Application of Machine Learning for Atmospheric Characterization Using University of Calcutta ST Radar

Abstract

An active phased-array Stratosphere-Troposphere (ST) VHF Doppler Radar operating at 53 MHz is being installed at Ionosphere Field Station (22.94°N, 88.51°E, and 34°N geomagnetic latitude) of University of Calcutta in the eastern part of India adjoining northern Bay of Bengal. This radar is unique, being the only one operational at this frequency in the entire eastern and north-eastern part of the country and also in the south-east Asian longitude sector. It is designed to measure atmospheric winds, which could be very severe particularly during tropical cyclones which often affect this region. In this paper, efforts have been made to develop machine learning solutions that are customized to the needs of weather and climate modelling. With this and other radars in India, models of various atmospheric parameters may be developed using measured data over a period of time as training dataset. Applying this model, it can predict the values of the atmospheric parameters for a period outside the training interval. So, the model outputs may help override huge infrastructure and resources necessary for establishing radars.