Search

Flash floods pose a major challenge in various regions of the world, causing serious damage to life and property. Here we investigated the Izvorul Dorului river basin from Romania, to identify slope surfaces with a high potential for flash-flood employing a combination of fuzzy logic algorithm with the following four machine learning models: classification and regression tree, deep learning neural network, XGBoost and naïve Bayes. Ten flash-flood predictors were used as independent variables to determine the flash-flood potential index. As a dependent variable, we used areas with ttorrential phenomena divided into training (70%) and validating data set (30%). Predictive ability and the degree of correlation between factors were assessed through the correlation-based feature selection (CFS) method and through the confusion matrix, respectively. In the training phase, all ensemble models yielded good and very good accuracies of over 84%. The spatialization of flash-flood potential index (FFPI) over the study area showed that high and very high values of flash-flood potential occur in the northern half of the region and occupy the following weights within the study area: 53.11% (FFPI Fuzzy-CART), 45.09% (Fuzzy-DLNN), 45.58% (Fuzzy-NB) and 44.85% (Fuzzy-XGBoost). The validation of the results was done through the ROC curve method. Thus, according to success rate, Fuzzy-XGBoost (AUC = 0.886) is the best model, while in terms of prediction rate, the ideal one is Fuzzy-DLNN (AUC = 0.84). The novelty of this work is the application of the four ensemble models in evaluating this natural hazard. [ABSTRACT FROM AUTHOR]

There is an evident increase in the importance that remote sensing sensors play in the monitoring and evaluation of natural hazards susceptibility and risk. The present study aims to assess the flash-flood potential values, in a small catchment from Romania, using information provided remote sensing sensors and Geographic Informational Systems (GIS) databases which were involved as input data into a number of four ensemble models. In a first phase, with the help of high-resolution satellite images from the Google Earth application, 481 points affected by torrential processes were acquired, another 481 points being randomly positioned in areas without torrential processes. Seventy percent of the dataset was kept as training data, while the other 30% was assigned to validating sample. Further, in order to train the machine learning models, information regarding the 10 flash-flood predictors was extracted in the training sample locations. Finally, the following four ensembles were used to calculate the Flash-Flood Potential Index across the Bâsca Chiojdului river basin: Deep Learning Neural Network–Frequency Ratio (DLNN-FR), Deep Learning Neural Network–Weights of Evidence (DLNN-WOE), Alternating Decision Trees–Frequency Ratio (ADT-FR) and Alternating Decision Trees–Weights of Evidence (ADT-WOE). The model's performances were assessed using several statistical metrics. Thus, in terms of Sensitivity, the highest value of 0.985 was achieved by the DLNN-FR model, meanwhile the lowest one (0.866) was assigned to ADT-FR ensemble. Moreover, the specificity analysis shows that the highest value (0.991) was attributed to DLNN-WOE algorithm, while the lowest value (0.892) was achieved by ADT-FR. During the training procedure, the models achieved overall accuracies between 0.878 (ADT-FR) and 0.985 (DLNN-WOE). K-index shows again that the most performant model was DLNN-WOE (0.97). The Flash-Flood Potential Index (FFPI) values revealed that the surfaces with high and very high flash-flood susceptibility cover between 46.57% (DLNN-FR) and 59.38% (ADT-FR) of the study zone. The use of the Receiver Operating Characteristic (ROC) curve for results validation highlights the fact that FFPIDLNN-WOE is characterized by the most precise results with an Area Under Curve of 0.96. [ABSTRACT FROM AUTHOR]