跨尺度溢油数值模型的不确定性和可预报性研究

With the rapid development of marine oil transportation and offshore oil exploitation in China, the risk of oil spills to the marine environment has increased drastically. As a basic part of oil contigency plan and ecological risk assessment, oil spill numerical model has been confronted with the single scale in the research scope and the uncertainty factors of influence. Therefore, it is of great significance in making improvements in existing oil spill models. The project aims at studying the multi-scale characteristics of oil spill transport process and simulating the spilled oil trajectory by coupling the far-field and near-field models. On the basis, combination with hydrological and meteorological data, the stochastic simulation statistics model is set up for oil spill risk prediction by employing Latin Hypercube Sampling method, whose credibility is illuminated though uncertainty analysis combined with remote sensing data and drifting buoy path. By developing the largest Lyapunov exponent algorithm upon the unstructured mesh based on the nonlinear theories, the methods have been conducted to study the trajectory predictability of oil spill model. The research results not only can be used to forecast spill trajectory of real oil spill events, but also can be used for assessing the oil spill risk of waterways, ports and ocean platforms.

伴随着我国海洋石油运输业和石油开采业的迅猛发展，溢油污染的风险随之显著增强。溢油数值模型作为溢油应急计划和生态风险评估的基础，目前存在着研究尺度单一、不确定性因子干扰以及预报时间较短的问题，因此有对现有溢油数值模型的改进有重要意义。本项目针对溢油输运过程的多尺度特征进行研究，采用远场与近场动力模型相结合的方法模拟溢油的轨迹。在此基础上结合水文气象数据，采用拉丁超立方体抽样技术建立随机模拟统计模型进行溢油事故的风险预测，结合遥感数据和漂流浮标的轨迹不确定性分析统计结果的可信性。以非线性理论为支撑，通过计算非结构网格下的最大Lyapunov指数分布，研究改善溢油模型可预报性的方法。研究成果不仅可以用于真实溢油事故的轨迹预测，还可以用于航道、码头以及海洋平台的溢油风险评估。

突发性溢油事故会对海域生态环境构成严重破坏，随着海洋交通流量快速增加和海洋平台开采区域迅猛开拓，中国近海区域溢油污染风险日益加剧。溢油预警和应急技术离不开溢油行为与归宿数值模型的基础研究。本项目围绕溢油环境行为的数值模拟中不确定性和可预报性这两个基础科学问题，展开溢油输运模型研制和环境风险评价等方面的研究。.本项目构建了基于非机构网格并考虑波流耦合效应并三维近海水动力模型，实现了油膜从泄漏点近场到全海域远场跨尺度的精确追踪，定量阐释了表面波的Stokes漂移和三维辐射应力对油滴湍流扩散过程强化的物理机制。通过优化算法，对溢油模型中的关键参数进行优化率定，揭示导致预报不确定的主要因素，发现减小预报误差、提高模型可预报性的途径。提出不确定性条件下综合多元状态空间的环境风险维度的概念,构建涵盖多类型、多介质、多受体的环境风险综合评估模式。.在本项目的资助下，发表研究论文12篇，其中SCI检索论文10篇（包括国际高水平学术期刊Environmental Pollution 1篇，Science of The Total Environment 1篇，Marine Pollution Bulletin 2篇，Ocean Engineering 1篇）；联合培养博士生1名，硕士生2名；参加学术会议5次。