Yingying Zhao, Emanuele Di Lorenzo, Matthew Newman, Antonietta Capotondi, Samantha Stevenson

Published in Geophysical Research Letters, August 2023

Understanding and forecasting Tropical Pacific Decadal-scale Variability (TPDV) strongly rely on climate model simulations. Using a Linear Inverse Modeling (LIM) diagnostic approach, we reveal Coupled Model Intercomparison Project Phase 6 models have significant challenges in reproducing the spatial structure and dominant mechanisms of TPDV. Specifically, while the models' ensemble mean pattern of TPDV resembles that of observations, the spread across models is very large and most models show significant differences from observations. In observations, removing the coupling between extratropics and tropics reduces TPDV by ∼60%–70%, and removing the tropical thermocline variability makes the central tropical Pacific a key center of action for TPDV and El Niño Southern Oscillation variability. These characteristics are only confirmed in a subset of models. Differences between observations and simulations are outside the range of natural internal TPDV noise and pose important questions regarding our ability to model the impacts of natural internal low-frequency variability superimposed on long-term climate change.

Key Points

• A Linear Inverse Model is used to decompose and compare tropical Pacific decadal variability dynamics in observations and climate models
• Coupled Model Intercomparison Project Phase 6 models have significant challenges in reproducing the observed dynamics and spatial pattern of tropical Pacific decadal variability
• Many models (>50%) cannot reproduce the roles of extratropics-tropics coupling and thermocline variability in tropical Pacific variability

Zhao, Y., Di Lorenzo, E., Newman, M., Capotondi, A., & Stevenson, S. (2023). A Pacific tropical decadal variability challenge for climate models. Geophysical Research Letters, 50, e2023GL104037. https://doi.org/10.1029/2023GL104037