An artificial intelligence model capable of quantifying the damages of each new coal, gas, or oil project has been presented, offering the first detailed estimate of the carbon footprint and social repercussions of an exploitation still in the approval phase.
Announcement Details: Who, What, Precise Figures
A group of researchers has unveiled a predictive model that combines a deep neural network with satellite data, ECMWF reanalyses, and Copernicus observations. The system also integrates databases of industrial emissions and public health indicators to produce a mapping of impacts at the scale of the basin concerned.
The core of the tool relies on machine learning: it trains the neural network using decades of atmospheric observations, correlating variations in CO₂, fine particle, and greenhouse gas concentrations with extraction activities. Thanks to this approach, the model can extrapolate future effects even before exploitation begins, evaluating prediction uncertainty at each stage.
Results are presented as impact scores in terms of "potential damages" for ecosystems, air quality, and local populations. No absolute figures are communicated, but authors emphasize that each new project notably increases the risk of exacerbating extreme heat episodes, air pollution, and community displacements.