Storm Surge Prediction Research Articles

Using surrogate modeling to predict storm surge on evolving landscapes under climate change

Planners managing coastal flood risk under a constrained computational budget face a tradeoff. Simulating many time periods or scenarios limits how many storm simulations can be run on each landscape. In this analysis, we present a deep learning model to predict storm surge as a function of storm parameters but also landscape features and boundary conditions (e.g., sea level). It is trained on peak surge elevations from Advanced Circulation (ADCIRC) hydrodynamic simulations of coastal Louisiana in a 2020 baseline and decadal periods from 2030 to 2070 under two morphological and climate scenarios. Leave-one-landscape-out cross-validation yielded a 0.086-m RMSE and 0.050-m MAE over 90 storms per landscape and 94,013 geospatial locations. A two-sided Kolmogorov-Smirnov test comparing annual exceedance probability (AEP) estimates from the model predictions to ADCIRC simulations rejected the null hypothesis that the predicted and ADCIRC AEP values were drawn from the same distribution only 1.1% of the time.

Development and performance of a high-resolution surface wave and storm surge forecast model: application to a large lake

Abstract. A real-time forecast model of surface hydrodynamics in Lake Ontario (Coastlines-LO) was developed to automatically predict storm surges and surface waves. The system uses a dynamically coupled Delft3D–SWAN model with a structured grid to generate 48 h predictions for the lake that are updated every 6 h. The lake surface is forced with meteorological data from the High Resolution Deterministic Prediction System (HRDPS). The forecast model has been running since May 2021, capturing a wide variety of storm conditions. Good agreement between observations and modelled results is achieved, with root mean squared errors (RMSEs) for water levels and waves under 0.02 and 0.26 m, respectively. During storm events, the magnitude and timing of storm surges are accurately predicted at nine monitoring stations (RMSE <0.05 m), with model accuracy either improving or remaining consistent with decreasing forecast length. Forecast significant wave heights agree with observed data (1 %–12 % relative error for peak wave heights) at four wave buoys in the lake. Coastlines-LO forecasts for storm surge prediction for two consecutive storm events were compared to those from the Great Lakes Coastal Forecasting System (GLCFS) to further evaluate model performance. Both systems achieved comparable results with average RMSEs of 0.02 m. Coastlines-LO is an open-source wrapper code driven by open data and has relatively low computational requirements compared to GLCFS, making this approach suitable for forecasting marine conditions in other coastal regions.

Predicting storm surge extremes on the Southeast Brazilian Coast: Long-term projections with neural networks

Extreme sea level events, resulting from the confluence of tides and storm surges, pose a significant threat to coastal populations and economies. The escalating risks associated with these events are exacerbated by climate change, manifesting in heightened storm intensity, increased frequency, and rising sea levels. Precise estimation of the probability of extreme storm surges is crucial for effective coastal management and adaptation. However, utilizing historical storm data is challenging due to data scarcity and the imperative to consider potential non-stationarity induced by climate change in predicting such events. This study addresses these challenges by introducing two neural network-based machine learning systems: a Multilayer Perceptron (MLP) and a Long Short-Term Memory (LSTM). Leveraging local and remote atmospheric and oceanic conditions, these systems project storm surges until 2060, incorporating climate projections. Trained and evaluated using sea level data from Imbetiba Port in Macaé, Rio de Janeiro, Brazil, the models utilize dynamic regionalization data from the RegCM4 and WW3 models, forced by HadGEM2-ES and MPI climate models. Both neural network models exhibited similar performance patterns, demonstrating high agreement in predicting storm surge heights with a 100-year return value, based on Imbetiba Port data. Projections utilized peaks-over-threshold (POT) methods, and extremes were calculated using a generalized Pareto distribution (GPD). Long-term projections indicated a 28 % increase (MLP ANN) and a substantial 70 % increase (LSTM RNN) in estimating extreme values, surpassing the observed storm surge of 0.67 m. Projected mean values were 0.86 m for the MLP network and 1.15 m for the LSTM network, providing valuable insights into the potential amplification of extreme sea level events in the studied region.

Adaptive formulation for probabilistic storm surge predictions through sharing of numerical simulation results across storm advisories

As a tropical storm/cyclone approaches landfall, real-time probabilistic predictions for the anticipated surge provide valuable information for emergency preparedness/response decisions. These probabilistic predictions are made through an uncertainty quantification process that involves: (i) generating a sufficiently large ensemble of storm scenarios based on the nominal storm advisory and the anticipated forecast errors; (ii) performing high-fidelity numerical simulations to obtain surge predictions for each storm scenario; and (iii) estimating surge statistics of interest by assembling the simulation results. This process is repeated whenever the nominal storm advisory is updated. The number of storm scenarios utilized in the analysis directly impacts the statistical accuracy of the probabilistic predictions; a larger ensemble improves accuracy but requires greater computational resources to provide predictions with the desired expediency to guide real-time decisions. This paper revisits two recently proposed Monte-Carlo (MC) frameworks that aim to improve accuracy without increasing the computational burden: adaptive importance sampling (AIS) and adaptive multi-fidelity Monte Carlo (AMFMC). The foundational concept behind them is similar: share numerical simulation results across the probabilistic predictions performed for different storm advisories to accelerate the MC estimation. This is achieved differently for each approach, through adaptive development of an importance sampling proposal density (for AIS) or a surrogate model (for AMFMC). Here, a direct comparison between these frameworks is established, focusing on the mechanisms for the information sharing and the challenges encountered in tuning the algorithm adaptive characteristics to provide probabilistic estimates across a large number of quantities of interest (QoIs), corresponding to the surge predictions for different locations within the coastal region of interest. As this large number results in conflicting choices for the adaptive characteristics, a compromise solution needs to be promoted. The efficacy of the two frameworks is examined in detail in this setting, comparing the accuracy of idealized implementations (adaptive decisions independently made for each QoI) to the accuracy of practical implementations (single, compromise decision within the MC implementation). The study also showcases the importance of information sharing across storm advisories in real-time probabilistic storm surge predictions and provides guidelines for an efficient adaptive MC formulation in such settings.

On the compilation of the bathymetric data of the bay of Bengal domain in the nodal points of triangular mesh to be compatible for the implementation of finite element method

This study focuses on the compilation of bathymetric data for the Bay of Bengal domain, specifically for the area between 15°N to 23°N latitudes and 85°E to 95°E longitudes, to ensure compatibility with the Finite Element Method (FEM). A high-resolution color map of the region was generated using ArcGIS software, which served as the basis for discretizing the domain into a triangular mesh. Bathymetric data for the Bay of Bengal, including the Meghna River, was then compiled using a combination of recent data from the Bangladesh Inland Water Transport Authority (BIWTA) and generated data, which were interpolated using a two-way cubic spline interpolation method. The Modified Inverse Distance Weighted Interpolation (MIDWI) method was employed to map this bathymetric data to the nodal points of the triangular mesh, ensuring accurate representation of the seafloor’s varying depths. This work is essential for developing a reliable storm surge prediction model for the Bay of Bengal using FEM, as it provides the necessary geometric and bathymetric details required for precise numerical modeling.

Surge-NF: Neural Fields inspired peak storm surge surrogate modeling with multi-task learning and positional encoding

Storm surges pose a significant threat to coastal communities, necessitating rapid and precise storm surge prediction methods for long-time risk assessment and emergency management. High-fidelity numerical models such as ADCIRC provide accurate storm surge simulations but are computationally expensive. Surrogate models have emerged as an alternative option to alleviate the computational burden by learning from available numerical datasets. However, existing surrogate models face challenges in capturing the highly non-stationary and non-linear patterns of storm surges, resulting in over-smoothed response surfaces. Moreover, the dry–wet status of nearshore nodes has not been informatively considered in the training process.This study proposes Surge-NF, a novel point-based surrogate model inspired by Neural Fields (NF) from computer graphics. Surge-NF introduces two key innovations. A positional encoding module is proposed to mitigate over-smoothing of high-frequency peak storm surge spatial dependencies. A multi-task learning framework is proposed to simultaneously learn and predict the dry–wet status and peak surge values, leveraging task dependencies to improve prediction accuracy and data efficiency. We evaluate Surge-NF on the NACCS database with comparison to state-of-the-art alternative surrogate models. Surge-NF consistently reduces RMSE/MAE by 50% and achieves 4–5 times computational cost gain over baselines, requiring only 50 training storms to produce accurate predictions. The complementary benefits of the positional encoding and multi-task learning modules are evident from the improved prediction capability with their combined use.Overall, Surge-NF represents a significant advancement in storm surge surrogate modeling, offering its novel and unique ability to capture high-frequency spatial variations and leverage task dependencies. It has the potential to greatly enhance storm surge risk assessment and emergency response management, enabling effective decision-making and mitigation strategies to safeguard coastal communities from the devastating impacts of storm surges.

Explanations for the positive storm surges on the left side of landfall typhoons in China

The coastal regions of Southeast China frequently experience unusual positive storm surges on the left side of landfalling typhoons, a phenomenon historically overlooked and inadequately explained by conventional circular wind field models. In this study, a high resolution, two-dimensional storm surge model based on ADCIRC along with tide gauge data were used to investigate the spatiotemporal distribution of these surges and proposes underlying mechanisms, informed by a comparative analysis of circular and ERA5 reanalysis wind fields during typical typhoon event 9711 Winnie. Analyzing tide gauge data spanning from 1986 to 2016, the study uncovers a distinct pattern of left-side positive storm surges along the southeastern coast, notably on the Fujian coast and within the Taiwan Strait, which are found to be comparable to those on the cyclone’s right side. The research also documents a significant escalation in both the frequency and intensity of these left-side surges over the past three decades. Simulation results highlights the inadequacies of circular wind field models in operational forecasting and emphasizes the necessity of accounting for topographic influences and the structural complexity of wind fields in storm surge predictions. This is particularly pertinent in semi-enclosed seas with intricate hydrodynamics, such as the Taiwan Strait. The insights gleaned from this study are pivotal for enhancing the real-time simulation and prediction of storm surges, which are vital for coastal safety and disaster prevention measures.

Predictability of Hurricane Storm Surge: An Ensemble Forecasting Approach Using Global Atmospheric Model Data

Providing storm surge risk information at multi-day lead times is critical for hurricane evacuation decisions, but predictability of storm surge inundation at these lead times is limited. This study develops a method to parameterize and adjust tropical cyclones derived from global atmospheric model data, for use in storm surge research and prediction. We implement the method to generate storm tide (surge + tide) ensemble forecasts for Hurricane Michael (2018) at five initialization times, using archived operational ECMWF ensemble forecasts and the dynamical storm surge model ADCIRC. The results elucidate the potential for extending hurricane storm surge prediction to several-day lead times, along with the challenges of predicting the details of storm surge inundation even 18 h before landfall. They also indicate that accurately predicting Hurricane Michael’s rapid intensification was not needed to predict the storm surge risk. In addition, the analysis illustrates how this approach can help identify situationally and physically realistic scenarios that pose greater storm surge risk. From a practical perspective, the study suggests potential approaches for improving real-time probabilistic storm surge prediction. The method can also be useful for other applications of atmospheric model data in storm surge research, forecasting, and risk analysis, across weather and climate time scales.

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.

Prediction of Storm Surges in the East Coast of Peninsular Malaysia in Response to Climate Change

This study employed the MIKE 21HD modeling framework, integrating data from the future climate database, "Policy Decision Making for Future Climate Change (d4PDF)" on selected critical ensembles to investigate the impact of climate change on storm surge height (SSH) along the East Coast of Peninsular Malaysia during the Northeast Monsoon season. Three scenarios, including past-historical data, future climate with a 2°C increment (+2K), and future climate with a 4°C increment (+4K), were analyzed. The study had validated model accuracy through root mean square error (RMSE), demonstrating alignment with observed data. Results indicated a correlation between temperature increase and storm surge height (SSH), with higher temperature scenarios leading to more severe surge outcomes. The analysis identified the Future +4K scenarios as yielding the most critical SSH across all stations, with recorded SSH at Kuala Pahang (SSHmax = 1.379 m), Chendering (SSHmax = 1.344 m), Geting (SSHmax = 1.251 m), and Tanjung Sedili (SSHmax = 1.251 m) stations. Considering the findings, it was recommended to reassess the baseline provision for coastal engineering projects, suggesting a storm surge resilience design exceeding the observed maximum. The utilization of d4PDF data in this study laid the groundwork for targeted strategies to mitigate the impacts of climate-induced storm surges in vulnerable coastal regions.

Adaptive multi-fidelity Monte Carlo for real-time probabilistic storm surge predictions

Real-time, probabilistic predictions of the expected storm surge represent an important information source for guiding emergency response decisions during landfalling tropical storms/cyclones. The uncertainty quantification in these predictions is accomplished as follows: an ensemble of sample storm scenarios is generated based on the nominal storm advisory and a probabilistic description of the corresponding forecast errors; subsequently high-fidelity numerical simulations are performed to predict the surge for each of these scenarios; the simulation results are finally leveraged to estimate the statistics of interest. This process is repeated whenever a new storm advisory is issued. The need to establish predictions in real-time, to support emergency planning decisions, creates incentives for high computational efficiency for this probabilistic estimation. This paper investigates an adaptive Multi-Fidelity Monte Carlo (MFMC) framework for achieving such an objective. As a lower-fidelity model within the MFMC setup, a surrogate model is adaptively developed based on high-fidelity numerical simulations from the current or past storm advisories, achieving information sharing across them. MFMC leverages the correlation between the high- and low-fidelity models to establish unbiased predictions with high statistical accuracy, significantly improving abilities to provide timely and informative estimates. To facilitate the development of the low-fidelity model using only a small number of high-fidelity numerical simulations, a combination of physics-based and data-driven dimensionality reduction techniques are introduced for the input and the output, respectively, of the surrogate model. To accommodate the use of high-fidelity simulations from the current advisory in the surrogate model calibration, the MFMC implementation is established using leave-one-out surrogate model predictions. Finally, the challenge of establishing MFMC predictions for a large number of quantities of interest (QoIs), corresponding to the surge at different geographic locations, is discussed. These QoIs may advocate conflicting decisions for the optimal MFMC implementation, and an efficient search for a compromise solution is introduced. The efficiency of the proposed MFMC framework is showcased by considering its application to different historical storms.

Accurate storm surge prediction in hurricane area of the Atlantic Ocean using a new multi-recursive neural network based on gate recursive unit

This paper proposes an enhanced storm surge forecasting approach utilizing novel neural networks, specifically a model based on gated recurrent units (GRUs), and incorporates a distortion loss function known as Distortion Loss Including Shape and Time (DILATE) for improved prediction accuracy and reliability. Historical data from four stations in the hurricane-prone region of the Atlantic Ocean are utilized for both training and testing purposes. Various models are employed to predict storm surges with lead times of 3, 6, and 12 hours, and their performance is evaluated by comparing the predicted values with the observed ones. Furthermore, the influence of various physical factors on storm surge formation is examined. The findings indicate that the GRU-DILATE model is more suitable for storm surge prediction, particularly for longer lead times. This model effectively reduces prediction delays and accurately captures the trends in wave height associated with storm surges. Among the physical factors studied, wind speed and atmospheric pressure emerge as the most important variables influencing storm surges. Additionally, the significance of wind direction is highlighted, as it plays a crucial role in determining whether the sea level rises or falls during the initial stages of a storm surge. It is expected that the proposed model has the potential to enhance response and preparedness measures in hurricane-prone areas, finding applications in coastal planning, infrastructure design, disaster management, and other aspects of ocean engineering.

Economic loss assessment of typhoon-induced storm surge disasters in the South China Sea based on GSA-BP model

In the context of global climate warming and rising sea levels, the frequency of tropical cyclones in the South China Sea region has shown a significant upward trend in recent years. Consequently, the coastal areas of the South China Sea are increasingly vulnerable to storm surge disasters induced by typhoon, posing severe challenges to disaster prevention and mitigation in affected cities. Therefore, establishing a multi-indicator assessment system for typhoon storm surges is crucial to provide scientific references for effective defense measures against disasters in the region. This study examines 25 sets of typhoon storm surge data from the South China Sea spanning the years 1989–2020. A comprehensive assessment system was constructed to evaluate the damages caused by storm surges by incorporating the maximum wind speed of typhoons. To reduce redundancy among multiple indicators in the assessment system and enhance the stability and operational efficiency of the storm surge-induced disaster loss model, the entropy method and bootstrap toolbox were employed to process post-disaster data. Furthermore, the genetic simulated annealing algorithm was utilized to optimize a backpropagation neural network intelligent model (GSA-BP), enabling pre-assessment of the risks associated with storm surge disasters induced by typhoon and related economic losses. The results indicate that the GSA-BP model outperforms the genetic algorithm optimized BP model (GA-BP) and the simulated annealing algorithm-optimized BP model (SA-BP) in terms of predicting direct economic losses caused by storm surges. The GSA-BP model exhibits higher prediction accuracy, shorter computation time, and faster convergence speed. It offers a new approach to predicting storm surge losses in coastal cities along the South China Sea.

A framework for flexible peak storm surge prediction

A framework for flexible peak storm surge prediction

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.

Comparative Analysis of BALSSA and Conventional NWP Methods: A Case Study in Extreme Storm Surge Prediction in Macao

In coastal regions, accurate storm surge prediction is crucial for effective disaster management and risk mitigation. This study presents a comparative analysis between BALSSA (Bidirectional Attention-based LSTM for Storm Surge Architecture) and the Japan Meteorological Agency (JMA) numerical storm surge model, focusing on the Saola-induced storm surge in Macao, September 2023. To train and assess the model, we leverage an extensive dataset comprising meteorological and tide level information from more than 80 typhoon occurrences in Macao spanning the period from 2017 to 2023. The results provide evidence of BALSSA’s effectiveness in capturing the complex spatio-temporal dynamics of storm surges, with a lead time of up to 72 h, as reflected by its MAE of 0.019 and RMSE of 0.024. It demonstrates reliable accuracy in predicting storm surge magnitude, timing, and spatial extent, potentially contributing to more precise and timely warnings for coastal communities. Furthermore, the real-time data assimilation feature of BALSSA ensures up-to-date information, aligned with the latest observations, which is essential for effective emergency preparedness and response. The high-resolution grids enhance risk assessment, highlighting BALSSA’s potential for early warnings, emergency planning, and coastal risk management. This study contributes valuable insights to the broader field of storm surge prediction, guiding decision-making processes and supporting the development of effective strategies to enhance coastal resilience.

Storm surge prediction improvement using high resolution meso-scale model products over the Bay of Bengal

Storm surge prediction improvement using high resolution meso-scale model products over the Bay of Bengal

Assessment and Integration of ERA5 Reanalysis and Fujita−Takahashi Models for Storm Surge Prediction in the East China Sea

With global climate warming, the frequency and intensity of typhoons are increasing, highlighting the significance of studying storm surges for coastal engineering disaster mitigation. In this study, we assessed the predictive capabilities of the new ERA5 reanalysis model and the traditional Fujita−Takahashi model for storm surges. We found that the traditional Fujita−Takahashi model, utilizing a prelandfall typhoon wind field, exhibited higher accuracy in storm surge predictions, while the ERA5 reanalysis model, employing a postlandfall wind field, demonstrated superior performance. By considering the strengths and weaknesses of both wind field models and analyzing the impact of Typhoon In-fa (2021) on the East China Sea, we determined the influence of this typhoon on storm surge heights along the eastern coastal region. These research findings provide valuable insights for the development of effective protection strategies, offering valuable references for coastal resilience planning.

SHORT TERM SPATIALLY DENSE PREDICTION OF STORM SURGE ALONG THE NEW ZEALAND COASTLINE

Storm surge is the rise in water level generated by wind and atmospheric pressure changes associated with tropical or mid-latitude storms. In conjunction with tides, it is one major driver of coastal flooding associated with storms events. Because local inundation is strongly modulated by the local shape of the coastline and the bathymetric slope, accurate storm surge prediction by the mean of traditional numerical models requires the use of very fine grids and is hence very resource intensive. This means that the performance of a live prediction system based on such methods will likely be subject to a trade-off between prediction accuracy, prediction speed and cost (Wang et al., 2009). Several publications have demonstrated the potential of machine learning approaches for the prediction of storm surge (e.g. (Tiggeloven et al., 2021), (Cagigal et al, 2020)). However, the developed methods often focus on local predictors and aim at predicting storm surge at a single location at a time. In this study, we explore the use of several data driven methods as an alternative to numerical methods to predict storm surge along the coast of New Zealand.

PROBABILISTIC FORECASTING SYSTEM OF STORM SURGE USING A STOCHASTIC TYPHOON MODEL

This paper presents a newly developed real time probabilistic forecasting system of storm surge at an arbitrary target location when a typhoon is formed and approaches to the location. Such probabilistic information may be essential for business or public entities along the coast who need to make appropriate decision to what extent they should break their daily operations and prepare for the upcoming coastal hazard. Developed system, based on Monte Carlo simulation, consists of two parts: (i) generation of a number of virtual typhoons; and (ii) storm surge prediction for each typhoon.