Flood forecasting with machine learning models in an operational framework

Abstract

Abstract. Google's operational flood forecasting system was developed to provide accurate real-time flood warnings to agencies and the public with a focus on riverine floods in large, gauged rivers. It became operational in 2018 and has since expanded geographically. This forecasting system consists of four subsystems: data validation, stage forecasting, inundation modeling, and alert distribution. Machine learning is used for two of the subsystems. Stage forecasting is modeled with the long short-term memory (LSTM) networks and the linear models. Flood inundation is computed with the thresholding and the manifold models, where the former computes inundation extent and the latter computes both inundation extent and depth. The manifold model, presented here for the first time, provides a machine-learning alternative to hydraulic modeling of flood inundation. When evaluated on historical data, all models achieve sufficiently high-performance metrics for operational use. The LSTM showed higher skills than the linear model, while the thresholding and manifold models achieved similar performance metrics for modeling inundation extent. During the 2021 monsoon season, the flood warning system was operational in India and Bangladesh, covering flood-prone regions around rivers with a total area close to 470 000 km2, home to more than 350 000 000 people. More than 100 000 000 flood alerts were sent to affected populations, to relevant authorities, and to emergency organizations. Current and future work on the system includes extending coverage to additional flood-prone locations and improving modeling capabilities and accuracy.

Improving hydrologic models for predictions and process understanding using neural ODEs

“…Machine learning methods provide great versatility (Shen, 2018;Shen et al, 2018;Reichstein et al, 2019) and have demonstrated unprecedented accuracy in various modelling tasks like predictions in ungauged basins (PUB; e.g. Kratzert et al, 2019b;Prieto et al, 2019), transfer learning to data-scarce regions (Ma et al, 2021) or flood forecasting (Frame et al, 2022;Nevo et al, 2022). Nonetheless, deep learning remains a field of progress with gaps to fill.…”

Machine learning applications for weather and climate need greater focus on extremes

“…Five out of six studies evaluated neural network-based models, indicating that neural networks can be successful for this task. The other study, Nevo et al (2022), used a bespoke approach for their flood inundation modelling. No study tested alternative complex methods like random forests, so they provide no evidence about such methods.…”

“…To show how well ML-based systems perform in situations going beyond events seen in training, the most extreme events can be set aside in a second test dataset, as in Frame et al (2022) and Nevo et al (2022). This approach could be made even stronger by doing this before any model development is done, so the model structure and hyperparameters are chosen without being able to see the most extreme events beforehand.…”

Development of an Artificial Neural Network Algorithm Embedded in an On-Site Sensor for Water Level Forecasting

“…Second, to increase the lead time of the response, physically based models are used in the traditional framework for predictions. These systems consider different hydrological and inundation modeling components based on the specific target region, size of basins, available data and resources, and system development approach [ 19 ]. These physically based models considering detailed hydraulic processes (e.g., solving Saint-Venant equations) are complex and computationally intensive, causing limited applications in practical applications due to the availability of input data for parameterization and the detailed requirements of simulations [ 20 , 21 , 22 ].…”