The heat balance shapes deep convection in the Irminger Sea

The Irminger Sea is one of the key region in the North Atlantic where deep winter convection develops.

We use ARMOR-3D dataset (0.25×0.25˚, from 1993) for calculating of heat (and freshwater) content, oceanic heat fluxes in the upper 500-m layer and the vertical heat exchange with the lower layers through the 500 m level. The air-sea heat exchange was derived from the OAFlux dataset (1×1˚), the radiation balance was obtained from the ERA-Interim reanalysis (0.25×0.25˚).

Computation of the heat balance was done over a closed region covering the central and western Irminger Sea (58-62˚ N and 36-44˚ W). The computations were repeated for several similar rectangular areas to analyze sensitivity of the analysis to the choice of boundaries of the region. However, the results of the analysis were largely independent from these variations.

The upper ocean heat advection in the study region was 37 TW (integrated along all boundaries of the region, sign «+» means that flux is directed to the study region), and was the dominant term in the annual mean heat balance. The annual mean latent heat flux (-21 TW) and sensible heat flux (-5 TW) were directed from the ocean. The annual mean radiation balance was 8 TW (to the ocean), while vertical heat exchange with the lower layers was low (-0.1 TW). On average, heat balance of upper 500-m layer was positive, and not all heat fluxes might be considered. For example, the contribution of horizontal mesoscale eddy exchange could be important. However, the amplitude of the interannual variability of the heat balance of about 15 TW was close to that of the heat content (about 20 TW), while the correlation between the parameters was significant and high (0.79) (after removal of the quadratic trend 0.80). This suggests that the main heat fluxes, which affect the interannual variability of the heat content in the upper 500-m layer were taken into account.

Interannual variability of maximum convection depth in the central Irminger Sea was found to significantly correlate with the upper ocean heat content mean over September-November (-0.73); both parameters showed a similar long-term tendency. The correlation of the convection depth with the freshwater content (September-November) was significantly less and positive (0.49). The latter is counterintuitive, as we expect a decrease of the convective depth with an increase of the upper ocean freshwater contents. It can be assumed this correlation was induced by a high negative correlation between the upper ocean heat and freshwater contents in the region (-0.64). The analysis, thus, suggests that the long-term variability of deep convection in the Irminger Sea was shaped by variability of the main heat fluxes, entering the region.