The goal of this project is to develop dynamically adaptive grid techniques for future climate model predictions. Adaptive grid refinement techniques provide an attractive framework for atmospheric flows since they allow improved resolution in a limited region without requiring a fine grid resolution throughout the entire model domain. Therefore, the model domain to be resolved with higher resolution is kept at a minimum, greatly reducing computer memory and speed requirements.

The adaptive grid technique will be implemented into NASA's next generation climate model, which has recently been developed at the NASA Goddard Space Flight Center (GSFC). The model is parallelized using the so-called hybrid approach that combines shared memory and distributed memory paradigms. Parallel and high performance computing play an essential role in the project. The adaptive grid method will optimally exploit these computing resources, which involves not only a highly efficient computational design of the numerics and algorithms but also aspects concerning the scalability of the problem, the communication patterns and load balancing properties.

The available hardware comprises a heterogeneous distributed environment that includes -- besides multiprocessor workstations -- an IBM SP2 and a 16-processor SGI Origin. Therefore, the ACTS Toolkit will help utilize these resources and, in addition, addresses performance and visualization issues.

The research project is characterized by an interdisciplinary approach involving atmospheric science, computer science and mathematical/numerical aspects. The work is done in close collaboration between the University of Michigan, NASA GSFC and the National Center for Atmospheric Research (NCAR).