Forecasting Pressure Drop and Maximum Sustained Wind Speed Associated with Cyclonic Systems Over Bay of Bengal with Neuro – Computing

Abstract The present research intends to develop a neuro-computing based adaptive intelligent model to predict the pressure drop (PD) at centre and maximum sustained wind speed (MSWS) linked to vortical convective system at the stage of the utmost strength over the Bay of Bengal (BOB) of the North Indian Ocean (NIO). The vortical convective systems considered in this study incorporate the stages from deep depression to extreme severe cyclones. The low level vorticity (LLV), mid-tropospheric relative humidity (MRH), upward wind speed at 850, 500 and 200 hPa pressure levels are obtained as the most suitable input parameters through factor analysis. The adaptive neural network models are trained with the data from 1990 to 2015 to forecast the PD and MSWS over BOB. The result reveals that the multilayer perceptron (MLP) model provides good accuracy at 6 and 30 h lead time in forecasting the PD. But minimum error is obtained at 6 h time before in anticipating the PD at the highest intensity stage of vortical convective system. The result further shows that the MLP model is the most competent for projecting MSWS at the peak intensity stage of vortical convective systems with minimum forecast error at 60 h lead time. The model outputs are compared to the existing conventional models and subsequently the outcomes are supported with observations from 2016 to 2019.