German scientists said that their understanding of the climate and Earth science could be significantly improved by the use of Artificial Intelligence (AI).
AI can be applied to all data related to extreme events such as fire spreads or hurricanes, which are very complex processes that have a little influence on local conditions.
AI can also be applied to soil movement and vegetation dynamics data, atmospheric and ocean transport – some of the classic topics of Earth system science.
Markus Reichstein of the Max Planck Institute for Biogeochemistry in Jena, Germany, said that, from a large number of sensors, a deluge of Earth system data has become available, but we’ve been lagging behind in analysis and interpretation so far.
The co-author Joachim Denzler, from the Friedrich Schiller University in Jena (FSU) added, beyond classical machine learning applications such as image recognition, natural language processing or AlphaGo, this is where deep learning techniques become a promising tool.
However, deep learning approaches are difficult. All the statistical and data-driven approaches do not guarantee physical consistency per se and are highly dependent on data quality, according to the study published in the journal Nature.
The requirement for data processing and storage capacity is very high.
If so-called hybrid models are created by bringing both the techniques together, they can, for example, be used for modelling the motion of ocean water to predict sea surface temperature. While for now, the temperatures are modelled physically and the ocean water movement is represented by a machine learning approach.
Reichstein also explained that the idea is to combine the best of two worlds, the consistency of physical models with the versatility of machine learning, to obtain highly improved models.