LANS Informal Seminar
Abstract: Coupling methods have been a limiting factor to researchers’ ability to address science questions where the relevant processes are strongly coupled. Each model inside the coupled system is often associated with different time scales, posing a great challenge on time integration. While the time integration for each model is often well founded in theory, there is little work on characterizing the influence of different coupling strategies on the stability and accuracy of the the fully coupled system.
In this talk, I will discuss our work on analyzing the stability of different coupling strategies for multi-domain PDEs that arise in general circulation models used in climate simulations. I will focus on fully coupled ocean-atmosphere models that are needed to represent and understand their complicated interactions, becoming increasingly important in climate change assessment in recent years. In particular, I will address the stability of the coupled ocean-atmosphere models for various interface conditions such as Dirichlet-Neumann condition and bulk condition, which is unique to climate modelling. By analyzing a simplified model, I will show how the parameterization of the bulk condition and other physical parameters affect the coupling stability.