A tropical cyclone (TC) is one of the deadliest and most damaging natural disasters affecting people, livestock, agriculture and the economics of the coastal areas. Reductions in uncertainty are of great benefit for the disaster management authorities to plan for the evacuation and mitigation processes (Willoughby et al., 2007; Lazo et al., (2010). The major components of cyclone warnings are the forecasts of the track, winds and pressure in addition to the precise landfall point with the time of crossing the land. Predicting the track of a cyclone helps in knowing the direction in which it is moving and the area it is likely to affect. The intensity is primarily estimated from the maximum sustained wind speed, which provides a measure for the severity of a cyclone. The wind is one of the major hazards associated with a TC creating damage to houses, bridges, electrical poles, mangroves and the ecosystem. Although the damage in the coastal region is high, inland damage cannot be ruled out. Strong winds are present at the eyewall of a cyclone. The intensity of the cyclone together with the wind speed and the pressure helps in predicting the storm surge, although the spatial extent of the storm and the direction of travel are also important in predicting storm surge. Strom surge is the most devastating component of the cyclones, particularly for coasts, such as are typical for India, that have a highly varying bathymetry. Since bathymetry is one of the most critical components in estimating the storm surge, even a slight error in predicting the landfall point can lead to a different storm surge heights. The time of crossing is used to include the impacts of tides and helps in arranging the evacuation process. Thus, location, winds and the pressure of a cyclone, landfall point with the time of crossing the land are the critical components in predicting the storm surge.
Several dynamical, statistical and statistical-dynamic models have been developed to predict the cyclone parameters. Mohanty and Gupta (1997) and Gupta (2006) summarised different track prediction techniques. Bell (1979) described the operational forecasting models. Ali et al., (2007) summarised different approaches used in predicting the cyclones. They used Artificial Neural Network (ANN) technique to predict the position of the cyclone alone in terms of the latitude and longitude using the past 12 hours observations. In this paper, we attempt to use the same technique to predict winds, pressure and landfall as well, in addition to the storm location in terms of latitude and longitude.
ANN is one of the powerful data mining tools for computing input-output relationships. It is an information processing paradigm that works somewhat like hypothesized biological system in the human brain. ANN consists of an interconnected assembly of models, whose functionality is based on a neuron (Swain et al., 2006). The analysis can be used as a standalone application or as a complement to statistical analysis. This non-dynamic numerical model has been used in many oceanographic (Aliet al., 2004, Tolman et al., 2005, Jain and Ali 2006, Swain et al., 2006, Jain et al., 2007, Ali et al., 2012a, b, c) and in meteorological studies (Liu et al., 1997, Ali et al., 2007, Sharma and Ali., 2012). This technique is also found useful for satellite parameter retrievals (Krasnopolsky and Schiller., 2003, Ali et al., 2015, Sharma et al., 2013, Ali et al., 2016). ANN requires three sets of data, one for training, one for verification and the other for validation. The first dataset is used to train the model, the second dataset to test the model for any shortcomings and finally, the validation dataset is used in statistical parameter estimation. The validation dataset is an independent data that is not considered in developing the model. The popular ANN models are radial basis functions (RBF) and multilayer perceptions (MLP). In an ANN model, both the input and the output variables are normalized to vary between 0 and 1.