Assessment of surrogate models for flood inundation: The physics-guided LSG model vs. state-of-the-art machine learning models.

Hydrodynamic models can accurately simulate flood inundation but are limited by their high computational demand that scales non-linearly with model complexity, resolution, and domain size. Therefore, it is often not feasible to use high-resolution hydrodynamic models for real-time flood predictions or when a large number of predictions are needed for probabilistic flood design. Computationally efficient surrogate models have been developed to address this issue. The recently developed Low-fidelity, Spatial analysis, and Gaussian Process Learning (LSG) model has shown strong performance in both computational efficiency and simulation accuracy. The LSG model is a physics-guided surrogate model that simulates flood inundation by first using an extremely coarse and simplified (i.e. low-fidelity) hydrodynamic model to provide an initial estimate of flood inundation. Then, the low-fidelity estimate is upskilled via Empirical Orthogonal Functions (EOF) analysis and Sparse Gaussian Process models to provide accurate high-resolution predictions. Despite the promising results achieved thus far, the LSG model has not been benchmarked against other surrogate models. Such a comparison is needed to fully understand the value of the LSG model and to provide guidance for future research efforts in flood inundation simulation. This study compares the LSG model to four state-of-the-art surrogate flood inundation models. The surrogate models are assessed for their ability to simulate the temporal and spatial evolution of flood inundation for events both within and beyond the range used for model training. The models are evaluated for three distinct case studies in Australia and the United Kingdom. The LSG model is found to be superior in accuracy for both flood extent and water depth, including when applied to flood events outside the range of training data used, while achieving high computational efficiency. In addition, the low-fidelity model is found to play a crucial role in achieving the overall superior performance of the LSG model.

A noise adaptive approach for nodal water demand estimation in water distribution systems.

Hydraulic models have emerged as a powerful tool for simulating the real behavior of water distribution systems (WDSs). In using the models for estimating nodal water demands, measurement uncertainty must be considered. A common approach is to use the covariance of measurement noises to quantify the measurement uncertainty. The noise covariance is typically assumed constant and estimated a priori. However, such an assumption is frequently misleading as actual measurement accuracies are affected by measuring instruments and environmental noises. In this study, we develop a variational Bayesian approach for real-time estimation of noise covariance and nodal water demands. The approach can adaptively adjust the noise covariance with the variation of the noise intensity, thereby efficiently avoiding model overfitting. The measurement residual decomposition reveals that this new approach is effective in determining model structural errors caused by topological structure parameterization.

Python program for spatial reduction and reconstruction method in flood inundation modelling.

Fast and accurate modelling of flood inundation has gained increasing attention in recent years. One approach gaining popularity recently is the development of emulation models using data driven methods, such as artificial neural networks. These emulation models are often developed to model flood depth for each grid cell in the modelling domain in order to maintain accurate spatial representation of the flood inundation surface. This leads to redundancy in modelling, as well as difficulties in achieving good model performance across floodplains where there are limited data available. In this paper, a spatial reduction and reconstruction (SRR) method is developed to (1) identify representative locations within the model domain where water levels can be used to represent flood inundation surface using deep learning models; and (2) reconstruct the flood inundation surface based on water levels simulated at these representative locations. The SRR method is part of the SRR-Deep-Learning framework for flood inundation modelling and therefore, it needs to be used together with data driven models. The SRR method is programmed using the Python programming language and is freely available from https://github.com/yuerongz/SRR-method.•The SRR method identifies locations which are representative of flood inundation behavior in surrounding areas.•The representative locations selected following the SRR method have sufficient flood data for developing emulation models.•Flood inundation surfaces can be reconstructed using the SRR method with a detection rate of above 99%.