Enhancing Disaster Resilience through Innovative Fog Monitoring and Forecasting

Mr Prasad Deshpande
PhD Student, Department of Civil Engineering
Indian Institute of Technology Kanpur

Mr Roshan Banbariya
Project Employee, Indian Institute of Technology  Kanpur

Ms Kriti Bajaj
Data Scientist, Citi Bank, Pune

Mr Yash Patil
MS Research Student, Department of Computer Science and Engineering 
Indian Institute of Technology Kanpur

Frequent atmospheric fog in regions like the Indo-Gangetic Plains results in transportation and economic losses. Sparse visibility sensor networks hinder local fog and forecasting. This study proposes a two-part solution: i) fog detection using outdoor images as an alternative to expensive visibility sensors, and (ii) improved fog forecasting by assimilating various sources of fog information, namely, satellites, Weather Prediction models (e.g., Graphcast), in-situ visibility observations, and publicly available outdoor images. Integration of Graphcast and unifying diverse fog sources present methodological breakthroughs. Real-time forecasts are accessible at fog.iitk.ac.in. The methodology involves geotagged dataset utilization and uncertainty-aware neural networks.