By Johnny Lin
Presented at: UCLA Department of Atmospheric and Oceanic Sciences, AOS 271 Seminar, 19 May 2006, Los Angeles, CA.
Historically, climate models have been developed incrementally and in compiled languages like Fortran. This has limited their modularity and robustness. The code is difficult to maintain and keep from becoming brittle, particularly for climate scientists whose focus is physics rather than programming. In this talk I describe work by myself and collaborators at the University of Chicago's Climate Systems Center to use the open-source, object-oriented, interpreted language Python to experiment with various methods of making climate models more modular. Examples include a sea-ice model, ocean model, and a basic atmospheric model, and illustrate how Python can be used to construct a light and intuitive but powerful framework for coupling components, while providing computational flexibility like automatic parallelization and platform-independence.
Updated: May 24, 2006. Author: Johnny Lin <email address>. This work is licensed under a Creative Commons Attribution-ShareAlike 2.0 License.