You are here

Model Coupling: Challenges and Connections for Climate and Mechanics

Chad Sockwell, Sandia National Laboratories

Paul Kuberry, Sandia National Laboratories

Jeffrey Connors, University of Connecticut

Pavel Bochev, Sandia National Laboratories

Kara Peterson, Sandia National Laboratories

Earth System Models (ESMs) involve a large number of constituent components operating at different temporal and spatial scales and having different mathematical properties.  The multi-physics and multi-scale behavior of the ESM requires specialized discretizations for each component. Next Generation computer platforms promise enough to power to simulate a fully coupled ESM far beyond resolutions used in the past, leading to an emerging need for novel coupling and heterogeneous numerical methods (HNM) to accurately couple the components and preserve the properties from each specialized discretization. On the other hand, HNM and coupling methods have historically been a central topic in the mechanics community with much attention to fluid-fluid coupling and fluid-solid coupling that mirror the challenges in ESMs. This expertise in the mechanics community, combined with the emerging challenges in the climate community, leads to an opportunity for crosscutting advancements. 

This session will bring together numerical analysis, computational mechanics, and climate modeling experts who are working on a broad spectrum of numerical technologies and software technologies supporting the development of HNMs and coupling methods. The aim is to make the computational mechanics community aware of the challenges within the HNM and coupling methods existing in ESMs, while making the climate modeling experts more aware of the breadth of coupling and HNMs technologies that exist in the computational mechanics community. Topics include but are not limited to:  mathematical foundations for HNMs, heterogeneous domain decomposition, optimization-based couplings, generalized Schwarz methods, partitioned time-integration, multi-physics time-integration, structure preserving domain decomposition,  novel software technologies that implement such methods, or the application of such methods in a climate-relevant or mechanics context.