Compframe 2005

Atlanta, GA, June 22-23, 2005

Coupling Weather and Climate Models with the Earth System Modeling Framework.
Shujia Zhou, V. Balaji, Carlos Cruz, Arlindo da Silva, Chris Hill, Erik Kluzek, Shep Smithline, Atanas Trayanov, Weiyu Yang.
Abstract
A physical Earth system model typically consists of several model components, which are coupled throughexchanging data. For example, the well-publicized El Nino-Southern Oscillation (ENSO) phenomenon isthe outcome of atmosphere-ocean interactions, and is usually modeled by coupling atmosphere and ocean components. Most models solve partial differential equations on a large gridpoint set for a long time periodand consequently require significant amounts of high-performance computing resources. Since Earthsystem model components, such as global atmospheric circulation models, also contain many physical processes (radiation, cloud formation, precipitation, etc.), considerable time and manpower are needed todevelop a production-quality model. As more physical processes are added to the models, the models arebecoming increasingly complex and difficult to modify and maintain. It is even more challenging to compare models from different organizations and construct new models from model components developedin different organizations. A software framework to facilitate model development and coupling would be agreat benefit to the weather and climate communities.
To achieve that goal, NASAs Earth-Sun System Technology Office/Computational Technologies Projecthas funded the development of the Earth System Modeling Framework (ESMF). The ESMF project enablesclose collaboration from major U.S. Earth system modeling organizations: the NSF/National Center for Atmospheric Research (NCAR), the NASA/Goddard Global Modeling and Assimilation Office (GMAO),the Massachusetts Institute of Technology (MIT), the University of Michigan, the DOE/Argonne NationalLaboratory (ANL), the DOE/Los Alamos National Laboratory (LANL), the NOAA/Geophysical Fluid Dynamics Laboratory (GFDL), and the NOAA/National Centers for Environmental Prediction (NCEP).Recently the U.S. Navy, U.S. Air Force, and U.S. Army joined the ESMF collaboration through thecreation of the Battlespace Environments Institute, a virtual center that will couple codes from multiple services to create integrated forecasts.
The ESMF software consists of a superstructure for coupling and exchanging data between component models (e.g., atmosphere, ocean) and model subcomponents (e.g., physics, dynamics); and an infrastructureconsisting of (1) data structures for representing grids and fields and (2) an optimized, portable set of low-level utilities. The data constructs and low-level utilities are used by the coupling superstructure and may also be used separately to compose scientific applications. Conceptually, an application running underESMF may be thought of as a sandwich, with the upper coupling layer and lower utility layer provided byESMF and the middle layer provided by the application developer. The ESMF superstructure sits above the components of an application, controlling inter-component data transfer and sequencing. The ESMFinfrastructure lies below the components, offering integrated tools for intra-component communication,error handling, time management, profiling, and other standard modeling functions. More information on ESMF can be found at [1, 2].
To examine and test ESMF capabilities, we have specifically constructed several experiments involving model components from different organizations. In particular, we have coupled the GFDL FlexibleModeling System (FMS) B-grid atmosphere [3] with the MITgcm ocean [4] and the NASA-NCAR finite-volume Community Atmosphere Model (fvCAM) [5] with the NCEP Spectral Statistical Interpolation (SSI) analysis [6]. These experiments illustrate the ESMF component strategy and coupling services.
Access full paper (pdf file)