Title: HPC and Climate System Modelling : A User's Perspective

 

 Ravi S Nanjundiah
              
  Centre for Atmospheric & Oceanic Sciences,
              
  Indian Institute of Science, Bangalore 560 012
              
  email: ravi@caos.iisc.ernet.in



Abstract:



Climate change is a major topic of concern today. Climate change studies require
generation of various scenarios using numerical models of the climate system. 
The climate system model consists of four major sub-components viz. atmosphere,
ocean, land (with biosphere) and sea-ice. To simulate the behaviour of the
climate system, these sub-models need to continously interact with each other.
Since the impact of climate changes are relatively slow (of the order of decades
or more) and since the response time of some of the components (e.g.  oceans)
 is slow, these models have to be integrated over large intervals of time. 
These simulations require extremely large computational resources.

 From a computational point of view, these models need to have good
scalability. I will be presenting some of our experiences in conducting
simulations with a climate system model. These simulations were conducted
to study the variability of monsoons as coupled and evolving system. We
conducted a 100 year simulation on the Param Padma System of CDAC. The 
computations consumed about a 27 days of computational time on 104 processors.
I will also present about our current efforts in implementing a similar 
climate system on a computational grid.