Title :
A hybrid approach to improving rainfall forecasts
Author_Institution :
Massachusetts Inst. of Technol., MA
Abstract :
High-resolution rainfall forecasting has important benefits, such as enabling flood prediction, yet little progress has been made toward developing an effective strategy. The hybrid approach presented here combines weather physics, statistics, and artificial neural networks. The strategy is able to draw on all available information, account for and use aspects of the domain physics that are better understood, and exploit the strengths of the available data-dictated tools
Keywords :
geophysics computing; neural nets; rain; statistical analysis; time series; weather forecasting; flood prediction; neural networks; rainfall forecasting; statistical models; statistics; time series; weather physics; Extrapolation; Floods; Mathematical model; Neural networks; Numerical models; Physics; Predictive models; Spatial resolution; Statistics; Weather forecasting;
Journal_Title :
Computing in Science & Engineering
DOI :
10.1109/MCISE.2002.1014976
