Back Propogation(BP)-neural network for tropical cyclone track forecast

Tropical Cyclone (TC) track prediction is still a big and unsolved problem from the perspectives of theory and application due to the complicated and non-linear physical mechanisms and lack of calculating capabilities and observations. Neural Network works effectively and efficiently in simulating non-linear relationships. Therefore, the present study employs the BP-neural network to predict TC tracks. After the model is trained by historical TC track data (e.g., latitude and longitude), it perform relatively well in tropical cyclone prediction according to the verification.