Flood Susceptibility Modeling Research Articles

Flood susceptibility modelling using advanced ensemble machine learning models

Floods are one of nature's most destructive disasters because of the immense damage to land, buildings, and human fatalities. It is difficult to forecast the areas that are vulnerable to flash flooding due to the dynamic and complex nature of the flash floods. Therefore, earlier identification of flash flood susceptible sites can be performed using advanced machine learning models for managing flood disasters. In this study, we applied and assessed two new hybrid ensemble models, namely Dagging and Random Subspace (RS) coupled with Artificial Neural Network (ANN), Random Forest (RF), and Support Vector Machine (SVM) which are the other three state-of-the-art machine learning models for modelling flood susceptibility maps at the Teesta River basin, the northern region of Bangladesh. The application of these models includes twelve flood influencing factors with 413 current and former flooding points, which were transferred in a GIS environment. The information gain ratio, the multicollinearity diagnostics tests were employed to determine the association between the occurrences and flood influential factors. For the validation and the comparison of these models, for the ability to predict the statistical appraisal measures such as Freidman, Wilcoxon signed-rank, and t-paired tests and Receiver Operating Characteristic Curve (ROC) were employed. The value of the Area Under the Curve (AUC) of ROC was above 0.80 for all models. For flood susceptibility modelling, the Dagging model performs superior, followed by RF, the ANN, the SVM, and the RS, then the several benchmark models. The approach and solution-oriented outcomes outlined in this paper will assist state and local authorities as well as policy makers in reducing flood-related threats and will also assist in the implementation of effective mitigation strategies to mitigate future damage.

Flash flood risk modeling of swat river sub-watershed: a comparative analysis of morphometric ranking approach and El-Shamy approach

Globally, flash floods are the most damaging natural hazards, because of its sudden nature and difficulty in forecasting that restrains emergency responses. Flash floods triggered by thunderstorms are frequent in the high mountainous area of Hindukush Himalaya in the north of Pakistan. The current study employed two widely used approaches for flash flood risk modeling, i.e., morphometric ranking approaches (MRAs) and El-Shamy’s approach to analyzing their effectiveness for flash flood susceptibility modeling in Swat river watershed, district Swat, Pakistan. The digital elevation model was utilized to delineate the watershed and drainage network using the ArcHydro tool of ArcGIS. A total of 15 morphometric parameters have been used for flash flood modeling. The analysis suggests that the Swat river watershed consists of 17 sub-basins. The MRA-based flash flood risk assessment suggests that sub-basins B1 and B2 (12% of total sub-basins) are highly susceptible to flash flooding. While sub-basins B4, B6, B11, B13, and B17 (59% of the total sub-basin) are moderately susceptible to flash flood risk. El-Shamy-based modeling suggests that sub-basins B12, B14, and B16 (18% of total sub-basins) have high flood susceptibility while sub-basin B5 has the lowest. The field observation and 2010 flood reports by WHO and NDMA suggest that the flash flood susceptibility modeling results by the MRA are more accurate as compared with El-Shamy’s approach. The results of the study perceived to help execute appropriate remedial measures to extenuate the potential flash floods in the study area.

Ensemble fuzzy MCDM for spatial assessment of flood susceptibility in Ibadan, Nigeria

Ibadan, one of the largest cities in West Africa, experiences recurrent flood hazards, causing severe damages to lives and properties. Although various flood susceptibility models have evolved to mitigate flood hazards in different parts of the world, model performances vary according to the peculiarity of the study area and conditioning factors. Therefore, specialized studies are essential to determine model performances for specific locations. This study develops an integrated spatial multi-criteria decision-making model to classify flood susceptibility in Ibadan, Nigeria. For the improvement of the weighting accuracy of the flood causative criteria, Fuzzy AHP (FAHP) was integrated with GIS for the weight computation and overlay process. Ten flood causative factors were evaluated using the AHP and FAHP models. Results indicate rainfall, runoff, and distance to stream as the most significant flood causative factors with FAHP and AHP weights of 22%, 18%, 16%, and 23%, 23%, 18%, respectively. The FAHP and AHP maps identified the southern region as the most susceptible, with the FAHP also highlighting the susceptibility of the central region. The models’ accuracies were validated by overlaying the AHP and FAHP maps with locations of previous flood occurrences in the study area. The Fuzzy AHP map had a 91% alignment with the historical flood locations while the AHP model produced a 45% match, confirming the higher accuracy of the FAHP model and suitability for flood susceptibility mapping in Nigeria and other West African cities.

Flood susceptibility modeling in Teesta River basin, Bangladesh using novel ensembles of bagging algorithms

The flooding in Bangladesh during monsoon season is very common and frequently happens. Consequently, people have been experiencing tremendous damage to properties, infrastructures, and human casualties. Usually, floods are one of the devastating disasters from nature, but for developing nations like Bangladesh, flooding becomes worse. Due to the dynamic and complex nature of the flooding, the prediction of flooding sites was usually very difficult for flood management. But the artificial intelligence and advanced remote sensing techniques together could predict and identify the possible sites, which are vulnerable to flooding. The present work aimed to predict and identify the flooding sites or flood susceptible zones in the Teesta River basin by employing state-of-the-art novel ensemble machine learning algorithms. We developed ensembles of bagging with REPtree, random forest (RF), M5P, and random tree (RT) algorithms for obtaining reliable and highly accurate results. Twelve factors, which are considered as the conditioning factors, and 413 current and former flooding points were identified for flooding susceptibility modelling. The Information Gain ratio statistical technique was utilized to determine the influence of the factors for flooding. We applied receiver operating characteristic curve (ROC) for validation of the flood susceptible models. The Freidman test, Wilcoxon signed-rank test, Kruskal–Wallis test and Kolmogorov–Smirnov test were applied together for the first time in flood susceptibility modelling to compare the models with each other. Results showed that more than 800 km2 area was predicted as the very high flood susceptibility zones by all algorithms. The ROC curve showed that all models achieved more than 0.85 area under the curve indicating highly accurate flood models. For flood susceptibility modelling, the bagging with M5P performed superior, followed by bagging with RF, bagging with REPtree and bagging with RT. The methodology and solution-oriented results presented in this paper will assist the regional as well as local authorities and the policy-makers for mitigating the risks related to floods and also help in developing appropriate measures to avoid potential damages.

Novel hybrid intelligence models for flood-susceptibility prediction: Meta optimization of the GMDH and SVR models with the genetic algorithm and harmony search

Floods are among the deadliest natural hazards for humans and the environment. Identifying the most flood-susceptible areas is a fundamental step in the development of flood mitigation strategies and for reducing flood damage. There is an ongoing global debate regarding the most suitable model for flood-susceptibility modeling and predictions. There is also a growing interest in the development of parsimonious and precise models for flood-susceptibility prediction. This study proposed several novel hybrid intelligence models based on the meta-optimization of the support vector regression (SVR) and group method of data handling (GMDH) using different meta-heuristic algorithms, i.e., the genetic algorithm (GA) and harmony search (HS). In contrast to the traditional models, in the SVR model computational complexity does not depend on the dimensionality of the input space. GMDH model has also advantage of being appropriate to analyze multi-parametric data sets. The methodology was developed for the Haraz-Neka watershed, one of the most flood-prone areas in the coastal margins of the Caspian Sea. A total of nine geospatial parameters (slope degree, aspect, elevation, plan curvature, profile curvature, distance to the river, land use, lithology, and rainfall) were identified as the main flood-conditioning factors using information gain ratio (IGR) analyses. Based on existing reports, 132 flood locations were identified in the study area, 92 points (70%) were used together with geospatial data for flood-susceptibility modeling, and the remaining 40 points (30%) were used to validate the models. An initial flood-susceptibility model was constructed based on the SVR and GMDH models. The model parameters were optimized using the GA and HS to reproduce the flood-susceptibility maps. The prediction accuracy of the resultant maps was evaluated in terms of various statistical measures, i.e., mean square error (MSE), root mean square error (RMSE), receiver operating characteristic (ROC) curve, and the area under the ROC curve (AUC). The results showed that the SVR model had superior performance (AUC = 0.70–0.75, RMSE = 0.29–0.36, MSE = 0.08–0.13) compared to the GMDH model (AUC = 0.67–0.74, RMSE = 0.32–0.39, MSE = 0.1–0.15). Both the GA and HS remarkably improved the SVR and GMDH performance, with the SVR-GA model performing the best (AUC = 0.75, RMSE = 0.29, MSE = 0.08) followed by the SVR-HS model (AUC = 0.75, RMSE = 0.33, MSE = 0.11). Our results verify the efficiency of the proposed hybrid models for spatial flood-susceptibility prediction. The proposed models can be adopted for use in other regions with similar hydro-environment characteristics.

Flood Susceptibility Modeling of Megech River Catchment, Lake Tana Basin, North Western Ethiopia, Using Morphometric Analysis

The present study intended to flood susceptibility modeling by using drainage morphometric investigations of the Megech River catchment, Lake Tana Basin, North Western Ethiopia. Drainage morphometric criterion study performs a critical aspect in recognizing the factual aspects of river catchment with concern to floods. In the present study, we depicted the Megech River catchment into four sub-watersheds, followed by clipping off the drainage grid by using ALOS-PALSAR (Advanced Land Observing Satellite-Phased Array-Type L-Band Synthetic Aperture Radar) digital elevation model, toposheets, and Landsat-8 Operational Land Imager coupled with the Geographic Information System (GIS) program. The drainage morphometric parameters, such as basic, linear, areal and relief, were calculated in the current research by using the standard formula. The morphometric parameters, such as bifurcation ratio, drainage density, length of overland flow and drainage frequency, have a direct connection with flood susceptibility. Hence, rank 1 assigned to the highest values of the above-mentioned parameters followed by second-rank to second-highest value and rank third given the lowest value of the above parameters. The morphometric parameters, such as circulatory ratio, form factor, elongation ratio, drainage texture and compactness coefficient, have a reverse relation with flood proneness. Hence, rank 1 assigned to the lowest values of those parameters, followed by rank two to the second-lowest value and rank three given to the highest value of the above parameters. The compound factor is computed by aggregating the assigned ranks of the morphometric drainage parameters mentioned above and then dividing by the number of morphometric criteria used for sub-watersheds prioritization. The results of the present study displayed sub-watershed 3 scored very high prioritization (Ist), including a compound factor value of 1.89, and sub-watershed 2 got the highest compound factor value of 3.11, and it scored the lowest rank (IVth). The found sub-watershed 3 in the current research area was acquired as a high-priority grade one, and it demands urgent flood regulation remedies for competent water budget devising and administration in the Megech catchment area. The current study demonstrates the capability of sub-watershedwise drainage morphometric investigations in the flood susceptibility analysis using toposheets, optical remote sensing data and digital elevation model associated with GIS tools.

Optimization of state-of-the-art fuzzy-metaheuristic ANFIS-based machine learning models for flood susceptibility prediction mapping in the Middle Ganga Plain, India

This study is an attempt to quantitatively test and compare novel advanced-machine learning algorithms in terms of their performance in achieving the goal of predicting flood susceptible areas in a low altitudinal range, sub-tropical floodplain environmental setting, like that prevailing in the Middle Ganga Plain (MGP), India. This part of the Ganga floodplain region, which under the influence of undergoing active tectonic regime related subsidence, is the hotbed of annual flood disaster. This makes the region one of the best natural laboratories to test the flood susceptibility models for establishing a universalization of such models in low relief highly flood prone areas. Based on highly sophisticated flood inventory archived for this region, and 12 flood conditioning factors viz. annual rainfall, soil type, stream density, distance from stream, distance from road, Topographic Wetness Index (TWI), altitude, slope aspect, slope, curvature, land use/land cover, and geomorphology, an advanced novel hybrid model Adaptive Neuro Fuzzy Inference System (ANFIS), and three metaheuristic models-based ensembles with ANFIS namely ANFIS-GA (Genetic Algorithm), ANFIS-DE (Differential Evolution), and ANFIS-PSO (Particle Swarm Optimization), have been applied for zonation of the flood susceptible areas. The flood inventory dataset, prepared by collected flood samples, were apportioned into 70:30 classes to prepare training and validation datasets. One independent validation method, the Area-Under Receiver Operating Characteristic (AUROC) Curve, and other 11 cut-off-dependent model evaluation metrices have helped to conclude that the ANIFS-GA has outperformed other three models with highest success rate AUC = 0.922 and prediction rate AUC = 0.924. The accuracy was also found to be highest for ANFIS-GA during training (0.886) & validation (0.883). Better performance of ANIFS-GA than the individual models as well as some ensemble models suggests and warrants further study in this topoclimatic environment using other classes of susceptibility models. This will further help establishing a benchmark model with capability of highest accuracy and sensitivity performance in the similar topographic and climatic setting taking assumption of the quality of input parameters as constant.

An integrated approach of GIS and hybrid intelligence techniques applied for flood risk modeling

Flood is a typical natural disaster, which results in huge economic damage and human loss; therefore, accurately predicting flood-prone areas is important for preventing and mitigating the impacts of floods. The main objective of this study is to present new ensemble models, which are based on Index of Entropy (IOE), Fuzzy Membership Value (FMV), Frequency Ratio (FR), and Information Value (IV) for assessing flood susceptibility. For this purpose, data from a total of 78 flood events were taken into account as basic data for the training model and validation of results. Location and spatial characteristics of these historical flood events were used to identify the relevant criteria for flood susceptibility modeling (FSM) and in acquiring the contribution of each criterion in susceptibility of the region toward flood. The FMV-IOE, FR-IOE, and IV-IOE models were used to distinguish between presence and absence of flood and its mapping. These models were also employed to perform feature selection in order to reveal the variables, which may contribute for flood occurrence extensively. Finally, for the validation of results, the Area Under the Receiver Operating Characteristic (AUROC) was computed for each flood susceptibility map. The validation of results indicated that AUROC for three mentioned models varies from 0.963 to 0.969 (AUROC FMV-IOE = 96.9%, AUROC FR-IOE = 96.8%, and AUROC IV-IOE = 96.3%). Results acknowledged that the main drivers of flood occurrence were soil, land use, and SPI factors. The results of this research are of great importance for the task of mitigation and in reduction of the impacts of future losses, including land-use planning for the region under study. Current research makes a significant contribution to developing GISciences by means of proposing a new approach for GIS-based decision makings systems. From the methodological perspective, the results of this research are of great importance in analyzing the capability of hybrid intelligence techniques and their integration with fuzzy and GIS decision-making systems.HighlightsWe have developed three novel hybrid methods for flood modeling.Soil, land use, and distance to river were the effective factors.FMV-IOE model shows a better result in flood prediction.Using the results of this study, prevention measures can be taken in advance of a flood.

Flash flood susceptibility modelling using functional tree and hybrid ensemble techniques

The present research aims to assess and judge the capability of flash flood susceptibility (FFS) models considering hybrid machine learning ensemble techniques for the FFS assessment in the Gorgan Basin in Iran. Three novel intelligence approaches, namely, bagging–functional tree (BFT), dagging–functional tree, and rotational forest–functional tree are used for modelling, with consideration to 15 flood conditioning factors (FCFs) as independent variables and 426 flood locations as dependent variables. Three threshold-dependent and -independent approaches are used to evaluate the goodness-of-fit and prediction capability of the ensemble models with a single functional tree (FT). These approaches include the area under the receiver operating characteristic curve of the success rate curve (SRC) and prediction rate curve (PRC), efficiency (E) and true skill statistics (TSS). The random forest model is used to determine the relative importance of FCFs. Elevation, stream distance and normalized difference vegetation index (NDVI) have crucial roles in the study area during flash flood occurrences. According to the results of all threshold-dependent and -independent approaches (AUC of SRC = 0.933, AUC of PRC = 0.959, E = 0.76 and TSS = 0.72), the BFT ensemble model has the greatest accuracy in terms of modelling FFS. Results also show that the performance of the FT model is enhanced by three meta-classifiers. The seed cell area index technique is also used to check model classification accuracy and reliability. Results of this technique show that all the models demonstrate good performance and reliability. However, the FFS maps prepared by machine learning ensemble techniques have excellent accuracy and reliability, as per the results of validation methods. Thus, these FFS maps can be used as a convenient tool to reduce the effect of flood in flash flood-prone areas.

Spatial predicting of flood potential areas using novel hybridizations of fuzzy decision-making, bivariate statistics, and machine learning

The global warming and climate changes determined a considerable increase in the frequency of floods and their related damages. Therefore, the high accuracy prediction of flood susceptible areas plays a key role in flood warnings and risk reduction. The main objective of this study is to propose novel hybridizations of fuzzy Analytical Hierarchy Process (FAHP), Index of Entropy (IoE), and Support Vector Machine (SVM) for predicting the areas susceptible to floods. Buzău river catchment (Romania) was the area on which the present study was focused. In this regard, a database with 205 flooded locations, 205 non-flood locations and 12 flood predictors was established and used to train and validate the flood susceptibility models. The performance of the proposed models was evaluated using the Receiver Operating Characteristic (ROC) Curve and statistical metrics. The results show that all the hybrid models have a high prediction performance and outperform the stand-alone models. Among them, the SVM-IoE model (AUC = 0.979) has the highest performance, followed by the FAHP-IoE (AUC = 0.97), IoE (AUC = 0.969), SVM (AUC = 0.966) and FAHP (AUC = 0.947). These results highlight a very high efficiency of all the applied models. The application of the models mentioned above revealed that a percentage between 12.5% (FPIIoE) and 21.2% (FPIFAHP) of the study area is characterized by high and very high exposure to these hydrological hazards.

Flash Flood Susceptibility Modelling: A Review

Flash flood is common problem in the developed or developing country especially in urban areas. This is due to the changes of land use, poor planning of drainage system, deforestation and unplanned land use and etc. Flash flood susceptibility modelling (FFSM) for urban area is important to avoid the loss and etc. FFSM can be vital component and essential tools for planning and management of natural disaster and environmental. The objective of this paper is to review selected paper on flash flood susceptibility modelling using Geographic Information System (GIS) coupling empirical models. A lot of methods have been coupling with GIS to developed flood susceptibility modelling, for example, Weight of Evidence, Random Forest, etc. However, there is still lack of susceptibility model for flash flood. The GIS coupling method can improve the accuracy of flood susceptibility modelling. Thus, this paper will review the ability of GIS coupling with empirical models for the flash flood susceptibility modelling.

Flood susceptibility mapping by ensemble evidential belief function and binomial logistic regression model on river basin of eastern India

Flooding is the overflow of water from stream, river, lake and sea that occurs all over the world and has disastrous effects on human society and environment. Frequent severe flood event in eastern India cause of death and damages every year so, the development of flood susceptibility method is needed for identifying the flood vulnerability areas to reduce the damages. Techniques of Remote Sensing (RS) and Geographical Information System (GIS) can help to flood susceptibility modeling by the accrued and analyzing huge amount of data in short time. The main objectives of this study are to determine the effectiveness of Evidence Belief Function (EBF), binomial Logistic Regression (LR) and ensemble of EBF and LR (EBF-LR) model with RS and GIS techniques for flood susceptibility mapping and spatial prediction of flood-susceptible areas in the Koiya river basin of West Bengal, India. Eight flood conditioning factors; Land use and land cover (LULC), soil, rainfall, normalized differences vegetation index (NDVI), distance to river, elevation, topographic wetness index (TWI) and stream power index (SPI) have been used, and total 264 historical flooding points were mapped, and randomly divided in to training (70%) and validating (30%) dataset. Flood susceptibility map has been generated by applying EBF, LR and ensemble EBF-LR method with the help of training and eight causative factors dataset. The maps have been divided in to six classes; extremely low, very low, low, moderate, high and very high. The receiver operating characteristic (ROC) curve has been used to accuracy assessment of the susceptibility map, and the area under curve (AUC) disclosures 87.9%, 85.2% and 84.1% prediction rate for the EBF-LR, EBF and LR model, respectively. This study is helpful to flood management program, dissection makers and planning in local administrative level.

Spatial flood susceptibility prediction in Middle Ganga Plain: comparison of frequency ratio and Shannon’s entropy models

This work focuses on comparing results of flood susceptibility modelling in the part of Middle Ganga Plain, Ganga foreland basin. Following inclusivity rule, 12 major flood explanatory factors including a new one, geomorphology, have been utilized. Out of 1000 randomly generated flood-points from 2008 Landsat 5 TM image derived flood polygon, 70% have been utilized for the training purpose of Shannon’s entropy (SE) model and 30% for area under receiver operating characteristic (AUROC) method of validation of both, SE and frequency ratio (FR), models. Result from FR shows that the contributions of specific classes of different explanatory factors to flooding susceptibility vary whereas the SE model suggests that geomorphology is the most contributing factor. The AUROC curve for SE (0.90) was better than that for FR (0.85). Hence, the SE model predicts flood-susceptible areas more accurately than the multivariate statistical model FR in the study area.

Modeling flood susceptibility using data-driven approaches of naïve Bayes tree, alternating decision tree, and random forest methods

Floods are one of the most devastating types of disasters that cause loss of lives and property worldwide each year. This study aimed to evaluate and compare the prediction capability of the naïve Bayes tree (NBTree), alternating decision tree (ADTree), and random forest (RF) methods for the spatial prediction of flood occurrence in the Quannan area, China. A flood inventory map with 363 flood locations was produced and partitioned into training and validation datasets through random selection with a ratio of 70/30. The spatial flood database was constructed using thirteen flood explanatory factors. The probability certainty factor (PCF) method was used to analyze the correlation between the factors and flood occurrences. Consequently, three flood susceptibility maps were produced using the NBTree, ADTree, and RF methods. Finally, the area under the curve (AUC) and statistical measures were used to validate the flood susceptibility models. The results indicated that the RF method is an efficient and reliable model in flood susceptibility assessment, with the highest AUC values, positive predictive rate, negative predictive rate, sensitivity, specificity, and accuracy for the training (0.951, 0.892, 0.941, 0.945, 0.886, and 0.915, respectively) and validation (0.925, 0.851, 0.938, 0.945, 0.835, and 0.890, respectively) datasets.

Spatial prediction of flood potential using new ensembles of bivariate statistics and artificial intelligence: A case study at the Putna river catchment of Romania.

Flash-flood is considered to be one of the most destructive natural hazards in the world, which is difficult to accurately model and predict. The objective of the present research is to propose new ensembles of bivariate statistics and artificial intelligences and to introduce a comprehensive methodology for predicting flood susceptibility. The Putna river catchment of Romania is selected as a case study. In this regard, a total of six ensemble models were proposed and verified: Multilayer Perceptron neural network-Frequency Ratio (MLP-FR), Multilayer Perceptron neural network -Weights of Evidence (MLP-WOE), Rotation Forest-Frequency Ratio (RF-FR), Rotation Forest-Weights of Evidence (RF-WOE), Classification and Regression Tree-Frequency Ratio (CART-FR), and Classification and Regression Tree-Weights of Evidence (CART-WOE). In a first step, a geospatial database was created for the study area. This database includes 132 flood locations and 14 conditioning factors (lithology, slope angle, plan curvature, hydrological soil group, topographic wetness index, landuse, convergence index, elevation, distance from river, profile curvature, rainfall, aspect, stream power index, and topographic position index). In the next step, the Information Gain Ratio was used to evaluate the predictive ability of these factors. Subsequently, the database was used to train and validate the six ensemble models. The Receiver operating characteristic (ROC) curve, area under the curve (AUC), and statistical measures were used to evaluate the performance of the models. The results show that the prediction capability of the proposed ensemble models varied from 86.8% (the RF-FR model) to 93.9% (the RF-WOE model). These values indicate a high prediction performance for all the models. Therefore, we can state that the proposed ensemble models are new reliable tools which can be used for flood susceptibility modelling.

Flood susceptibility modeling and hazard perception in Rwanda

Flooding is a deleterious phenomenon that induces detrimental impacts on humans, properties and environment. As a result, the knowledge of susceptible places and hazard perception is increasingly pertinent. This study mainly aims at identifying areas susceptible to flood through the application of logistic regression model using remote sensing data (RS) and Geographical Information System (GIS). A flood inventory was generated using 153 historical flood locations and a total of 10 predicting factors (elevation, slope, aspect, profile curvature, distance from rivers, distance from roads, the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Soil Index (NDSI), the Topographic Wetness Index (TWI) and rainfall) were utilized. Flood points were randomly subdivided into training (75%) for model building and testing (25%) points for validation through the Area Under Curve (AUC) approach. The results indicated that NDVI and rainfall are the most influencing variables for estimating flood risk as they showed a high positive relationship with flood occurrence in the study area. Testing datasets disclosed 79.8% of prediction rate using the AUC. Moreover, the results have been linked with community perception on flood and the outcome revealed that the government is perceived as responsible for all flood mitigation measures instead of being a shared responsibility. This perception may contribute to the increase in susceptibility. The results of this study will be essential for upcoming development projects from different organizations operating in many developing countries and would assist as a baseline for flood risk reduction and management especially for Rwanda.

Flood susceptibility modelling using novel hybrid approach of reduced-error pruning trees with bagging and random subspace ensembles

Flooding is a very common natural hazard that causes catastrophic effects worldwide. Recently, ensemble-based techniques have become popular in flood susceptibility modelling due to their greater strength and efficiency in the prediction of flood locations. Thus, the aim of this study was to employ machine learning-based Reduced-error pruning trees (REPTree) with Bagging (Bag-REPTree) and Random subspace (RS-REPTree) ensemble frameworks for spatial prediction of flood susceptibility using a geographic information system (GIS). First, a flood spatial database was constructed with 363 flood locations and thirteen flood influencing factors, namely altitude, slope angle, slope aspect, curvature, stream power index (SPI), sediment transport index (STI), topographic wetness index (TWI), distance to rivers, normalized difference vegetation index (NDVI), soil, land use, lithology, and rainfall. Subsequently, correlation attribute evaluation (CAE) was used as the factor selection method for optimization of input factors. Finally, the receiver operating characteristic (ROC) curve, standard error (SE), confidence interval (CI) at 95%, and Wilcoxon signed-rank test were used to validate and compare the performance of the models. Results show that the RS-REPTree model has the highest prediction capability for flood susceptibility assessment, with the highest area under (the ROC) curve (AUC) value (0.949, 0.907), the smallest SE (0.011, 0.023), and the narrowest CI (95%) (0.928–0.970, 0.863–0.952) for the training and validation datasets. It was followed by the Bag-REPTree and REPTree models, respectively. The results also proved the superiority of the ensemble method over using these methods individually.

Flash flood susceptibility modelling using geomorphometric approach in the Ushairy Basin, eastern Hindu Kush

This study focuses on flash flood susceptibility modelling using geomorphometric ranking approach in the Ushairy Basin. In the study area, flash floods are highly unpredictable and the worst hydrometeorological disaster. An advanced spaceborne thermal emission and reflection radiometer global digital elevation model was used as input data in a geographic information system environment to delineate the target basin. A total of 17 sub-basins were delimited using a threshold of $$4 \hbox {km}^{2}$$ . The attribute information of each sub-basin was analysed to compute the geomorphometric parameters by applying Hortonian and Strahler geomorphological models. The results were analysed and categorised into five classes using statistical techniques, and the rank score was assigned to each class of all parameters depending on their relation with flash flood risk. In this study, 16 parameters were analysed to quantify the geomorphometric number of each sub-basin depicting the degree of flash flood susceptibility. The geomorphometric number of each sub-basin was linked to the geo-database for spatial visualisation. The analysis reveals that extremely high, very high, high and moderate sub-basins susceptible to flash floods were spread over an area of 55%, 8.5%, 23.7%, and 11.5%, respectively. It was found that out of total settlements, 53% are located in the extremely highly and very highly susceptible sub-basins. In the study area, the upper reaches are characterised by snow-covered peaks, steep slopes and high drainage densities ( $${>}1.7 \hbox { km/km}^{2})$$ . The analysis further indicated that the flash flood susceptibility increases with the increase in area, relief and relief ratio of the sub-basins. Model accuracy was assessed using primary data regarding past flood damages and human fatalities. Similarly, socio-demographic conditions of each sub-basin were also compared and linked to the extent of flash flood susceptibility. This study may assist the district government and district disaster management authority of Dir upper to initiate flood risk reduction strategies in highly susceptible zones of the Ushairy Basin.

A comparative assessment of flood susceptibility modeling using Multi-Criteria Decision-Making Analysis and Machine Learning Methods

Floods around the world are having devastating effects on human life and property. In this paper, three Multi-Criteria Decision-Making (MCDM) analysis techniques (VIKOR, TOPSIS and SAW), along with two machine learning methods (NBT and NB), were tested for their ability to model flood susceptibility in one of China’s most flood-prone areas, the Ningdu Catchment. Twelve flood conditioning factors were used as input parameters: Normalized Difference Vegetation Index (NDVI), lithology, land use, distance from river, curvature, altitude, Stream Transport Index (STI), Topographic Wetness Index (TWI), Stream Power Index (SPI), soil type, slope and rainfall. The predictive capacity of the models was evaluated and validated using the Area Under the Receiver Operating Characteristic curve (AUC). While all models showed a strong flood prediction capability (AUC > 0.95), the NBT model performed best (AUC = 0.98), suggesting that, among the models studied, the NBT model is a promising tool for the assessment of flood-prone areas and can allow for proper planning and management of flood hazards.

Flash flood susceptibility modeling using an optimized fuzzy rule based feature selection technique and tree based ensemble methods

The main objective of the present study was to provide a novel methodological approach for flash flood susceptibility modeling based on a feature selection method (FSM) and tree based ensemble methods. The FSM, used a fuzzy rule based algorithm FURIA, as attribute evaluator, whereas GA were used as the search method, in order to obtain optimal set of variables used in flood susceptibility modeling assessments. The novel FURIA-GA was combined with LogitBoost, Bagging and AdaBoost ensemble algorithms. The performance of the developed methodology was evaluated at the Bao Yen district and the Bac Ha district of Lao Cai Province in the Northeast region of Vietnam. For the case study, 654 floods and twelve geo-environmental variables were used. The predictive performance of each model was estimated through the calculation of the classification accuracy, the sensitivity, the specificity, the success and predictive rate curve and the area under the curves (AUC). The FURIA-GA FSM compared to a conventional rule based method gave more accurate predictive results. Also, the FURIA-GA based models, presented higher learning and predictive ability compared to the ensemble models that had not undergone a FSM. Based on the predictive classification accuracy, FURIA-GA-Bagging (93.37%) outperformed FURIA-GA-LogitBoost (92.35%) and FURIA-GA-AdaBoost (89.03%). FURIA-GA-Bagging showed also the highest sensitivity (96.94%) and specificity (89.80%). On the other hand, the FURIA-GA-LogitBoost showed the lowest percentage in very high susceptible zone and the highest relative flash-flood density, whereas the FURIA-GA-AdaBoost achieved the highest prediction AUC value (0.9740), based on the prediction rate curve, followed by FURIA-GA-Bagging (0.9566), and FURIA-GA-LogitBoost (0.8955). It can be concluded that the usage of different statistical metrics, provides different outcomes concerning the best prediction model, which mainly could be attributed to sites specific settings. The proposed models could be considered as a novel alternative investigation tools appropriate for flash flood susceptibility mapping.