海洋环流与波动重点实验室知识库

中国东南沿海区域经济富庶、城镇密集，同时也长期遭受台风灾害的侵扰。分析台风以及台风引起的风暴潮的危险性，预测可能发生的极值风速与风暴增水，对这些地方结构的抗风设计以及防灾减灾至关重要。

本文首先以中国东南沿海城市深圳市为例，采用传统的Monte-Carlo方法分析其台风危险性。基于CMA西北太平洋热带气旋最佳路径数据集的历史台风数据，提取了对深圳市有影响的台风的关键参数，并确立了每个关键参数最优的概率模型。以这些概率模型为基础采用Monte-Carlo方法进行随机抽样，产生1000年虚拟台风事件。本文采用了Yan Meng（YM）风场模型，并分析了该模型对一些风场参数的敏感性，得出该模型对地面粗糙度以及Holland气压参数B非常的敏感，两者对计算风速的大小具有相反的作用。采用YM风场模型对虚拟台风的风场进行模拟，并提取虚拟台风的极值风速序列。利用不同的极值分布对极值风速序列进行拟合，通过拟合优度检验得出Weibull分布要优于Gumbel分布。最后预测了深圳市不同重现期的极值风速，并与结构规范中推荐的风速以及其他一些参考文献的结果进行对比，得出产生差异的主要原因是Holland气压参数B模型的不同。

其次由点到面，基于Monte-Carlo方法，本文分析了整个中国东南沿海区域的台风危险性。首先将整个东南沿海区域分成0.25° ´ 0.25°的网格点，然后利用Monte-Carlo方法产生每个网格点1000年间的虚拟台风事件。本文采用 YM 风场模型模拟了100个历史台风的最大风速，通过使这些最大风速与观测的最大风速误差和最小，建立了一组新的计算Holland气压剖面参数B和最大风速半径R_max的公式。最后利用新的台风参数计算方案、YM 风场模型、特定点的台风衰减模型以及极值分布模型，预测了每个网格点不同重现期的极值风速，为中国东南沿海台风多发区域绘制了不同重现期的设计风速图。

由于Monte-Carlo方法依赖于一定区域内气候保持一致的假定，因此较为适合研究单个站点的台风危险性。为了在更大区域上研究台风危险性，接下来本文采用较为先进的简化经验路径方法构造了整个西北太平洋海域1000年的热带气旋事件集，并预测了每个站点不同重现期的极值风速，形成了中国沿海台风多发区新的设计风速分布图。将经验路径方法与Monte-Carlo方法预测的极值风速进行对比，发现两种方法预测结果的差异主要是由两种方法构造的虚拟台风的中心压强存在差异以及模型本身的不确定性造成的。我们还研究了台风衰减模型、路径模型、Holland气压剖面参数、最大风速半径和极值分布对预测的极值风速的影响。发现不同的台风衰减模型对预测的极值风速影响最小，这主要是由于不同的衰减模型得到的登陆台风的压强相差不大；在我国东南沿海大部分地区，非简化经验路径模型预测的风速值要大于简化路径模型预测的结果，这主要是由于两种路径模型构造的台风中心压强以及台风路径距研究点的最小距离存在差异；不同的气压剖面参数模型会对极值风速的预测产生较大的影响，当参数值偏大时，预测的极值风速也偏大；不同极值分布预测的极值风速有很大的差异，这主要是由极值分布本身的特点所造成的，一般概率分布具有“长尾”特征的分布，预测的极值风速偏大。基于构造的1000年热带气旋事件集，估算了近年来影响中国最强的4个台风，分别是Meranti（2016）、Hato（2017）、Mangkhut（2018）和Lekima（2019），在我国东南沿海站点引起的极值风速的重现期，评估了它们的危险性。

最后，以采用台风经验路径模型产生的1000年热带气旋事件集为基础，结合YM台风风场模型以及SWAN+ADCIRC耦合的风暴潮模型，研究了深圳市台风的风-潮-浪的危险性。对模拟得到的台风风-潮-浪数据进行统计分析，发现深圳市最大增水的频率分布直方图有“长尾”的特征，而最大风速以及最大有效波高的频率分布具有“短尾”的特征。采用广义帕雷托分布（GPD）来估计台风风-潮-浪的上尾分布，分别得到了三个量不同重现期的预测值。为了考虑风-潮-浪的综合效应，建立了深圳市台风风-潮-浪联合灾害图。我们可以将本文对深圳市的风险评估方法应用于其他沿海地区，并且可以将其扩展到考虑未来气候变化的影响。

Southeast coastal region is the most developed and populated area in China, and is also one of the regions most seriously impacted by typhoons in the world. Analyzing typhoon and storm surge risks and predicting maximum wind speeds and surge levels of typhoons are vital for the design of critical structures and typhoon disaster mitigation in these areas.

Firstly, taking the southeastern coastal city of Shenzhen as an example, this study employ the traditional Monte-Carlo simulation to analyze of the typhoon wind hazard of Shenzhen. Based on the historical typhoons obtained from the China Meteorological Administration(CMA) Best Track Dataset for Tropical Cyclones over the Western North Pacific, typhoon key parameters are extracted and optimal statistical distributions are established for these parameters in relation to Shenzhen. Then, the Monte-Carlo method is employed to sample from each distribution to generate 1000-year virtual typhoons. The Yan Meng (YM) wind field model is introduced, and the sensitivity of the YM model to several parameters is discussed. We find that the model is very sensitive to the surface roughness and Holland pressure profile parameter B, and the two have opposite effects on the wind speed. We use the YM model to simulate the wind field of the virtual typhoons and extract the extreme wind speeds. Different extreme value distributions are used to fit the extreme wind speed series, and the Weibull distribution is better than Gumbel distribution through the goodness-of-fit test. Finally, the extreme wind speeds for different return periods are predicted and compared with the results from current structural code and other references. We find the main reason for the difference is the different B model.

From single point to entire area, this paper describes a technique for analyzing the areawide typhoon risk for the southeast China coastal region based on the Monte-Carlo method. The whole region is divided into 0.25° ´0.25° grid cells and each grid became a site of interest. A Monte-Carlo method is used to generate virtual typhoons and 1000 years of typhoons are simulated for the different grid cells. YM wind field model is adopted to simulate the maximum wind speeds of 100 historical typhoons. By minimizing the errors between these maximum wind speeds and the observed maximum wind speeds, a new set of formulas is established to calculate the radius to maximum winds (R_max) and Holland pressure profile parameter (B). Using this newly developed scheme which incorporated the YM wind field model, region-specific statistical models for the decay rate of typhoons after reaching land, and the appropriate extreme value distribution, we predict the site-specific extreme wind speeds associated with various return periods and propose a new map of wind speeds for the typhoon-prone coasts of China.

Because the Monte-Carlo method relies on an assumption of uniform climatology over the subregion, it is more suitable to study the typhoon risk at a single point. In order to properly model typhoon wind risk over large regions, a simplified empirical track model approach is used to simulate 1000-year storms in the whole Northwest Pacific Basin. The extreme wind speeds of different return periods for different locations are predicted and a new map of wind speeds for the typhoon-prone coasts of China is proposed which can provide the completely new reference for assessing the risk of large-scale systems. Besides we compare the extreme wind speed predicted by the empirical track model and Monte-Carlo method. And we find the difference of extreme wind speed is mainly caused by the difference of the central pressure of the virtual typhoons constructed by two methods and the uncertainty of the model itself. We investigate the influence of typhoon decay model, track model, Holland pressure profile parameter, the radius to maximum winds, and the extreme value distribution on the predicted extreme wind speed. We find the different typhoon decay models have least influence on the predicted extreme wind speed. This is mainly because the central pressures obtained by the different decay models have little difference. Over most of the southeast coast of China, the predicted wind speed by the non-simplified empirical track model is larger than that from the simplified tracking model. This is mainly due to the difference in the typhoon central pressure and minimum approch distance constructed by the two track models. Different B models will have a greater impact on the prediction of extreme wind speed. When the B value is larger, the predicted extreme wind speed is also larger. The extreme wind speed predicted by different extreme value distribution is quite different which is caused by the characteristics of the extreme value distribution itself. Generally, the extreme wind speed predicted by the distribution with the "thick tail" is larger. Based on the generated 1000-year virtual typhoons for Northwest Pacific basin, we predict the the return periods of typhoon wind speeds along the China southeast coast caused by four super typhoons Meranti (2016), Hato (2017), Mangkhut (2018) and Lekima (2019) in order to assess the typhoon wind hazard.

Finally, based on the 1,000 years of full-track typhoon events by empirical track model, the YM wind field model and SWAN+ADCIRC (Simulating Waves Nearshore +Advanced Circulation) coupled model, we investigate typhoon wind-surge-wave risk for Shenzhen City. Statistical analysis is carried out on the simulated wind-surge-wave data. It is observed that the probability distribution of maximum surge heights at the Shenzhen exhibited a heavy tail, and that of the peak wind speeds and SWHs exhibited a thin tail. The generalized Pareto distribution (GPD) is applied to estimate the upper tail of the storm wind-surge-wave. The resulting return periods of wind-surge-wave are predicted, respectively. The joint typhoon wind-surge-wave hazard maps for Shenzhen are also developed to consider the combined effects of the wind, surge and wave. This risk assessment methodology may be applied to other coastal areas and can be extended to consider the effect of future climate change.