Oceanography

Turbulence and Structure Functions #

In general, I’m interested in the connections between our analytical and numerical understanding of the ocean, focusing on ocean mixing and turbulence. My goal is to improve our understanding of how energy and other flow properties move between scales. This is important for ensuring the accuracy of large-scale ocean models, as they can’t directly simulate turbulence at smaller scales. Many of my current research questions are related to structure functions, which connect energy dissipation and related variables to the differences between flow variables (such as velocity and advection) at pairs of grid points at a constant distance. These methods have a long history in isotropic flows, but there is substantial work remaining to develop and improve them for flow fields with directional dependence.

I am currently working on taking transformation methods developed for 2-dimensional flows (as vertical motion is typically very small at large scales in the ocean) and extending them to three dimensions. I am also developing more efficient methods of calculating structure functions, including through the use of vectorized operations on CUDA GPUs. In addition, I have begun working on the scaling properties of Oceananigans across multiple GPUs, which can allow for larger, faster ocean simulations than a single GPU can provide, while using less resources than a comparable simulation on a large CPU cluster.