Title :
Satellite-derived estimation of rainfall in forward-backward thunderstorm propagation model using neural network expert system techniques
Author :
Zhang, Ming ; Scofield, Roderick A.
Author_Institution :
NOAA/NESDIS, Washington, DC, USA
Abstract :
An artificial neural network expert system technique for rainfall estimation from satellite data is presented. Estimation results of rainfall from a forward-backward thunderstorm propagation model using neural network expert system techniques are given. Using artificial neural network expert system techniques, estimation or computation of rainfall amounts is 10 times faster. The average error of rainfall estimates for the total precipitation event is reduced to less than 30%. The average error of rainfall estimates for the forward-backward thunderstorm propagation model for cases from May to August is about 1.45%. After all information has been received, the satellite-derived estimation of rainfall needs about 2 s of HDS 9000 CPU time to execute
Keywords :
expert systems; feedforward neural nets; geophysics computing; learning (artificial intelligence); rain; thunderstorms; artificial neural network expert system; forward-backward thunderstorm propagation model; neural network expert system techniques; rainfall; satellite derived estimation; total precipitation event; Artificial neural networks; Atmospheric modeling; Computer architecture; Computer networks; Expert systems; Intelligent networks; Neural networks; Pattern recognition; Satellites; Weather forecasting;
Conference_Titel :
Neural Networks, 1992. IJCNN., International Joint Conference on
Conference_Location :
Baltimore, MD
Print_ISBN :
0-7803-0559-0
DOI :
10.1109/IJCNN.1992.226951
