The ice from the Arctic disappears faster, and a model based on AI tries to predict at what pace the melting will occur
Marine ice in Arctic (photo source, dreamstime.com)
The most recent report of the World Meterorology Organization shows that the Arctic ice continues to decrease. Further losses are provided in the seas in Barents, Bering and Ohotsk. In this context, a model based on AI tries a much better forecast at the seasonal level.
Marine ice has an impact on the solar energy absorbed by the Arctic Ocean, and its disappearance accelerates regional heating. The forecasting of marine ice ice extension has significant implications for climate change and navigation in the region.
The problem is that seasonal forecasts for ice in September often face a difficulty known as the “spring predictability barrier”.
To address this issue, a research team from the Institute of Oceanology of the Chinese Academy of Science has developed a new artificial intelligence model called Sicnetseason, intended for seasonal scale forecasts. The study was published in the Geoscientific Model Development magazine and is quoted by Phys.org.
The traditional methods of forecasting marine ice, based on numerical models, frequently face this “predictability barrier” (Spring Predictability Barrier) in the case of Arctic ice. This is a concept that describes the increased difficulty of predicting El Niño and La Niña during the spring of the northern hemisphere, especially in March -May.
Basically, this “Predictability Barrier” Spring is a period in the spring in which the weather models cannot accurately predict the development of the future EL Niño or Laña phenomena, due to the instability of the atmosphere-ocean system at that time.
The Chinese Sicnetseash model promises more accurate forecasts compared to the existing numerical and statistical models.
Experimental results show that when April and are used as a beginning for the forecast of the extension of the September ice, the Sicnetseash model obtains an improvement of the predictive ability of 7-10% and an accuracy of over 14% higher in the forecasts regarding the limits of the ice, significantly reducing this precitability barrier in the spring.
Photo source: dreamstime.com
