Variability of convection processes in the Nordic Seas from a high-resolution coupled Earth System Model

Recent studies of convection in the Nordic Seas have challenged the conventional view that Lower North Atlantic Deep Water (NADW) primarily forms in the Greenland and Iceland Seas, suggesting that the formation of deep waters is a more gradual process that initiates in the Norwegian Sea, Lofoten Basin, and possibly the Arctic Ocean. Furthermore, the variability of convection, whose interannual signal is for instance observed in the Denmark Strait overflow transport (e.g., Macrander et al. 2005), is possibly influenced by sea-ice extent and atmosphere variability, both locally and remotely. Here, we investigate these issues from a modeling perspective, using a high-resolution, fully-coupled Earth System Model simulation: the Energy Exascale Earth System Model (E3SM, version 0.1) pre-industrial simulation, which features a 1/10° horizontal resolution in the ocean/sea-ice components and a 1/4° resolution in the atmosphere/land components. Since greenhouse gases are fixed at pre-industrial levels, any variability that is detected from the simulated trajectory is solely due to internal variability within the model (atmosphere, ocean, and sea-ice components). Results show that the deepest convection sites are found next to the ice edge in the Greenland and Iceland Seas, while more intermediate water formation is found upstream in the Lofoten and Norwegian basins. Based on a mixed layer depth metric in the Greenland Sea, we identify two modes of convection: a low-convection composite and a high-convection composite. The latter is triggered by a high surface heat flux tongue near the continental slope, which is more extensive during years when sea-ice coverage is limited to the shelf area rather than extended to the Greenland gyre interior. Furthermore, we investigate an atmospheric connection to the convection variability in the Nordic Seas through a self-organizing maps (SOMs) analysis. Relationships of leading, lagging, or simultaneous correlations between the SOMs atmospheric circulation (in terms of lowest model-level winds and sea-level pressure) and the oceanic convection will be identified to better understand the atmospheric-oceanic connections and feedbacks.