This study presents a risk prediction of coastal hazards induced by typhoons, which are a severe natural hazard that often occur in coastal regions. Taking the coastal hazards happened in Shenzhen as a case study, where is a southeast coastal city of China, we described a methodology to predict the typhoon wind-surge-wave hazard. A typhoon empirical tracking model was adopted to construct full-track typhoon events for 1000 years, based on the statistical characteristics of observed typhoons from satellite imageries. For each individual typhoon, a wind-field model is applied to compute the wind speeds, while the Simulating Waves Nearshore and Advanced Circulation (SWAN+ADCIRC) coupled model is applied to simulate the significant wave heights (SWHs) and storm surge heights. By frequency distribution histogram, it is noted that there exhibits a heavy tail in the probability distribution of maximum surge heights and a thin tail of the peak wind speeds and SWHs in the coastal area of Shenzhen, China. Using the Generalized Pareto Distribution (GPD) model, the extreme values of typhoon wind-surge-wave associated with various return periods can be predicted. Taking account into the combined effects of the wind, surge and wave, the joint hazard maps of typhoon wind-surge-wave can be produced for the study area. The methodology of this case study can provide a new reference for risk prediction of coastal hazards induced by typhoon in similar coastal regions like Shenzhen, China.

The Yellow Sea cold water mass (YSCWM) occupies a wide region below the Yellow Sea (YS) thermocline in summer which is the most conservative water and may contain clearer climate signals than any other water masses in the YS. This study investigated the low-frequency variability of the southern YSCWM (SYSCWM) and established the “forcing mechanism bridge” using correlation analysis and singular value decomposition. On the interannual timescale, the southern oscillation can affect the SYSCWM through both the local winter monsoon (WM) and the sea surface net heat flux. On the decadal timescale, the Pacific decadal oscillation (PDO) can affect the SYSCWM via two “bridges”. First, the PDO affects the SYSCWM intensity by Aleutian low (AL), WM, and surface air temperature (SAT). Second, the PDO affects the SYSCWM by AL, WM, Kuroshio heat transport, and Yellow Sea warm current. The Arctic oscillation (AO) affects the SYSCWM by the Mongolian high, WM, and SAT. Before and after the 1980s, the consistent phase change of the PDO and the AO contributed to the significant decadal variability of the SYSCWM. Finally, one simple formula for predicting the decadal variability of SYSCWM intensity was established using key influencing factors.

Typhoons have caused serious economic losses and casualties in coastal areas all over the world. The big size of the tropical cyclone sample by stochastic simulation can effectively evaluate the typhoon hazard risk, and the typhoon full-track model is the most popular model for typhoon stochastic simulation. Based on the advantages of machine learning in dealing with nonlinear problems, this study uses a backpropagation neural network (BPNN) to replace the regression model in the empirical track model, reestablishes the neural network model for track and intensity prediction in typhoon stochastic simulation, and constructs full‐track typhoon events of 1000 years for Northwest Pacific basin. The validation results indicate that the BPNN can improve the accuracy of typhoon track and intensity prediction.

Mesoscale eddies occur frequently in the Luzon Strait and its adjacent area, and accurate prediction of eddy structure changes is of great significance. In recent years, artificial neural network (ANN) has been widely applied in the study of physical oceanography with the continuous accumulation of satellite remote sensing data. This study adopted an ANN approach to predict the evolution of eddies around the Luzon Strait, based on 25 years of sea level anomaly (SLA) data, 85% of which are used for training and the remaining 15% are reserved for testing. The original SLA data were firstly decomposed into spatial modes (EOFs) and time-dependent principal components (PCs) by the empirical orthogonal function (EOF) analysis. In order to calculate faster and save costs, only the first 35 PCs were selected as predictors, whereas their variance contribution rate reached 96%. The results of predicted reconstruction indicated that the neural network-based model can reliably predict eddy structure evaluations for about 15 days. Importantly, the position and variation of four typical eddy events were reconstructed, and included a cyclone eddy event, an eddy shedding event, an anticyclone eddy event, and an abnormal anticyclone eddy event.

Zhejiang Province in China is one of the most typhoon-prone regions globally, making typhoon and storm surge hazard analysis critically important for disaster mitigation. This study integrates the Tropical Cyclone Risk Model (TCRM) with a machine learning-based storm surge forecasting model to analyze typhoon hazards and storm surge risks at four representative coastal sites in Zhejiang Province: Haimen, Ruian, Wenzhou, and Zhapu. Firstly, the input database of the TCRM has been updated and subsequently used to generate a 1000-year synthetic typhoon event catalog for the Northwest Pacific region. Secondly, four machine learning models—Long Short-Term Memory (LSTM), Back Propagation (BP), Support Vector Regression (SVR), and Random Forest (RF)—were developed to forecast storm surge component at the four sites, with sensitivity analysis conducted on the input parameters. Among the four models, RF consistently outperformed the others across all four sites. Thirdly, by integrating the storm surge forecasting model with the Yan Meng (YM) typhoon wind field model, extreme wind speed sequences and extreme surge component sequences were derived for the four coastal sites. Finally, four extreme value distribution models—empirical distribution, Weibull, Gumbel, and Generalized Pareto Distribution (GPD)—were applied to fit the extreme wind and surge sequences. Goodness-of-fit tests indicated that the GPD best captured extreme wind speeds at all four sites and extreme surge levels at Haimen, Ruian, and Wenzhou. Using the optimal distributions, return periods (10-, 50-, 100-, and 200-year) for extreme wind speeds and surge components were calculated, providing actionable references for disaster risk management authorities.