Swarm intelligence and neural nets in forecasting the maximum sustained wind speed along the track of tropical cyclones over Bay of Bengal

Tropical cyclones are well-known extreme weather and the cause of considerable damages, injuries and loss of life. The assessment of the maximum sustained wind speed along the track of the tropical cyclones is very important for estimating the strength of the cyclones. The swarm intelligence in the form of ant colony optimization (ACO) technique is introduced in this study to compute the pheromone deposition along the track of tropical cyclones followed by neural nets to forecast the maximum sustained wind speed of the cyclones occurring over the Bay of Bengal of North Indian Ocean. The ACO is a nonlinear problem-based meta-heuristic optimization method for finding approximate solutions to discrete optimization problems and simulates the decision-making processes of ant colony similar to other adaptive learning techniques. The method has shown its application potential in various fields including the prediction of monsoon rainfall. In this study, the amount of pheromone deposition during the successive stages of the cyclones has been estimated. A range of minimum central pressure (MCP), central pressure drop (PD), maximum sustained wind speed (MSWS) and intensity (T-No) associated with the cyclones of Bay of Bengal are utilized to form the input matrix of the neural nets. The neural nets are trained to forecast the maximum sustained wind speed along the track of the tropical cyclones over Bay of Bengal. The result reveals that the errors in forecasting the MSWS along the track of tropical cyclones with 6, 12, 18 and 24 h lead time are 2.6, 2.9, 3.1 and 4.8, respectively. The result is compared with the existing dynamical, statistical and adaptive models to evaluate the skill of the present model. The result is well validated with observation.