Understanding drivers and impacts of CGCM biases in representing the decadal variability of Labrador Sea convection

The Labrador Sea (LS) is one of the few regions in the world ocean where deep convection occurs. The intense air-sea interaction drives the convective mixing and the site acts as a window through which anthropogenic carbon is sequestered into the interior ocean. Recent work highlights that buoyancy forcing over the Labrador Sea is key in controlling the Atlantic Meridional Overturning Circulation (AMOC) and that AMOC inter-annual signals are closely related to the variability of the Labrador Sea convection. Historic observations collected over the past 60 years show that the LS convective activity undergoes dramatic interannual-to-decadal variability and –within the limitation of the available measurements – no statistically significant trend. A model integration performed using a regional ocean model (ROMS, Regional Oceanic Modeling System) run at 5km horizontal resolution over the LS can reproduce the observed variability. However, coupled general circulation models (CGCMs) from the Coupled Model  Intercomparison Project Phase 5 (CMIP5) are not yet capable of representing the extent and  statistical properties of the LS convection, while often displaying a weakening trend for the past  50 years. Model biases hamper the representation of the AMOC and of the inventories of dissolved inorganic carbon, and limit our ability to project their future changes.  The overarching objectives of this project are to diagnose the sources of CGCMs biases in the LS focusing on a subset of CMIP5 runs and to quantify the impacts of those biases on the representation of carbon uptake and inventories in the basin. They will be achieved through a sensitivity study to be performed using regional ocean-only ROMS simulations covering most of the North Atlantic forced by momentum, heat and freshwater fluxes, and/or boundary conditions from the CMIP5 runs.
This project will establish cause-effect linkages between the representation of mesoscale processes, of the atmospheric forcing fields, and of the gyre circulation, and the (modeled) Labrador Sea circulation, its variability and carbon uptake characteristics.  The regional simulations will include an ocean biogeochemical and carbon cycling module. The interpretation of all model analysis will be aided by careful comparisons with shipboard and Argo measurements in the Labrador Sea, and along 53N. In this regard we will build upon our ongoing collaboration with Dr I. Yashayeav at the Bedford Institute of Oceanography. This project will contribute a better understanding of the potential predictability of Labrador Sea convection and of the natural and anthropogenically forced variability of the AMOC.
This proposal addresses the objective of the NOAA funding opportunity, CVP AMOC-Climate Linkages in the North and/or South Atlantic (NOAA-OAR-CPO-2016-2004413) to ‘refine the  current scientific understanding of the AMOC state, variability and change’ by focusing on the  interannual and decadal variability for the LS branch. The proposed work contributes to three  priorities identified in the US AMOC 2014 report and advances the NOAA’s Next-Generation  Strategic Plan to ‘improve scientific understanding of the changing climate system by diagnosing  the physical and biogeochemical biases in the CGCMs that are used in the future prediction and  projections by the Intergovernmental Panel on Climate Change’.