Earth science

Spread thin in the Arctic

The northernmost segment of the global mid-ocean-ridge system, the Gakkel ridge, is a bit of an oddity. Provocative data are emerging from an international geophysical and geochemical study of the region.

In 1893, Fridtjof Nansen and members of his expedition set out to conduct a remarkable scientific study in the Arctic. Their ship, the Fram, was designed to become encased in pack ice and to move with it, and over a three-year period the Fram's stately motion demonstrated the drift of the polar ocean current. Just over a century later, a fresh chapter in the scientific study of the northern high latitudes opened with a two-ship, international expedition to study the ocean floor beneath the Arctic Ocean. Some results from this study — AMORE, for Arctic Mid-Ocean Ridge Expedition — appeared in Nature early this year1, and others are described by Michael et al.2 and Jokat et al.3 on pages 956 and 962 of this issue. The Arctic Ocean is covered by sea ice, so the studies were carried out on scientifically equipped icebreakers, the US Coast Guard Cutter Healy and the German PFS Polarstern.

At mid-ocean ridges, partial melting of upwelling mantle from deep within the Earth produces magmas that erupt and cool to become basalt lavas. This magmatic activity creates new ocean crust that spreads on either side of the ridge, at different rates depending on the location. Ridges may also be sites of intense hydrothermal activity, as sea water penetrates the crust, becomes heated by its proximity to hot magma at depth, and re-emerges into the water column through vents as element-rich fluids.

The ridge beneath the Arctic Ocean is the 1,800-km-long Gakkel ridge (Fig. 1), and it has some unique characteristics. Given these characteristics, studies of other parts of the global mid-ocean-ridge system allowed certain predictions to be made about what would be found there. For instance, almost three decades ago aeromagnetic surveys had shown that the Gakkel ridge is spreading at the slowest rate of any mid-ocean ridge4, with a 'full rate' (that is, accounting for motion on either side) of about 0.3–1.0 cm per year in the eastern part. That compares with about 6 cm per year for 'intermediate-spreading' ridges elsewhere. Spreading rate had been shown to correlate in a general way with a host of ridge characteristics, such as the balance between magmatism and tectonic faulting, the composition of the lavas, the architecture of the ocean crust, and the nature and distribution of hydrothermal activity.

Figure 1: On top of the world — the Gakkel ridge, subject of the detailed surveys carried out from the Healy and Polarstern.


The zones delineated by Michael et al.2 and Jokat et al.3 are a western volcanic zone (WVZ); a central, sparsely magmatic zone (SMZ); and an eastern volcanic zone (EVZ). Only the western 1,000 km of the 1,800-km ridge were accessible for study. Brown is land; light green, continental shelf; blue, deep ocean.

From these correlations, it was predicted that the Gakkel ridge would have generally 'anaemic' magmatism that would progressively decrease in volume as the spreading rate decreased towards the east, where there is a thicker, conductively cooled cap on mantle upwelling and melting5. In addition, the extrapolation from fast-spreading to slow-spreading rates suggested that there would be little or no hydrothermal activity along ultraslow-spreading sections such as the Gakkel ridge, because of decreased magmatism and heat (which together drive hydrothermal circulation).

Surprisingly, both of these predictions were found to be incorrect. Based on their spectacular bathymetric and tectonic maps of the western 1,000 km of the ridge, as well as the amounts of lava recovered by dredging, Michael et al.2 show that the eruption of magma is surprisingly robust in the western zone, only sparse in the central zone, and robust again in the eastern zone (Fig. 1), where the spreading rate is slowest. Furthermore, hydrothermal plume activity turns out to be among the most vigorous of any ocean ridge yet studied. The authors speculate that this is a consequence of the highly focused nature of the magmatism there, which leads to long-lived, discrete volcanic structures, combined with episodes of crustal extension and faulting. These are the two factors required for hydrothermal activity: faulting provides the pathway for sea water to penetrate into the sea floor, and magmatism provides the heat to drive fluid circulation.

The Gakkel ridge lies about 5 km beneath the ocean surface, which is unusually deep — about twice the average — for ocean ridges worldwide. The implication is that the crust here is unusually thin6, possibly as a result of a combination of cool mantle temperatures and ultraslow spreading. Using the methods of seismic refraction, Jokat et al.3 show decisively that the crust here is indeed exceedingly thin. In fact, dredging in the deepest parts of the central zone of the ridge recovered an abundance of rocks that are typical of the mantle, but few or no basalts — that is, basaltic crust that usually overlies mantle seems to be completely absent in some places.

Finally, the new results2,3 provide essential data for examining the relationship between the chemical composition of basalt lavas and the crustal thickness of the ridge from which they were collected. Figure 2 shows the inverse correlation between regional averages of crustal thickness and the 'normalized' sodium contents of lavas recovered from ridges worldwide7,8. This correlation may be explained by assuming that regional differences in the extent of mantle melting produce corresponding variations in basalt composition. Sodium is most abundant in magmas produced by small amounts of melting. And if less melt is produced, the crust created by that melt will be thinner.

Figure 2: Correlation of the chemical composition of ocean-ridge basalt lavas with the crustal thickness of the ridge.

Shown here are regional averages of the crustal thicknesses associated with various locations on mid-ocean ridges around the world, as inferred from seismic data, plotted against a measure of the sodium content of the lavas (Na8.0, which is Na2O normalized to 8% MgO)7. The new measurements2,3 from the Gakkel ridge — from the eastern volcanic zone (EVZ) and the sparsely magmatic zone (SMZ) — extend this correlation to the very thinnest crust yet identified. Determinations of crustal thickness are from ref. 7, augmented or superseded by data from many sources; chemical analyses are primarily from ref. 9.

The extreme position of the Gakkel ridge on Fig. 2 suggests that it does indeed represent a global 'end-member' in terms of its small extent of melting. Although most researchers agree that variations in mantle temperature lead to differences in the extent of melting from region to region, some have argued that spreading rate may play a significant or even dominant role at ridges that spread at ultraslow rates5. Michael et al. argue that if spreading rate were the primary factor governing the extent of melting, then the normalized sodium in the lavas (Na8.0)7 should increase, and magmatic output should decrease, on moving from the western volcanic zone, through the central zone, into the eastern zone as spreading rate decreases. In fact, although the western zone is lowest in Na8.0, consistent with greatest extents of melting, the high Na8.0 content of the central zone is equivalent to that of the more eastern zone. Here, then, is evidence against a simple relationship between spreading rate and magmatism.

These are just some of the wealth of results to emerge from studies of the Gakkel ridge. Perhaps the most important finding is that scientific preconceptions continue to be challenged with each new project in Earth science, underscoring the necessity of continued exploration of the structure, chemistry and behaviour of our planet.


  1. 1

    Edmonds, H. N. et al. Nature 421, 252–256 (2003).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Michael, P. J. et al. Nature 423, 956–961 (2003).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Jokat, W. et al. Nature 423, 962–965 (2003).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Vogt, P. R., Taylor, P. T., Kovacs, L. C. & Johnson, G. L. J. Geophys. Res. 84, 1071–1089 (1979).

    ADS  Article  Google Scholar 

  5. 5

    Reid, I. & Jackson, H. R. Mar. Geophys. Res. 5, 165–172 (1981).

    Google Scholar 

  6. 6

    Coakley, B. J. & Cochran, J. R. Earth Planet. Sci. Lett. 162, 81–95 (1998).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Klein, E. M. & Langmuir, C. H. J. Geophys. Res. 92, 8089–8115 (1987).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Langmuir, C. H., Klein, E. M. & Plank, T. in Mantle Flow and Melt Generation at Mid-Ocean Ridges: AGU Monograph 71 (eds Phipps Morgan, J., Blackman, D. K. & Sinton, J. M.) 183–280 (Am. Geophys. Un., Washington DC, 1992).

    Google Scholar 

  9. 9

    Lehnert, K., Su, Y., Langmuir, C. H., Sarbas, B. & Nohl, U. Geochem. Geophys. Geosyst. 1, 1999GC000026.2000 (2000).

Download references

Author information



Corresponding author

Correspondence to Emily M. Klein.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Klein, E. Spread thin in the Arctic. Nature 423, 932–933 (2003).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing