In this video from PASC17, Christoph Schär (ETH Zurich, Switzerland) presents: How to Escape from the Data Avalanche of High Resoultion Climate Models.
“There are large efforts towards refining the horizontal resolution of climate models to O(1 km) with the intent to represent convective clouds explicitly rather than using semi-empirical parameterizations. This refinement would move the governing equations closer to first principles and is expected to reduce the uncertainties of climate models. However, the output volume of climate simulations would dramatically grow, and storing it for later analysis would likely become impractical, due to limited I/O bandwidth and mass-storage capacity. In this presentation we discuss possible solutions to this challenge. Key elements of the proposed strategy are the exploitation of online analysis in combination with rerunning the simulation, as well as the use of bit-reproducible codes and a data virtualization layer. The work is based on an ongoing project that uses a GPU-enabled version of the COSMO model as a regional climate model at 2.2 km resolution on a European domain.”
The main objective of Prof. Christoph Schär’s group is to improve the understanding of the climate system and its interactions with the water cycle on time-scales from one day to 100 years. Our mission is to better understand the underlying mechanisms, trends, variations and extremes; and to improve the predictive capabilities and exploitation of weather and climate models.
In the following video from PASC17, Dr. Christoph Schär describes the computational challenges of climate modeling.
Thanks to Rich Brueckner from insideHPC Media Publications for recording the video.