智慧城市是城市发展的新兴模式。然而,近年来频发的城市洪涝给智慧城市的管理和发展带来了严峻的挑战。暴雨洪涝模拟是城市防洪减灾的关键技术之一,也是智慧城市风险应急管理中重要的决策支持依据。本文首先分析了智慧城市以及智慧水务的内涵,探讨了智慧城市发展和管理对于城市洪涝模拟新的需求;在此基础上梳理了智慧城市发展给城市洪涝模型带来的新的数据支撑;面向智慧城市应急管理决策,对比分析了3种现有的水文模型方法,认为在智慧城市大数据和技术的支持下,基于物理基础的分布式模型是未来城市洪涝模拟的主流发展方向。最后,探讨了建立城市洪涝模型的建模机制、尺度、基础数据库的建立、更新和管理、地表与地下管网模拟并重等关键问题,分析认为精细化、与RS和GIS技术融合、注重时空过程和智慧服务是智慧城市背景下城市洪涝模拟发展的必然趋势。

Smart city is the emerging pattern of city development. However, in recent years, frequent urban floods have brought serious challenges to the management and development of smart cities. Urban flooding simulation is not only one of the key technologies of city flood control and mitigation, but also an important decision support tool for smart city disaster risk and emergency management. This article first indicates the connotation of smart city and smart water, and analyzes the new requirements of smart city development and management on urban flood inundation models. Second, it discusses the newly available data support of urban flood simulation models brought by the development of smart cities, and compares three existing hydrological models from the perspective of city emergency management decision-making. Third, this article points out that under the support of big data and technology of smart cities, the hydraulic-based distributed urban flood simulation models have a promising prospect in application. By analyzing the development of urban flood simulation models, we conclude that micro-level simulation, integration with remote sensing and GIS technologies, and emphasis on spatiotemporal process of flood inundation and smart service are the inevitable trend of development of urban flood simulation models.

以海口市为例，利用1960—2017年的66场台风资料，包括台风日降雨和风暴潮的极值水位，建立复合洪涝组合情景。基于多个情景，对海口市在海平面上升背景下极端雨潮复合洪涝灾害的潜在风险进行深入研究。结果表明：1）在台风事件中，风暴潮为复合洪涝灾害的关键致灾因子，受灾最严重的地区主要位于南渡江入海口和北部滨海地区；2）在最大雨潮复合洪涝组合情景中，海口市受淹面积估计为148 km²，相较于最小雨潮复合洪涝组合情景增加了约15倍，淹没区域中一半以上的区域积水深度超过1 m；3）在极端雨潮复合情景下，海甸岛、新埠岛及江东新区一带的滨海地区受到海平面上升影响最显著。根据RCP8.5情景预测，到2100年，海口市受极端复合洪涝灾害的影响区域预计达到约203 km²。

In the context of climate change, the escalating frequency of extreme weather phenomena has exacerbated the severity of compound floods in the southeastern coastal regions of China. Rising sea levels significantly contribute to the inundation of low-lying coastal urban areas. The quantitative assessment of compound flood risk offers scientific support for disaster prevention and reduction in coastal cities and for coastal management initiatives. Using Haikou City as a case study, the daily precipitation and maximum storm surge tide data from 66 typhoons that affected Haikou between 1960 and 2017 were utilized to construct compound flood combination scenarios. Based on the quantitative method of D-Flow FM (Delft3D-FLOW Flexible Mesh) numerical simulation, the potential risks of extreme rainfall and storm surge compound flood disasters under sea level rise scenarios were thoroughly investigated by integrating various scenarios. The findings revealed the following: 1) Storm surge was the primary factor contributing to compound flooding during typhoons, with the estuary of the Nandu River and the northern coast being the most affected. 2) In the scenario of maximum rainfall and storm surge combination, the inundation area of Haikou is about 148 km², which is approximately 15 times larger than the minimum rainfall and storm surge combination scenario. Moreover, in more than half of the inundated areas, the water depth exceeds 1 meter. 3) Under extreme rainfall and storm surge compound scenarios, the areas encompassing Haidian Island, Xinbu Island, and Jiangdong New Area were significantly affected by sea level rise. By 2100, the total flooding area is projected to reach about 203 km² under the RCP8.5 scenario. Sea level rise significantly amplifies urban flood risks, implying that coastal cities are poised to encounter heightened threats and manage future challenges. Through comprehensive comparisons of multiple rainfall and storm surge compound flooding scenarios under sea level rise, the temporal and spatial characteristics of the compound flooding risk were systematically evaluated. The results provide an important scientific basis for sustainable regional development, effective management, and prevention.

The rise of machine learning (ML) has significantly advanced the field of coastal oceanography. This review aims to examine the existing deficiencies in numerical predictions of storm surges and the effort that has been made to improve the predictive accuracy through the application of ML. The readers are guided through the steps required to implement ML algorithms, from the first step of formulating problems to data collection and determination of input features to model selection, development and evaluation. Additionally, the review explores the application of hybrid methods, which combine the bilateral advantages of data-driven methods and physics-based models. Furthermore, the strengths and limitations of ML methods in predicting storm surges are thoroughly discussed, and research gaps are identified. Finally, we outline a vision toward a trustworthy and reliable storm surge forecasting system by introducing novel physics-informed ML techniques. We are meant to provide a primer for beginners and experts in coastal ocean sciences who share a keen interest in ML methodologies in the context of storm surge problems.

The damage caused by floods is increasing worldwide, and if floods can be predicted, the economic and human losses from floods can be reduced. A key parameter of flooding is water level data, and this paper proposes a water level prediction model using long short-term memory (LSTM) and a gated recurrent unit (GRU). As variables used as input data, meteorological data, including upstream and downstream water level, temperature, humidity, and precipitation, were used. The best results were obtained when the LSTM–GRU-based model and the Automated Synoptic Observing System (ASOS) meteorological data were included in the input data when experiments were performed with various model structures and different input data formats. As a result of the experiment, the mean squared error (MSE) value was 3.92, the Nash–Sutcliffe coefficient of efficiency (NSE) value was 0.942, and the mean absolute error (MAE) value was 2.22, the highest result in all cases. In addition, the test data included the historical maximum water level of 3552.38 cm in the study area, and the maximum water level error was also recorded as 55.49, the lowest result. Through this paper, it was possible to confirm the performance difference according to the composition of the input data and the time series prediction model. In a future study, we plan to implement a flood risk management system that can use the predicted water level to determine the risk of flooding, and evacuate in advance.

Stream gauge height (water level) is a significant indicator for forecasting future floods. Flooding occurs when the water level exceeds the flood stage. Predicting imminent floods can save lives, protect infrastructure, and improve road traffic management and transportation. Deep neural networks have been increasingly used in this domain due to their predictive capabilities in capturing complex features and interdependencies. This study employs four distinct models—Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM), transformer, and LSTNet—with MLP serving as the baseline model to forecast water levels. The models are trained using data from 20 distinct river gages across the state of Missouri to ensure consistent performance. Random search optimization is employed for hyperparameter tuning. The prediction intervals are set at 4, 6, 8, and 10 (each interval equivalent to 30 min) to ensure that performance results are robust and not due to random weight initialization or suboptimal hyperparameters and are consistent throughout different prediction intervals. The findings of this study indicate that the LSTNet model leads to a better performance than the other models, with a median RMSE of 0.00724, 0.00959, 0.01204, and 0.01230 for the 4, 6, 8, and 10 intervals, respectively. As climate change leads to localized storms driven by atmospheric shifts, water level fluctuations are becoming increasingly extreme, further exacerbating data drift in real-world datasets. The LSTNet model demonstrates superior performance in terms of RMSE, MAE, and the correlation coefficient across all prediction intervals when forecasting water levels under data drift conditions.

Bangladesh is in the floodplains of the Ganges, Brahmaputra, and Meghna River delta, crisscrossed by an intricate web of rivers. Although the country is highly prone to flooding, the use of state-of-the-art deep learning models in predicting river water levels to aid flood forecasting is underexplored. Deep learning and attention-based models have shown high potential for accurately forecasting floods over space and time. The present study aims to develop a long short-term memory (LSTM) network and its attention-based architectures to predict flood water levels in the rivers of Bangladesh. The models developed in this study incorporated gauge-based water level data over 7 days for flood prediction at Dhaka and Sylhet stations. This study developed five models: artificial neural network (ANN), LSTM, spatial attention LSTM (SALSTM), temporal attention LSTM (TALSTM), and spatiotemporal attention LSTM (STALSTM). The multiple imputation by chained equations (MICE) method was applied to address missing data in the time series analysis. The results showed that the use of both spatial and temporal attention together increases the predictive performance of the LSTM model, which outperforms other attention-based LSTM models. The STALSTM-based flood forecasting system, developed in this study, could inform flood management plans to accurately predict floods in Bangladesh and elsewhere.