Prediction of Landfall Intensity, Location, and Time of a Tropical Cyclone

The prediction of the intensity, location and time landfall a tropical cyclone well advance in with high accuracy can reduce human material loss immensely. In this article, we develop Long Short-Term memory based Recurrent Neural network model to predict intensity (in terms maximum sustained surface wind speed), (latitude longitude), hours after observation period) which originates North Indian ocean. takes as input best track data consisting its location, pressure, sea temperature, for certain (from 12 36 hours) anytime during course series then provide predictions accuracy. For example, using 24 course, provides state-of-the-art results by predicting time, latitude, longitude mean absolute error 4.24 knots, 4.5 hours, 0.24 degree, 0.37 degree respectively, resulted distance 51.7 kilometers from location. We further check efficacy on three recent devastating cyclones Bulbul, Fani, Gaja, achieved better than test dataset.