The global climate is constantly changing and global warming is progressing faster than the best models can cope with.  Assessing the future impact of influencing factors requires hundreds of years of simulations as we move further and further away from the reference points (see “Climate models can't explain what's happening to the Earth”).  Despite vast climate research, the models often disagree about past and future global warming.  The latest AI machine learning approach has been adapted to train a convolutional neural network that predicts the time and its uncertainty based on the current state of the climate system until various regional warming thresholds are reached.  Limited observations are then added to fine-tune the climate model-trained network to improve predictions.

Barnes, Elizabeth A, Noah S Diffenbaugh and Sonia I Seneviratne. "Combining Climate Models and Observations to Predict the Time Remaining Until Regional Warming Thresholds are Reached". IOP Publishing. December 10, 2024. https://iopscience.iop.org/... (Contributed by Gregory Autin).

Posted on 27/01/25

Share this! 
Recent Abstracts