Google is trying to take advantage of artificial intelligence and machine learning algorithms to forecast the weather faster and more accurately, and the company has developed a new tool, which it called “nowcasting,” which showed initial success in being able to accurately predict weather patterns with almost immediate results, and was able to produce forecasts For up to six hours in advance in just 5 to 10 minutes, these numbers confirm that they outperform traditional models even in the early stages.
According to the British newspaper “Daily Mail”, while some traditional forecasts generate huge amounts of data, it may take hours to complete.
Google said: “A big advantage of machine learning is that inference is mathematically inexpensive given the existence of a model already trained, providing almost instant predictions and in the original high accuracy of input data.”
She added: “If it takes 6 hours, for example, to calculate the forecast, this only allows three to four times a day and produces forecasts based on historical data that exceeds 6 hours, which limits our knowledge of what is happening now.”
But on the contrary, instant broadcasting is especially useful for instant decisions, from directing traffic and logistics to evacuation planning.
The ability to make rapid forecasts can be particularly useful with global climate change, causing weather patterns to become more disorganized and sometimes more extreme.
Although speed is among the main advantages of the Google tool, the company says that the immediate broadcast is still superior to it in traditional prediction methods when it comes to long-term forecasts, and as a result, it is unlikely that the new Google system will replace the traditional prediction models Totally, but it will be used to complement it and fill gaps for short-term coverage.