Enhanced Climate Simulations and Earth System Modeling using High-End Computing Systems
College: Liberal Arts and Sciences
Award year: 2008-2009
Collaborators: John Towns, Division Director of Persistent Infrastructure Jay Alameda, Senior Technical Program Manager Mark Straka, Senior Research Programmer Peter Bajcsy, Research Scientist NCSA
Improvement of climate models and climate forecast is of fundamental importance to the U.S. Climate Change Research Program. While continued research on development of more detailed biogeochemical and climate models are essential, the major challenges and uncertainties are in treatment of biogeochemical cycle and climate feedbacks in an Earth System Model. The overarching objective of the proposed research is to reduce uncertainties in our understanding of how Earth's climate, biogeochemical systems, and human activities interact over the 21st century and beyond through simulations using a more accurate climate system model that includes the full range of human and natural climate feedbacks with increased realism and spatial resolution. To achieve this objective, I will pursue the following specific tasks (1) extend the capabilities of the NCAR's Community Climate System Model (CCSM) to include representations of biogeochemical processes related to carbon cycle and human activities interactions with climate system; (2) identify and utilization of data for evaluating the performance of the biogeochemical cycles model in coupled carbon-climate model, and (3) Enhance the computational performance and scalability of coupled biogeochemical-climate simulations on available and future computing architectures for use in national and international assessments of climate change.
The vastly increased capabilities and improvements of models such as what we are proposing will push existing compute resources past the bounds of what is possible today. A vital part of our research under this proposal is discovering new and more flexible algorithms for the component models to increase the simulation throughput of the CCSM and the future biogeochemical cycles-CCSM coupling. We will be working towards a highly scalable and optimized simulation code that will enable us to push the NCSA's current hardware farther than we previously have been able to, and to then easily adapt our models to take full advantage of petascale computing once it becomes available to allow higher resolution simulations with more comprehensive treatment of processes and feedbacks. In order to achieve our computer science related challenges, I plan to interact with the PI and CET Divisions at NCSA, which has expertise in integrating codes, extending the codes, parallelizing the codes and executing them on high-performance computing systems. In collaboration with NCSA, we will bring to the community a new well-tested version of CCSM that will run efficiently on tomorrow's large-scale compute infrastructure, including NCSA's flagship Blue Waters Petascale Computing System.