Skip to main content
U.S. flag

An official website of the United States government

Publication Date
26 October 2023

Revisiting a Constraint on Equilibrium Climate Sensitivity From a Last Millennium Perspective

Using past 1000 global temperature variability to constrain climate sensitivity.
Print / PDF
Powerpoint Slide

Scientists at UCLA and USC have evaluated whether a previously proposed emergent constraint between global interannual temperature variability and equilibrium climate sensitivity (ECS) holds up when the former is derived over the past millennium rather than the past century. They find that the relationship between interannual temperature variability and ECS is stronger when the temperature variability is computed over a longer time frame. Paleoclimate reconstructions of temperature are then used in conjunction with this relationship to reduce uncertainty in model estimates of ECS. Notably, this constraint suggests an ECS value of 2.6 ± 0.7K.


Substantial doubt had previously been cast on the use of global temperature variability metrics to constrain inter-model spread in ECS. This study shows that when temperature variability metrics are computed over much longer time periods (1150 years instead of 100), a meaningful relationship with ECS does exist. This relationship can then be used to make informed predictions about the likely values of ECS. Our estimates of ECS are slightly weaker than the best estimate of 3K from the Intergovernmental Panel on Climate Change.


Despite decades of effort to constrain equilibrium climate sensitivity (ECS), the latest generation of climate models still exhibits a large spread. Past studies have sought to reduce ECS uncertainty through a variety of methods including emergent constraints, a model evaluation technique that offers the potential to reduce inter-model spread in projections of climate change. One example uses global temperature variability over the past century to constrain ECS. While this method shows promise, it has been criticized for its susceptibility to the influence of anthropogenic forcing and the limited length of the instrumental record used to compute temperature variability. Here, we investigate the emergent relationship between ECS and two metrics of global temperature variability using model simulations and paleoclimate reconstructions over the last millennium (850–1999). We find empirical evidence in support of these emergent relationships. Observational constraints suggest a central ECS estimate of 2.6–2.8 K, consistent with the Intergovernmental Panel on Climate Change's consensus estimate of 3K. Moreover, they suggest ECS “likely” ranges of 1.8–3.3 K and 2.0–3.6 K.

Point of Contact
Chad Thackeray
University of California Los Angeles (UCLA)
Funding Program Area(s)