| 摘要: |
| 本文基于时间分布参数设置，利用伴随同化方法，反演了Ekman模型中随时间变化的风应力拖曳系数，并在孪生实验和实际实验中对该方法进行了验证。在孪生实验中，研究了参数反演结果对不同影响因素的响应，包括：风速分布、风应力系数分布、风应力系数初始猜测值、风应力系数独立变量个数、观测数据误差和观测的深度。孪生实验结果验证了伴随同化方法反演Ekman模型中时变风应力系数的有效性，具体包括如下五个方面结论：1）不同风速分布下均能成功反演出不同风应力拖曳系数分布；2）反演结果对初始猜测值较为敏感，风应力系数初始猜测值越接近给定值，反演结果越好；3）风应力系数独立点个数的选取会显著影响反演结果，合理的选择有利于提高反演效率及减小观测数据误差；4）观测误差能够影响反演结果，观测数据误差在20%以下时能取得合理的反演结果；5）反演结果对观测数据的表层和次表层流速更为敏感，这是由Ekman流的物理性质决定的。实际实验，利用百慕大锚系试验平台的风速和流速数据，去除周期性潮流和地转流成分后得到Ekman流成分，并作为观测输入到该同化模型，反演出了适用于该区域和该时段的随时间变化的风应力系数。通过比较模拟流速和观测流速，证明利用伴随同化方法能从实测数据中反演出合理的时变风应力系数，对于海洋模型风应力系数的确定是一项有益的尝试。 |
| 关键词: 伴随同化 风应力系数 时间变化 Ekman模型 参数估计 |
| DOI：10.11693/hyhz20190100029 |
| 分类号:P731 |
| 基金项目:国家重点研发计划“全球变化及应对”重点专项，2017YFA0604100号；国家重点研发计划“海洋环境安全保障”重点专项，2017YFC1404000号；国家自然科学基金，41876086号；中央高校基本科研业务费专项资金。 |
|
| ADJOINT PARAMETER ESTIMATION OF TIME-VARYING WIND DRAG COEFFICIENT FOR AN EKMAN MODEL |
|
LI Guo-Qing1, GAO Yan-Qiu2, ZHANG Ji-Cai1
|
|
1.Institute of Physical Oceanography, Ocean College, Zhejiang University, Zhoushan 316021, China;2.State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou 310012, China
|
| Abstract: |
| Based on the time-varying parameters, we used the adjoint assimilation method to invert the time-varying wind drag coefficients (WDCs) in the Ekman model. The method is testified by running twin experiments and practical experiments. In the twin experiments, the response of estimated WDCs to different influencing factors is studied, including:distributions of wind speed, distributions of WDCs, initial guess values, number of independent parameters, observation errors, and layers of observations. The results of twin experiments verified the effectiveness of the time-varying WDCs in the Ekman model. The results are followed. 1. The different distributions of WDCs can be successfully inverted under different wind speed distributions. 2. The inversion result is sensitive to the values of initial guess; therefore, the initial guess should be as reasonable as possible to improve the results and reduce the convergence time. 3. The selection of the number of the independent WDCs can significantly affect the inversion results. 4. The observation error can affect the inversion results, and reasonable inversion results can be obtained with maximum error below 20%. 5. The inversion results are more sensitive to the surface and subsurface observations, which is determined by the physical dynamics of the Ekman model. In practice, the observed Ekman current components are obtained from Bermuda Testbed Moorings by removing the periodic tidal components and the geostrophic components. Then the observed Ekman currents are assimilated into the model to invert the time-varying WDCs during the observation. By comparing the simulated with the observed velocities, we proved that the adjoint assimilation method could derive reasonable time-varying WDCs from measured data, which is a useful attempt to determine the WDCs for ocean models. |
| Key words: adjoint assimilation wind drag coefficient time varying Ekman layer model parameter estimation |
