Storm Surge Forecasting Research Articles

A comparison of global and regional ocean tide models with tide gauges in the East Asian marginal seas

To enhance storm surge forecasting and establish open boundary conditions for sophistical numerical simulation of tides and tidal currents in the Bohai Sea, Yellow Sea, and East China Sea, this study evaluated the accuracy of eight global and regional tide models (EOT20, FES2014, NAO.99Jb, TPXO9, DTU16, HAMTIDE12, OSU12, and GOT4.10c) based on 41 tide gauges in the region. The differences between tide models and 8 offshore tide gauges were notable, with the root mean square (RMS) values ranging from 12.67 to 25.97 cm for M2, 6.09–13.73 cm for S2, 2.96–6.56 cm for K1, and 2.50–4.20 cm for O1. For the offshore stations, the NAO.99Jb model demonstrated superior performance with the root square sum (RSS) value of 12.63 cm among the eight tide models. For the island and coastal stations, the EOT20 model performed the best for M2 (RMS 13.03 cm), the FES2014 model for S2 (RMS 6.26 cm), and the DTU16 model for K1 and O1 (RMS 2.73 cm and 2.26 cm, respectively). Overall, the EOT20 model also exhibited the lowest RSS value (15.22 cm) among the eight models for 24 island and coastal stations. The EOT20 model slightly outperformed others with the RSS of 14.88 cm across all 32 tide gauges. Regarding the Sa tidal component, great discrepancies were found between TIDAL CONSTANTS (TICON) harmonic constant data and tide gauges. The reason for this is that the Doodson number for the Sa tidal component derived from tide gauge data, which is influenced by meteorological factors, should be (0,0,1,0,0,0), while the TICON harmonic constant data employed the Doodson number of (0,0,1,0,0,−1) corresponding to the basic astronomical variables. A correction method was proposed to adjust tidal constants from inconsistent Doodson numbers. Comparing with the harmonic constants of the Sa tidal component at 28 coastal tide stations revealed large errors in the NAO.99b, FES2014, and EOT20 models, with RMS values of 19.03, 22.08, and 13.61 cm, respectively. Consequently, caution should be taken when using the Sa results from these ocean tide models.

Comparative evaluation of meteorological inputs for improved storm surge modeling: A case study of tropical Cyclone Vayu

The selection of meteorological inputs in storm surge is crucial, with wind and pressure fields playing a significant role in energy transfer and the study area's bathymetry. While using observed track information for wind and pressure drop has been the standard approach for the past decade, recent studies have highlighted the need for atmospheric inputs from models like the Weather Research and Forecast model (WRF) for storm surge forecast. This study aims to compare the performance of a DELFT-3D FM storm surge model forced with inputs from IMD (India Meteorological Department) observed the best track and pressure drop (EXP-1) and wind and pressure fields from WRF (EXP-2) for Tropical Cyclone Vayu, which occurred in the southern Arabian Sea basin in June 2019. The study compares the simulated storm water levels and highlights the benefits of using time and space-varying wind and pressure input for improved surge representation. Results show that the WRF-DELFT setup outperforms the IMD-DELFT setup, particularly for tide gauge locations close to the storm eye. The simulated storm surge captures the intensified stage of Vayu and provides a more realistic representation than the model forced with IMD track data. However, biases and limitations, such as inadequate representation of land surface parameterization, are identified. The study suggests further exploring wave-induced effects on total water level and multiple cyclone scenarios to enhance wind speed and track displacement prediction accuracy and improved land-sea parameterization can help address these limitations.

Prediction of storm surge in the Pearl River Estuary based on data-driven model

Storm surges, a significant coastal hazard, cause substantial damage to both property and lives. Precise and efficient storm surge models are crucial for long-term risk assessment and guiding emergency management decisions. While high-fidelity dynamic models offer accurate predictions, their computational costs are substantial. Hence, recent efforts focus on developing data-driven storm surge surrogate models. This study focuses on the Pearl River Estuary in Guangdong Province. Initially, the dynamic ADvanced CIRCulation (ADCIRC) model was utilized to construct storm surge data for 16 historical typhoons, serving as training, validation, and testing data for data-driven models. Subsequently, Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Informer deep learning (DL) models were employed for forecasting of storm surge over the next 1h, 3h, 6h, 12h, and 18h. Finally, Shapley Additive exPlanations (SHAP) values were used for interpretability analysis of the input factors across different models. Results indicated that the proposed DL storm surge prediction model can effectively replicate the dynamic model’s simulation results in short-term forecasts, significantly reducing computational costs. This model offers valuable scientific assistance for future coastal storm surge forecasts in the Greater Bay Area.

Formulating a warning threshold for coastal compound flooding: A copula-based approach

To calculate warning thresholds for compound-flooding events triggered by heavy rainfall coupled with storm tides in Taiwan’s coastal urban areas, we applied copula-based analysis to observation data collected from 2001 till 2022 for Taipei City and New Taipei City and developed an empirical formula that accounts for both the capacity of the drainage infrastructure, which partially depends on the coastal sea level and varies over time, and the amount of precipitation. Compared against observation data from flood detectors, our predictions exhibited an accuracy of 85.2 % and 78.8 % for Taipei City and New Taipei City, respectively, thus improving upon the 62.8 % and 68.5 % success rates for thresholds estimated using only the hourly accumulated rainfall. These promising preliminary results suggest that reliable flood warnings for tidal-basin regions can be expedited by employing our formula and inputting rainfall and sea-level values from ensemble typhoon and storm-surge forecasts.

Examining the roles of visualizations in people's understanding of uncertain storm surge forecasts, interpretation of risk, and decision-making

Maps and other forms of geovisualizations play a central yet complex role in communicating weather hazard risks and threats to the public. Through designing storm surge risk communication prototypes and testing these visualizations in focus groups with residents of coastal communities in Georgia and South Carolina, this study investigates how different aspects of geospatial visualization influence individuals' understanding and relatability of storm surge information and how different elements of geospatial risk visualization influence people's intent to take protective actions. The focus group discussions revealed common themes and unique perspectives in the interpretation of hurricane storm surge flooding forecasts and illustrated the complexity of communicating risk and uncertainty to diverse populations. This study found a sequential pattern of visual information processing, beginning with understanding the content and its perceived accuracy and uncertainty, followed by personalizing the information and deciding on protective action. Showing storm surge forecast maps at multiple spatial scales was valuable, combined with realistic visualizations in the form of images or 3D animations of flooded buildings. Maps with recognizable places and labels helped study participants locate themselves and assess storm surge risks, although residents of urban, barrier island, and rural coastal communities interpreted the information in their local and personal context. Additional research is needed to understand how people interact with different types of storm surge forecast information and how to improve visual risk communication in ways that enhance people's understanding and responses to storm surge risk in the context of their local environment and prior experiences.

Advancing storm surge forecasting from scarce observation data: A causal-inference based Spatio-Temporal Graph Neural Network approach

Rapid and precise forecasting of storm surge in coastal regions is crucial for ensuring safety of coastal communities’ life and property. Yet, learning a data-driven forecasting model from observation data such as gauges and post-event reconnaissance remains challenging, due to the observation data scarcity and the real-world complexity. Recently, deep learning has received increasing attention, but existing deep learning approaches solely focus on individual site scenarios, ignoring the value of information contained in neighboring sites’ observations. In this study, we propose to integrate graph neural networks (GNN) and gated recurrent unit (GRU) to capture the spatial and temporal storm surge dependencies across multiple observation stations. GNN provides the unique capability to model non-Euclidean complex spatial relationship across observation stations, while GRU enhances the data efficiency of temporal dependency modeling. To account for the effect of complex coastline topography, the Liang–Kleeman information flow theory is employed to establish a causal-inference based graph edge scheme connecting multiple observation stations. The Causal-inference based Spatio-Temporal Graph Neural Network (CSTGNN) were trained and evaluated on 13-year observation data from 4 observation stations along Florida coastline. Experiments affirm the competence of CSTGNN, which outperformed six commonly used competitive baselines across different metrics and observation stations, under lead times up to six hours. Furthermore, benefits of capturing the spatial dependency and leveraging causal inference are also comprehensively examined. To conclude, we believe that this novel spatio-temporal forecasting framework can result in enhanced coastal resilience by its improved storm surge forecasting capability.

Storm surge forecasting based on physics-informed neural networks in the Bohai Sea

Physics-informed neural networks (PINN), as a new method of integrating artificial neural networks (ANN) and physical laws, have been considered and applied in the fields of ocean forecasting and ocean research. In this paper, the simplified two-dimensional (2D) storm surge governing equation is introduced into an ANN to establish a PINN-based storm surge forecast model. The numerical simulation results of 14 storm surge events in the Bohai Sea are selected as the PINN training set, and 6.3% of the training set data are randomly selected to reconstruct the storm surge field information. The storm surge reconstructed at each tide station is nearly identical to the storm surge curve simulated by the numerical model, with the root mean square error (RMSE) less than 0.12 m and absolute error of maximum storm surge less than 0.2 m. The analysis of the storm surge field at key moments (storm surge height lager than 1 m) shows that the difference in storm surge field between the PINN reconstruction and the numerical model is generally less than 0.4 m. Two storm surge events in the Bohai Sea are selected as forecast cases, and the same network structure, parameters, and storm surge data assimilation scheme are used for predictions by the ANN, PINN, and numerical model. The results show that compared to the ANN and numerical models, the average relative error of the maximum storm surge predicted by the PINN is reduced by approximately 25%, which significantly improves the forecast accuracy, therefore, the PINN is suitable for storm surge forecasting and research due to its advantages in small sample data training and strong physical meaning.

Bias correction of operational storm surge forecasts using Neural Networks

Storm surges can give rise to extreme floods in coastal areas. The Norwegian Meteorological Institute (MET Norway) produces 120 h regional operational storm surge forecasts along the coast of Norway based on the Regional Ocean Modeling System (ROMS), using a model setup called Nordic4-SS. Despite advances in the development of models and computational capabilities, forecast errors remain large enough to impact response measures and issued alerts, in particular, during the strongest storm events. Reducing these errors will positively impact the efficiency of the warning systems while minimizing efforts and resources. Here, we investigate how forecasts can be improved with residual learning, i.e., training data-driven models to predict the residuals in forecasts from Nordic4-SS. A simple error mapping technique and a more sophisticated Neural Network (NN) method are tested. The simple error mapping technique provides a reduction in the Root Mean Square Error (RMSE) of less than 4%. Using the NN residual correction method, the RMSE in the Oslo Fjord is reduced by 36% for lead times of one hour, 9% for 24 h, and 5% for 60 h. Therefore, the residual NN method is a promising direction for correcting storm surge forecasts, especially on short timescales. Moreover, it is well adapted to being deployed operationally, as (i) the correction is applied on top of the existing model and requires no changes to it, (ii) all predictors used for NN inference are already available operationally, (iii) prediction by the NNs is very fast, typically a few seconds per station, and (iv) the NN correction can be provided to a human expert who may inspect it, compare it with the model output, and see how much correction is brought by the NN, allowing to capitalize on human expertise as a quality validation of the NN output. While no changes to the hydrodynamic model are necessary to calibrate the neural networks, they are specific to a given model and must be recalibrated when the numerical models are updated.

Hindcast Insights from Storm Surge Forecasting of Super Typhoon Saola (2309) in Hong Kong with the Sea, Lake and Overland Surges from Hurricanes Model

Super Typhoon Saola (2309) skirted past south-southeast of Hong Kong within 40 km on the night of 1 September 2023, posing a significant storm surge threat to Hong Kong. Given the close proximity of Saola with a peak intensity of about 210 km/h within 300 km of Hong Kong, a close call of the “super typhoon direct-hit” scenario, this case provides valuable insights from a hindcast review of storm surge forecasts and warning operation using the Sea, Lake and Overland Surges from Hurricanes (SLOSH) model, which is the operational storm surge model adopted by the Hong Kong Observatory (HKO). The performance of the HKO’s PRobabilistic Inundation Map Evaluation System (PRIMES) using both statistical and model ensemble approaches was also reviewed in this paper. Saola was a challenging case for operational forecasting of a compact TC structure with changes in storm size and intensity when it came close to Hong Kong. With major observations of storm structure using weather radar and dense automatic weather station, tide gauge and water level gauge networks, the high sensitivity of storm surge forecasts to the storm size parameter and the distance of closest approach was clearly revealed in the case of Saola. Even with a circularly symmetric TC parametric model like SLOSH, the hindcast review results illustrated that the model outputs were reasonably accurate during the closest approach of Saola given an accurate storm size and distance of closest approach were input, and using a highly computationally efficient storm surge model made it possible for the nowcasting of storm surges to handle compact and intense TC direct-hit cases in operational TC forecasting. Taking a nowcasting approach not only helps provide more reliable storm tide forecasts, but also facilitates the formulation of a better warning strategy when making final-call decisions in emergency response actions, based on the more frequent real-time analysis of TC position, intensity and storm size and the more accurate prediction of these parameters. A nowcasting workflow for storm surge operation was proposed in this paper.

Assessing Storm Surge Multiscenarios Based on Ensemble Tropical Cyclone Forecasting

AbstractEnsemble forecasting is a promising tool to aid in making informed decisions against risks of coastal storm surges. Although tropical cyclone (TC) ensemble forecasts are commonly used in operational numerical weather prediction systems, their potential for disaster prediction has not been maximized. Here we present a novel, efficient, and practical method to utilize a large ensemble forecast of 1,000 members to analyze storm surge scenarios toward effective decision making such as evacuation planning and issuing surge warnings. We perform the simulation of TC Hagibis (2019) using the Japan Meteorological Agency's (JMA) nonhydrostatic model. The simulated atmospheric predictions were utilized as inputs for a statistical surge model named the Storm Surge Hazard Potential Index to estimate peak surge heights along the central coast of Japan. We show that Pareto‐optimized solutions from an ensemble storm surge forecast can describe potential worst (maximum) and optimum (minimum) storm surge scenarios. These solutions exemplify a diversity of trade‐off surge outcomes across diverse coastal locations, reflecting variations in coastal geometry, including bathymetry. For example, some of the Pareto‐optimized solutions that illustrate worst surge scenarios for inner bay locations are not necessarily accountable for bringing severe surge cases in open coasts. We further emphasize that an in‐depth evaluation of Pareto‐optimal solutions can shed light on how meteorological variables such as track, intensity, and size of TCs influence the worst and optimum surge scenarios, which is not clearly quantified in current multiscenario assessment methods such as those used by JMA/National Hurricane Center in the United States.

Operational Storm Surge Forecasting at the National Hurricane Center: The Case for Probabilistic Guidance and the Evaluation of Improved Storm Size Forecasts Used to Define the Wind Forcing

Abstract The primary source of guidance used by the Storm Surge Unit (SSU) at the National Hurricane Center (NHC) for issuing storm surge watches and warnings is the Probabilistic Tropical Storm Surge model (P-Surge). P-Surge is an ensemble of Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model forecasts that is generated based on historical error distributions from NHC official forecasts. A probabilistic framework is used for operational storm surge forecasting to account for uncertainty related to the tropical cyclone track and wind forcing. Previous studies have shown that the size of a storm’s wind field is an important factor that can affect storm surge. A simple radius of maximum wind (RMW) prediction scheme was developed to forecast RMW based on NHC forecast parameters. Verification results indicate this scheme is an improvement over the RMW forecasts used by previous versions of P-Surge. To test the impact of the updated RMW forecasts in P-Surge, retrospective cases were selected from 25 storms from 2008 to 2020 that had an adequate number of observations. Evaluation of P-Surge forecasts using these improved RMW forecasts shows that the probability of detection is higher for most probability of exceedance thresholds. In addition, the forecast reliability is improved, and there is an increase in the number of high probability forecasts for extreme events at longer lead times. The improved RMW forecasts were recently incorporated into the operational version of P-Surge (v2.9), and serve as an important step toward extending the lead time of skillful and reliable storm surge forecasts.

Assessing the Risk of Extreme Storm Surges from Tropical Cyclones under Climate Change Using Bidirectional Attention-Based LSTM for Improved Prediction

Accurate prediction of storm surges is crucial for mitigating the impact of extreme weather events. This paper introduces the Bidirectional Attention-based Long Short-Term Memory (LSTM) Storm Surge Architecture, BALSSA, addressing limitations in traditional physical models. By leveraging machine learning techniques and extensive historical and real-time data, BALSSA significantly enhances prediction accuracy. Utilizing a bidirectional attention-based LSTM framework, it captures complex, non-linear relationships and long-term dependencies, improving the accuracy of storm surge predictions. The enhanced model, D-BALSSA, further amplifies predictive capability through a doubled bidirectional attention-based structure. Training and evaluation involve a comprehensive dataset from over 70 typhoon incidents in Macao between 2017 and 2022. The results showcase the outstanding performance of BALSSA, delivering highly accurate storm surge forecasts with a lead time of up to 72 h. Notably, the model exhibits a low Mean Absolute Error (MAE) of 0.0287 m and Root Mean Squared Error (RMSE) of 0.0357 m, crucial indicators measuring the accuracy of storm surge predictions in water level anomalies. These metrics comprehensively evaluate the model’s accuracy within the specified timeframe, enabling timely evacuation and early warnings for effective disaster mitigation. An adaptive system, integrating real-time alerts, tropical cyclone (TC) chaser, and prospective visualizations of meteorological and tidal measurements, enhances BALSSA’s capabilities for improved storm surge prediction. Positioned as a comprehensive tool for risk management, BALSSA supports decision makers, civil protection agencies, and governments involved in disaster preparedness and response. By leveraging advanced machine learning techniques and extensive data, BALSSA enables precise and timely predictions, empowering coastal communities to proactively prepare and respond to extreme weather events. This enhanced accuracy strengthens the resilience of coastal communities and protects lives and infrastructure from the escalating threats of climate change.

FORECAST VERIFICATION OF AN ENSEMBLE TROPICAL CYCLONE STORM SURGE SYSTEM

Tropical Cyclone (TC) induced storm surges can be especially damaging and require emergency preparations. A unique challenge of providing TC storm surge advice to emergency services relates to the sensitivity of storm surge forecasts to uncertainty in the TC forecast. Probabilistic forecast products are a good fit for such advice, as long as forecast skill characteristics are known and understood. The Australian Bureau of Meteorology operates a dynamical ensemble prediction system to provide forecasts of storm surge driven by TCs. This is a step change from the legacy parametric and/or scenario-based techniques previously used at the Bureau. There are a number of benefits to be gained from dynamic storm surge ensemble forecasts over more traditional systems. Most importantly, the surge forecast is very sensitive to errors in the TC location, velocity, intensity, and size. Ensemble storm surge forecasting is required in order to take account of this uncertainty in the TC forecast.

DEVELOPMENT OF MESHES FOR COMPOUND FLOOD SIMULATION FOR THE TEXAS COAST

The Texas coast is subjected to frequent storms leading to floods in the yearly hurricane season. To provide forecasts of hurricane storm surge, a shallow water equation solver, the Advanced CIRCulation model, Luettich (1992) is used. The current state-of-the-art ADCIRC meshes for the Texas coast used in forecasting were developed about a decade ago, see, Hope (2013). ADCIRC uses unstructured meshes that require significant tailoring to ensure accurate and physically relevant solutions. The current mesh used in forecasting has very high resolution on the Texas coast for accuracy and numerical stability, whereas areas far away from Texas are of low resolution for computational efficiency. The applicability and accuracy of these meshes is regularly verified to be used in operational storm surge forecasting.

Bridging the Gap: Enhancing Storm Surge Prediction and Decision Support with Bidirectional Attention-Based LSTM

Accurate storm surge forecasting is vital for saving lives and avoiding economic and infrastructural damage. Failure to accurately predict storm surge can have catastrophic repercussions. Advances in machine learning models show the ability to improve accuracy of storm surge prediction by leveraging vast amounts of historical and realtime data such as weather and tide patterns. This paper proposes a bidirectional attention-based LSTM storm surge architecture (BALSSA) to improve prediction accuracy. Training and evaluation utilized extensive meteorological and tide level data from 77 typhoon incidents in Hong Kong and Macao between 2017 and 2022. The proposed methodology is able to model complex non-linearities between large amounts of data from different sources and identify complex relationships between variables that are typically not captured by traditional physical methods. BALSSA effectively resolves the problem of long-term dependencies in storm surge prediction by the incorporation of an attention mechanism. It enables selective emphasis on significant features and boosts the prediction accuracy. Evaluation has been conducted using real-world datasets from Macao to validate our storm surge prediction model. Results show that accuracy and robustness of predictions were significantly improved by the incorporation of attention mechanisms in our models. BALSSA captures temporal dynamics effectively, providing highly accurate storm surge forecasts (MAE: 0.0126, RMSE: 0.0003) up to 72 h in advance. These findings have practical significance for disaster risk reduction strategies, saving lives through timely evacuation and early warnings. Experiments comparing BALSSA variations with other machine learning algorithms consistently validate BALSSA’s superior predictive performance. It offers an additional risk management tool for civil-protection agencies and governments, as well as an ideal solution for enhancing storm surge prediction accuracy, benefiting coastal communities.

Member Formation Methods Evaluation for a Storm Surge Ensemble Forecast System in Taiwan

The forecast of typhoon tracks remains uncertain and is positively related to the accuracy of the storm surge forecast. The storm surge prediction error increases dramatically when the forecast track error is larger than 100 km. This study aims to develop an ensemble storm surge prediction system using parametric weather models to account for the uncertainty in typhoon track prediction. The storm surge model adopted in this study is COMCOT-SS storm surge forecast system. Two methods are introduced and analyzed to generate the ensemble members in this study. One is from the weather ensemble prediction system (WEPS), and the other is from the error distribution of the deterministic forecasts (EDF). The ensemble prediction results show that the ensemble mean of WEPS performs similarly to the deterministic forecast. However, the maximum surge height of WEPS is often lower than one from EDF. The verification results suggest that, for disaster prevention, EDF provides stronger warnings to the coastal region than WEPS. However, it may provide overestimated forecasts in some cases.

Machine-Learning-Based Model for Hurricane Storm Surge Forecasting in the Lower Laguna Madre

During every Atlantic hurricane season, storms represent a constant risk to Texan coastal communities and other communities along the Atlantic coast of the United States. A storm surge refers to the abnormal rise of sea water level due to hurricanes and storms; traditionally, hurricane storm surge predictions are generated using complex numerical models that require high amounts of computing power to be run, which grow proportionally with the extent of the area covered by the model. In this work, a machine-learning-based storm surge forecasting model for the Lower Laguna Madre is implemented. The model considers gridded forecasted weather data on winds and atmospheric pressure over the Gulf of Mexico, as well as previous sea levels obtained from a Laguna Madre ocean circulation numerical model. Using architectures such as Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) combined, the resulting model is capable of identifying upcoming hurricanes and predicting storm surges, as well as normal conditions in several locations along the Lower Laguna Madre. Overall, the model is able to predict storm surge peaks with an average difference of 0.04 m when compared with a numerical model and an average RMSE of 0.08 for normal conditions and 0.09 for storm surge conditions.

Accurate storm surge prediction using a multi-recurrent neural network structure

This paper considers storm surge prediction using a neural network and considering multiple physical characteristics. Based on the factors that influence storm surges and historical observation data, we divide the input to the neural network into time features extracted from the prediction target and the auxiliary features that affect storm surges, and construct a feature gate within multiple recurrent neural network (RNN) cells. Historical hurricane data are used to assess the effectiveness and accuracy of the proposed model. Comparative analysis against a long short-term memory (LSTM) storm surge prediction model is conducted to verify the prediction performance of the proposed method. The comparison results show that the multi-RNN model is superior to the LSTM model in terms of four evaluation metrics and for all lead times. In particular, the multi-RNN model accurately predicts the maximum storm surge water level, and the prediction results are more consistent with the rise and fall of the water. A comparison of the storm surge forecasts using inputs from different time intervals under different evaluation indices confirms the generalization and stability of our proposed model. The experiments of storm surge prediction at six stations further confirm the wide applicability of the model.

Warning water level determination and its spatial distribution in coastal areas of China

Abstract. The warning water level is the default water level at which storm surges may occur along a coast and indicates a stage of alert. This level forms the basis for storm-surge forecasting, and prewarning is an important reference for governments and aids in the decision-making process for storm-surge prevention and disaster risk mitigation. The warning water level has four categories (blue, yellow, orange, and red) based on water level observational data. Taking into account the actual defense capability of the shore, we determined the warning water level by comprehensively analyzing factors, including the high water level at the typical return period of each shore section, wave exposure degree and defense capability of storm-surge protection facilities, and the shore-section importance level. Here, we proposed a quantitative method for determining the warning water level, and the application of this method was introduced by determining the warning water level at the shore section of Zhifu District (Yantai City, Shandong Province, China) as an example. We analyzed the spatial distribution characteristics of the warning water levels for 259 shore sections along the coast and revealed their current marine disaster prevention capabilities. Our findings provide a valid direction for determining future warning water levels and a reliable scientific reference for redetermining warning water levels in coastal areas while improving marine disaster prevention and protection capabilities.

Storm Surge Forecast Using an Encoder–Decoder Recurrent Neural Network Model

This study presents an encoder–decoder neural network model to forecast storm surges on the US North Atlantic Coast. The proposed multivariate time-series forecast model consists of two long short-term memory (LSTM) models. The first LSTM model encodes the input sequence, including storm position, central pressure, and the radius of the maximum winds to an internal state. The second LSTM model decodes the internal state to forecast the storm surge water level and velocity. The neural network model was developed based on a storm surge dataset generated by the North Atlantic Comprehensive Coastal Study using a physics-based storm surge model. The neural network model was trained to predict storm surges at three forecast lead times ranging from 3 h to 12 h by learning the correlation between the past storm conditions and future storm hazards. The results show that the computationally efficient neural network model can forecast a storm in a fraction of one second. The neural network model not only forecasts peak surges, but also predicts the time-series profile of a storm. Furthermore, the model is highly versatile, and it can forecast storm surges generated by different sizes and strengths of bypassing and landfalling storms. Overall, this work demonstrates the success of data-driven approaches to improve coastal hazard research.