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

Abstract

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

FFM: Flood Forecasting Model Using Federated Learning

“…The Hydraulic models of flow have been practiced forecasting rainfall, storms and tsunamis [36]. These models are also used to predict impact of climatic change [37], ocean waves [38] and floods [39]. It was observed that long term prediction models were more successful than short term prediction models in hazard prediction [40] [41].…”

Flood Susceptibility Mapping Using Image-Based 2D-CNN Deep Learning: Overview and Case Study Application Using Multiparametric Spatial Data in Data-Scarce Urban Environments

“…Methods based on numerical modelling and multicriteria decision-analysis (MCDA) that integrate GIS and remote sensing have been proposed for zoning of food vulnerability or susceptibility maps within urban areas [8,9]. MCDA approaches, however, have the disadvantage that they rely on subjective expert knowledge and do not apply food event data, and therefore their outputs are likely to be biased and prone to error [10]. On the other hand, hydrodynamic numerical models are intrinsically limited by the discretization of the physical domain of the watershed and the governing physical equations.…”

“…ANNs have high prediction ability but cannot determine the infuence and role of the food causative factors [30]. Due to the shortcomings of the standalone algorithms, hybrid ML algorithms have been proposed for improved food susceptibility mapping [10].…”

“…Deep learning algorithms ofer advantages such as enabling adequate training of large datasets, performing unsupervised and semisupervised learning, and the ability to resolve complex and nonlinear relationships. Several investigations have introduced DL-based algorithms for food susceptibility mapping, with Convolutional Neural Networks (CNN) as the most preferred DL approach [10,13,27,35,36]. Compared with other DL algorithms, CNN utilizes the information from adjacent pixels of the input image (raster) with much fewer trainable parameters and is, therefore, extremely suitable for solving 2-D imagebased problems.…”

“…Compared with other DL algorithms, CNN utilizes the information from adjacent pixels of the input image (raster) with much fewer trainable parameters and is, therefore, extremely suitable for solving 2-D imagebased problems. Furthermore, due to the spatial inductive bias in CNNs, they are best suited for processing the 2Draster data used in food analysis [10]. Wang et al [35] indicated that CNNs can get rid of the tedious feature engineering steps and extract useful information directly from the original data, and they can play an important role in FSM.…”