Skip to main content
U.S. flag

An official website of the United States government

Publication Date
16 May 2022

Predicting Slowdowns in Decadal Climate Warming Trends With Explainable Neural Networks

Subtitle
Explainable machine learning methods predict temporary slowdowns in decadal warming and uncover potential drivers.
Print / PDF
Powerpoint Slide
Science

Collaborators Dr. Zachary Labe and Prof. Elizabeth Barnes (Colorado State University) demonstrated that an artificial neural network (ANN) trained with CESM2 Large Ensemble maps of ocean heat content can skillfully predict simulated onsets of temporary decadal slowdowns in global-mean temperature, as well as the warming slowdown observed in the early 2000s. These events are often triggered by transitions to a negative Interdecadal Pacific Oscillation (IPO) phase (as identified via explainable AI methods), but other precursors, such as an aerosol forcing, may exist.

Impact

Although the ANN is trained only on climate model data from one climate model, we find that it also produces skillful predictions of the early  2000s  warming slowdown in observational data. This work shows that temporary warming slowdowns may have some predictability from Pacific climate variability and highlights the promising future for using machine learning tools in a wide variety of decadal climate prediction problems.

Summary

Long-term observations reveal that Earth's average temperature is rising due to human-caused climate change. Along with this warming trend are also variations from year to year and even over multiple decades. This temperature variability is often tied to regional patterns of heat in the deep ocean, which can then modulate weather and climate extremes over land. In an attempt to better predict temperature variability on decadal timescales, Labe and Barnes use a machine learning method called artificial neural networks (ANNs) and data from a climate model experiment, which was designed to compare climate change and variability. Here, the ANN uses maps of ocean heat to predict the onset of temporary slowdowns in the rate of global warming in both the climate model and in real-world observations. The study then uses a visualization technique to find which areas of ocean heat that the ANN is using to make its correct predictions, which are found to be mainly across the Pacific Ocean. In agreement with recent research, this study finds that new data science methods, like machine learning, can be useful tools for predicting variations in global climate.

Point of Contact
Elizabeth A. Barnes
Institution(s)
Colorado State University
Funding Program Area(s)
Publication