Assessing and Understanding Atlantic Multidecadal Variability in a Suite of GFDL Climate Models: Roles of Climate Feedback and Teleconnection

The Atlantic Multidecadal Variability (AMV) is one of the most prominent decadal variability modes and has a worldwide climate impact. Among all the decadal variability modes, AMV likely has the greatest potential of predictability because of its hypothesized close association with deep ocean dynamics, in particular, the Atlantic Meridional Overturning Circulation (AMOC). In spite of some advances in the last three decades, however, many fundamental questions on the mechanism of AMV remain unclear, in particular, regarding the roles of climate feedback and oceanic teleconnections. What are the roles of climate feedbacks of different regions on AMV, and do those feedbacks contribute to different mechanisms of variability in the subpolar and subtropical North Atlantic? What are the roles of various oceanic teleconnections on AMV and how is the time scale of the AMV associated with ocean dynamics?
We propose to study the mechanism of AMV in the Atlantic in a suite of GFDL climate models using a combined statistical and dynamic approach, with focus on the roles of climate feedback and oceanic teleconnections. First, we will perform statistical analyses on the available model simulations, including newly available multi-millennial simulations, to assess climate feedbacks and climate teleconnections. Climate feedback will be assessed with various lead-lag feedback analysis methods, including the multi-variate Generalized Equilibrium Feedback Analysis (GEFA). The temporal evolution of the AMV, especially in the subsurface ocean, will be examined using lead-lag correlation analyses and the Linear Inverse Modeling (LIM) analysis. The interaction between the tropical and subpolar North Atlantic will also be examined using LIM. Second, we will perform systematic modeling surgery sensitivity experiments in the newly developed GFDL model “SPEAR”. The goal of the sensitivity experiments will be to explicitly assess the roles of ocean-atmosphere feedback and oceanic teleconnections on the characteristics of simulated AMV. We will perform “Partial Coupling” (PC) experiments by suppressing ocean atmosphere coupling in specific regions, notably the global ocean outside the North Atlantic region, and then, the subtropical North Atlantic, to explicitly assess the roles of ocean-atmosphere feedback outside the North Atlantic and subpolar North Atlantic, respectively. The role of cloud feedback will be studied in experiments that isolate cloud feedbacks. We will also perform “Partial Blocking” (PB) experiments by blocking oceanic teleconnections with a sponge wall in the ocean component model. Westward teleconnection associated with planetary waves will be assessed by a set of PB experiments with meridional PB walls across the middle North Atlantic in the subtropical and subpolar regions, while southward teleconnection along the western boundary will be assessed in a set of PB experiments with zonal PB walls across the western boundary current at several latitudes.
Our proposed work will be a close collaboration between OSU and GFDL. Our research will also be coordinated closely with the ongoing modeling activity at GFDL and forms a part of a comprehensive research strategy at GFDL on decadal climate variability and predictability, including decadal climate prediction and global warming research.