Harnessing AI, machine learning to improve forecasts: IMD

NEW DELHI: Indian weather scientists have started harnessing the power of artificial intelligence and machine learning to enhance weather forecasts, India Meteorological Department Director General Mrutyunjay Mohapatra said.

Over the next few years, the emerging technologies would also complement numerical weather forecasting models which are widely used at present to predict weather, he said, adding that the weather office has been increasing observational systems to make mesoscale weather forecasts at the panchayat level or over 10 sq km area at a faster rate.

The department has deployed a network of 39 Doppler weather radars that cover 85 per cent of the country’s landmass and enable hourly forecasts for prominent cities, Mohapatra said.

“We have started using Artificial Intelligence in a limited way but within the next five years, AI will significantly enhance our models and techniques,” he said.

Mohapatra said the IMD has digitised weather records for the country dating back to 1901 and artificial intelligence could be used to sift through this plentiful information to generate knowledge about weather patterns.

Artificial intelligence models are data science models which do not go into the physics of the phenomena but utilise past data to generate knowledge that can be used to make better forecasts, the IMD director general said.

He said expert groups have been formed in the Ministry of Earth Sciences and the IMD to harness artificial intelligence.

“Both artificial intelligence and numerical forecasting models will complement each other to improve forecast accuracy. Both will work hand in hand and nobody can replace the other,” Mohapatra said.

Addressing the need for hyper-localised forecasts, Mohapatra acknowledged IMD’s challenges in delivering village-level predictions for specific hazards.

“We aim to provide forecasts at the Panchayat or village level... tailoring weather information to sector-specific needs in agriculture, health, urban planning, hydrology, and environment,” he said. The IMD chief stressed the importance of data-driven decision-making in the era of information abundance.

“Incorporating AI and machine learning allows us to harness past data to extract valuable insights and improve forecasting accuracy without solely relying on traditional physics-based models,” he said.