The Relationship Between the Pacific Decadal Oscillation and the Atlantic Multi-Decadal Oscillation

This study explores the interactions between climate variability in the North Pacific and North Atlantic basins. The dominant climate modes in each basin (the Pacific Decadal Oscillation (PDO) for the North Pacific, the Atlantic Multi-decadal Oscillation (AMO) for the North Atlantic) have regional teleconnections downstream that affect weather patterns in North America and Europe respectively. These teleconnections can also reach the subsequent ocean basins and force variability in these climate modes. A better understanding of these connections can improve models and provide better long-term predictions of these modes’ impacts.

Previous studies explore potential links between the two basins and suggest that significant correlation exists between the PDO and the AMO, but only when the AMO leads by roughly 12 years or the PDO leads by 1 year. They briefly mention a high correlation at a period of 18 years but do not delve further nor propose a mechanism. Other studies find connections between the two basins on millennial timescales. The mechanism for this connection is a shut-down of the Atlantic Meridional Overturning Circulation (AMOC) via freshwater input. The subsequent teleconnections reach and affect north Pacific climate variability.

Our observational results suggest that the two basins and their respective modes are linked at multi-decadal timescales. These results are statistically significant with 99% confidence. Further, we find that the correlation is dependent on the method of calculating the AMO index, namely how the global warming signal is removed (by subtracting the global mean time-series or via linear detrending). The 0-lag correlation is consistently negative when using the former, while the latter provides r values of roughly 0. There is also considerable spread in the magnitude of the signal depending on which sea-surface-temperature dataset is analyzed.

Certainty in this signal is difficult to ascertain due to relatively short observational records and the potential that external forcing confounds it in a way that cannot be controlled for. Thus, the physical nature of this link will be explored using climate model outputs, specifically comparing pre-industrial control runs to historical runs. We will utilize an “ensembles of ensembles” dataset to compare model results to observational results and to determine what role external forcing plays in this inter-basin relationship. This new dataset provides ensembles of runs from different model configurations that can help to definitively answer the questions posed above.