Weather forecasting has gradually been getting more and more sophisticated. It has also got far more important as the climate gets more unpredictable and extreme events threaten to cause massive economic damage and loss of life. So an early warning system is vital.
Ever larger computer systems making millions of calculations over many hours are now part of the daily forecasting in most developed countries. Sadly large parts of the world, many very vulnerable to dangerous climate events, do not have the money, personnel or computing power to develop the 10-day forecasting system they need.
But researchers at Cambridge University think they have found a solution by harnessing artificial intelligence. They use AI to create advanced weather forecasts which they claim outperform supercomputers and are thousands of times faster, needing only the power of a laptop.
Aardvark Weather, which has the backing of the Alan Turing Institute, Microsoft Research and the European Centre for Medium-Range Weather Forecasts, claims its system could replace current weather forecasting methods altogether. It would be able to give local forecasts, for example temperature extremes for African crops or wind speeds for European windfarms. Most importantly, it would give every developing country and thinly populated region a reliable forecast and an early warning system of potential disasters.
