Vigyan Yuva awardee Prashant Kumar: Satellite data for weather forecast
The ISRO scientist assimilates satellite and ground data optimally and customises forecasting models to Indian conditions.
Prashant Kumar of the Indian Space Research Organisation (ISRO), winner of one of this year’s Vigyan Yuva Shanti Swarup Bhatnagar awards for Space Science and Technology, describes his work on assimilation and optimisation of weather data for more precise forecasting.

What I do
I work on assimilation of weather data, which is an optimal combination of first-guess (a forecast based on model runs) and observations by satellites and on the ground, and their associated errors. This is aimed at making the data as robust as possible.
How I do it
For optimisation of satellite data, we performed sensitivity experiments to decide which kind of satellites are needed in future. For precise forecasting, an accurate initial state of the atmosphere is needed. To determine this, we used weather models available from international agencies and customised them to Indian conditions.
We have developed a hybrid forecasting system for Bihar Mausam Sewa Kendra (BMSK), Patna. We corrected model predictions using artificial intelligence and machine learning techniques, and previous ground observations. Another example of our work has been customising a Weather Research and Forecasting model using Indian Doppler Weather Radar observations in collaboration with the India Meteorological Department and the Ministry of Earth Science.
We have also developed long-term high-resolution satellite rainfall products (a kind of measurement of rainfall systems) over India in collaboration with JAXA, Japan’s space agency. We published a comparison showing that its performance is better than that of IMERG, a rainfall product from NASA. Our product is now incorporated in the KrishiDSS portal of the Space Applications Centre (SAC), ISRO, for agriculture drought monitoring.
We conducted the first experiment to use all-sky thermal infrared observations from an Indian satellite using a unique technique. In collaboration with NASA’s Jet Propulsion Laboratory, we developed software for RAiDER, a package that contains tools to calculate tropospheric corrections for Radar.
