Title :
Hurricane tracking via backpropagation neural network
Author :
Johnson, Gregory P. ; Lin, Frank C.
Author_Institution :
U.S. Dept. of Commerce, NOAA, Wallops, VA, USA
Abstract :
A backpropagation neural network is trained using meteorological data for the North Atlantic Ocean Basin. The resulting network is utilized to predict future tracks of hurricanes six hours in advance. Comparison with historical tracking data supplied by the National Oceanic and Atmospheric Administration shows surprisingly good agreement between prediction and actual tracks. The methodology appears quite promising for this vital task
Keywords :
backpropagation; neural nets; storms; tracking; weather forecasting; wind; ARIMA; Hurricane Allen; Hurricane Emily; Hurricane Gloria; Hurricane Jenanne; North Atlantic Ocean Basin; backpropagation neural network; hurricane tracking; meteorological data; time series; Atmospheric modeling; Backpropagation; Hurricanes; Large-scale systems; Meteorology; Neural networks; Predictive models; Storms; Tropical cyclones; Weather forecasting;
Conference_Titel :
Neural Networks, 1995. Proceedings., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2768-3
DOI :
10.1109/ICNN.1995.487576
