Aimed at benefiting all countries in the tropical region, India has become the first country globally to develop a forecasting system — named the Bharat Forecast System (BFS) — that can predict weather with greater accuracy and detail. In common parlance, it is down to the village level. The system has improved the prediction skill by up to 67 per cent.
Indigenously developed by the Indian Institute of Tropical Meteorology (IITM), Pune, the BFS was handed over to the India Meteorological Department (IMD) on Monday by Union Earth Sciences Minister Jitendra Singh for operational use. The system had been under experimental testing since 2022, and data generated from it helped improve the accuracy of weather forecasts by 30–67 per cent, depending on the event, such as monsoon, nowcast (forecasts for the next two hours), extreme rainfall events, or cyclones.
Feature
The IMD will now use this system to provide latest weather forecasts and alerts to farmers with more precision.
On the feature of the system, M. Ravichandran, Secretary in the Ministry of Earth Sciences, said: “Earlier, we were able to give the forecast for 4-5 panchayats together. Now we can give the forecast for at least 1 or 2 panchayats, if those are very small. If there are some panchayats within a radius of 6 km, those panchayats can be covered by this model. Within 6 km radius, you can concentrate your forecast.”
He said that while many meteorological organisations around the world are working towards achieving this level of precision, India has become the first to reach the 6-kilometre resolution milestone. “This is why I call BFS a globally important but locally relevant forecasting system,” the secretary said.
‘More accurate’
Stating that earlier models were not able to predict extreme rainfall events with higher accuracy, IITM director A Suryachandra A Rao said there is a 30 per cent improvement in the accuracy now during extreme rainfall events. “In the monsoon core zone (rainfed areas of mainly Central meteorological subdivision), the prediction has improved by 67 per cent. So, in other events also, including cyclones, there is improvement in accuracy in forecast.,” he said.
Ravichandran said: “By implementing this particular TCO code, we are going to improve the forecast. We will now assimilate more and more data from radar and other systems, which will help us improve further in the coming years.”
TCO or Triangular Cubic Octahedral is a sophisticated method used to divide the Earth’s surface (or atmosphere) into a grid of points for running global numerical weather prediction models. It has to be thought of as a specific way to map the globe for a computer to perform calculations.
Ravichandran, who also worked at IITM from 1988 to 1997, said that by implementing the BFS, both the spatial pattern of forecasts and the volume of observations have improved due to the TCO code framework.
Accepting the BFS from IITM, M Mohapatra, Director General of IMD, said: “This is a milestone, as India will be recognised for achieving a 6-kilometre resolution. This will help in fulfilling the 2019 vision of making weather forecasts accessible to every household and ensuring that they are actionable.”
Rao said as the BFS is suitable for the tropical region, countries that fall between 30 degrees South and 30 degrees North Latitudes can take the benefit of it. India’s position lies in northern hemisphere, between 8.4 degrees and 37.6 degrees latitudes.
Earlier in his presentation, the IITM director said the previous supercomputer ‘Pratyush’ used to take up to 10 hours to run the Bharat forecasting model, whereas the newly-installed (in 2024) supercomputer Arka performs the same data-crunching within 4 hours.
Pointing out that the tropical region experiences many chaotic weather events, Ravichandran stressed the need to develop much higher resolution models than those used in other parts of the world.
Published on May 26, 2025
