Replacing a third-generation wave model with a fetch based parametric solver in coastal estuaries
Replacing a third-generation wave model with a fetch based parametric solver in coastal estuaries
复制标题
在沿海河口用基于提取的参数求解器替换第三代波浪模型
DOI：
10.1016/j.ecss.2021.107192
复制
发表时间：
2021
影响因子：
2.8
通讯作者：
R. Weaver
中科院分区：
文献类型：
--
作者：
Samuel C. Boyd;R. Weaver
在受限的河口环境中，使用第三代波浪模型来预测波高的计算成本可能很高，这表明需要开发模型来降低现有耦合水动力模型的计算成本。本研究重点开发和测试参数化波浪求解器，该求解器包含四种波高公式（SMB、SPM、TMA 和 CEM），用于预测受限河口环境中的波浪特性。重点是在不影响准确性的情况下提高效率，从而允许在桌面计算资源上执行集合波涌预报。对参数求解器性能的评估是双重的，首先将参数求解器与位于佛罗里达州东海岸的印第安河泻湖域的第三代波浪模型“近岸模拟波浪”(SWAN) 进行比较。然后将参数求解器和 SWAN 与河口域中某一点的现场 ADCP 波浪数据进行比较。创建三个不同的合成风场可以进行模型比较，风场可以测试参数模型，以便重现 (1) 完全发展的条件，(2) 风速变化，以及 (3) 热带风暴级风事件中的风向变化。为了进行一致性比较，同一域内的波高解由 SWAN 和参数模型生成。参数模型性能与 SWAN 的比较表明，4 成员参数模型的全局准确度在 87% 以内，与 SWAN 相比，运行时间提高了两个数量级以上。参数模型的集合平均波高解比实测波高低5.5%； SWAN 解比同一位置测量的波高高 5.5%。因此，参数波浪模型被证明是运行昂贵的第三代波浪模型来预测封闭河口系统中波浪的可行替代方案。
Within a restricted estuarine environment, the use of third-generation wave models for predicting wave heights can be computationally expensive, signaling a need for model development that reduces the computational costs of existing coupled hydrodynamic models. This study focuses on the development and testing of a parametric wave solver that incorporates four wave height formulations (SMB, SPM, TMA, and CEM) for predicting wave properties in a restricted estuarine environment. The emphasis is on improved efficiency without affecting accuracy, allowing for ensemble wave-surge forecasting to be performed on desktop computational resources. Evaluation of the performance of the parametric solver is twofold, first the parametric solver is compared to a third-generation wave model, Simulating Waves Nearshore (SWAN), for the Indian River Lagoon domain, which lies on Florida's east coast. Then both the parametric solver and SWAN are compared toin-situADCP wave data at a point in the estuarine domain. The creation of three different synthetic wind fields allows for model comparison, with wind fields permitting testing of the parametric model in order to reproduce (1) fully developed conditions, (2) wind speed variability, and (3) wind direction variability in tropical storm level wind events. For consistency comparison, wave height solutions over the same domain are generated by SWAN and the parametric models. Comparisons made between the parametric model performance and SWAN show a 4-member parametric model is accurate to within 87% globally, with a runtime improvement of over two orders of magnitude compared to SWAN. The parametric model's ensemble average wave height solution was 5.5% less than thein-situmeasured wave height; the SWAN solution was 5.5% greater than the measured wave height at the same location. Therefore, the parametric wave model proves to be a viable alternative to running an expensive third-generation wave model for predicting waves in an enclosed estuarine system.