A Statistical Model for Predicting Tropical Cyclone Intensity in North Indian Ocean (NIO)

A tropical cyclone (TC) is a rapidly rotating atmospheric system which has a low-pressure center, namely an eye, strong winds and a spiral arrangement of thunderstorms that produces heavy rain and causes severe destruction. Every year TC of the Northern Indian Ocean (NIO) basin affect South-eastern and South-western India significantly. Supper cyclone (1999), Mala (2006), Gonu (2007), Nargis (2008), Aila (2009), Giri (2010), Phailin (2013), Hudhud (2014), Fani (2019), Pabuk (2019), Amphan (2020), Yaas (2021) are the few such cyclones which affected the Indian coastal region at large extent. Therefore, reliable forecasts of these events are very essential. It is the intensity, maximum wind speed of a storm which causes damages to properties and lives. Therefore, along with track, intensity prediction of TCs should also be emphasized. In this study, a simple statistical model is proposed for predicting the intensity of cyclones over the Northern Indian Ocean basin with very few numbers of climatological predictors. The model parameters are estimated from the cyclone database that is developed over the NIO basin during the period 2001-2019. Minimum Sea level Pressure (MSLP), Sea Surface Temperature(SST), Initial Tropical Cyclone Intensity(ITCI) and Intensity Change During 12 hours(IC12) are selected as the predictors in the model. The model was fitted for dependent samples of 60 tropical cyclones and tested for independent samples of 15 tropical cyclones forecast up to 72 hours.The mean absolute errors (MAE) and root mean square errors (RMSE) of cyclones during 2001 to 2016 are less than 2 and 3 knots respectively for the forecast hour from 12h to 72h. When the model tested with the cyclones formed in the year 2017-2019, mean absolute error and root mean square error are found less than 7 knots and 9 knots for forecast hour from 12h to 72h. The results indicate suitability of the model for the operational use.