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

Fraehr, Niels; Wang, QJ; Wu, Wenyan; Nathan, Rory

doi:10.1029/2022wr032248

Cited by 23 publications

(53 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the following, we describe the details of the methodology of the LSG model for predicting water depth, and at the end of this section, we explain how the methodology differs from the one previously proposed by Fraehr et al (2022). have to be set up for the specific study area.…”

Section: Methodology Of the Lsg Modelmentioning

confidence: 99%

“…This has led to the concept of developing hybrid approaches that combine the benefits of multiple models whilst overcoming some of the limitations. One of the most recent hybrid models is the Lowfidelity, Spatial analysis, and Gaussian Process learning (LSG) model developed by Fraehr et al (2022). The LSG model accurately simulates the dynamic behaviour (e.g.…”

Section: Introductionmentioning

confidence: 99%

“…Although Fraehr et al (2022) demonstrated the good performance of the LSG model, the methodology was only applied to the simulation of inundation extent and timing of a flood event and not the water depth. Information on water depth is important to correctly represent the degree of hazard associated with a predicted inundation extent of a flood event (Hunter et al, 2007), and is a key indicator in warning systems and for estimating flood losses (Antony et al, 2021;Chang et al, 2019;Zischg et al, 2018).…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, in the study by Fraehr et al (2022) the LSG model was only applied to a coupled 1D-2D hydrodynamic model that had a structured quadratic grid (i.e. one which provided predictions at regular spacing throughout the model domain without increasing grid resolution in areas of interest).…”

Section: Introductionmentioning

confidence: 99%

“…The versatility of the LSG model is explored by applying the LSG model to a 2D hydrodynamic model with an unstructured grid to simulate flood inundation on the flat and complex Chowilla floodplain in Australia. In this application, we use a low-fidelity model that is considerably coarser than the model used previously in the study by Fraehr et al (2022) to test the upskilling and speed-up capabilities of the LSG methodology even further. Through this application, it is also investigated how weighting before the EOF analysis affects the accuracy of the LSG model and whether this should be used when applying the LSG model on an unstructured grid.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Development of a fast and accurate hybrid model for floodplain inundation

Fraehr

Wang

et al. 2023

Preprint

View full text Add to dashboard Cite

High computational cost is often the most limiting factor when running high-resolution hydrodynamic models to simulate spatial-temporal flood inundation behaviour. To address this issue, a recent study introduced the hybrid Low-fidelity, Spatial analysis, and Gaussian Process learning (LSG) model. The LSG model simulates the dynamic behaviour of flood inundation extent by upskilling simulations from a low-resolution hydrodynamic model through Empirical Orthogonal Function (EOF) analysis and Sparse Gaussian Process (Sparse GP) learning. However, information on flood extent alone is often not sufficient to provide accurate flood risk assessments. In addition, the LSG model has only been tested on hydrodynamic models with structured grids, while modern hydrodynamic models tend to use unstructured grids. This study therefore further develops the LSG model to simulate water depth as well as flood extent and demonstrates its efficacy as a surrogate for a high-resolution hydrodynamic model with an unstructured grid. The further developed LSG model is evaluated on the flat and complex Chowilla floodplain of the Murray River in Australia and accurately predicts both depth and extent of the flood inundation, while being 12 times more computationally efficient than a high-resolution hydrodynamic model. In addition, it has been found that weighting before the EOF analysis can compensate for the varying grid cell sizes in an unstructured grid and the inundation extent should be predicted from an extent-based LSG model rather than deriving it from water depth predictions.

show abstract

Section: Methodology Of the Lsg Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Development of a fast and accurate hybrid model for floodplain inundation

Fraehr

Wang

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

Prevention/mitigation of natural disasters in urban areas

Chai

2023

Smart Constr. Sustain. Cities

View full text Add to dashboard Cite

Preventing/mitigating natural disasters in urban areas can indirectly be part of the 17 sustainable economic and social development intentions according to the United Nations in 2015. Four types of natural disasters—flooding, heavy rain-induced slope failures/landslides; earthquakes causing structure failure/collapse, and land subsidence—are briefly considered in this article. With the increased frequency of climate change-induced extreme weathers, the numbers of flooding and heavy rain-induced slope failures/landslides in urban areas has increased in recent years. There are both engineering methods to prevent their occurrence, and more effectively early prediction and warning systems to mitigate the resulting damage. However, earthquakes still cannot be predicted to an extent that is sufficient to avoid damage, and developing and adopting structures that are resilient against earthquakes, that is, structures featuring earthquake resistance, vibration damping, and seismic isolation, are essential tasks for sustainable city development. Land subsidence results from human activity, and is mainly due to excessive pumping of groundwater, which is a “natural” disaster caused by human activity. Countermeasures include effective regional and/or national freshwater management and local water recycling to avoid excessive pumping the groundwater. Finally, perspectives for risk warning and hazard prevention through enhanced field monitoring, risk assessment with multi-criteria decision-making (MCDM), and artificial intelligence (AI) technology.

show abstract