The impact of polar mesoscale storms on northeast Atlantic Ocean circulation

Atmospheric processes regulate the formation of deep water in the subpolar North Atlantic Ocean and hence influence the large-scale ocean circulation1. Every year thousands of mesoscale storms, termed polar lows, cross this climatically sensitive region of the ocean. These storms are often either too small or too short-lived to be captured in meteorological reanalyses or numerical models2–4. Here we present simulations with a global, eddy-permitting ocean/sea-ice circulation model, run with and without a parameterization of polar lows. The parameterization reproduces the high wind speeds and heat fluxes observed in polar lows as well as their integrated effects, and leads to increases in the simulated depth, frequency and area of deep convection in the Nordic seas, which in turn leads to a larger northward transport of heat into the region, and southward transport of deep water through Denmark Strait. We conclude that polar lows are important for the large-scale ocean circulation and should be accounted for in short-term climate predictions. Recent studies3,4 predict a decrease in the number of polar lows over the northeast Atlantic in the twenty-first century that would imply a reduction in deep convection and a potential weakening of the Atlantic meridional overturning circulation. Polar lows are mesoscale (<1,000 km diameter) low-pressure systems that occur throughout theworld’s subpolar seas, but tend to cluster in the North Atlantic over the Nordic (Greenland, Iceland, Norwegian) and Irminger seas4–7. Here minor differences in atmospheric forcing are sufficient to alter the density of the surface waters, such that the rate at which water sinks to depth is changed1. This ventilation process, known as deep open-ocean convection, is one of the fundamental mechanisms responsible for the renewal of North Atlantic Deep Water (NADW)—the major deep-water mass driving theAtlanticmeridional overturning circulation8 (AMOC). Polar lows are not generally well resolved in global meteorological analyses, reanalyses or climatemodels owing to their small scales (typically ∼250 km) and short lifetimes (typically 24–48 h; refs 2– 4). Yet, the most intense polar lows are associated with localized gale force winds and heat losses from the ocean >1,000Wm−2, sufficient to significantly change the underlying ocean9–12. For brevity we will use the term polar low to encompass all polar mesoscale cyclones. We note, however, that in the literature13 the term polar low is usually reserved for more intense polar mesoscale cyclones with wind speeds>15m s−1. Many climate models predict that anthropogenic increases in atmospheric carbon dioxide will slow down the AMOC (ref. 14). However, in omitting polar lows, climatemodels are, at present, deficient at forcing the ocean over the critical deep-water formation regions of the subpolar seas, limiting confidence in their predictions.