Journal of Tropical Oceanography >
Storm surge simulations of the coastal area of Shenzhen using different types of typhoon meteorological fields—a case study of Typhoon Mangkhut*
Copy editor: LIN Qiang
Received date: 2023-02-10
Revised date: 2023-05-04
Online published: 2023-05-10
Supported by
Shenzhen Science and Technology Innovation Committee(WDZC20200819105831001)
Guangdong Basic and Applied Basic Research Foundation(2022B1515130006)
Storm surges caused by typhoon seriously affect life and business in coastal areas, which is one of the most serious marine disasters that cause economic losses. Shenzhen is located on the edge of the northern South China Sea, which is vulnerable to typhoon induced storm surges. The study of Shenzhen offshore storm surges can not only promote understanding of the physical mechanisms of storm surges, but also has an important significance for the effective disaster prevention and reduction warning of coastal cities. In the process of storm surge modelling studies, a typhoon meteorological field is the key factor for the accuracy of storm surge model simulations. Based on the FVCOM (finite volume community ocean model) current model and SWAN (simulation wave nearshore) wave model, a regional storm surge and wave coupling model is established for the offshore area of Shenzhen. We use reanalysis of meteorological data (European center for medium weather forecasting, ECMWF), ideal typhoon model (Holland) and atmospheric model (weather research and forecast, WRF) as driving field conditions to simulate the storm surge process during Typhoon Mangkhut. The main conclusions are as follows: the low resolution ECMWF reanalysis meteorological data is difficult to accurately reflect the horizontal structures of typhoon, which leads to simulation errors. Overall, Holland meteorological field can accurately simulate Typhoon Mankhut, but it cannot reproduce the structural deformation of typhoon in the coastal region, which results in high simulated storm surge water levels in and around Shekou (Shenzhen Bay, inside the Pearl River Estuary). WRF has a good simulation effect on wind speeds, air pressure fields, water levels and waves as a whole. WRF is a good solution to the problem of high storm surge levels in Holland near the typhoon landfall. The quantitative improvement of WRF in the Pearl River Estuary and Shenzhen Bay area can reach about 20%~30%. In the future storm surge study, if the Holland meteorological field is used, care should be taken into simulation results of the above areas. In addition, both Holland and WRF have good wave simulation results.
ZHANG Zheran , HU Junyang , ZHOU Kai , ZHANG Penghui , XING Jiuxing , CHEN Shengli . Storm surge simulations of the coastal area of Shenzhen using different types of typhoon meteorological fields—a case study of Typhoon Mangkhut*[J]. Journal of Tropical Oceanography, 2023 , 42(6) : 1 -14 . DOI: 10.11978/2023017
表1 WRF模型配置及参数化方案选择Tab. 1 WRF model configuration and parameterization scheme selection |
模型有关参数 | 具体配置及参数选择 |
---|---|
驱动数据 | 美国NCEP的全球气象再分析数据FNL |
数据分辨率 | 空间分辨率: 1°×1°; 时间分辨率: 6h |
计算时段 | 201809140000—201809180000(UTC) |
计算时长 | 4d (96h) |
计算区域 | 104°—134°E、15°—30°N |
网格分辨率 | dx=dy=9km |
网格数目 | 346×192 |
计算时间步长 | 30s |
结果输出频率 | 5min一次 |
微物理过程方案 | WSM 3 类简单冰方案 |
长波辐射方案 | RRTM方案 |
短波辐射方案 | Dudhia方案 |
调用辐散物理方案的时间间隔 | 30min |
近地面层方案 | Monin-Obukhov方案 |
陆面过程方案 | Noah方案 |
边界层方案 | YSU方案 |
积云参数化方案 | 浅对流Kain-Fritsch方案 |
积云参数化方案的调用时间间隔 | 5min |
图6 台风“山竹”期间各测站不同风场下风速时间序列对比图Fig. 6 Comparisons of wind speed at each measuring station during Typhoon Mangkhut |
图7 台风“山竹”期间各测站地表气压时间序列对比图Fig. 7 Comparisons of surface air pressure at each measuring station during Typhoon Mangkhut |
表2 各气象场各测站平均均方根误差计算表Tab. 2 Calculation table of the mean square root error of each station in each meteorological field |
气象场名称 | 风速平均误差/(m·s-1) | 气压平均误差/hpa |
---|---|---|
ECMWF | 3.0 | 5.5 |
Holland | 7.2 | 12.1 |
WRF | 4.5 | 4.6 |
图10 台风“山竹”期间各气象场各测站风暴潮水位模拟结果对比图Fig. 10 Comparison of storm surge levels under various meteorological forcing at each measuring station during Typhoon Mangkhut |
表3 各气象场各测站风暴潮水位误差计算表Tab. 3 Calculation table of storm surge level error at each station in each meteorological field |
气象场 | 蛇口站 | 南澳站 | 东山站 | ||||||
---|---|---|---|---|---|---|---|---|---|
RMSE/m | 极值误差/m | 相位误差 | RMSE/m | 极值误差/m | 相位误差 | RMSE/m | 极值误差/m | 相位误差 | |
ECMWF | 0.37 | 0.48 | 0 | 0.37 | 0.72 | 0 | 0.36 | 0.80 | 0 |
Holland | 0.42 | 1.09 | 0 | 0.30 | 0.18 | 0 | 0.29 | 0.14 | 0 |
WRF | 0.34 | 0.10 | 2h | 0.32 | 0.20 | 1~2h | 0.33 | 0.18 | 1~2h |
图11 台风“山竹”期间各气象场模拟的深圳近海风暴潮水位随时间变化图Fig. 11 Variations of Shenzhen offshore storm surge water level simulated by various meteorological fields during Typhoon Mangkhut |
图12 台风“山竹”期间各气象场模拟的最高风暴潮水位空间分布图Fig. 12 Maps of maximum storm surge level simulated using various meteorological fields during Typhoon Mangkhut |
图14 台风“山竹”期间各气象场各测站波浪有效波高模拟结果对比图Fig. 14 Comparison of significant wave height using various meteorological fields at each measuring station during Typhoon Mangkhut |
表4 各气象场各测站的波浪有效波高均方根误差计算表Tab. 4 Calculation table of the mean square root error of each station in each meteorological field |
气象场 | 各测站平均均方根误差/m | 平均相对误差/% |
---|---|---|
ECMWF | 0.71 | 20.51 |
Holland | 0.30 | 10.04 |
WRF | 0.38 | 13.12 |
[1] |
曹越男, 张增海, 原野, 等, 2014. 南海台风风暴潮数值模拟及其对气象强迫场的敏感性试验[J]. 热带气象学报, 30(6): 1119-1126.
|
[2] |
邓国通, 刘敏聪, 邢久星, 等, 2022. 深圳近海风暴潮影响因素分析[J]. 热带海洋学报, 41(3): 91-100.
|
[3] |
冯士筰, 李凤岐, 李少菁, 1999. 海洋学导论[M]. 北京: 高等教育出版社 (in Chinese).
|
[4] |
高钦钦, 朱建荣, 端义宏, 等, 2012. 对称和非对称台风对东海南海风暴潮影响比较[J]. 华东师范大学学报(自然科学版), 6: 57-72.
|
[5] |
蒋小平, 钟中, 张金善, 等, 2007. 台风浪模拟预报中的风场比较研究[J]. 海洋通报, 2: 11-19.
|
[6] |
梁必骐, 梁经萍, 温之平, 1995. 中国台风灾害及其影响的研究[J]. 自然灾害学报, 1: 84-91.
|
[7] |
林伟, 方伟华, 2013. 西北太平洋台风风场模型中Holland B系数区域特征研究[J]. 热带地理, 33(2): 124-132.
|
[8] |
毛献忠, 姜茜, 2012. 深圳香港海域可能最高潮位和浪高计算分析[J]. 海洋工程, 30(2): 129-135.
|
[9] |
唐燕玲, 2018. 基于波流耦合模型的洋山海域风暴潮水动力特性研究[D]. 杭州: 浙江大学.
|
[10] |
文萍, 许映龙, 柳龙生, 2019. 台风“山竹”(1822)引发华南暴雨过程机制分析[J]. 海洋气象学报, 39(3): 29-35.
|
[11] |
于福江, 傅赐福, 郭洪琳, 等, 2020. 现代风暴潮预报技术及应用[M]. 北京: 科学出版社 (in Chinese).
|
[12] |
曾祥锋, 2018. 基于WRF中尺度数值模式台风风场模拟及验证[D]. 广州: 广州大学.
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
[29] |
|
[30] |
|
[31] |
|
/
〈 | 〉 |