 |
| Google uses artificial intelligence to provide weather forecasts in almost real time |
On Monday, Google released a new study "Artificial Intelligence" for short-term weather forecasts, given the difficulty in forecasting weather forecasts for climate change.
The integration of machine learning technology in meteorology is still in its early stages, so it has not been built in any commercial system again, but the initial results seem promising. Google researchers describe how they can produce accurate forecasts of precipitation with an accuracy of 1 km 6 hours before precipitation in minutes. In comparison, current weather forecasting techniques take hours of computing.
Researchers estimate that rapid forecasts can be "an important tool for adapting effectively to climate change, and above all." In a world characterized by increasingly unpredictable weather conditions, researchers believe that short-term forecasts are needed to achieve "dealing with crises and reducing loss of life and property".
Google uses artificial intelligence to provide weather forecasts in almost real time
The problem with traditional prediction methods (especially physical simulations) is that their computing power is incredible. For example, weather simulations by US federal authorities have to handle up to 100 terabytes of data from weather stations and operate them for hours on expensive supercomputers. The Google method does not provide results within a few minutes because it does not attempt to design sophisticated weather systems, but it provides predictions about simple radar data as a means of precipitation.
 |
| Google uses artificial intelligence to provide weather forecasts in almost real time |
The company's researchers trained artificial intelligence models using historical radar data collected by the National Oceanic and Atmospheric Administration between 2017 and 2019 in neighboring countries in the United States. They said their expectations are better or better than the three common methods of forecasting the same data, but when they try to make forecasts 6 hours before the due date, their model works well.
Although not aware of the full importance of artificial intelligence in the field of weather forecasting, many other companies, including IBM and Monsanto, have made progress in this area. As Google researchers explain, these predictive methods are becoming increasingly important in people's lives as the impacts of climate change are becoming increasingly visible.