Mechanisms of low-frequency variability in North Atlantic Ocean heat transport

Dylan Oldenburg, University of Washington Seattle Campus, Oceanography, Seattle, United States, Robert Jnglin Wills, ETH Zürich, Institute for Atmospheric and Climate Science, Zürich, Switzerland, Kyle Armour, University of Washington, Dept. of Atmospheric Sciences and School of Oceanography, Seattle, United States and LuAnne Thompson, University of Washington, School of Oceanography, Seattle, United States
Abstract:
We study the drivers of ocean heat transport (OHT) variability in the North Atlantic Ocean using a low-frequency component analysis that highlights variability at decadal and longer timescales. This analysis calculates low frequency patterns (LFPs) of OHT, which are linear combinations of the leading Empirical Orthogonal Functions that maximise the ratio of low-frequency variance to total variance. A major advantage of this approach is that time resolution is not lost, because the data are not lowpass filtered.

We analyse long pre-industrial control simulations of two climate models – CCSM4 (1300 years long) and GFDL ESM2M (500 years long). We choose these models because they differ in their primary deep convection sites – the Labrador Sea dominates in CCSM4, whereas the Greenland-Iceland-Norwegian (GIN) Seas dominate in ESM2M. We compute lead-lag regressions between the leading LFP and relevant variables such as the sea-level pressure, mixed-layer depth, overturning streamfunction, upper-ocean density and surface water mass transformation. We find that in both models, low-frequency OHT variability is primarily driven by variations in the Atlantic Meridional Overturning Circulation (AMOC). This variability is preceded by an anomalous sea-level pressure pattern linked to anomalous winds off eastern North America, cooling the Labrador Sea waters and increasing convection. Water mass transformation changes are dominated by anomalous heat fluxes. A strengthened AMOC carries anomalous warm water northward into the subpolar gyre, which shuts down the anomalous convection, weakening AMOC. Surprisingly, ESM2M shows similar mechanisms to CCSM4, with Labrador Sea convection variability contributing substantially to its AMOC and OHT variability, despite its climatological deep-water formation being focussed in the GIN Seas. It thus seems that Labrador Sea deep water formation may play a key role in decadal AMOC and OHT variability, even if the GIN Seas dominate the mean-state.

Back to: Atlantic Ocean Variability in a Changing Climate: Observations, Modeling, and Theories I