban flooding in Can Tho city, Vietnam. Hydrol. Earth Syst. Sci., 17: 379-394. DOI: 10.5194/hess-17-379-2013\u003c/li\u003e\n\u003cli\u003eIbitoye, M.O., Komolafe, A. A., Adegboyega, S. A., and Adebola, A. O, 2020. Analysis of vulnerable urban properties within river Ala floodplain in Akure, Southwestern Nigeria. Spatial Information Research, 28(4):431\u0026ndash;445 DOI 10.1007/s41324-019-00298-6. \u003c/li\u003e\n\u003cli\u003eIghile, E.H. Shirakawa, H.; Tanikawa, H. (2022). Application of GIS and Machine Learning to Predict Flood Areas in Nigeria. \u003cem\u003eSustainability 14\u003c/em\u003e, 5039. https://doi.org/10.3390/su14095039\u003c/li\u003e\n\u003cli\u003eIshaya, S., Ifatimehin, O. O., \u0026amp; Abaje, I. B. (2009). Mapping Flood Vulnerable Areas in a Developing Urban Centre of Nigeria. \u003cem\u003eJournal of Sustainable Development in Africa\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(4).\u003c/li\u003e\n\u003cli\u003eKomolafe, Akinola \u0026amp; Adegboyega, Suleiman \u0026amp; Akinluyi, Francis. (2015). A Review of Flood Risk Analysis in Nigeria. \u003cem\u003eAmerican journal of environmental sciences. 11.\u003c/em\u003e 157 -166. 10.3844/ajessp.2015.157.166.\u003c/li\u003e\n\u003cli\u003eKomolafe, A. A., Srikantha H., and Avtar R., 2018. Development of generalized loss functions \u003c/li\u003e\n\u003cli\u003efor rapid estimation of flood damages: a case study in Kelani River basin. Applied Geomatics, 10 (1): 13-30, \u003c/li\u003e\n\u003cli\u003eKomolafe, A. A., Awe, B. S., Olorunfemi, I. E., \u0026amp; Oguntunde, P. G. (2020). Modelling flood-prone area and vulnerability using integration of multi-criteria analysis and HAND model in the Ogun River Basin, Nigeria. Hydrological Sciences Journal, 65(10), 1766\u0026ndash;1783. https://doi.org/10.1080/02626667.2020.1764960\u003c/li\u003e\n\u003cli\u003eLuino, F. (2016) \u0026lsquo;Floods\u0026rsquo;, Encyclopedia of Earth Sciences Series, pp. 1\u0026ndash;6. doi:10.1007/978-3-319-12127-7_126-1.\u003c/li\u003e\n\u003cli\u003eMosavi, A., Ozturk, P., \u0026amp; Chau, K. W. (2018). Flood Prediction Using Machine Learning Models: Literature Review. \u003cem\u003eWater\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(11), 1536. https://doi.org/10.3390/w10111536 \u003c/li\u003e\n\u003cli\u003eOlatona, O. O., Obiora-Okeke O. A., and Adewumi J. R. (2018). Mapping of flood risk zones in Ala River Basin Akure, Nigeria. \u003cem\u003eAmerican Journal of Engineering and Applied Sciences\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), 210\u0026ndash;217. doi:10.3844/ajeassp.2018.210.217\u003c/li\u003e\n\u003cli\u003eOyinloye, M.A. and Olamiju, O.I. (2011) \u0026lsquo;Flood risk mapping and vulnerability analysis using GIS: empirical evidences from New Town area, Ondo, Ondo State, Nigeria\u0026rsquo;, Int. J. Society Systems Science, Vol. 3, No. 3, pp.291\u0026ndash;304. \u003c/li\u003e\n\u003cli\u003ePedersen, A.N., P.S. Mikkelsen and K. ArnbjergNielsen, 2012. Climate change-induced impacts on urban flood risk influenced by concurrent hazards. J. Flood Risk Manage., 5: 203-214. DOI: 10.1111/j.1753-318X.2012.01139.x \u003c/li\u003e\n\u003cli\u003ePoesen, J., Nachtergaele, J., Verstraeten, G. and. Valentin, C., Gully erosion and environmental change: importance and research needs, Catena, 50 (2-4), pp. 91-133, 2003. \u003c/li\u003e\n\u003cli\u003eP. Sayers, Y. L.i, G. Galloway, E. Penning-Rowsell, F. Shen, K. Wen, Y. Chen, and T. Le Quesne. 2013. Flood Risk Management: A Strategic Approach. Paris, UNESCO. \u003c/li\u003e\n\u003cli\u003eRahman, A. J. A., \u0026amp; Ramli, N. A. (2024). Flood Prediction Using Artificial Neural Networks: A Case Study in Temerloh, Pahang. \u003cem\u003eQeios\u003c/em\u003e. https://doi.org/10.32388/tuz29y\u003c/li\u003e\n\u003cli\u003eSchanze, J., 2006. Flood Risk Management- a Basic Framework. In: Flood Risk Management-Hazards, Vulnerability and Mitigation Measures, Schanze, J., E. Zeman and J. Marsalek (Eds.), Springer, pp: 149-167. \u003c/li\u003e\n\u003cli\u003eScheuer, S. and V. Meyer, 2007. FloodCalc. Software tool for the calculation of multicriteria flood damage and risk maps \u003c/li\u003e\n\u003cli\u003eSowmya, K., C.M. John and N.K. Shrivasthava, 2015. Urban flood vulnerability zoning of Cochin City, southwest coast of India, using remote sensing and GIS. Nat. Hazards, 75: 1271-1286. DOI: 10.1007/s11069-014-1372-4 \u003c/li\u003e\n\u003cli\u003eSuleiman, Y.M, M.B. Matazu, A.A. Davids and M.C. Mozie, 2014. The application of geospatial techniques in flood risk and vulnerability mapping for disaster management at Lokoja, Kogi State, Nigeria. J. Environ. Earth Sci., 4: 54-61. \u003c/li\u003e\n\u003cli\u003eSuresh, R. (2012) Soil and water conservation engineering. Delhi: Standard Publishers Distributors.\u003c/li\u003e\n\u003cli\u003eUNISDR, 2015. Disaster Risk Reduction in the Post-2015 Development Agenda: Transforming Our World: The 2030 Agenda fpr Sustainable Development., \u003cem\u003ein\u003c/em\u003e UNISDR, ed.: Geneva, Switzerland, UNISDR. Available at: https://www.unisdr.org/we/inform/publications/45417.\u003c/li\u003e\n\u003cli\u003eUNOCHA. Nigeria: Humanitarian Response Plan; United Nations Office for the Coordination of Humanitarian Affairs: Abuja, Nigeria, 2018\u003c/li\u003e\n\u003cli\u003eZehra, N. (2020). Prediction Analysis of Floods Using Machine Learning Algorithms (NARX \u0026amp; SVM). \u003cem\u003eInternational Journal of Sciences: Basic and Applied Research\u003c/em\u003e, \u003cem\u003e49\u003c/em\u003e(2), 24\u0026ndash;34.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Flood prediction, Geographic Information System, Multi-Criteria Decision-Making Analysis, Machine Learning, Flood Susceptibility, River Basin","lastPublishedDoi":"10.21203/rs.3.rs-4863685/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4863685/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFlooding is a recognized form of natural disaster that can lead to loss of life, destruction of critical infrastructure with consequences impacting critical sectors including agriculture and health. This study aims to map out flood susceptible areas within the Ala River basin of Ondo State, Nigeria by integrating the Analytical Hierarchy Process (AHP) Multi-Criteria Decision Analysis (MCDA) technique and Support Vector Machines (SVM) Machine Learning (ML) model. Nineteen factors including elevation, slope, aspect, curvature (profile and plan), roughness, flow direction, flow accumulation, drainage density, distance from the river, TWI, STI, SPI, soil, geology, NDVI, NDMI, LULC, and rainfall were considered as input parameters. Flood susceptibility maps generated from each of these approaches were combined to create a more comprehensive flood susceptibility map of the study area. The AHP analysis has a consistency ratio of 1.8%. Precision, recall, f1-score, accuracy score, and ROC-AUC curve were used in evaluating the AHP-MCDA and SVM-ML model. Based on the evaluation, the combined flood susceptibility map result showed the best performance with the AUC score 0.74, SVM-ML with a score 0.73, and the AHP-MCDA having the least score of 0.59. As these results demonstrate, multiple approaches are required to mitigate flooding.\u003c/p\u003e","manuscriptTitle":"Integrated GIS-Based MCDA and Machine Learning Techniques in Flood Susceptibility Mapping in Ala River Basin, Nigeria","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-02 06:36:47","doi":"10.21203/rs.3.rs-4863685/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"30064b2d-744c-4246-ace3-7d3b7348660e","owner":[],"postedDate":"September 2nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-09-25T23:47:54+00:00","versionOfRecord":[],"versionCreatedAt":"2024-09-02 06:36:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4863685","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4863685","identity":"rs-4863685","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}