Bengaluru: Researchers at the Indian Institute of Science (IISc), Bengaluru, are working through the AI & Robotics Technology Park (ARTPARK) to develop an AI-powered weather model, which aims to provide heat stress predictions at the sub-district (taluk) level in Karnataka. IISc plans to extend the coverage to Tamil Nadu and eventually across India. With extreme heat events becoming more frequent due to climate change, this initiative aims to deliver up to 10-day localised heat-risk forecasts, equipping authorities and the public with a vital tool to mitigate health risks. IISc and ARTPARK researchers are generating these predictive signals, which now require thorough validation.
The idea for this initiative emerged after India experienced extreme heatwave events last year. There was a clear need to inform local authorities in advance, allowing them to better prepare and implement strategies to manage such extreme conditions. Traditionally, heatwave assessments have relied primarily on temperature. However, a more comprehensive approach—incorporating factors such as humidity, wind speed, temperature anomalies, and nighttime temperatures—is essential for developing accurate thermal discomfort indices. This index will be further refined by leveraging public health data through collaborative efforts.
As a pilot project, it was first decided to test the utilities and limitations of this approach in Karnataka state. Additionally, there is active collaboration with the Department of Health and Family Welfare to assess the added value of these forecasts for health officials.
A more comprehensive model
Vijay Chandru, a visiting professor at NCBS and executive advisor for healthcare at ARTPARK, explained that this AI-driven approach significantly enhances existing weather prediction models.
“Our goal is to improve forecast lead times and develop more localized predictions, starting at the taluk level and eventually reaching the Gram Panchayat level,” he stated in an interview with ETV Bharat.
The Indian Meteorological Department (IMD), the Karnataka State Natural Disaster Monitoring Centre (KSNDMC), and the Karnataka Department of Health and Family Welfare are key collaborators in this initiative. Researchers are providing heat maps to KSNDMC for real-time analysis and feedback. The team is utilising AI models developed globally by various researchers and adapting them to the Indian context, particularly for extreme heat conditions. Chandru noted that over the next few months, state authorities will provide field observations to evaluate the model’s performance. Based on their feedback, the model will be further refined.
Leveraging AI to enhance forecast accuracy
Professor Ravi Nanjundiah, former Director of the Indian Institute of Tropical Meteorology and a Professor at IISc’s Centre for Atmospheric and Oceanic Sciences (CAOS), while discussing heat conditions, said that heat stress is not just about temperature—it also depends on factors like humidity. He explained that in places like Mangaluru, where temperatures may not reach extreme highs but humidity remains elevated, the combination can still lead to severe heat stress.
Professor Nanjundiah also highlighted the transformative role of AI in weather forecasting. Traditional systems rely on large supercomputers to generate forecasts, whereas this AI model can produce high-quality predictions with significantly lower computational resources. "In principle, anyone with a GPU-equipped computer can create a forecast. We are leveraging AI techniques to enhance forecast accuracy while reducing computational costs,” he said.
The 10-day forecasts approach
“Forecasts typically become less reliable beyond five days, making it a significant challenge to develop a model that can provide accurate heat wave predictions for up to 10 days. However, we are hopeful that AI models can overcome this limitation without compromising accuracy. A clearer picture will emerge once we analyse the ground truth data at the end of the season," said Professor Nanjundiah.
Vybhav GR, lead researcher at ARTPARK, further explained that the decision to focus on a 10-day forecast at the taluk level was driven by key requirements from government authorities, who need sufficient time to prepare and take action at the local level.
Next step and future expansion
Once the AI model has been thoroughly assessed and refined, the IISc-ARTPARK team will propose its integration into experimental use by IMD, KSNDMC, and state authorities. If successful, the system will move toward operational deployment, making it accessible for public use. This tool is expected to benefit weather service agencies, disaster management teams, and public health professionals working on climate, weather, and health-related initiatives.
Vijay Chandru mentioned that the team plans to expand this tool to other states, with Tamil Nadu likely to be next. Discussions with the Tamil Nadu Public Health Department are already underway to implement the system across more regions in the near future.
This initiative—backed by key meteorological and health organisations, including IMD, the Indian Institute of Tropical Meteorology (IITM) Pune, KSNDMC, and the Karnataka Department of Health and Family Welfare—is set to make a significant impact on public health and climate preparedness in India.
