Meteorological Variables Nowcasting using Machine Learning and Deep Learning Techniques

Abstract

In this work, we address two nowcasting problem statements. The first is the PM2.5 concentration temporal nowcasting over Delhi NCR. It has been observed that the high levels of PM2.5 concentrations in Delhi, particularly for the winter season, are affecting many people’s health and causing various respiratory diseases. In this work, we try different Machine Learning and Deep Learning techniques for PM2.5 nowcasting with a lead time of up to six hours. The second is Precipitation nowcasting using Bhopal Radar Data. It is essential because it provides critical information to protect people, property, and infrastructure from the impacts of extreme weather events. In this work, we develop a deep learning-based ConvLSTM model to perform spatiotemporal nowcasting.