Rosenberg, N.A., Amrhein, D.E., and Thompson, L. (2026)
Presented at:
Ocean Sciences Meeting 2026Upper ocean heat content in the Southern Ocean has exhibited surprising trends and patterns of variability over the last several decades, including notable cooling in the Pacific sector and a strong zonal asymmetry. Investigators have attempted to understand the complex roles of heat flux, wind, and ocean circulation in setting the spatial and temporal patterns of variability, including through the use of forward modeling experiments and statistical analyses of observational data. In this work, we use adjoint modeling to investigate the upper ocean heat content variability in this region using a different approach. The adjoint solution to the ECCOV4r4 State Estimate, a data-constrained ocean model simulation, is run via the free-to-use EMU software suite, and provides the sensitivity of the ocean state to stochastic atmospheric forcing. We combine this sensitivity with realistic atmospheric forcing patterns to derive Dynamics-Weighted Principal Components, introduced in Amrhein et al. (2024) and Stephenson et al. (2024). With these patterns, we run a suite of forward model simulations using the MITgcm to disentangle the roles of specific patterns of heat flux and wind stress in driving variability within the ocean.