In a recent (of Sept. 16, 2005) publication in Science, Hatun et al. find that record-high salinities have been observed over the past decade in the region where water from the Atlantic flows into the northern oceans. They combine an analysis of observations with simulations using an ocean model, concluding that the salinity of the inflow to the northern oceans is controlled by ocean dynamics and the circulation in the sub-polar gyre. The observations by Hatun et al. may suggest that at the moment the warm and salty waters from the south are especially warm and salty.

In another publication paper in Science from June 17th 2005, on the other hand, Curry & Mauritzen conclude that as a whole the northern North Atlantic has become significantly fresher (less salty) in recent decades. The latter study was based entirely on observations (hydrographic data between Labrador and Europe in the past 50 years). The recent evidence for salinification provided by Hatun et al. has been interpreted by some as being inconsistent with the evidence for high-latitude North Atlantic freshening found in previous reports. So what is really happening? Is the salinity increasing or decreasing? And can the two recent Science studies be consistent with each other?

Northern Europe experiences a mild climate relative to other regions of the same latitude. For instance, Oslo is at approximately the same latitude as the southern tip of Greenland, but has a substantially more mild climate. Though the atmosphere plays an important role here too, the heat transported by the ocean is a key factor responsible for the mild conditions in many parts of Northern Europe. A component of this transport is tied to a current system where salty surface water near the Arctic loses heat, becomes denser, and thus sinks into the deep ocean. This sinking is believed to be part of a large-scale ocean circulation known as the 'global conveyor belt', sometimes referred to as the thermohaline circulation because the circulation is driven by density variations related to variations in temperature and salinity.

Many of the surface currents of the world oceans (i.e., the ocean 'gyres' which appear as rotating horizontal current systems in the upper ocean) are driven by the wind, however, the sinking in the Arctic is related to the buoyancy forcing (effects that change either the temperature or salinity of the water, and hence its buoyancy). The sinking is mainly driven by the saltiness of the water, which is affected by evaporation of fresh water from the surface or, particularly in the Arctic, freezing seawater which leaves salt behind in the water beneath the ice.

Theory and modelling suggest that if the sinking of the salty surface waters in the North Atlantic slowed down or stopped, there would be a reduction in the heat transport by the ocean, which would have implications for the climate of northern Europe. Projections of potential future climates indicate that this may occur in response to increasing greenhouse gas forcing, though the degree of the change is highly variable from one model simulation to another.

Freshening of the ocean can result from numerous factors--the melting of ice, freshwater discharge from rivers, or increased precipitation at high latitudes. The salinity levels of the northern ocean region are also influence by the inflow of warm and salty water from lower latitudes in the Atlantic Ocean. Thus, the salinity of the water is the result of a delicate balance between multiple competing influences.

It is important to keep in mind that different regions of the ocean in the Arctic, often separated by sharp fronts, can have different characteristics and that they may undergo different changes. These oceanic regions are often referred to by different names, but we will try to simplify the discussion here by using the term 'northern oceans' when referring to the basin between Greenland, Iceland, and Norway; Hatun et al. (2005) use 'Arctic Mediterranean' whereas Curry & Mauritzen (2005) use the label 'Nordic Seas'.

The northern ocean region.

There has been increasing evidence over the last 50 years that the waters of the Subpolar and Nordic Seas have indeed become fresher. What Curry & Mauritzen (2005) did was to quantify when, where, how fast, and in what quantities this freshwater entered these seas. It appears from their analysis that the amount of freshwater added in recent decades was much larger than previously assumed. They also estimated how long it would take before the deep return current in the high latitudes would cease if the freshening continued at more or less the same rate, and arrived at an estimate around 100-200 years. This suggests the possibility that a slowing of the conveyor belt in the forseeable future could be a real possibility, and not just a theoretical curiosity.

Hatun et al. agree that large areas in the northern oceans are freshening. Furthermore, Curry & Mauritzen state that the salinity was lowest in the mid-1990s and that the seas have become more salty since then. Both also describe an outstanding event in the late 1960s early 1970s where a region of very low salinity was observed (known as "the Great Salinity Anomaly" - GSA). The GSA can also be seen in Fig. 2 of Hatun et al., and is a well-known feature in Atlantic oceanography. Curry & Mauritzen did find that the freshening trend peaked in the 1990s, an indication that warm and salty water from the south is "tending to counteract" the freshening influence. So there is no inconsistency with the Hatun et al. paper.

But does that mean that the freshening of the subpolar gyre has ceased? No: The northern oceans are significantly fresher than they were in the 1960s - the extra 1.8 meter of freshwater in the Nordic Seas has lost 10 cm, and the extra 3 meters of freshwater in the Subpolar Gyre has, in an upper limit, lost 1 meter. The last timeframe in the Sub-polar Gyre was not published in Curry & Mauritzen's paper because they had very little data from the western Sub-polar Gyre in that period - the volume budget would therefore be biased towards the salty eastern atlantic, where the warm, salty subtropical waters reside. The (upper limit) 1 meter estimate is therefore expected to be reduced when more data from the western Sub-polar Gyre enters the database.

The different focus in the two studies may simply be giving descriptions of the same situation, even while some media reports portrayed the more recent paper as proof against the possibility of freshening of the northern oceans. Curry & Mauritzen analysed the salinity throughout the depth of the ocean. They observe an accumulation of fresh water in the entire column of the sub-polar ocean basins, especially at intermediate depths. They also find indications of higher salinities in the region where the Atlantic waters flow into the northern oceans. Hatun et al. focus on near-surface salinity, as it is near the surface the ocean circulation is strongest. However, part of their analysis also include observations from a transsect penetrating down to 800m at the Rockall Trough. The analysis of Curry & Mauritzen was based on 3-dimensional gridded oceanic observations whereas that of Hatun et al. focused on three inflow points into the Arctic Ocean. Hatun et al. also used altimeter data (local sea level height measurements from satellite observations) to diagnose the norther oceans gyre circulation. The altimeter data provides a measure of the heat and salt content of the water, but because the water density is a complicated non-linear function of temperature and salinity, it is difficult to invert the measurements to infer the salinity. Nevertheless, the local height profiles give indications of the currents that arise from sea level height differences.

Gridding sparse ocean observations onto a very high (in this case, 1-by-1 degree latitude x longitude) resolution is prone to producing some apparent structures that are simply artifacts of mathematical interpolation, even when isopycnal methods are utilised (this is common for gridding of data). On the other hand, the budgeting of salinity implicit in the ocean model used by Hatun et al. may not properly account for river run-off (freshens the water), transport from the Pacific, the Canadian Archipelago, the East Greenland current, or melting processes. The ocean model used by Hatun et al. has a northern latitude limit of 78N, where an artificial boundary is imposed with the salinity, temperatures and velocities all prescribed at that boundary by results from another model. The salt transport at this boundary is not well-known. If the prescribed salt transport is not correct, then the salt budget of the model will not represent reality.

Hatun et al. examined the possibilities that [i] a change in rain falling over the ocean (freshens the water) and evaporation (increases the salinity by removing water and leaving salt behind), [ii] increased salinity in the sub-tropical gyre (in the main part of the North Atlantic), [iii] increased salinity in the sub-polar gyre, or [iv] dynamical changes in the relative contributions from the two gyres could explain the high salinities in the in-flow regions. Of these processes, they concluded that it was the latter that was responsible for the high salinity in the region where Atlantic water flows into the northern oceans. This explanation involves a displacement of an oceanic front and hence a change in the circulation structure. Comparisons with observations in these regions show good agreement between the model and the observations. Thus, although their conclusions about the salinity within the inflow region being closely related to the dynamics of the sub-polar gyration circulation, the model may not give a representative account of the total salt content in the entire northern oceans. But their focus was on the increase in salinity in a certain region, not a general decrease, and hence they did not examine as other factors such as river-run off and melting. Finally, the results by Curry & Mauritzen focused on longer time scales for which they had collected ocean observations, while only model results were provided for long-term evolution by Hatun et al. Even if the model agreed well with respect to saline anomalies in the inflow region, it has not yet been established whether it provides representative values for the absolute salinity in the entire ocean basin.

Given the uncertainties and caveats implicit these two studies, their conclusions regarding trends in salinity of the northern oceans may not be as inconsistent as they might appear. The use of the term "record-high" in the paper by Hatun et al. may be misleading, as this only refers to a limited region since the 1960s (southwest off Iceland) or a very short interval (one decade) and doesn't reflect the general degree of salinity in the entire basin over a longer period. The long-term observations suggest that the last decade has been a mere blip in a long-term trend towards fresher conditions in the northen oceans. 1995 was when the salt content in the norther oceans was at its lowest. It is hard to see how the relationship suggested by Hatun et al. can explain how the 19000 cubic km of fresh water reported by Curry & Mauritzen could be removed. However, these papers are probably not the final word on this.

To conclude: The Subpolar Gyre and the Nordic Seas are probably still a lot fresher than they ever were in the decades before the 1990s (even if the warm and salty Gulf Stream system is now putting up a good fight). If one gets another impression from the Hatun paper it is wrong. Now the question is: what will happen in the future? Will the general freshening trend continue, or will we return to the 1960s levels? None of us have a crystal ball, so no-one really knows. There will always be this battle between the saline waters from the south and the freshwaters from the north, so swings in the time evolution of the overall freshwater loading of the northern seas, as we see right now, should always be expected. It is reasonable to assume that the freshwater input will continue to increase in the future because the earth is warming, causing increasing ice melt and increased precipitation (both over ocean and over land, which yields larger river runoff to the ocean). On the other hand, the subtropical waters can be expected to become saltier in the future, for the same reason (increased hydrological cycle gives more evaporation in the subtropics, thus increased salinities in subtropical waters). The question is which component will win. Neither paper has made quantitative estimates of future scenarios for the freshwater potential associated with the various components (evaporation, precipitation, ice melt).

References:

Hatun H., Sandø A.B, Drange H., Hansen B. & Valdimarsson H. (2005) 'Inlfuence of the Atlantic Subpolar Gyre on the Thermocline Circulation', Science, vol 309, 1841-1844

Curry R. & Maurtizen C. (2005) 'Dilution of the Northern North Atlantic Ocean in Recent Decades', Science, vol 308, 1772-1774