ANN and ARMA based thunderstorm prediction over Andhra Pradesh using INSAT 3D Satellite

Abstract

Extreme events such as severe thunderstorms, heat waves, cyclones, heavy rainfall events are increasing day by day in recent years over India. Out of all extreme events, thunderstorms are causing more damage and deaths when compared to others. Thunderstorms are tougher to be predicted in advance due to their faster development. In this paper, we tried to analyse two severe thunderstorm cases in premonsoon season for the time period 2017 and 2018 by using INSAT-3D satellite data. This satellite data helps us to monitor the convective cloud system every 30 minutes. Using this satellite data we are able to calculate the atmospheric indices like LI, KI, TTI and HI for every 30 minutes. After being analyzed by INSAT-3D satellite data, we tried to predict the peak stage of thunderstorms using ANN and ARMA techniques. The atmospheric based stability indices have been used as inputs for ANN & ARMA models inorder to achieve prediction. ANN prediction was better than ARMA prediction when compared to INSAT-3D satellite data

Thunderstorms Prediction Using Satellite Images

“…However, future work utilizing computer vision techniques may be able to address this issue. However, In the future, they [42,68] plan that the number of case studies will be expanded to encompass numerous hail events. Comparably, this research [43,69] raises several doubts concerning the spatial size employed in this model [70], and additional research into the model's spatial scale could lead to improved prediction models for outages caused by thunderstorms.…”