By Kanjyik Ghosh

MUMBAI, Dec 22 (Reuters) – India is testing artificial intelligence (AI) to build climate models to improve weather forecasting as torrential rains, floods and droughts proliferate across the vast country, a top weather official said.

Global warming has triggered more intense clashes of weather systems in India in recent years, increasing extreme weather events, which the independent Centre for Science and Environment estimates have killed nearly 3,000 people this year.

Weather agencies around the world are focussing on AI, which can bring down cost and improve speed, and which Britain’s Met Office says could “revolutionise” weather forecasting, with a recent Google-funded model found to have outperformed conventional methods.

Accurate weather forecasting is particularly crucial in India, a country of 1.4 billion people, many impoverished, and the world’s second-largest producer of rice, wheat and sugar.

The India Meteorological Department (IMD) provides forecasts based on mathematical models using supercomputers. Using AI with an expanded observation network could help generate higher-quality forecast data at lower cost.

The department expects the AI-based climate models and advisories it is developing to help improve forecasts, K.S. Hosalikar, head of climate research and services at IMD, told Reuters.

The weather office has used AI to generate public alerts regarding heatwaves and such diseases as malaria, Hosalikar said. It plans to increase weather observatories, providing data down to village level, potentially offering higher-resolution data for forecasts, he said.

The government said on Thursday it wants to generate weather and climate forecasts by incorporating AI into traditional models, and has set up a centre to test the idea through workshops and conferences.

“An AI model doesn’t require the high cost involved in running a supercomputer – you can even run it out of a good quality desktop,” said Saurabh Rathore, an assistant professor at Indian Institute of Technology-Delhi.

Better data is also needed to make the most out of AI, experts say.

“Without having high-resolution data in space and time, no AI model for location-specific magnification of existing model forecasts is feasible,” said Parthasarathi Mukhopadhyay, a climate scientist at the Indian Institute of Tropical Meteorology.

Reporting by Kanjyik Ghosh; Editing by Shivam Patel and William Mallard

Our Standards: The Thomson Reuters Trust Principles.