Skip to main content
U.S. flag

An official website of the United States government

Publication Date
14 April 2023

Volcanic Forcing Degrades Multiyear-to-Decadal Prediction Skill in the Tropical Pacific

Print / PDF
Powerpoint Slide

Twin sets of decadal prediction hindcasts using the CESM1 model are compared, one that includes volcanic aerosol forcing (Volc) and one that does not (NoVolc). The experiments are similar to those proposed for component C of the Decadal Climate Prediction Project (DCPP) protocol for CMIP6, but here the NoVolc set includes all initializations between 1954-2015, thus permitting a first-ever assessment of global skill impacts as a function of lead time. Unexpectedly, the NoVolc set shows significantly higher skill for detrended tropical Pacific decadal variability (TPDV). A budget analysis reveals that unrealistic central-eastern tropical Pacific cooling in the Volc set in response to large eruptions results from a sequence of thermodynamic and dynamic (ENSO-like) processes that appear to overwhelm the realistic, internal warming signals simulated in NoVolc. Results suggest that more skillful decadal predictions may be possible by improving model response to volcanic forcing and its interaction with internal climate variability.


The work has already spurred significant interest from groups involved in DCPP, several of which intend to replicate the experiment to explore robustness across different decadal prediction systems. Depending on these findings, the DCPP panel could choose to prioritize no-information-from-the-future (nif) hindcast experiments for CMIP7. The broader impacts for coupled climate modeling are significant, as the results suggest that validating coupled model response to volcanic aerosols by comparing global mean surface temperature with observations is insufficient for establishing the fidelity of model behavior. 


The finding that volcanic aerosol forcing degrades decadal prediction skill in the tropical Pacific highlights a model and/or forcing deficiency that, if addressed, could lead to significant advances in the realism of both initialized and uninitialized climate simulations that include volcanic forcing.

Point of Contact
Stephen Yeager
National Center for Atmospheric Research (NCAR)
Funding Program Area(s)