Determining Flood Zonation Maps, Using New Ensembles of Multi-Criteria Decision-Making, Bivariate Statistics, and Artificial Neural Network

Abstract

Golestan Province is one of the most vulnerable areas to catastrophic flood events in Iran. The flood severity in this region has grown dramatically during the last decades, demanding a major investigation. Accordingly, an authentic map providing detailed information on floods is required to reduce future flood disasters. Three ensemble models produced by the combination of Evaluation Based on Distance from Average Solution (EDAS) and Multilayer Perceptron Neural Network (MLP) with Frequency Ratio (FR), and Weights of Evidence (WOE) are used to quantify the map flood susceptibility in Golestan Province, in the north of Iran. Ten flood effective criteria, namely altitude, slope degree, slope aspect, plan curvature, distance from rivers, Topographic Wetness Index (TWI), rainfall, soil type, geology, and land use, are considered for the modeling process. The flood zonation maps are validated by the receiver operating curve (ROC). The results show that the most precise model is MLP-FR (AUROC = 0.912), followed by EDAS-FR-AHP (AUROC = 0.875), and EDAS-WOE-AHP (AUROC = 0.845). The high accuracies of all methods applied to illustrate their capability in predicting flood susceptibility in future studies.

Flood Susceptibility in the Lower Course of the Coyuca River, Mexico: A Multi-Criteria Decision Analysis Model

“…In this study, five conditioning factors and one flood triggering facto (Figure 4). By means of an exhaustive bibliographic review and field work tory map of floodable and non-floodable areas was obtained to valid [25,26,[72][73][74]. A multicollinearity analysis was performed between the in by calculating the variance inflation factor (VIF) and the tolerance (TOL) flood influence factors were subjected to the AHP methodology to calcula tor weights and category weights for each flood influence factor [76][77][78].…”

“…In this study, five conditioning factors and one flood triggering factor were selected (Figure 4). By means of an exhaustive bibliographic review and field work, a flood inventory map of floodable and non-floodable areas was obtained to validate the results [25,26,[72][73][74]. A multicollinearity analysis was performed between the influence factors by calculating the variance inflation factor (VIF) and the tolerance (TOL) [25,26,75].…”

“…In the study area, there were no official records of flooding points that would help in the validation of the results (areas susceptible to flooding); for this reason, an inventory of recent floods was developed [25,72]. Journalistic sources and scientific journals (both international and national) were consulted [64, 65,68,69], semi-structured interviews were conducted with key actors, a survey on the perceptions of the local population was applied, and satellite observations of the area were made with Google Earth Pro to collect data on previous floods in the study area [25,26,35].…”

“…Rainfall is the main source of surface runoff and triggers flooding [26,72]. The precipitation data used as input were twelve-month cumulative precipitation data from NASA-Daymet 2021 [83].…”

Comparison of hydrological modeling, Artificial Neural Networks and multi-criteria decision making approaches for determining Flood Source Areas

“…One of the most important tasks in flood control is identifying the best methods for locating the primary sources of flooding across a catchment, in order to improve flood prevention techniques that is also known as flood source areas (FSA) (Singh et al, 2021), through accurate simulation and analysis of runoff generation processes at the sub-basin scale. In an attempt to solve this problem, many systematic methods such as hydrological models (Abdulkareem et al, 2018;Dehghanian et al, 2019;Maghsood et al, 2019), GIS-based methods (Cabrera & Lee, 2019;Hong & Abdelkareem, 2022;Mukherjee & Singh, 2020;Osei et al, 2021), remote sensing methods (Sadiq et al, 2022;Sharma et al, 2019;Syifa et al, 2019), multi-criteria decision methods (Ajjur & Mogheir, 2020;Hadian et al, 2022;Pham et al, 2021;Roy et al, 2021), and machine learning and data mining methods (Ha & Kang, 2022;Luu et al, 2021;Rahman et al, 2021) have been used.…”