Category Archives: News

By Bill Johns, Chief Scientist

Our cruise is now about halfway through, and spirits are high as we steadily make progress on accomplishing our cruise objectives.  Yesterday the final University of Miami mooring was deployed in the Iceland Basin, completing the turn-around operations on all of the seven moorings deployed along the eastern flank of the Reykjanes Ridge. We have now turned the mooring operations over to the NIOZ team, who at this moment are deploying their westernmost Irminger Sea mooring (see photo), and if all goes well will follow that with the deployment of their second (of five) moorings this afternoon. 

It is a rather nice morning in the Irminger Sea, with relatively calm conditions and the sun even peeking through the clouds from time to time. Apart from a following swell that is occasionally putting some water on the aft deck, it is a fine day for mooring work, seriously challenging our usual mantra that “it is nicer in the Iceland Basin”. Our mooring operations there were mostly in overcast and often cold, drizzly conditions and so the weather today in the Irminger Sea is a pleasant surprise (however, not to be discussed further..).  The only major glitch we have encountered so far is that one of the U. Miami moorings unexpectedly returned to the surface after its deployment, while communicating with the acoustic mooring releases to put them into “sleep mode” for the next two years until we come back again, and that mooring had to be recovered and redeployed again, basically costing us a day of ship time.  This is one of the reasons, besides unpredictable weather, that we build a little extra time into our cruises for contingencies.

The U.S. and Dutch teams continue to work very well together, and this is particularly evident on deck during the mooring operations.  During the first half of the cruise, when the U. Miami mooring work was taking place, the two lead technicians from NIOZ (Leon Wuis and Toon Koopman) ran the mooring winch for the U. Miami team, who were busy loading instruments and floats onto the mooring line as it was paid out.  Now the lead U. Miami technicians (Eduardo Jardim and Cedric Guigand) are doing the same for the Dutch team as they recover and deploy their moorings.  It is always interesting to see how different research groups approach their mooring operations, with somewhat different mooring designs and different methods of deployment. For example, the NIOZ team uses a “mooring table” (pictured in the photos) that has a built-in stopper system for changing mooring wires and attaching instruments, whereas for the U. Miami moorings this is all done at deck level using a traditional double line stopper system.

The mooring winch system we are using is a specially designed system called a LEBUS (aka “double barrel”) winch system (see photo) that is somewhat complicated but highly efficient for performing rapid mooring turnaround operations. It takes two people to run it with a third assisting, and with at four people on the aft deck making mooring connections, and several others bringing instruments in and out from the lab, it requires a large team and a carefully choreographed operation.

With the change-over in mooring operations, the night-time CTD watches have also changed at mid-cruise to free up personnel from the Dutch team for their mooring work.  During the first half of the cruise, these watches were mostly manned by NIOZ scientists but now they are being mostly manned by the U. Miami team.  Everyone is adapting to these changes well and pitching in wherever needed. It has been a great pleasure sailing with this entire team thus far, and – as we often say – “so far so good”, with our fingers crossed.

A simple Python code that computes the meridional overturning circulation (MOC) from a model dataset interpolated onto the OSNAP section. More information can be found on the o-snap website under the tabs Outreach/For Modelers and the code can be downloaded here.

As revealed by the first 21-month record (2014-2016) from the OSNAP array [Lozier et al. 2019], the subpolar overturning occurs primarily in the eastern subpolar region of the North Atlantic Ocean.  This finding contrasts the view of Labrador Sea variability dominating the overturning circulation variability across a range of latitudes via the southward propagation of density anomalies formed by convection in the basin interior.  The small overturning in the Labrador Sea was surprising since there was strong convection in the Labrador Sea during the 2014-2016 winters.  However, given the scale of the expected delay in the overturning response, the 2014-2016 time series may have been too short to pick up this response.  In short, the signals in the Labrador Sea may take a few years to imprint on the OSNAP array.

Continuous observations from the OSNAP array are now available for a 4-year period.  The 4-year time series (2014-2018) of the estimated overturning across the section is shown in the figure below.  Here the strength of the overturning circulation is defined as the maximum of the streamfunction in density coordinates, and is estimated across the full array, OSNAP West (Labrador Sea) and East (eastern subpolar gyre between Greenland and Scotland), separately.  As shown in Lozier et al. [2019], this decomposition allows us to make relative comparisons of the overturning contributions within the subpolar subbasins.  Clearly, there is a consistent pattern throughout the records: the overturning circulation takes place primarily across the OSNAP East section in the eastern subpolar region.  The extended time series proves the validity of the first 21-month record and supports a diminish contribution from the Labrador Sea, over the observational time scales. 

30-day time series of the strength of the meridional overturning circulation (MOC) across the full OSNAP array, and the OSNAP East and West sections, separately. Shading indicates the uncertainty in each 30-day estimate, obtained from Monte Carlo simulations. The 4-year mean MOC is 16.6 Sv# (uncertainty of 0.7 Sv; standard deviation of 3.5 Sv) across the full array, 16.8 Sv (uncertainty of 0.6 Sv; standard deviation of 2.8 Sv) across OSNAP East, and 2.6 Sv (uncertainty of 0.3 Sv; standard deviation of 1.2 Sv) across OSNAP West. # 1 Sv= 106 m3 s-1

Further analysis has been performed to assess the role of Labrador Sea and Irminger Sea convection in creating western boundary density anomalies in the region and the contributions of those boundary anomalies to the subpolar overturning variability [Li et al. 2020].  Results from this analysis shed light on the causes of North Atlantic overturning variability downstream from the subpolar region.

Of course, a question remains as to which subpolar basin(s) holds the key to the overturning variability over interannual and longer time scales.  Needless to say, maintaining high-quality trans-basin observations is critical for tackling this question.

Finally, we are pleased to announce that all of the seven 2020 OSNAP cruises have been carried out successfully. Special thanks to the PIs from the observational groups who worked tirelessly in (re)planning the cruises to accommodate the extraordinary challenges caused by the pandemic.  All cruises had to reduce their working time at the mooring sites because of the need for more transit time to and from their home port.  All observational groups outperformed expectations: everyone was safe and the data return was superb.  We also applaud the captain, the crew and everyone on the science teams!  With all of these efforts, we will be adding another two years of transport estimates in 2021 without significant delays.

References

Lozier, M.S., F. Li, S. Bacon, F. Bahr, A.S. Bower, S.A. Cunningham, M.F. de Jong, L. de Steur, B. deYoung, J. Fischer, S.F. Gary, B.J.W. Greenan, N.P. Holliday, A. Houk, L. Houpert, M.E. Inall, W.E. Johns, H.L. Johnson, C. Johnson, J. Karstensen, G. Koman, I.A. Le Bras, X. Lin, N. Mackay, D.P. Marshall, H. Mercier, M. Oltmanns, R.S. Pickart, A.L. Ramsey, D. Rayner, F. Straneo, V. Thierry, D.J. Torres, R.G. Williams, C. Wilson, J. Yang, I. Yashayaev, J. Zhao, A Sea Change in Our View of Overturning in the Subpolar North Atlantic, Science, 363(6426), 516-521 (2019). 

Li, F., M.S. Lozier, S. Bacon, A. Bower, S.A. Cunningham, M.F. de Jong, B. deYoung, N. Fraser, N. Fried, G. Han, N.P. Holliday, J. Holte, L. Houpert, M.E. Inall, W.E. Johns, S. Jones, C. Johnson, J. Karstensen, I.A. LeBras, P. Lherminier, X. Lin, H. Mercier, M. Oltmanns, A. Pacini, T. Petit, R.S. Pickart, D. Rayner, F. Straneo, V. Thierry, M. Visbeck, I. Yashayaev, C. Zhou, Subpolar North Atlantic Western Boundary Density Anomalies and the Meridional Overturning Circulation, submitted to Nature Communications (2020).

by Amy bower

While the rest of the OSNAP team has been busy preparing the first continuous time series of the AMOC in the high-latitude North Atlantic from the OSNAP array measurements (link to Science paper), the Bower Lab has been steadily processing data from more than 120 deep-sea floats that were released in the deep currents of the AMOC between 2014 and 2017 as part of “OSNAP Floats”. These so-called RAFOS floats (RAFOS stands for RAnging and Fixing Of Sound) were tracked continuously underwater using an array of 13 moored sound beacons spread throughout the subpolar region. Once a day during each float’s mission, the float “listened” for signals from the sound beacons, and internally recorded the time it heard each one (usually only 2-4 beacons were in range on any given day). After two years of collecting these “times-of-arrival” from the sound beacons, each float dropped a ballast weight, popped to the surface and transmitted all its stored data via satellite. We have been reconstructing each float trajectory by converting the acoustic travel times to distances, and triangulating float position for every day. It is always an exciting moment when the complete trajectory pops onto the computer screen, usually revealing some incredibly interesting pattern of the flow deep in the ocean.  

Nearly every diagram of the pathways of the lower limb of the AMOC in the subpolar North Atlantic show continuous lines of flow snaking around the rim of its three sub-basins:  the Labrador, Irminger and Iceland Basins (e.g., Fig. 1). This view of the deep circulation has been shaped largely by our observations of the subpolar ocean around its edges, which almost all show a relatively strong deep boundary current transporting modified versions of the dense overflows (Iceland-Scotland Overflow Water and Denmark Strait Overflow Water) that feed the lower limb of the AMOC. When we track the pathways of individual fluid parcels, or large groups of them, we see a somewhat different view of the pathways of the deep AMOC (Fig. 2). This “Lagrangian” view—by which we mean the ocean circulation as observed by following individual fluid parcels, in contrast to the “Eulerian” view, which is constructed from current measurements at fixed locations—highlights the turbulent nature of the currents in some regions of the deep ocean. The float trajectories show a much messier view because deep ocean currents are not always orderly streams of water flowing smoothly along, but rather consist of blobs or patches of water 10-100 km in diameter swirling around like slow-motion hurricanes.

In some regions, the Lagrangian and Eulerian views look similar, at least on the large scale. For example, most (but not all) of the floats released east of Greenland more or less follow the boundary around the southern tip of Greenland. In other regions, however, the boundary-following pipe-like view seems completely inadequate to describe the deep pathways of the AMOC. For example, floats coming through the Charlie-Gibbs Fracture Zone (a deep gap in the mid-Atlantic ridge that separates the Iceland and Irminger Basins) do not, in general, follow the boundary of the Irminger Basin. Instead, the floats spread westward and even southward in a somewhat chaotic fashion. Also, some floats released in the deep boundary current transporting Iceland-Scotland Overflow Water southward in the Iceland Basin never make it to the Irminger Basin through gaps in the ridge, but instead drift southward within the eastern North Atlantic, apparently having escaped from the deep boundary current system altogether.

So why do we care what pathways these deep waters take? For one practical reason—if we want to accurately monitor the strength of the deep AMOC, we better have a good idea where its major branches are so we can place our instruments judiciously. Also, the pathways of these deep water masses have an impact on how they are modified. A branch that carries deep waters along a slow, meandering interior pathway may lead to more stirring and modification of the water properties than would occur along a more direct, boundary-following path. This in turn affects how heat and other properties are re-distributed throughout the ocean.

The Bower Lab is looking forward to the upcoming General Assembly of the European Geosciences Union in Vienna, Austria next week, where we will have the opportunity to showcase the many and varied discoveries emerging from OSNAP Floats.

Fig 1. Recent diagram illustrating the circulation of the deep currents of the subpolar North Atlantic, From Daniault et al. (2016)

Fig. 2:  Trajectories of 123 RAFOS floats, tracked at depths from 1800 m to 2800 m between 2014 and 2018. Colors indicate individual float tracks.

Daniault, N. et al., 2016. The northern North Atlantic Ocean mean circulation in the early 21^st century. Progress in Oceanography 146, 142-158

by Clark Pennelly

The Labrador Sea in the northwestern Atlantic Ocean is a bit like salad dressing: the multiple water layers present will not happily mingle with others nearby, just like oil and vinegar. Of course, this all changes when the salad bowl that is the Labrador Sea gets stirred up! Strong winter storms pass through the Labrador Sea, cooling the surface water and making it more dense. Should this cooling process continue, the density will increase until the surface layer will have the same density as the subsurface layer, allowing them to efficiently mix together – just like shaking a bottle of dressing. We call this ‘shaking’ deep convection.

However, the Labrador Sea is not simply a 2-ingredient salad dressing mixture. A fresh and cold water cap is at the surface; this water originates from the Arctic and could be from melting of Greenland glaciers, sea-ice from the Beaufort Sea, or even runoff from one of the many rivers which discharge into the Arctic Ocean from Canada, northern Europe, and Russia. A warm and salty layer of water from the Atlantic Ocean exists below this. Further below is yet another water mass which is cold and salty, a product of mixing the two layers above it during the cold winter; we call this layer Labrador Sea Water. There are even more layers beneath, but I won’t get into them for now. These top three layers are important to the deep convection and water mass formation that previous blog posts have discussed.

My PhD thesis revolves around the ocean layers within the Labrador Sea and how they are modified by a variety of sources, such as an input of freshwater, heat, or perhaps changes in the weather. As this region is notorious for its rough seas, particularly during the stormy winter, research cruises (see https://www.o-snap.org/fleur-de-sel-life/ ) tend to occur/gather data after the winter has passed when seas are calmer and the 12-point Beaufort sea state scale (see https://en.wikipedia.org/wiki/Beaufort_scale) goes from 10’s and 11’s down to more manageable single digits. Unfortunately, research cruises may therefore miss clues about the mixing and formation of water layers that happen during the convective season. I use computer models to investigate what may be occurring when research cruises are not around to gather data. While they are not without their own set of problems, models allow us a peek into the unknown from the comfort of a Beaufort 0 sea-state office.

One portion of my research explores rotating circular features, which we call eddies, that form off the west coast of Greenland. These eddies are produced via changes in the seafloor nearby. As the West Greenland Current moves alongside the west coast of Greenland, it carries both fresh and cold water at the surface as well as salty and warm water below. These eddies carry both water layers away from the coast and into the heart of the Labrador Sea. Both of these water masses impact the Labrador Sea in a different way, so these eddies are rather interesting and important features. However, these small events are notoriously difficult to represent within numerical models, especially multi-decade simulations. We had to get a little creative in order to make sure we would resolve these eddies as modelling such eddies requires a lot of computing resources. We set up a complex simulation which has a resolution of 1/60 of a degree, meaning this simulation should be able to resolve features where are larger than a few kilometers. The simulation requires over 600 CPUs and about 72 hours of real time to produce 40 days of simulation time. To put that into perspective, our regular simulations may use 60 CPUs and finish 2 years of simulation time in 72 real hours; quite the difference! We are interested in running this expensive simulation across 17 years, from 2002 to 2018, so we can capture variability across many years, though some quick math says that this simulation will take about a year to finish. But what do we gain by resolving the Labrador Sea at high resolution?

Well-defined eddies is what we get! You can see those individual eddies within the red circle. From running this model we’ve so far learned that those eddies can last up to two years, meaning that as they travel southwest from the coast of Greenland they may encounter multiple winters and therefore be involved in multiple rounds of deep convection. I’m currently exploring how these eddies evolve through their lifetime in the Labrador Sea. We know they bring substantial heating into the region, limiting deep-water formation. Thus being able to detect them with our model is crucial in order to achieve an accurate spatial representation of convection. Lower resolution simulations may fail to resolve these features and thus their importance. Since they persist for multiple convective years, they may have a more complicated story than previously thought which I hope my research will help uncover. Numerical modelling allows for the scrutiny of such events as observations within the Labrador Sea may not provide the full information we seek.

by Helen Johnson

with input from Sam Cornish and Yavor Kostov

The first results from the OSNAP array, published this month in Science (https://science.sciencemag.org/content/363/6426/516), are incredibly exciting! Over the 21 months of data collected so far, it has been the conversion of warm, salty, shallow Atlantic water into colder, fresher, deep water east of Greenland that has dominated the overturning circulation and its variability.  This challenges the prevailing view that deep water formation in the Labrador Sea to the west of Greenland is the major player in determining overturning variability.  And it gives us some clues about how changes further north, in the Arctic Ocean, might affect things.

A major motivation for measuring the overturning circulation with the OSNAP and other observational arrays arises from the expectation that the overturning will change as a result of human-induced climate change. The overturning circulation is predicted to weaken over the coming century, due to a warming and freshening of the high latitude North Atlantic. Part of the freshening expected in these critical deep water formation regions is due to changes in the amount of freshwater exported from the Arctic to the Atlantic, on both sides of Greenland. The Arctic Ocean has recently accumulated a large amount of freshwater, but we do not know if or when it will be exported to the Atlantic, or at what rate.

My group in Oxford have been investigating changes in the amount of freshwater stored in the Arctic, and the reasons for them, in the hope that this will teach us something about the changes we might expect in freshwater export (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2017GL076870). We have deduced the relationship between winds over the Arctic and total Arctic freshwater storage in a climate model. We now know that if the winds over the Arctic change, it takes the system at least a decade to come into a new equilibrium, with a different amount of freshwater stored.  Based on the relationship we deduce between winds and freshwater content, we estimate changes in Arctic freshwater content over the last century – and our time series agrees well with the limited observational data available (see figure), giving us confidence that the relationship is a useful description of the real world!

Our results suggest that the large increase in Arctic freshwater content since 1992 can largely be explained by historical changes in the winds driving the Arctic Ocean circulation (although we can’t rule out a smaller contribution from sea-ice melt).  What’s more, this increase doesn’t seem exceptional compared to variability in our time series over the rest of the century, suggesting that it may simply be natural variability.

Of course, changes in freshwater exported from the Arctic, natural or not, still have the potential to affect the overturning circulation! The long memory of atmospheric conditions revealed by our results is important because, provided we know what the winds have done, we can potentially predict changes in Arctic freshwater content a few years ahead. Based on our work so far, we expect Arctic freshwater content to decrease over the coming decade, and freshwater export to the Atlantic to increase.

The OSNAP array is ideally-placed to detect any impact of this change on the high-latitude overturning circulation. Based on the OSNAP results so far, we might expect that Arctic freshwater export through Fram Strait to the Nordic Seas (rather than through the Canadian Archipelago to the Labrador Sea) will have the most impact. So we have several hypotheses that our continued OSNAP observations and modelling efforts will put to the test in the coming years!