Treffer: THE MODEL COUPLING TOOLKIT: A NEW FORTRAN90 TOOLKIT FOR BUILDING MULTIPHYSICS PARALLEL COUPLED MODELS.

The article focuses on the computational complexity of individual models, which combined with the popularity of the distributed-memory parallel programming model used on commodity microprocessor-based clusters, results in a parallel coupling problem when building a coupled model. Until recently, computer simulation of physical, chemical, biological, and environmental systems has focused on individual subsystems that are part of a greater whole. Study of the Earth's climate system, for example, is undertaken by scientists concentrating in one of the following fields--atmospheric physics, oceanography, sea ice modeling, and land-surface modeling. Parallel coupled models incur a high computational cost from running numerous compute-intensive algorithms to integrate the equations of evolution for each subsystem of the coupled system. The software engineering of a parallel coupled model requires solutions to many challenging problems in the implementation of a large system comprising many mutually interacting and separately developed parts. In this article, the authors focus on the data interactions between the models, and consider issues regarding build system and language interoperability.