Flood susceptibility mapping and assessment using a novel deep learning model combining multilayer perceptron and autoencoder neural networks

Journal of Flood Risk Management, Год журнала: 2020, Номер 14(1)

Опубликована: Дек. 18, 2020

Abstract Floods are one of the most destructive natural disasters causing financial damages and casualties every year worldwide. Recently, combination data‐driven techniques with remote sensing (RS) geographical information systems (GIS) has been widely used by researchers for flood susceptibility mapping. This study presents a novel hybrid model combining multilayer perceptron (MLP) autoencoder models to produce maps two areas located in Iran India. For cases, nine, twelve factors were considered as predictor variables mapping, respectively. The prediction capability proposed was compared that traditional MLP through area under receiver operating characteristic (AUROC) criterion. AUROC curve autoencoder‐MLP were, respectively, 75 90, 74 93% training phase 60 91, 81 97% testing phase, India results suggested outperformed and, therefore, can be powerful other studies

Язык: Английский

---

Flood hazards susceptibility mapping using statistical, fuzzy logic, and MCDM methods

Soft Computing, Год журнала: 2021, Номер 25(14), С. 9325 - 9346

Опубликована: Май 26, 2021

Язык: Английский

---

Examining LightGBM and CatBoost models for wadi flash flood susceptibility prediction

Geocarto International, Год журнала: 2021, Номер 37(25), С. 7462 - 7487

Опубликована: Авг. 31, 2021

This study presents two machine learning models, namely, the light gradient boosting (LightGBM) and categorical (CatBoost), for first time predicting flash flood susceptibility (FFS) in Wadi System (Hurghada, Egypt). A inventory map with 445 sites was produced randomly divided into groups training (70%) testing (30%). Fourteen controlling factors were selected evaluated their relative importance occurrence prediction. The performance of models assessed using various indexes comparison to common random forest (RF) method. results show areas under receiver operating characteristic curves (AUROC) above 97% all that LightGBM outperforms other terms classification metrics processing time. developed FFS maps demonstrate highly populated are most susceptible floods. present proves employed algorithms (LightGBM CatBoost) can be efficiently used mapping.

Язык: Английский

---

DEM resolution effects on machine learning performance for flood probability mapping

Journal of Hydro-environment Research, Год журнала: 2021, Номер 40, С. 1 - 16

Опубликована: Ноя. 9, 2021

Floods are among the devastating natural disasters that occurred very frequently in arid regions during last decades. Accurate assessment of flood susceptibility mapping is crucial sustainable development. It helps respective authorities to prevent as much possible their irreversible consequences. The Digital Elevation Model (DEM) spatial resolution one most base layer factors for modeling Flood Probability Maps (FPMs). Therefore, main objective this study was assess influence DEMs 12.5 m (ALOS PALSAR) and 30 (ASTER) on accuracy probability prediction using three machine learning models (MLMs), including Random Forest (RF), Artificial Neural Network (ANN), Generalized Linear (GLM). This selected 14 causative independent variables, 220 locations were dependent variables. Dependent variables divided into training (70%) validation (30%) modeling. Receiver Operating Characteristic Curve (ROC), Kappa index, accuracy, other statistical criteria used evaluate models' accuracy. results showed resolving DEM alone cannot significantly affect regardless applied MLM independently model performance In contrast, such altitude, precipitation, distance from river have a considerable impact floods region. Also, evaluation RF (AUC12.5,30m = 0.983, 0.975) more accurate preparing FPM than ANN 0.949, 0.93) GLM 0.965, 0.949) models. study's solution-oriented findings might help water managers decision-makers make effective adaptation mitigation measures against potential flooding.

Язык: Английский

---

Field based index of flood vulnerability (IFV): A new validation technique for flood susceptible models

Geoscience Frontiers, Год журнала: 2021, Номер 12(5), С. 101175 - 101175

Опубликована: Фев. 23, 2021

The flood hazard management is one of the major challenges in floodplain regions worldwide. With rise population growth and spread infrastructural development, level risk has increased over time. Therefore, prediction susceptible area a key challenge for adoption plans. Flood susceptibility modeling technically common work, but it still very tough job to validate models rigorous scientific manner. present work Atreyee River Basin India Bangladesh was planned establish artificial neural network (ANN), radial basis function (RBF), random forest (RF) their ensemble-based models. were constructed based on nine conditioning parameters. validated conventional way using receiver operating curve (ROC). To flood-susceptible models, two dimensional (2D) hydraulic simulation model developed. Also, index vulnerability developed applied validating which unique predictive Friedman test Wilcoxon Signed rank employed compare generated Results showed that 11.95%–12.99% entire basin (10188.4 km2) comes under high zones. Accuracy evaluation results have shown performance ensemble outperforms other standalone machine learning IFV also spatially adjusted with study recommended validation along ways.

Язык: Английский

---

Spatial Prediction of Future Flood Risk: An Approach to the Effects of Climate Change

Geosciences, Год журнала: 2021, Номер 11(1), С. 25 - 25

Опубликована: Янв. 5, 2021

Preparation of a flood probability map serves as the first step in management program. This research develops for floods resulting from climate change future. Two models Flexible Discrimination Analysis (FDA) and Artificial Neural Network (ANN) were used. optimistic (RCP2.6) pessimistic (RCP8.5) scenarios considered mapping future rainfall. Moreover, to produce occurrence maps, 263 locations past events used dependent variables. The number 13 factors conditioning was taken independent variables modeling. Of total locations, 80% (210 locations) 20% (53 model training validation. Receiver Operating Characteristic (ROC) curve other statistical criteria validate models. Based on assessments validated models, FDA, with ROC-AUC = 0.918, standard error (SE 0.038), an accuracy 0.86% compared ANN 0.897, has highest preparing study area. modeling results also showed that distance River, altitude, slope, rainfall have greatest impact Both models’ susceptibility maps area is related very low class. lowest high

Язык: Английский

---

Risk Assessment of Resources Exposed to Rainfall Induced Landslide with the Development of GIS and RS Based Ensemble Metaheuristic Machine Learning Algorithms

Sustainability, Год журнала: 2021, Номер 13(2), С. 457 - 457

Опубликована: Янв. 6, 2021

Disastrous natural hazards, such as landslides, floods, and forest fires cause a serious threat to resources, assets human lives. Consequently, landslide risk assessment has become requisite for managing the resources in future. This study was designed develop four ensemble metaheuristic machine learning algorithms, grey wolf optimized based artificial neural network (GW-ANN), random (GW-RF), particle swarm optimization ANN (PSO-ANN), PSO RF modeling rainfall-induced susceptibility (LS) Aqabat Al-Sulbat, Asir region, Saudi Arabia, which observes frequently. To obtain very high precision robust prediction from algorithms were integrated new techniques. Subsequently, LS maps produced by training dataset validated using receiver operating characteristics (ROC) curve on testing dataset. Based area under (AUC) value of ROC curve, best method selected. We developed curve-based sensitivity analysis investigate influence parameters modeling. The Gumble extreme distribution employed estimate rainfall at 2, 5, 10, 20, 50, 100 year return periods. Then, hazard prepared different periods integrating model estimated theory danger pixels prepare final have been exposed landslide. results showed that 27–42 6–15 km2 predicted zones algorithms. ROC, GR-ANN (AUC-0.905) appeared areas gradually increased over progression time (26 2 period 40 zone, 6 20 zone). Similarly, pixel also (37 62 km2). Various scrubland, built up, sparse vegetation, identified zone due hazards. In addition, these would be extensively landslides advancement Therefore, outcome present will help planners scientists propose management plans protecting landslides.

Язык: Английский

---

Flood susceptibility modeling based on new hybrid intelligence model: Optimization of XGboost model using GA metaheuristic algorithm

Advances in Space Research, Год журнала: 2022, Номер 69(9), С. 3301 - 3318

Опубликована: Фев. 22, 2022

Язык: Английский

---

Flood hazard zone mapping incorporating geographic information system (GIS) and multi-criteria analysis (MCA) techniques

Journal of Hydrology, Год журнала: 2022, Номер 612, С. 128268 - 128268

Опубликована: Июль 30, 2022

Язык: Английский

---

Revisiting 2013 Uttarakhand flash floods through hydrological evaluation of precipitation data sources and morphometric prioritization

Geomatics Natural Hazards and Risk, Год журнала: 2022, Номер 13(1), С. 646 - 666

Опубликована: Фев. 17, 2022

With advancements in computational technology, data assimilation techniques, high-resolution remote sensing, and complex climate models, numerous precipitation products are available with different spatiotemporal resolutions; however, their evaluation, especially the Himalayan region, is unexplored. Therefore, this study attempts to assess four sources (gridded observation dataset, reanalysis, satellite, numerical weather prediction models) of through hydrological modelling for catastrophic 2013 floods Uttarakhand, India. The Upper Ganga Basin located Western Himalayas selected as area consisting Alaknanda Bhagirathi streams eastern western parts. Hydrologic Engineering Center's Modeling System (HEC-HMS) employed rainfall-runoff modelling. rainfall from IMD, ERA-5, GPM-IMERG-Final, WRF model outputs forced into calibrated HEC-HMS assessing performance simulations. correlation coefficient simulations respect observed flow 0.89, 0.88, 0.55, respectively, whereas corresponding Modified Kling-Gupta Efficiency (KGE) 0.66, 0.72, 0.48, 0.71. Flash flood prioritization sub-watersheds based on morphometric characteristics suggests that basin relatively more vulnerable flash due elongated nature, highest relative relief, high mean slope.

Язык: Английский

---