Climate Modeling on GARUDA
Climate models are numerical representations of various parts of the Earth's climate system such as the incoming and outgoing radiation, the way the air moves, the way clouds form and precipitation falls, the way the ice sheets grow or shrink.
The Indian summer monsoon is a manifestation of complex interactions between land, ocean and atmosphere and the simulation of its mean pattern and its variability on inter-annual scales is one of the challenging problems in climate studies. The correct prediction of this complex phenomenon is vital to national planning and economic policy making. Improving the prediction capability of the models require high-end computing systems. Grid computing paradigm promises doing distributed computing transparently across multiple administrative domains. We are studying the feasibility of applying Grid computing for the said purpose of Monsoon modeling and simulations.
Seasonal Forecast Model (SFM) is implemented on GARUDA to predict the Indian summer monsoon rainfall in advance of a season. Initially, the low resolution configuration of SFM model is simulated at T62 resolution (approx. 200 km x 200 km global grid) on different clusters of GARUDA. Later, a framework is developed using the existing middleware services to conduct the ensemble forecast experiments using the high resolution configuration of SFM (T320 - approx. 37 km x 37 km) on GARUDA. The ensemble forecasting problem can be seen as a set of independent tasks, where a single application is run many times with different parameters and/or input files. By conducting ensemble experiments on GARUDA’s HPC clusters, the processing time significantly improved (From 1 month using 64 CPUs of a single cluster to less than 6 days on five clusters). This experiment gave us a foundation on which we build a reliable Grid environment for climate applications.
Aerosol modeling system for climate change studies
Atmospheric pollution and its impacts in India have been addressed through regulatory air quality measurements, which show particulate matter exceedence in most cities from the 1990s. Through their effects on radiation and the structure of clouds, aerosols have the potential to influence precipitation including its spatial patterns and trends. Aerosol distributions and their effects on radiation from global climate model studies at resolutions of 50-100 km, reveal the importance of carbonaceous aerosols in the Indian region, in addition to sulphate and fly ash. Atmospheric models need attention to improve mathematical representation of processes affecting formation, atmospheric transport and removal, and radiation interactions of ultrafine particles. A validated atmospheric simulation platform based on state-of-science meteorology and chemical transport models will help understand the role of aerosols in influencing regional climate and public health.
For high resolution simulations of aerosol interaction over India, we use the University, of Iowa STEM model interfaced with the Weather Research and Forecast model. The STEM model includes multiple aerosol species like elemental carbon, organic carbon and inorganic soluble ions, and a detailed treatment of aerosol formation, microphysics and deposition processes. This model has been used extensively in regional air quality and climate studies in Asia. To capture seasonal variability in pollutant surface concentrations and column contents which influence their climate effects, the model will be run for a one year period using a modeling domain over India with a grid size of 5-30 km.