floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time
Using machine learning for pluvial flood prediction tasks has gained growing attention in the past years. In particular, data-driven models using artificial neuronal networks show promising results, shortening the computation times of physically based simulations. However, recent approaches have use...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-08-01
|
| Series: | Water |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4441/13/16/2255 |
| _version_ | 1797521715109036032 |
|---|---|
| author | Julian Hofmann Holger Schüttrumpf |
| author_facet | Julian Hofmann Holger Schüttrumpf |
| author_sort | Julian Hofmann |
| collection | DOAJ |
| description | Using machine learning for pluvial flood prediction tasks has gained growing attention in the past years. In particular, data-driven models using artificial neuronal networks show promising results, shortening the computation times of physically based simulations. However, recent approaches have used mainly conventional fully connected neural networks which were (a) restricted to spatially uniform precipitation events and (b) limited to a small amount of input data. In this work, a deep convolutional generative adversarial network has been developed to predict pluvial flooding caused by nonlinear spatial heterogeny rainfall events. The model developed, floodGAN, is based on an image-to-image translation approach whereby the model learns to generate 2D inundation predictions conditioned by heterogenous rainfall distributions—through the minimax game of two adversarial networks. The training data for the floodGAN model was generated using a physically based hydrodynamic model. To evaluate the performance and accuracy of the floodGAN, model multiple tests were conducted using both synthetic events and a historic rainfall event. The results demonstrate that the proposed floodGAN model is up to 10<sup>6</sup> times faster than the hydrodynamic model and promising in terms of accuracy and generalizability. Therefore, it bridges the gap between detailed flood modelling and real-time applications such as end-to-end early warning systems. |
| first_indexed | 2024-03-10T08:17:26Z |
| format | Article |
| id | doaj.art-48cfcacb559f41fc8b80753645bfd7c2 |
| institution | Directory Open Access Journal |
| issn | 2073-4441 |
| language | English |
| last_indexed | 2024-03-10T08:17:26Z |
| publishDate | 2021-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Water |
| spelling | doaj.art-48cfcacb559f41fc8b80753645bfd7c22023-11-22T10:14:59ZengMDPI AGWater2073-44412021-08-011316225510.3390/w13162255floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real TimeJulian Hofmann0Holger Schüttrumpf1Institute of Hydraulic Engineering and Water Resources Management, RWTH Aachen University, 57074 Aachen, GermanyInstitute of Hydraulic Engineering and Water Resources Management, RWTH Aachen University, 57074 Aachen, GermanyUsing machine learning for pluvial flood prediction tasks has gained growing attention in the past years. In particular, data-driven models using artificial neuronal networks show promising results, shortening the computation times of physically based simulations. However, recent approaches have used mainly conventional fully connected neural networks which were (a) restricted to spatially uniform precipitation events and (b) limited to a small amount of input data. In this work, a deep convolutional generative adversarial network has been developed to predict pluvial flooding caused by nonlinear spatial heterogeny rainfall events. The model developed, floodGAN, is based on an image-to-image translation approach whereby the model learns to generate 2D inundation predictions conditioned by heterogenous rainfall distributions—through the minimax game of two adversarial networks. The training data for the floodGAN model was generated using a physically based hydrodynamic model. To evaluate the performance and accuracy of the floodGAN, model multiple tests were conducted using both synthetic events and a historic rainfall event. The results demonstrate that the proposed floodGAN model is up to 10<sup>6</sup> times faster than the hydrodynamic model and promising in terms of accuracy and generalizability. Therefore, it bridges the gap between detailed flood modelling and real-time applications such as end-to-end early warning systems.https://www.mdpi.com/2073-4441/13/16/2255flood modellingmachine learningdeep learninggenerative adversarial networksreal-time flood forecasting |
| spellingShingle | Julian Hofmann Holger Schüttrumpf floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time Water flood modelling machine learning deep learning generative adversarial networks real-time flood forecasting |
| title | floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time |
| title_full | floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time |
| title_fullStr | floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time |
| title_full_unstemmed | floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time |
| title_short | floodGAN: Using Deep Adversarial Learning to Predict Pluvial Flooding in Real Time |
| title_sort | floodgan using deep adversarial learning to predict pluvial flooding in real time |
| topic | flood modelling machine learning deep learning generative adversarial networks real-time flood forecasting |
| url | https://www.mdpi.com/2073-4441/13/16/2255 |
| work_keys_str_mv | AT julianhofmann floodganusingdeepadversariallearningtopredictpluvialfloodinginrealtime AT holgerschuttrumpf floodganusingdeepadversariallearningtopredictpluvialfloodinginrealtime |