Short Term Rain Prediction with Artificial Neural Networks

Virtually all research concerning short-term rain prediction to date makes use of spatial rainfall data, sometimes also taking another variable, such as wind direction, into account. This thesis explores the possibility of using neural networks for short-term rain predictions from several meteorological variables from only one gauging station. The input variables examined, besides rainfall, are wind speed and direction, temperature, humidity, and barometric pressure. The method cannot compete with spatial methods, and the results are indeed impractical, but they show some correlation which could be used to improve the spatial methods.