Estimating the Key Parameter of a Tropical Cyclone Wind Field Model over the Northwest Pacific Ocean: A Comparison between Neural Networks and Statistical Models

Estimation of maximum wind speed associated with tropical cyclones (TCs) is crucial to evaluate potential destruction. The Holland B parameter the key TC parametric field models. It plays an essential role in describing radial distribution characteristics a and has been widely used disaster risk evaluation. In this study, backpropagation neural network (BPNN) developed estimate (Bs) surface model. inputs BPNN include different combinations minimum center pressure difference (Δp), latitude, radius speed, translation intensity change rate from best-track data Joint Typhoon Warning Center (JTWC). We find that exhibits best performance when only inputting central difference. Bs estimated are compared those calculated previous statistical Results indicate proposed can describe well nonlinear relation between Δp. also found combination Holland’s model significantly improve underestimation overestimation two existing models (AH77 KZ07) for super typhoons.