A comparative assessment of multi-criteria decision-making analysis and machine learning methods for flood susceptibility mapping and socio-economic impacts on flood risk in Abela-Abaya floodplain of Ethiopia

Abstract

Floods have a terrible impact on people's lives and property all around the world. In this study, we evaluated the modeling capabilities of two Machine Learning (ML) approaches such as Naive Bayes Tree (NBT) and Naive Bayes (NB) and four Multi-Criteria Decision-Making (MCDM) analysis techniques MABAC (multi-attributive border approximation area comparison), TOPSIS (Technique for Order Preference by Similarity to Ideal Solution), VIKOR (Vise Kriterijumska Optimizacijaik Ompromisno Resenje), and SAW (Simple Additive Weighting) were applied in Abela-Abaya floodplain of Ethiopia. Sixteen flood influencing factors such as elevation, land use/cover (LULC), soil, aspect, geomorphology, normalized difference vegetation index (NDVI), rainfall, distance from river (DR..), topographic wetness index (TWI), sediment transport index (STI), stream power index (SPI), geology, curvature, flow accumulation (FA), slope and flow direction (FD) were used as input parameters. Area Under the Receiver Operating Characteristic Curve was used to assess and validate the models' predictive power (AUC). The NB model performed the best (AUC = 0.92), indicating that it is a viable strategy for determining flood-prone locations in order to properly plan and control flood hazards. Face to face interactive sessions were conducted with 160 respondents to find a coordinated analysis between the impact issues and the community's apprehension of flood danger and to analyze the socioeconomic impact of flood risk. The findings revealed that respondents' socio-demographic traits, flood experience, flood awareness, flood prevention responsibility, and government confidence building were all cohesively related to their opinion of flood danger. The estimated damage for households and farmland were $6249 and $5326 respectively by 2016 flood. The compilation of the findings of this study on flood susceptibility mapping and socioeconomic impact of flood risk seeks to enhance human perceptions of flood risk to minimize flood risk by enhancing communication about the issues and inspiring people of flood-prone areas to take measures for mitigating flood damage.

Keywords

• Abela-Abaya
• Flood susceptibility
• ML
• MCDM
• Ethiopia
• Socio-economic