Identifying the Timing of Regional Summertime Minimum Temperature Threshold Crossings and the Potential Subsequent Climate Evolutions

Marybeth C. Arcodia, Elizabeth A. Barnes

Earth s Future · 01 July 2026 · Colorado State University (incl. CIRA)

In one sentence. CNNs with probabilistic outputs predict timing of summertime minimum temperature threshold crossings in Central North America; observation transfer learning shifts timing 4 to 10 years later.

Abstract

Abstract Regional disparities in the magnitude and rate of warming drive uncertainties in not only the timing of critical temperature threshold crossings, but also potential climate system evolutions beyond threshold crossings. Here, convolutional neural networks are trained to predict the time remaining until the Central North American region will remain above specific temperature thresholds for summertime minimum temperature. Predictions of the parameters of non‐normal probability distributions provide uncertainty quantification for the time‐to‐threshold predictions using multiple CMIP6 climate models and four projection scenarios. While prior CNN‐based frameworks have predicted regional annual‐mean temperature crossings, here we extend the approach to summertime minimum temperatures including both internal and forced variability. Transfer learning is used to fine‐tune networks with additional training from observations, narrowing the time‐to‐threshold projection average spread by 44% and shifting predictions 4–10 years later than model‐only training. By identifying the timing of these crossings, we analyze both the ensemble‐mean climate response and the evolution of individual storylines aligning with the predicted crossing time. This targeted approach allows for extraction of physically‐viable trajectories of future climate change, grounded in the timing of threshold exceedance. Storylines show substantial spatial variability in post‐threshold warming across thresholds, scenarios, and ensemble members, with northern CNA emerging as a consistent hot spot. The framework is illustrated with one sector‐relevant application to cattle heat stress in Central North America. Identifying when regional temperature thresholds are crossed and how climate states could evolve afterward can inform future planning and adaptation efforts, especially for sectors vulnerable to climate change like agriculture.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.