Researchers say that the Aardvark Weather system, using artificial intelligence, needs thousands of times less processing than current systems, and is much faster. This new technology can provide precise climate predictions only using a desktop computer and by a single researcher, while traditional systems require powerful superconductors and large teams of professionals.
A transformation in the projections process
Currently, forecasting atmospheric status requires a complex steps that each take several hours and require the use of special superconductors and specialized teams for development, maintenance and implementation. But Aardvark Weather can completely transform this process. The system teaches its prediction models using raw data collected from meteorological stations, satellites, atmospheric balloons, ships and aircraft around the world, replacing traditional methods.
Improves speed, accuracy and reduction of costs
According to a study published on Thursday in the journal Nature, this new technology can significantly improve the speed, accuracy and cost of atmospheric forecasts. The study was conducted by the University of Cambridge, the Alan Turing Institute, Microsoft Rexch and the European Middle Ages Center (ECMWF).
“This method provides custom forecasts for different industries and regions,” said Richard Turner, a professor of machine learning at the University of Cambridge.
In contrast, traditional weather forecasting systems need large teams of researchers and years of time to create a custom model. In addition, current supercomputers spend hours spending time processing real -world data and producing models.
Revolutionary
The Turner described this as a completely different approach to the past and predicted that the technology would become a new and standard way in atmospheric predictions. He also stated that the Aardvark model will be able to provide precise eight -day predictions in the future, while current systems are often limited to five -day forecasts. Also, this technology also provides very accurate local predictions.
Democratic
Dr. Scott Hasking, director of environmental science and innovation at the Alan Turing Institute, described the progress as a major step towards “democratizing atmospheric predictions.” He believes that the technology could provide powerful prediction tools to developing countries, as well as to help policymakers, crisis managers and atmospheric forecasts.
Better prediction of natural disasters and climate change
Dr. Anna Allen, the main author of the study from the University of Cambridge, emphasized that the findings of the study could pave the way for improvement of natural disasters such as sea storms, forest fires and tornadoes. In addition, the technology also has the potential to improve air quality forecast, ocean dynamics and sea ice changes.
This approach previously used by ECMWF has led to faster and more accurate predictions.
According to the researchers, the Aardvark model has provided better performance than the US GFS global forecasting system, while only 2 % of traditional systems input data.
This progress could initiate a major change in climate forecasting, especially by reducing dependence on expensive superconductors and increasing access to atmospheric forecasts for different regions of the world.
Latest Passing over countries : Spain | Dominica | United Arab Emirates
RCO NEWS
