Oceanography: Arctic freshwater - Nature

news & views age of oceanic crust forming what is known as the Cretaceous quiet zone, rates of plate spreading are difficult to assess. Granot et al.1 report that the quiet zone is not so quiet after all. They employed deep-tow magnetometer measurements, taken only about 1,000 m above the sea floor, to obtain a high-resolution magnetic record from the oceanic crust created in the central Atlantic Ocean during the Cretaceous Normal Superchron. Rather than the flat, monotonous signal that might be expected, their magnetometer data recorded variations that might reflect changes in the intensity of the geomagnetic field throughout the superchron. Ruling out a tectonic or magnetic origin for the signal, Granot et al.1 propose that the behaviour of the geomagnetic field evolved throughout the superchron. In their scenario, the beginning of the superchron was marked by stability in the intensity of the geomagnetic field. Variability increased gradually, reaching a peak in the middle of the superchron, between 110 and 100 million years ago, and returned to stability by the end. The authors also isolated a few prominent intensity variations that may be useful as time markers within the Cretaceous quiet zone, which should aid estimates of seafloor spreading during this time. The proposed pattern of geomagnetic field behaviour doesn’t provide an easy answer to the question of why

the Cretaceous Normal Superchron started, and why it ended. The authors’ interpretations also challenge results of numerical simulations of the geodynamo, which so far have predicted lower field variation during superchrons2. These models also predict a higher geomagnetic field strength during superchrons, and such behaviour is supported by detailed palaeointensity analyses using single silicate minerals3 and basaltic glass4, which are more reliable than bulk samples of igneous rocks5. But testing field variability of the kind proposed by Granot et al. is beyond the current temporal resolution of these specimen-based experimental data sets. The discrepancy between the interpretations of Granot et al.1 and current simulations could be explained by the models’ lack of changes in core/mantle boundary conditions on a timescale of ten million years. However, the discrepancy may be more fundamental: numerical simulations predict that superchrons are associated with low heat flow across the core/mantle boundary. However, geologic evidence suggests the opposite may be the case, at least for the Cretaceous Normal Superchron. As well as the presence of a magnetically quiet zone, mid-Cretaceous times were extraordinary because of the eruption of unusually large volumes of magma on land, in the form of flood basalts, and across the sea floor. The seafloor eruptions led to the emplacement

of numerous oceanic plateaux (Fig. 1), most notably the giant Ontong Java Plateau6. These extreme volcanic outpourings are often linked to mantle plumes, and it is thought that plumes can be responsible for significant heat transfer across the core/ mantle boundary 7. The time markers and geomagnetic field variability reported by Granot et al.1 should be recognizable in deep-tow records from elsewhere in the Atlantic and other ocean basins. If confirmed, this would force us to revisit the question of what core and mantle conditions lead to superchrons. ❐ John A. Tarduno is in the Department of Earth & Environmental Sciences and the Department of Physics & Astronomy, University of Rochester, Rochester, New York 14627, USA. e-mail: [email protected] References 1. Granot, R., Dyment, J. & Gallet, Y. Nature Geosci. 5, 220–223 (2012). 2. Driscoll, P. & Olson, P. Geophys. Res. Lett. 38, L09304 (2011). 3. Tarduno, J. A., Cottrell, R. D. & Smirnov, A. V. Proc. Natl Acad. Sci. USA 99, 14020–14025 (2002). 4. Tauxe, L. & Staudigel, H. Geochem. Geophys. Geosyst. 5, Q02H06 (2004). 5. Aubert, J., Tarduno, J. A. & Johnson, C. L. Space Sci. Rev. 155, 337–370 (2010). 6. Tarduno, J. et al. Science 254, 399–403 (1991). 7. Schuberth, B. S. A., Bunge, H.-P., Steinle-Neumann, G., Moder, C. & Oeser, J. Geochem. Geophys. Geosyst. 10, Q01W01 (2009). 8. Ogg, J. & Smith, A. G. in A Geologic Time Scale (eds Gradstein, F., Ogg, J. & Smith, A.) 63–86 (Cambridge Univ. Press, 2004). 9. Müller, R. D., Sdrolias, M., Gaina, C. & Roest, W. R. Geochem. Geophys. Geosyst. 9, Q04006 (2008). 10. Coffin, M. F. & Eldholm, E. Rev. Geophys. 32, 1–36 (1994).

OCEANOGRAPHY

Arctic freshwater

The Arctic Ocean has become less saline, perhaps in response to climate change. Satellite and in situ observations reveal changes in the regional wind patterns that have re-routed freshwater and prevented it from leaving the Arctic Ocean in the past decades.

Cecilie Mauritzen

T

he hydrological cycle, which transports freshwater to almost every corner of the world, is an endless loop. It encompasses the atmosphere, mountain tops and glaciers, river valleys and lakes, groundwater and the ocean itself. In the ocean, the freshwater cycle leaves its signature in the distribution of salinity: surface salinity is high in regions where evaporation dominates rainfall, typically in the subtropics. On the other hand, low surface salinity characterizes regions where precipitation exceeds evaporation, such as the tropics and the subpolar and 162

polar regions. In a changing climate, both evaporation and precipitation are projected to intensify. Indeed, over the past decades a decline in salinity has been observed in the — already relatively fresh — western Arctic Ocean. Two studies, published in Nature1 and Nature Geoscience2, explore where the freshwater came from, and how it was captured in the Canadian section of the Arctic Ocean. As the Earth’s climate changes and the atmosphere becomes warmer, the air is also likely to become moister. In fact, for every 1 °C warming the atmosphere can hold

another 7% of water vapour. Therefore it is commonly thought that the hydrological cycle will strengthen in a warming climate. So far, the most convincing evidence for this effect is found in the ocean, where the regions that are typically saline have become even more so, and vice versa. This suggests that both evaporation and precipitation have increased. In the absence of direct measurements of these processes over most of the world’s ocean, these observations of ocean salinity are our clearest evidence. The Arctic Ocean, where precipitation dominates evaporation, is expected to

NATURE GEOSCIENCE | VOL 5 | MARCH 2012 | www.nature.com/naturegeoscience

© 2012 Macmillan Publishers Limited. All rights reserved

news & views receive far more precipitation in a warmer climate. Also, rivers flow into the Arctic Ocean from the Canadian and Russian landmasses, with large drainage areas that also are expected to receive increased precipitation. Indeed, observations indicate that Russian river discharge has increased by 20% over the past 20 years3. So it is logical to expect evidence of the changing freshwater cycle in the Arctic Ocean as well. It turns out that the western Arctic, and in particular the Canadian Basin, has seen an increase in freshwater content in the 2000s compared with earlier decades4. However, the Arctic, like the North Atlantic, is exposed to enormous decadal variability 5 that makes it hard to extract a long-term trend in the freshwater cycle. It is therefore important to decipher the causes of the observed salinity changes in the Arctic. The studies by Morison and colleagues1 and by Giles and colleagues2 aim to do just that, using satellite measurements of sea surface elevations. Both teams use the fact that at high latitudes, where the ocean is very cold, the density of the ocean is determined much more by salinity than by temperature. That means that fresher water, almost regardless of its temperature, will tend to be more buoyant than saline water (by contrast, at mid-latitudes the warmest water tends to be most buoyant). The teams use this to find an approximate relationship between changes in sea surface elevation and freshwater content. Surface-elevation changes, together with temperature and salinity measurements of the ocean, are also used to calculate changes in ocean velocity. Giles and co-workers2 set out to understand the role of clockwise circulating winds over the Canadian basin in the accumulation of freshwater. In the annual mean, there is a prominent region of high sea-level pressure centred in the Beaufort Sea of the Canadian Basin where the largest salinity decrease is found. The clockwise surface wind field associated with the Beaufort Sea High causes a clockwise circulation in the ocean, and it had been hypothesized that stronger clockwise winds had caused the accumulation of freshwater 6. Giles and colleagues find that the wind patterns have indeed changed towards stronger clockwise rotation over the 15-year period they study. In response, the ocean gyre circulation has strengthened, and the freshwater content has continuously increased. They note that this relationship is not a linear one: the trend in the wind field has been nearly constant since 1995, whereas the freshwater increase only starts in 2002. As Giles and colleagues point out, other mechanisms

a

b Beaufort Sea Canada

Siberia

Canada Basin Eurasian Basin

Greenland

Norway

Beaufort Sea Canada

Siberia

Canada Basin Eurasian Basin

Greenland

Norway

Figure 1 | Arctic circulation changes. a, During the negative phase of the Arctic Oscillation, river runoff from Siberia is quickly swept out of the Arctic Ocean towards the Atlantic, as outlined by Morison and colleagues1. This phase is, however, typically associated with a stronger atmospheric high pressure system in the Canadian Basin, the Beaufort High, leading to a faster-rotating ocean gyre (circle) that draws freshwater more efficiently into the Beaufort Sea, as discussed by Giles and colleagues2. b, Conversely, a high Arctic Oscillation Index brings freshwater from Siberia towards the Canadian Basin, but the accumulation in the Beaufort Sea is less efficient, because the atmospheric Beaufort High is typically weaker, leading to a slower rotation of the ocean gyre. By a combination of these processes, freshwater can build up in the Beaufort Sea for many years, only to be suddenly released if an abrupt and persistent drop in the Beaufort High should follow. Together, the studies aim to explain the accumulation of freshwater in the Canadian Basin of the Arctic Ocean in the past decade or so.

must have contributed to the change in freshwater content, but they also note that the winds are probably becoming more efficient at spinning up the ocean in the 2000s due to the thinner and less extensive ice cover. Morison and co-workers1 have looked for other mechanisms than Beaufort Gyre strength to explain the freshwater increase from 2005 to 2008 in the Canadian Basin. They find evidence in satellite altimetry, hydrography and hydrochemical properties that the Russian river outflows have been steered in a more eastward, alongshore, direction towards the Canadian Basin. Furthermore, they infer that this alteration in freshwater pathways was forced by a strengthening of the west-toeast hemispheric atmospheric circulation that is associated with low atmospheric pressure over the pole and relatively high pressure in mid-latitudes, termed a positive Arctic Oscillation index. They find that the accumulation of freshwater in the Canadian Basin between 2005 and 2008 is primarily due to this re-routing of Eurasian river runoff. The two studies seem to be complementary: the changes in freshwater content cannot be fully explained by either the Beaufort High2 (see also Fig. 2 of ref. 2)



or by the Arctic Oscillation (see Fig. 3a of ref. 1). Giles and co-workers find that although the centre of the Beaufort Gyre (defined by the maximum sea surface height) remained in place between 1995 and 2010, the dome — and therefore the region of freshwater accumulation — has extended toward Eurasia. The supply route of the Eurasian rivers identified by Morison and co-workers for the period 2005–2008 yields an increased supply of freshwater to the gyre that can explain this eastward extension, as well as where much of the extra freshwater is coming from. The Beaufort High and the Arctic Oscillation are not independent entities, and yet they are not identical8. In fact, the Beaufort High projects onto several atmospheric modes. A strong Beaufort High is typically part of the atmospheric pressure pattern associated with a low Arctic Oscillation index, which would direct river runoff towards the Atlantic. Conversely, a weak Beaufort High is often associated with a high Arctic Oscillation index and steers river runoff towards the Canada Basin (Fig. 1). It is reasonable to assume, therefore, that the source of the freshwater – ice melt, river runoff or other – and the rate of accumulation into the Canadian Basin will depend on the strength 163

news & views of the Arctic Oscillation pressure pattern, whereas the strength of the Beaufort High is responsible for the accumulation of freshwater into the Beaufort Gyre proper. The findings are potentially important for the global ocean circulation. When the clockwise Beaufort wind fields weaken, the excess freshwater can be drained rapidly from the Arctic. As both studies point out, a rapid inflow of freshwater from the Arctic to the North Atlantic, via the East Greenland Current, could potentially affect the overturning circulation and thus the transport of warm waters towards the Arctic. The amount of freshwater currently stored in the Arctic is comparable to the amount 9 released in the North Atlantic in the late 1960s and early 1970s, which apparently did not cause significant changes in the ocean circulation. But if the accumulation of freshwater in the Beaufort Sea continues for many years yet before

collapsing, a significant change in the North Atlantic is possible. Neither study discusses the changes in the climatic boundary conditions in much detail. Precipitation over the Arctic region — oceans as well as land-based river drainage basins — has risen over the past decades. At the same time, the salinity of the inflowing warm Atlantic waters has gone up during the past decade7. So when Morison and co-workers, for instance, found that over the entire Arctic Ocean, upper ocean freshwater content has changed very little between 2005 and 2008, it is probably a result of these two opposite effects cancelling each other out. The studies by Morison et al.1 and Giles et al.2 identify atmospheric processes that have led to a remarkable re-routing and accumulation of low-salinity water in the Arctic Ocean. Nevertheless, many mysteries remain before we can fully

explain and attribute the changes in ocean salinity in the Arctic, an environment of high interannual and decadal variability. To resolve these questions, long-term basin-wide monitoring will be of utmost importance. ❐ Cecilie Mauritzen is at the R&D Division, Norwegian Meteorological Institute, PO Box 43, Blindern, 0313 Oslo, Norway. e-mail: [email protected] References 1. Morison, J. et al. Nature 481, 66–70 (2012). 2. Giles, K. A., Laxon, S. W., Ridout, A. L., Wingham, D. J. & Bacon, S. Nature Geosci. 5, 194–197 (2012). 3. Shiklomanov, A. I. & Lammers, R. B. Environ. Res. Lett. 4, 045015 (2009). 4. Rabe, B. et al. Deep-Sea Res. I 58, 173–185 (2011). 5. Hurrell, J. W., Kushnir, Y. & Visbeck, M. Science 291, 603–605 (2001). 6. Proshutinsky, A., Bourke, R. H. & McLaughlin, F. A. Geophys. Res. Lett. 29, 2100 (2002). 7. Holliday, N. P. et al. Geophys. Res. Lett. 35, L03614 (2008). 8. Serreze, M. C. & Barrett, A. P. J. Climate 24, 159–182 (2011). 9. Curry, R. & Mauritzen, C. Science 308 1772–1774 (2005).

GEOMORPHOLOGY

Flow and form

Dune fields often exhibit complex patterns of vegetation and morphology over relatively short distances. An analysis of the White Sands dune field in New Mexico attributes the shift in dune form to the development of an internal boundary layer over the rough dune-field surface.

Keld R. Rasmussen

A

eolian dunes form in areas where sand is abundant, and wind flow is sufficient to mobilize and transport it. On Earth, a confluence of these two factors typically occurs in some coastal regions and in desert basins. Dune fields, particularly those in deserts, can extend from tens to hundreds of kilometres and form over long time-spans that often cover substantial changes in regional climate. It is therefore unclear whether the shape of dune fields that we observe today is the result of current environmental conditions1. Writing in Nature Geoscience, Jerolmack et al.2 show that the complexity of the dune field at White Sands, New Mexico, can be explained by modern processes, specifically the development of an internal boundary and its effects on sand redistribution. The moderately sized dune field in White Sands National Monument covers about 350 km2 (Fig. 1). Winds come from the southwest, and the field is covered by two types of dunes: barchanoid ridges and parabolic dunes. Both forms are typical of environments with airflow from a single, prevailing direction. Upwind of the dunes 164

is the White Sands Alkali Flat, a vegetationfree plain stripped of all loose sand above the shallow water-table. The transition from flat to dune field is marked by the presence of a dune ridge. The ridge marks the start of a series of barchanoid ridges, which continues for about 7 km. The subsequent transition to parabolic dunes occurs over a short distance. The dune field abruptly terminates in a plain marked by herbs and shrubs. The field essentially has a closed sediment budget, with the downwind progression of dunes leaving behind a sand-barren addition to the Alkali Flat. Airflow is driven by regional pressure gradients, and it is the pattern of irregularities on the underlying surface — known as roughness — that controls the friction exerted on the surface by that flow. In dune fields, the transition from the relatively flat plain to the roughness of the sand hills causes air speed near the surface to drop. The decline in speed will slowly move upwards through the air column as the flow progresses across the field, triggering the development of an internal boundary layer 3.

Jerolmack et al.2 hypothesize that it is the presence of this internal boundary layer that drives the morphology of the White Sands field. In their scenario, the internal boundary layer generated by the transition from the Alkali Flat to the ridges causes the energy available for sand transport to decline gradually as the air flows downstream. Based on the well-established fluid dynamics of internal boundary layers4, Jerolmack et al. created an analytical expression that describes the atmospheric dynamics over the dune field. Their calculations suggest that bed stress declines with distance from the upwind border of the field. They then used data from meteorological stations near the White Sands to calibrate atmospheric parameters for this setting. Once they determined bed shear stress, it was straightforward5 to predict the variability of sand transport across the dune field. The study relies on a number of simplifications, but nevertheless the predicted sand transport compares well to measurements of transport rate derived

