Forecasting Thunderstorms over Eastern India by Utilizing Meteorology-Informed Machine Learning Approach
Implementing Organization
NIT Rourkela
Principal Investigator
Bhishma Tyagi
NIT Rourkela
tyagib@nitrkl.ac.in
CO-Principal Investigator
Shri. SS Kundu, NESAC
NIT Rourkela
Project Overview
Eastern and North-Eastern India receives severe pre-monsoon thunderstorms, causing serious damage to life and property every year. The observational network over these regions has significantly improved over the year and researchers are attempting to give accurate forecasts using numerical models for these thunderstorms over the years. However, the utilization of soft computing techniques, including machine learning models are less explored for thunderstorm predictions. The proposed project will allow thunderstorm nowcasting with advanced machine learning models using Random Forest (RF), Gradient Boosting (GB) regression, and Tree-based Pipeline Optimization Tool (TP) Auto ML works based on meta-heuristic via genetic algorithms. The use of available meteorological parameters will be the first step in developing these models, and input of aerosols will be the next step in the development. Finally, aerosols and meteorology informed machine learning modeling system will be developed for the thunderstorm predictions over eastern and north-eastern India, and the developed modeling system will be delivered to ISRO.
Achievements
The project being undertaken by ISRO in collaboration with academia is directly or indirectly beneficial to the society.