A global Lagrangian eddy dataset based on satellite altimetry (GLED v1.0)

Mesoscale eddies, defined as rotating structures ranging typically from tens to hundreds of kilometers and lasting for several weeks to months, are ubiquitous in the global ocean. Isolated mesoscale eddies are generally considered as coherent structures with a material barrier that can trap the fluid within the eddy interior. Methods employed to identify coherent eddies can be classified into Eulerian and Lagrangian frameworks. Eddy datasets based on Eulerian methods, especially the eddy census of Chelton et al. (2011), have been used in a huge range of applications, from physics to biology. However, recent works have shown that Eulerian eddies are not necessarily coherent because there is strong and persistent water exchange across the Eulerian eddy boundary. In this study, millions of Lagrangian particles are advected by satellite-derived surface geostrophic velocities over a period of 1993-2019. Using the method of Lagrangian-averaged vorticity deviation by Haller et al. (2016), we present a global Lagrangian eddies dataset (GLED v1.0). This open-source dataset contains not only general features (eddy center position, equivalent radius, rotation property, etc.) of eddies with lifespans of 30, 90, and 180 days, but also the trajectory of particles trapped by coherent eddy boundaries over the lifetime. The greatest strength of GLED v1.0 is that the identified eddies are all material objects by construction. Our eddy dataset provides an additional option for oceanographers in studying the interaction between coherent eddies and other physical or biochemical processes in the Earth system.中尺度涡旋（Mesoscale eddies）通常指尺度范围为数十至数百公里、存续数周至数月的旋转环流结构，广泛分布于全球海洋中。孤立的中尺度涡旋通常被视为具有物质屏障的相干结构，可将流体困于涡旋内部。用于识别相干涡旋的方法可分为欧拉（Eulerian）框架与拉格朗日（Lagrangian）框架两类。基于欧拉方法的涡旋数据集，尤其是Chelton等人2011年发布的涡旋普查数据集，已被广泛应用于物理学到生物学等诸多研究领域。然而，近期研究表明，欧拉方法识别的涡旋未必属于相干结构，因为欧拉涡旋边界处存在强烈且持续的水体交换过程。本研究利用1993年至2019年间的卫星反演海面地转流场，驱动数百万个拉格朗日粒子进行平流输运。基于Haller等人2016年提出的拉格朗日平均涡度偏差（Lagrangian-averaged vorticity deviation）方法，本研究构建了全球拉格朗日涡旋数据集（GLED v1.0）。该开源数据集不仅包含存续时长分别为30天、90天和180天的涡旋的通用特征（如涡旋中心位置、等效半径、旋转特性等），还涵盖了相干涡旋边界在其存续期内困住的粒子的运动轨迹。GLED v1.0的最大优势在于，其识别出的涡旋从构建逻辑上均为物质实体。本涡旋数据集为海洋学家研究相干涡旋与地球系统中其他物理、生化过程之间的相互作用提供了全新的选择。