Award: OCE-1357373

The overarching objectives of this project are to simulate oxygen and nutrient cycling in the Labrador Sea, one of the regions of deep water formation in the North Atlantic, using a hierarchy of models. The Labrador Sea is also a region of extreme seasonality and intense biological productivity, thus oxygen cycling there likely reflects multiple physical and biological processes. The outcomes of this study promote a better understanding of the interannual variability of oxygen and nutrients in the Labrador Sea, and ultimately contribute to knowledge on how a changing climate impacts these cycles. The hierarchy of models cover a wide range of spatial scales from the scales of convective mixing (~100m) to the basin scale (1,000km). Historic observations show strong variance of oxygen on the interannual to decadal timescales. However, there are relatively few mechanistic studies of the oxygen cycling in the Labrador Sea, and the link between climate variability and oxygen cycling in this basin is not yet fully understood. A suite of sensitivity simulation using non-hydrostatic configuration of the MITgcm coupled with a simple ocean biogeochemistry scheme. This task is primarily performed by a PhD student (D. Sun). The high-resolution (250m grid spacing) model can reproduce ARGO-O2 float data in the convective region of the Labrador Sea (Fig 1), and it help explain the mechanisms behind the uptake of oxygen into the deep waters. Sensitivity experiments varying the air-sea heat flux and the rate of diffusive gas exchange support the theoretical prediction that under a strong convective mixing, diffusive and bubble-mediated gas fluxes work together to significantly increase the oceanic O2 uptake (Fig 2). To correctly model the variability of dissolved oxygen in the Labrador Sea, it is necessary to capture the regional features of the deep convection and its temporal variability. The Regional Ocean Model System (ROMS) configured at 5km horizontal resolution over the LS shows a convective region in the central Labrador Sea (Fig 3 left) and the vertical velocity exhibits significant interannual variability (Fig 3 right). The variability of potential temperature over the water column compares reasonably well with the existing observations (Fig 4) and the agreement is also good for salinity. This work was performed by a PhD student (F. Tagklis) supported by this study. He also performed the analysis of the CMIP5 experiments, to determine the pattern of ocean deoxygenation among the models. While the CMIP5 models reproduce climatological distribution of O2 reasonably well (Fig 5), representation of convective areas/strengths are widely different among the CMIP5 models (Fig 6). Detailed analysis of the model fields are performed. This project provided training opportunities and professional development for two graduate students including participation in oceanographic cruises and scientific conferences. The scientific outcome of this project resulted in four peer-review journal publicaitons. Last Modified: 05/01/2018 Submitted by: Takamitsu Ito