Forecast Modeling and Supercomputing


The air quality forecasting is a highly specialized area. It requires huge computational power on regular basis (see Figure below). This air quality forecasting model is known as “Atmospheric chemistry transport model”. To forecast the air quality of various pollutants along with weather parameters, IITM uses four nested domain starting from near global to the local city level covering PMR and neighbouring region. The inner PMR domain will have a resolution of 1.67 km x 1.67 km which means that we can get the information of air quality at each 1.67 km grid interval within PMR. All these 4 domains will run interactively and feedback of meteorology to chemistry and vice-versa will be accounted. This model requires several key inputs for accurate forecasting. Major amongst them are - emission inventory of pollutants from various sources, weather parameters, topographical data, land use-land cover data, initial and lateral boundary conditions, etc. The initial and lateral boundary conditions for the outermost domain in meteorological model will be taken either from NCEP reanalysis or from internally generated CFS of NCMRWF, Noida whereas for the chemical forecast model, it will be taken from MACC (Monitoring Atmospheric Composition and Climate), a project of European union under MoU between IITM and EU- project partners. An intensive field campaign by involving more than 100 college students and researchers has been conducted for a few months at PMR to collect the primary and secondary activity data to generate the emission inventory of pollutants which has constituted a key input to the forecasting model.

HPC supercomputing IITM, Pune.