Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Attention is all you need for an improved CNN-based flash flood susceptibility modeling. The case of the ungauged Rheraya watershed, Morocco

Akram Elghouat, Ahmed Algouti, Abdellah Algouti, Soukaina Baid

TL;DR

This study addresses the need for accurate flash flood susceptibility mapping in ungauged basins by integrating the convolutional block attention module (CBAM) into CNN backbones (ResNet18, DenseNet121, Xception). By evaluating CBAM placed at head, tail, or every block, the authors show that embedding CBAM in each convolutional block (In) yields the best performance, with DenseNet121+CBAM-In achieving an accuracy of 0.95 and an AUC of 0.98 on the test set. Using 16 conditioning factors over topographic, hydrological, meteorological, and environmental domains, and a 522-point inventory, the study also identifies distance to river and drainage density as the most influential predictors. The resulting flash flood susceptibility map highlights high-risk zones along rivers and in touristic areas, providing actionable insight for disaster management in the Rheraya watershed. The work demonstrates the practical value of attention mechanisms in improving DL-based hazard mapping and sets the stage for broader validation across basins and data sources.

Abstract

Effective flood hazard management requires evaluating and predicting flash flood susceptibility. Convolutional neural networks (CNNs) are commonly used for this task but face issues like gradient explosion and overfitting. This study explores the use of an attention mechanism, specifically the convolutional block attention module (CBAM), to enhance CNN models for flash flood susceptibility in the ungauged Rheraya watershed, a flood prone region. We used ResNet18, DenseNet121, and Xception as backbone architectures, integrating CBAM at different locations. Our dataset included 16 conditioning factors and 522 flash flood inventory points. Performance was evaluated using accuracy, precision, recall, F1-score, and the area under the curve (AUC) of the receiver operating characteristic (ROC). Results showed that CBAM significantly improved model performance, with DenseNet121 incorporating CBAM in each convolutional block achieving the best results (accuracy = 0.95, AUC = 0.98). Distance to river and drainage density were identified as key factors. These findings demonstrate the effectiveness of the attention mechanism in improving flash flood susceptibility modeling and offer valuable insights for disaster management.

Attention is all you need for an improved CNN-based flash flood susceptibility modeling. The case of the ungauged Rheraya watershed, Morocco

TL;DR

This study addresses the need for accurate flash flood susceptibility mapping in ungauged basins by integrating the convolutional block attention module (CBAM) into CNN backbones (ResNet18, DenseNet121, Xception). By evaluating CBAM placed at head, tail, or every block, the authors show that embedding CBAM in each convolutional block (In) yields the best performance, with DenseNet121+CBAM-In achieving an accuracy of 0.95 and an AUC of 0.98 on the test set. Using 16 conditioning factors over topographic, hydrological, meteorological, and environmental domains, and a 522-point inventory, the study also identifies distance to river and drainage density as the most influential predictors. The resulting flash flood susceptibility map highlights high-risk zones along rivers and in touristic areas, providing actionable insight for disaster management in the Rheraya watershed. The work demonstrates the practical value of attention mechanisms in improving DL-based hazard mapping and sets the stage for broader validation across basins and data sources.

Abstract

Effective flood hazard management requires evaluating and predicting flash flood susceptibility. Convolutional neural networks (CNNs) are commonly used for this task but face issues like gradient explosion and overfitting. This study explores the use of an attention mechanism, specifically the convolutional block attention module (CBAM), to enhance CNN models for flash flood susceptibility in the ungauged Rheraya watershed, a flood prone region. We used ResNet18, DenseNet121, and Xception as backbone architectures, integrating CBAM at different locations. Our dataset included 16 conditioning factors and 522 flash flood inventory points. Performance was evaluated using accuracy, precision, recall, F1-score, and the area under the curve (AUC) of the receiver operating characteristic (ROC). Results showed that CBAM significantly improved model performance, with DenseNet121 incorporating CBAM in each convolutional block achieving the best results (accuracy = 0.95, AUC = 0.98). Distance to river and drainage density were identified as key factors. These findings demonstrate the effectiveness of the attention mechanism in improving flash flood susceptibility modeling and offer valuable insights for disaster management.
Paper Structure (20 sections, 7 equations, 10 figures, 3 tables)

This paper contains 20 sections, 7 equations, 10 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Study area
  4. Flash flood inventory
  5. Flash conditioning factors
  6. Topographical factors
  7. Hydrological and meteorological factors
  8. Environmental and anthropogenic factors
  9. Methodology
  10. Factor optimization
  11. Models
  12. Model validation
  13. Sensitivity analysis
  14. Results
  15. Factor optimization
  16. ...and 5 more sections

Figures (10)

  • Figure 1: Locations of the study area and historical flash floods
  • Figure 2: Flash flood events and damages in the study area
  • Figure 3: (a) Elevation, (b) slope angle, (c) Curvature, (d) Aspect, (e) Distance to river, (f) TRI
  • Figure 4: (g) Convergence index, (h) Ruggedness index, (i) Flow accumulation, (j) Rainfall, (k) Drainge density, (l) SPI
  • Figure 5: (m) TWI, (n) NDVI, (o) Land cover, (p) Distance to roads
  • ...and 5 more figures