GENERAL circulation models (GCMs) are used by meteorological agencies around the world for computer simulation of the Earth’s atmosphere and oceans. These models, however, are not very good at forecasting rainfall more than a few days in advance. The next really big breakthrough could come from using machine learning to mine historical climate data, build models based on clever algorithms, and use these to forecast droughts and floods months in advance.
While machine learning is now a well-established discipline, and artificial neural networks (ANNs) a well understood subcomponent, this technology is only beginning to be applied to rainfall forecasting. Most of the research effort has been concentrated in China, India – and in Australia by John Abbot and myself.
We review the research and summarize our work in chapter 3 of a new book ‘Engineering and Mathematical Topics in Rainfall’ just published by IntechOpen – the first publisher of open access scientific books.
Entitled, ‘Forecasting of Medium-term Rainfall Using Artificial Neural Networks: Case Studies from Eastern Australia’, chapter 3 explains how and why rainfall forecasting is amenable to machine learning. We also explain how we developed and improved our forecasting technique from 2012 through to 2016.
Our method clearly specifies how much rain will fall at specific locations on a month by month basis. Contrast this with seasonal forecasts from General Circulation Models that are usually presented as vague probabilities.
Of course, quantifying the actual amount of rainfall forecast is fundamental to assessing the skill of a technique. One of the potential problems with current forecasting by meteorological agencies around the world is that there is no quantification, or clear goal for improvement. Then again, for the most part these agencies are monopolies.
To read about our monthly rainfall forecasting technique, download the chapter here:
Keywords: rainfall, forecasting, artificial neural network, Australia
Short link: https://wp.me/p9TIoh-5z