IISc's AI model to offer heat-stress predictions

Bengaluru: An AI-powered model that provides taluk level heat-stress predictions in Karnataka is being developed by Bengaluru's IISc, which could be a game-changer in predicting high temperature events that pose major health risks.

Researchers from the Indian Institute of Science's (IISc) AI & Robotics Technology Park (ARTPARK) are the brains behind the model that provides heat predictions for up to 10 days.

AI-based models generate weather forecasts faster and with significantly lower computational resources, unlike current forecasting systems that require large supercomputers.

ARTPARK researchers, who are testing the model, in the next couple of weeks will hand it over to the Karnataka State Natural Disaster Monitoring Centre (KSNDMC), one of the key stakeholders along with the IMD and state health and family welfare department. As of now, researchers are sharing heat maps with the KSNDMC.

In the IISc model, people can understand maximum temperature distribution, anomalies and daily temperature variations. They can also calculate thermal comfort and a heat risk index, apart from getting a 10-day forecast, which will show how the temperature feels (human perception) combining humidity levels, geographical conditions and wind speeds.

"There are two reasons a 10-day forecast approach has been adopted: to provide forecasts at a higher spatial granularity (sub-district level) compared to the current district-level forecasts, and to improve the lead time of predictions from the current three to five days to an extended 10-14 days,” said Vybhav G R, lead researcher at ARTPARK.

Although heatwaves are considered 'silent killers' by experts, there is a limited understanding of heat impacts on health.

Heat stress occurs when the human body is exposed to prolonged high temperatures and humidity, raising core body temperatures. Health issues such as fatigue, dehydration, chest pain, dizziness and muscle cramps experienced during high temperatures are still not widely understood as being potentially heat-related. 

Climate action plans have several gaps, offering generic advisories such as avoiding afternoon hours and not accounting for heat stress in migrant workers or tourists, for example, the team explained.

The collaboration will aid respective agencies working on climate forecasting, disaster management and health to take preventive action during peak summer.

Prof Ravi Nanjundiah from the Centre for Atmospheric and Oceanic Sciences (CAOS), IISc, who is offering the climate modelling support to the team, noted that the system might also integrate extreme rainfall events going forward, once the existing model is validated in a more rigorous fashion.

Retired IISc professor Dr Vijay Chandru, who is the executive advisor for healthcare to ARTPARK, said a better prediction of heat conditions will help state health authorities take preventive action such as check medication stocks, improve health resources at a hyperlocal level, in light of unprecedented conditions expected due to climate change.