Connecting Atmospheric Dynamics to Abyssal Ocean Geometry on Paleoclimate Time Scales

Proxy measurements of ocean tracers at the Last Glacial Maximum (LGM, ca. 23-19 ka) suggest that the western Atlantic Ocean was filled dominantly by Antarctic Bottom Water (AABW), with a lesser role for North Atlantic Deep Water (NADW) relative to the modern. Adjoint sensitivities calculated in ocean models provide recipes for how changes in atmospheric boundary conditions (e.g. wind stress, surface air temperature, and precipitation) can alter abyssal watermass distributions and flow pathways. A challenge in interpreting adjoint sensitivities is that their spatiotemporal structure is governed by the dynamics of the ocean model, not by atmospheric dynamics. As a result, an apparent dominant dynamical link between the atmosphere and ocean diagnosed by the adjoint may never be realized because it is precluded by destructive interference from coherent atmospheric behavior.

Here we bring atmospheric dynamics to bear on adjoint sensitivities by finding coherent patterns of atmospheric glacial-interglacial differences that maximize the projection onto adjoint sensitivities, with a focus on investigating atmospheric origins of abyssal watermass geometry at the LGM. Atmospheric circulation patterns will be identified using LGM scenarios from coupled model simulations. Results suggest a strong role for Southern Ocean atmospheric circulation changes in controlling abyssal watermass properties on millennial time scales and beyond. This approach holds promise for improving ongoing efforts towards ocean state estimation during paleoclimates.