Probing the Ocean's Multiscale Pathways

洋流是一个典型的多尺度系统，涉及整个空间长度和时间周期的过程和结构。从数千公里的全盆地环流到毫米级的湍流涡旋的尺度之间的非线性耦合是理解和模拟海洋环流和混合的主要困难，也限制了我们预测气候的能力。该项目的主要目标是利用卫星和高分辨率模型的数据，探索从10000公里到10公里左右的全球海洋中不同尺度的能量循环耦合。该项目使用了一种有点新奇的“粗粒化”方法来分析多尺度的相互作用，这种方法比经典的“均值-涡流”分解更通用、更强大。对于数值模型，几乎永远不可能直接将所有运动分解到最小的尺度。取而代之的是，小尺度对感兴趣的大尺度环流的影响是使用参数来估计的，参数的选择通常取决于分辨率尺度的截止点在哪里，有时取决于可能确定哪些物理过程重要的特定位置。这项研究与“比例感知”和“位置感知”参数的搜索是一致的，这两个参数将普遍适用，不需要针对特定条件进行调整。该项目可以对气候科学中的一个基本问题产生直接影响：在多大程度上，时间变化在流动系统本身内自然出现，或者是对外部强迫的反应。这项工作还有望为当前方案的参数调整提供先验约束，为可能应用于允许涡流的海洋模式的拟议方案提供先验约束，并将有助于发展一类新的海洋参数化，这些参数化是时间、位置和分辨率的函数。这项工作还将展示一种自洽的综合方法来分析和模拟多尺度系统的动力学，这可能对气候以外的许多领域产生重要影响。为这项研究开发的多尺度分析代码将在Github上提供，以允许采用开放的开发方法。该项目将支持两名初级科学家，一名在她开始博士学位时，另一名在他职业生涯的初级阶段。最后，该项目的研究将通过罗切斯特博物馆和科学中心整合到推广工作中。粗粒化具有严格的数学基础，与成熟的物理技术密切相关，包括宏观电磁学、重整化群和大涡模拟。此外，与经典分解不同，粗粒化与粗分辨率气候模拟的参数化要求是一致的。可以相对容易地推导出控制球体上不同尺度的动力学的方程，为研究海洋流动中的多尺度路径开辟了一种新的、可能具有变革性的方式，包括能量转移、动量和示踪物的传输以及不同尺度上的强迫--所有这些都可以在地理和时间上进行探测。该项目将分析不同路径和过程之间的地理和时间相关性，包括它们对大气强迫变化的幅度和频率响应。气候模式中需要参数化的物理过程取决于网格分辨率，也取决于对任何给定长度尺度上的主导过程的理解和量化。在这方面，该项目理想地结合了一种新的系统方法来推进反映潜在亚网格物理的“尺度感知”和“位置感知”参数化，以及对不同长度尺度和过程的幅度和频率响应的更深入了解。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Oceanic flow is a quintessentially multiscale system, involving processes and structures over an entire continuum of spatial lengths and time periods. The nonlinear coupling between scales ranging from the many thousands of kilometers of basin-wide circulation down to the millimeter size of turbulent eddies presents a major difficulty in understanding and modeling oceanic circulation and mixing, and also in limiting our predictive ability to forecast climate. The main objective of this project is to probe the energy cycle coupling different scales from the order of 10000 km down to around 10 km in the global ocean, using data from satellites and high-resolution models. The project utilizes a somewhat novel ‘coarse-graining’ approach to analyze multiscale interactions that is more versatile and powerful than the classical ‘mean-eddy’ decomposition. It numerical models, it is almost never possible to directly resolve all motions down to the smallest scales. Instead, the influence of the smaller scales on the larger scale circulation of interest is estimated using parameterizations, whose choice typically depend on where the cut off for resolved scales are and sometimes on the particular location which may determine what physical processes are important. This research is aligned with the search for ‘scale-aware’ and ‘location-aware’ parameterizations, which would apply universally without needing to be tuned for specific conditions. The project can have a direct bearing on a fundamental problem in climate science: the extent to which temporal variability is naturally emergent within the flow system itself or is a response to external forcing. The work also promises to offer a priori constraints on parameter tuning of current schemes, on proposed schemes that may be applied to eddy permitting ocean models, and will help in the development of a new class of ocean parameterizations that are a function of time, location, and resolution. This work will also demonstrate a self-consistent integrated methodology to analyze and model the dynamics of multiscale systems, which can have an important impact on many fields beyond climate. The multiscale analysis codes developed for this study will be made available on Github to allow for an open development approach. The project will support two junior scientists, one at the beginning of her PhD and another at the threshold of his career. Finally, the project’s research will be integrated into outreach efforts through the Rochester Museum and Science Center.Coarse-graining has a rigorous mathematical foundation and is closely related to well-established physics techniques, including macroscopic electromagnetism, renormalization group, and large eddy simulation. Moreover, unlike the classical decomposition, coarse-graining is consistent with the parameterization requirements of coarse-resolution climate simulations. Equations governing the dynamics of different scales on the sphere can be derived relatively easily, opening up a new and potentially transformative way to studying multiscale pathways in oceanic flows, including the transfer of energy, transport of momentum and tracers, and forcing at different scales — all of which can be probed both geographically and temporally. This project will analyze the geographic and temporal correlations between different pathways and processes, including their amplitude and frequency response to changes in atmospheric forcing. The physical processes that require parametrization in climate models depend on the grid resolution and also on an understanding and quantification of the dominant processes at any given length-scale. In this respect, the project is ideally aligned to advancing a new systematic approach to ‘scale-aware’ and ‘location-aware’ parameterizations that reflect the latent subgrid physics, along with a deeper understanding of the amplitude and frequency responses of different length-scales and processes.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.