| 07 Sep 2026
Seasonal reversal of circulation in Qeqertarsuup Tunua (Disko Bay), Greenland
Abstract. Qeqertarsuup Tunua (Disko Bay), located in West Greenland, is a highly productive ecosystem and a major site of glacier-ocean interaction with the Kangia (Ilulissat Icefjord), yet its circulation and the mechanisms governing its seasonal variability have not been previously investigated. An empirical orthogonal function (EOF) analysis of upper-ocean velocity from the GLORYS12 reanalysis (1993–2026) provides the first quantitative description of the bay's dominant circulation mode, EOF1, which explains 55.8 % of the variance and is coherent through the upper ~130 m. EOF1 reveals a seasonal reversal between a cyclonic regime, in which an intensified West Greenland Current establishes cyclonic circulation in the bay, and an anticyclonic regime, in which the flow within the bay reverses. The seasonal cycle accounts for 49.5 % of the mode's variance, with the cyclonic regime prevailing from approximately May–June to November–December (with a peak in August), and the anticyclonic regime dominating in January–April. The alternation is consistent with a reversal of the along-shelf ocean surface stress forcing, strongly correlated with the circulation mode. The bay's circulation is, in turn, coherent with the West Greenland Current system at Davis Strait, where observed near-surface along-strait velocities flowing into Baffin Bay approximately double during the cyclonic regime and fall to near zero during the anticyclonic regime. Temperature transport towards the bay along the West Greenland shelf is an order of magnitude larger during the cyclonic regime than during the anticyclonic regime (+7.6 against +0.7 TW). Independent evidence from seven drifting-iceberg trajectories (June–July 2025) and an Argo float (June–October 2022) is qualitatively in line with these regimes. By governing the seasonal delivery of the warmest, fastest-flowing West Greenland shelf water into the bay, this circulation mode affects the bay's heat budget and the near-surface transport of freshwater, nutrients, and larvae that support the region's fisheries.
This paper looks at the circulation in Qeqertarsuup Tunua (Disko Bay), using the 1/12 degree GLORYS12 ocean reanalysis, in combination with some other data sources for the region. An EOF analysis finds the main mode being a seasonal reversal between a cyclonic and anti-cyclonic regime, coherent through the top ~130 m of the Bay. The authors find the cyclonic mode is established by an intensified West Greenland Current. The alternation of the modes is strongly linked to changes in the along-shore wind stress outside the Bay. The switch between the modes has major impacts on the temperature transport into the Bay, impacting its heat budget, and likely, the ecosystem of the region.
This is an interesting and well written paper, with generally good quality figures. Given the importance linkage between the circulation in the region and the important icestream at Kangia (Ilulissat Icefjord), there are potentially important larger scale implications related to understanding the circulation of the Bay. Thus, the paper deserves eventual publication in Ocean Science. That said, there are ways the manuscript could be improved, with my big picture, as well as specific, comments, given below.
Major Comments:
The introduction feels a little short and could provide a bit more background, including on the big picture significance of the region, and why understanding its circulation is important. It also feels like all the relevant papers and reports that have examined the oceanography of this region in the past are not mentioned. Providing a more complete background will help put the present work in context for readers. For example, there are some detailed studies of Greenland waters, including of Disko Bay, in a series of DMI reports by Buch and then Ribergaard. There are also so older works in various ICES reports, as well as the J. Northw. Atl. Fish. Sci. There are also some more recent papers by Mortensen, and Rysgaard that look at the oceanography of the region. Especially, Rysgaard et al. (2020) explores the possibility a seasonal southward current in the region, which might be related to the two modes the authors discuss in this paper. Also, in the figure 1 caption, the authors mention Hansen et al. (2012) related to their surface circulation schematic but this paper and its representation of the circulation is not otherwise discussed.
I think it would be good for readers to understand more about the tool that authors use – the GLORYS12 reanalysis, and how it behaves in the study region. Which version of the reanalysis is used? And when were the fields extracted/downloaded? How was it forced and how does that forcing compare with the atmospheric forcing product used elsewhere in this study? How is Greenland discharge represented as that might be relevant for the Disko Bay circulation? Does the version of GLORYS12 used include the NEMO iceberg module? Or is all freshwater converted to liquid and input at the surface? Does the version of GLORYS12 used include tidal forcing? In more technical aspects, how many grid cells are there in the Disko Bay region? What is the bathymetric product used and how robust do the authors feel it is for the study region? What fields are assimilated in the study region? All these are details that will help readers understand the strengths and weaknesses of the product and how they may impact the results presented.
Also, although the authors state “The GLORYS12 product has been shown to perform well in studies focusing on West Greenland”, has it been evaluated in Disko Bay? If not, some evaluation would strengthen the study.
At different points in the paper, the authors present averages over different depth ranges: 0 to 450 m, 0 to 100 m, etc. Justifying these choices would be good. And given that the circulation in the region has multiple watermasses and layers, wouldn’t some additional discussion of the vertical structure of the circulation be relevant. This is hinted at in figure 3, but could be further expanded on. At the very least, the vertical structure of the region should be introduced somewhere, so readers can understand better the results – such as the substantial change in spatial coherence below 130 m.
Finally, at the end of the discussion, I think it would be good to have a larger discussion on the significance of the results. What do they really mean and what might it tell us about the Disko Bay/WGC system going forward. Might we expect to see one more or another more in a future climate? And if so, how might they impact Greenland melt and/or the provision of freshwater to the surrounding ocean.
Specific Comments:
Line 20 – In describing the geographical situation, it would be good to mention the passage leaving the region to the north. It is mentioned later in the paper, but would be good to at least introduce early on.
Line 79 – I don’t believe the resolution of GLORYS12 is ~9 km by 3-4 km in the study region. Where did the authors get those numbers? Given the tri-polar grid that NEMO uses, the scale factors e1 and e2 should be very close in size to each other.
Line 102 – Would be good to also reference Curry et al. (2014) with respect to the Davis Strait array.
Section 2.1.5 – Argo floats aren’t normally deployed on shelves, nor typically have parking depths of ~90 dbar. Was this float then specially deployed for some project? It would be good to add some background/explanation. The authors also mention 2 floats in Disko Bay itself, but that their parking depths were too deep to use in this study – but what about the displacements from those floats when they were at the surface?
Line 149 – Since the zonal and meridional velocities are not co-located on the NEMO grid, where they interpolated to the same point (e.g. a T or f grid point) for the EOF analysis?
Line 151 – The comment about weighting the grid point area confuses me (see by comment above related to Line 79). NEMO/GLORYS uses a tri-polar grid.
Figure 2 – Long term averaging can make values very small in places with high variability. Might some snapshots of the velocity field also be useful. Discrete color bars, like in most of the other figures in the paper, would be good. Making the panels bigger would also help.
Line 312/Line 362 – Would be good to try to link the southward flow to the discussion in Rysgaard et al. (2020).
Line 319 – Not sure if deseasoning is a real word in English.
Figure 9 – Discrete color contouring would be good.