Deep learning methods for flood mapping: a review of existing applications and future research directions

Hydrology and earth system sciences, Journal Year: 2022, Volume and Issue: 26(16), P. 4345 - 4378

Published: Aug. 25, 2022

Abstract. Deep learning techniques have been increasingly used in flood management to overcome the limitations of accurate, yet slow, numerical models and improve results traditional methods for mapping. In this paper, we review 58 recent publications outline state art field, identify knowledge gaps, propose future research directions. The focuses on type deep various mapping applications, types considered, spatial scale studied events, data model development. show that based convolutional layers are usually more as they leverage inductive biases better process characteristics flooding events. Models fully connected layers, instead, provide accurate when coupled with other statistical models. showed increased accuracy compared approaches speed methods. While there exist several applications susceptibility, inundation, hazard mapping, work is needed understand how can assist real-time warning during an emergency it be employed estimate risk. A major challenge lies developing generalize unseen case studies. Furthermore, all reviewed their outputs deterministic, limited considerations uncertainties outcomes probabilistic predictions. authors argue these identified gaps addressed by exploiting fundamental advancements or taking inspiration from developments applied areas. graph neural networks operators arbitrarily structured thus should capable generalizing across different studies could account complex interactions natural built environment. Physics-based preserve underlying physical equations resulting reliable speed-up alternatives Similarly, resorting Gaussian processes Bayesian networks.

Language: Английский

---

Towards better flood risk management: Assessing flood risk and investigating the potential mechanism based on machine learning models

Journal of Environmental Management, Journal Year: 2021, Volume and Issue: 293, P. 112810 - 112810

Published: May 21, 2021

Language: Английский

---

Novel ensemble machine learning models in flood susceptibility mapping

Geocarto International, Journal Year: 2021, Volume and Issue: 37(16), P. 4571 - 4593

Published: Feb. 19, 2021

The research aims to propose the new ensemble models by combining machine learning techniques, such as rotation forest (RF), nearest shrunken centroids (NSC), k-nearest neighbour (KNN), boosted regression tree (BRT), and logitboost (LB) with base classifier adabag (AB) for flood susceptibility mapping (FSM). proposed were implemented in central west coast of India, which is vulnerable events. For inventory mapping, a total 210 localities identified. Twelve effective factors selected using boruta algorithm FSM. area under receiver operating characteristics (AUROC) curve other statistical measures (sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), absolute (MAE)) employed estimate compare success rate approaches. validation results individual terms AUC value AB (92.74%) >RF (91.50%) >BRT (90.75%) >LB (89.07%) >NSC (88.97%) >KNN (83.88%), whereas showed that AB-RF (94%) was highest prediction efficiency followed by, AB-KNN (93.33%), AB-NSC (93.02%), AB-LB (92.83%), AB-BRT (92.64%). outcomes established more appropriate increase accuracy different single models. Therefore, this study can be useful proper planning management hazard alike geographic environment.

Language: Английский

---

Flash-flood hazard using deep learning based on H2O R package and fuzzy-multicriteria decision-making analysis

Journal of Hydrology, Journal Year: 2022, Volume and Issue: 609, P. 127747 - 127747

Published: March 24, 2022

Language: Английский

---

A novel flood risk management approach based on future climate and land use change scenarios

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 921, P. 171204 - 171204

Published: Feb. 23, 2024

Language: Английский

---

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

Soft Computing, Journal Year: 2021, Volume and Issue: 25(14), P. 9325 - 9346

Published: May 26, 2021

Language: Английский

---

Development of flood hazard map and emergency relief operation system using hydrodynamic modeling and machine learning algorithm

Journal of Cleaner Production, Journal Year: 2021, Volume and Issue: 311, P. 127594 - 127594

Published: May 27, 2021

Language: Английский

---

Application of GIS and Machine Learning to Predict Flood Areas in Nigeria

Sustainability, Journal Year: 2022, Volume and Issue: 14(9), P. 5039 - 5039

Published: April 22, 2022

Floods are one of the most devastating forces in nature. Several approaches for identifying flood-prone locations have been developed to reduce overall harmful impacts on humans and environment. However, due increased frequency flooding related disasters, coupled with continuous changes natural social-economic conditions, it has become vital predict areas highest probability ensure effective measures mitigate impending disasters. This study predicted flood susceptible Nigeria based historical records from 1985~2020 various conditioning factors. To evaluate link between incidence fifteen (15) explanatory variables, which include climatic, topographic, land use proximity information, artificial neural network (ANN) logistic regression (LR) models were trained tested develop a susceptibility map. The receiver operating characteristic curve (ROC) area under (AUC) used both model accuracies. results show that techniques can areas. ANN produced higher performance prediction rate than LR model, 76.4% 62.5%, respectively. In addition, highlighted those low-lying regions southern extremities around water From study, we establish machine learning effectively map serve as tool developing mitigation policies plans.

Language: Английский

---

Flood susceptibility mapping of Northeast coastal districts of Tamil Nadu India using Multi-source Geospatial data and Machine Learning techniques

Geocarto International, Journal Year: 2022, Volume and Issue: 37(27), P. 15252 - 15281

Published: June 30, 2022

Flooding is one of the most challenging and important natural disasters to predict, it becoming more frequent intense. The study area badly damaged by devastating flood in 2015. We assessed susceptibility northern coastal Tamil Nadu using various machine learning algorithms such as Gradient Boosting Machine (GBM), XGBoost (XGB), Rotation Forest (RTF), Support Vector (SVM), Naive Bayes (NB). Google Earth Engine (GEE) used demarcate flooded areas Sentinel-l other multi-source geospatial data generate influential factors. Recursive Feature Elimination (RFE) removes weak factors this study. resultant map classified into five classes: very low, moderate, high, high. GBM algorithm attained high classification accuracy with an under curve (AUC) value 92%. urbanized vulnerable identifying inundation useful for effective planning implementation.

Language: Английский

---

A novel framework for addressing uncertainties in machine learning-based geospatial approaches for flood prediction

Journal of Environmental Management, Journal Year: 2022, Volume and Issue: 326, P. 116813 - 116813

Published: Nov. 23, 2022

Globally, many studies on machine learning (ML)-based flood susceptibility modeling have been carried out in recent years. While majority of those models produce reasonably accurate predictions, the outcomes are subject to uncertainty since (FSMs) may varying spatial predictions. However, there not attempts address these uncertainties because identifying agreement projections is a complex process. This study presents framework for reducing disagreement among four standalone and hybridized ML-based FSMs: random forest (RF), k-nearest neighbor (KNN), multilayer perceptron (MLP), genetic algorithm-gaussian radial basis function-support vector regression (GA-RBF-SVR). Besides, an optimized model was developed combining models. The southwest coastal region Bangladesh selected as case area. A comparable percentage potential area (approximately 60% total land areas) produced by all Despite achieving high prediction accuracy, discrepancy observed, with pixel-wise correlation coefficients across different ranging from 0.62 0.91. exhibited accuracy improved number classification errors. presented this might aid formulation risk-based development plans enhancement current early warning systems.

Language: Английский

---