Collaborative Research: Submesoscale-Resolving Large Eddy Simulations Using Reduced Biogeochemical Models

点击翻译按钮获取中文摘要

Improved understanding of upper ocean biogeochemistry requires a comprehensive look at the interactions between chemical tracers and turbulence at scales smaller than one kilometer (also termed "submesoscale"), but modeling chemical tracers and processes at these small scales requires a tremendous amount of computing power and time. As a result, new reduced biogeochemical models, solved in novel ways, must be developed to perform large eddy simulations (LES) of coupled biogeochemistry and physical processes at these scales. In the proposed project, techniques adapted from the field of combustion for the reduction of large chemical kinetics mechanisms will, for the first time, be used to reduce the size of large ocean biogeochemical models. This interdisciplinary research effort will be undertaken by a collaborative team consisting of an expert in numerical simulations of both reacting and oceanographic flows, a biological oceanographer with extensive knowledge of ocean biogeochemistry, and an expert in chemical model reduction and solution. The tools developed in this project will be made available to the broader oceanographic community, and involvement of interdisciplinary PhD students will expose a new generation to computational methods in oceanography and the Earth sciences. The project will provide research experience and mentorship to undergraduate students, particularly those under-represented in STEM, by designing projects derived from, and complementary to, the proposed research. Ultimately, this project will benefit society through improvements to Earth system models (ESMs) used to study climate, resulting in more accurate predictions of future climate impacts on human health, safety, and property. Solution of the reduced models will be performed on graphical processing units (GPUs) using a high-order Runge-Kutta-Chebyshev (RKC) time integration scheme. This project will culminate in the simulation of realistic ocean scenarios and comparisons will be made with observational data from the Drake Passage to determine the role of submesoscale processes in generating small-scale patchiness in the partial pressure of carbon dioxide. Ultimately, insights obtained from the LES will be used to develop a better understanding of the interactions between small-scale turbulence and biogeochemistry in the upper ocean, including the characteristics, dynamical origins, and effects of tracer patchiness. Integration of complex biogeochemical models within high-fidelity LES has previously been exceptionally difficult, but the proposed model reduction, as well as the use of GPUs and the high-order RKC integrator, will enable high-resolution studies of fully-coupled turbulent and biogeochemical processes at submesoscales. These improvements will be made possible by leveraging techniques from chemical kinetics modeling for combustion, where reduction and integration of large chemical mechanisms in high-fidelity simulations has been common for nearly a decade. The proposed simulation effort will provide insights into the effects of submesoscale turbulence, including wave-driven Langmuir turbulence, on the upper-ocean carbon cycle, and will inform the future development of improved ESMs. In particular, interactions between submesoscale turbulence and biogeochemical tracers are thought to be the cause of tracer patchiness and require substantial further study to develop more accurate subgrid-scale parameterizations for ESMs. Moreover, simulations of the Drake Passage will provide concrete insights and understanding of tracer patchiness for realistic conditions.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.