A PCA spatial pattern based artificial neural network downscaling model for urban flood hazard assessment

Carreau, Julie; Guinot, Vincent

doi:10.1016/j.advwatres.2020.103821

Cited by 23 publications

(25 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…EOF analysis can be used for downscaling data from low‐resolution to high‐resolution, thus making it appealing for use with low‐ and high‐fidelity flood inundation modeling. For this reason, Carreau and Guinot (2021) recently predicted high‐resolution water depths and discharge using a hybrid surrogate approach that combined a low‐resolution hydrodynamic model with Artificial Neural Network (ANN) emulator models to predict ECs from a high‐resolution hydrodynamic model. Carreau and Guinot (2021) demonstrated the value of using EOF analysis and emulator models to downscale the results from low‐fidelity models, and they obtained higher resolution predictions of water depth and discharge for flooding events in urban environments.…”

Section: Introductionmentioning

confidence: 99%

“…For this reason, Carreau and Guinot (2021) recently predicted high‐resolution water depths and discharge using a hybrid surrogate approach that combined a low‐resolution hydrodynamic model with Artificial Neural Network (ANN) emulator models to predict ECs from a high‐resolution hydrodynamic model. Carreau and Guinot (2021) demonstrated the value of using EOF analysis and emulator models to downscale the results from low‐fidelity models, and they obtained higher resolution predictions of water depth and discharge for flooding events in urban environments. They derived the “low‐fidelity model results” by averaging the high‐fidelity results over selected subdomains.…”

Section: Introductionmentioning

confidence: 99%

“…An emulator, such as the ANN used by Carreau and Guinot (2021), is well suited to describe the complex functional relationships that exists between the ECs. Nevertheless, in recent years a probabilistic treatment of predictions has increased in popularity and with it, interest in Gaussian Process (GP) models.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Upskilling Low‐Fidelity Hydrodynamic Models of Flood Inundation Through Spatial Analysis and Gaussian Process Learning

Fraehr

Wang

et al. 2022

Water Resources Research

View full text Add to dashboard Cite

Floods are some of the most destructive natural disasters in the world and they are projected to become more severe and frequent with climate change (IPCC, 2021). During a flood event normally dry areas are inundated until a maximum inundation extent is reached (flooding period), whereafter the water recedes back to the normal state (recession period). Capturing the dynamics of this behavior is of great importance for risk management and has led to the development of advanced hydrodynamic models. Hydrodynamic models can represent different levels of complexity and precision. For simulating the dynamics of flood inundation, two-dimensional hydrodynamic models that numerically solve the depth-averaged Navier-Stokes equations on a high-resolution grid is normally applied (Teng et al., 2017). These high-resolution hydrodynamic models are often referred to as high-fidelity models, where the fidelity refers to the model's degree of realism (Razavi et al., 2012). However,

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Upskilling Low‐Fidelity Hydrodynamic Models of Flood Inundation Through Spatial Analysis and Gaussian Process Learning

Fraehr

Wang

et al. 2022

Water Resources Research

View full text Add to dashboard Cite

show abstract

“…Flow modeling results are often used as input for complementary analyses, such as damage modeling, solute [46] or sediment transport and morphodynamic modeling. It is, therefore, necessary to assess whether the porosity models succeed in predicting, at the right scale, the flow variables needed for these complementary analyses and which postprocessing steps may be necessary [53].…”

Section: Directions For Further Researchmentioning

confidence: 99%

Porosity Models for Large-Scale Urban Flood Modelling: A Review

Dewals

Bruwier

Pirotton

et al. 2021

Water

View full text Add to dashboard Cite

In the context of large-scale urban flood modeling, porosity shallow-water models enable a considerable speed-up in computations while preserving information on subgrid topography. Over the last two decades, major improvements have been brought to these models, but a single generally accepted model formulation has not yet been reached. Instead, existing models vary in many respects. Some studies define porosity parameters at the scale of the computational cells or cell interfaces, while others treat the urban area as a continuum and introduce statistically defined porosity parameters. The porosity parameters are considered either isotropic or anisotropic and depth-independent or depth-dependent. The underlying flow models are based either on the full shallow-water equations or approximations thereof, with various flow resistance parameterizations. Here, we provide a review of the spectrum of porosity models developed so far for large-scale urban flood modeling.

show abstract

“…There is a considerable body of literature on regression models (see, among others, Cabral et al, 2019;Filgueiras et al, 2020;García-Nieto et al, 2020;Huang et al, 2020;Knappett et al, 2020;Zarei & Mahmoudi, 2020). Literature related to neural network includes Bomers et al (2019), Pyo et al (2020), Kumar et al (2020), Yang et al (2021) and Carreau & Guinot (2021), among many others.…”

Section: Introductionmentioning

confidence: 99%

Water demand prediction optimization method in Shenzhen based on the zero-sum game model and rolling revisions

et al. 2021

View full text Add to dashboard Cite

In this study, a deep learning model based on zero-sum game (ZSG) was proposed for accurate water demand prediction. The ensemble learning was introduced to enhance the generalization ability of models, and the sliding average was designed to solve the non-stationarity problem of time series. To solve the problem that the deep learning model could not predict water supply fluctuations caused by emergencies, a hypothesis testing method combining Student's t-test and discrete wavelet transform was proposed to generate the envelope interval of the predicted values to carry out rolling revisions. The research methods were applied to Shenzhen, a megacity with extremely short water resources. The research results showed that the regular bidirectional models were superior to the unidirectional model, and the ZSG-based bidirectional models were superior to the regular bidirectional models. The bidirectional propagation was conducive to improving the generalization ability of the model, and ZSG could better guide the model to find the optimal solution. The fluctuations in water supply were mainly caused by the floating population, but the fluctuation was still within the envelope interval of the predicted values. The predicted values after rolling revisions were very close to the measured values.

show abstract

A PCA spatial pattern based artificial neural network downscaling model for urban flood hazard assessment

Cited by 23 publications

References 46 publications

Upskilling Low‐Fidelity Hydrodynamic Models of Flood Inundation Through Spatial Analysis and Gaussian Process Learning

Upskilling Low‐Fidelity Hydrodynamic Models of Flood Inundation Through Spatial Analysis and Gaussian Process Learning

Porosity Models for Large-Scale Urban Flood Modelling: A Review

Water demand prediction optimization method in Shenzhen based on the zero-sum game model and rolling revisions

Contact Info

Product

Resources

About