Earth science

Through thick and thin

The sea floor around the Hawaiian island chain is unusually shallow. New seismic evidence suggests that this up-raised ‘swell’ is partly due to heating and thinning of the lithosphere beneath.

Earth scientists revere the Hawaiian islands as the inspiration for some of the most fundamental concepts of their discipline. More than 40 years ago, J. T. Wilson suggested that the linear increase in the age of the islands along the chain reflected the steady motion of the Pacific plate over a source of heat, or hotspot, fixed in the mantle beneath1. W. Jason Morgan subsequently proposed that hotspots were in fact the tops of mantle plumes — rising currents of hot buoyant material originating deep in the mantle2. Meanwhile, shipboard surveys of the Pacific ocean floor had mapped out the Hawaiian ‘swell’, a large area of anomalously shallow sea floor that surrounds the Hawaiian chain (Fig. 1). What created the Hawaiian swell? The fact that it doesn't simply collapse under its own weight indicates that there must be an upward force that counteracts the pull of gravity. On page 827 of this issue, Xueqing Li et al.3 weigh in with new data and fresh insight on the origin of the swell-supporting force.

Figure 1: The Hawaiian swell.

The sea floor surrounding the chain of islands is anomalously shallow, relative to normal sea floor of the same age, over an area about 1,200 km wide and 3,000 km long. The maximum height of the elevated region (excluding the islands themselves) is about 1,400 m.

There are two principal hypotheses. The first is lithospheric thinning4, or ‘rejuvenation’ (Fig. 2a). Here attention focuses on the lithosphere, the cold and mechanically stiff outermost layer of the Earth that includes the crust and part of the upper mantle. Under oceans, lithosphere a few kilometres thick is continuously created at mid-ocean ridges, and gradually thickens as it moves away and cools in contact with the ocean. The lithosphere under the Big Island of Hawaii, for example, has been cooling for some 90 million years, making it about 100 km thick. According to the rejuvenation model, however, normal oceanic lithosphere that happens to move over a hotspot is thinned, or rejuvenated, during the time it remains over the hotspot — about five million years in the case of Hawaii, based on the distance over which the initial uplift of the swell occurs. The lithosphere then slowly thickens again by conductive cooling as it moves away from the hotspot. The result is an elongated trough at the base of the lithosphere, filled with rock that is hotter and less dense than the colder normal lithosphere to either side. By Archimedes' principle, this lighter material pushes the thinned lithosphere upwards to create a swell.

Figure 2: The origin of the Hawaiian swell.

Two reasonably successful but incomplete models have been proposed. a, The first is the ‘rejuvenation’ model, in which the swell is compensated by buoyant material in a trough at the base of the lithosphere. The trough is created when heat from a fixed hotspot thins the moving lithosphere. b, According to the ‘dynamic support’ model, the swell is supported by hot material from an ascending mantle plume, which spreads below the base of the lithosphere in the form of a broad pancake. c, Li et al.'s3 results support a hybrid of the two: the ‘dynamic thinning’ model. The swell as a whole is compensated dynamically, but convection currents within the hottest central part of the pancake cause the lithosphere to thin over an area about 300 km wide, starting about 100 km downstream from the vertical plume conduit.

The rejuvenation model explains qualitatively the rapid uplift and the subsequent slow decay of the Hawaiian swell moving along the island chain to the northwest. It is also consistent with the observed anomalies in the gravitational potential over the swell, which imply that the compensating upward force originates at shallow depths. But the model does not propose a credible mechanism by which a hotspot can thin the fast-moving lithosphere by several tens of kilometres in a few million years.

In contrast to the purely static rejuvenation hypothesis, the second, ‘dynamic support’ model proposes that the swell is held up by the flow associated with an ascending mantle plume. On reaching the base of the lithosphere, the buoyant plume material spreads laterally, forming a broad ‘pancake’ which pushes the lithosphere upwards (Fig. 2b). A simple physical model for the gravitational spreading of viscous fluid poured onto a moving plate5,6 predicts that the pancake will have a broad quasi-parabolic shape that agrees well with the shape of the Hawaiian swell. The model also predicts quantitatively the swell's rapid uplift and subsequent slow decay. But because the compensating density anomalies are below the base of normal lithosphere, the model has trouble accounting for the observed gravitational-potential anomalies.

Which of these models is correct? Li et al.3 propose that the answer might be both. They have used variations in the travel times of seismic waves generated by earthquakes to create a topographic map of the base of the lithosphere beneath the Hawaiian swell. The waves used are of two types: compressional (P) waves, with particle motion along the direction of propagation (like sound waves in air); and shear (S) waves with transverse particle motion. Li et al. start from the well-known principle that, when a seismic wave of a given type (P or S) arrives obliquely at a boundary, some of its energy will be converted into a wave of the other type. The time required for this new wave (the converted phase) to reach a seismometer at the Earth's surface is a measure of the depth of the boundary at the ‘piercing point’ where the conversion took place. Using S-to-P converted waves, Li et al. have shown that the thickness of the lithosphere decreases from 100–110 km beneath Big Island to 50–60 km beneath the island of Kauai, 500 km to the northwest. The thinned region is about 300 km wide.

These results3, together with basic fluid-mechanical principles, suggest a new hybrid model for the origin of the Hawaiian swell — ‘dynamic thinning’ (Fig. 2c). In this picture, an ascending mantle plume strikes the base of normal oceanic lithosphere and spreads out laterally, as happens in the dynamic support model. However, as the plume material is carried northwest by the moving lithosphere, its hottest central portions begin to undergo thermal convection, induced by conductive heat loss to the cooler lithosphere above. Such convection — which may take the form of ‘rolls’ aligned with the plate motion direction7,8 (Fig. 2c) — transfers heat vertically much more efficiently than does conduction alone9. It will therefore erode and thin the lithosphere, producing a trough filled with hot plume material whose amplitude increases the further it gets from the hotspot. Accordingly, the swell as a whole will be supported dynamically by material in motion. But the material compensating the swell's inner portion will be substantially shallower than predicted by the traditional dynamic support model.

Following on from Li and colleagues' work, perhaps the most important task is to extend their analysis to greater distances along the Hawaiian island chain, with the help of new seismometers deployed further to the northwest. This might reveal the rethickening of the lithosphere that should eventually occur as the plume material cools. A second task is to test the conclusions of Li et al. against new measurements of heat flux on the Hawaiian swell. Thermally induced thinning of the lithosphere to 50–60 km should produce measurable anomalies of the heat flux at the ocean floor, starting about 10 million years later, or after about 900 km of motion by the Pacific plate. However, attempts to detect such anomalies10 have so far been unsuccessful. It seems the Hawaiian swell still has much to teach the Earth scientist.


  1. 1

    Wilson, J. T. Can. J. Phys. 41, 863–870 (1963).

    ADS  Article  Google Scholar 

  2. 2

    Morgan, W. J. Nature 230, 42–43 (1971).

    ADS  Article  Google Scholar 

  3. 3

    Li, X., Kind, R., Yuan, X., Wölbern, I. & Hanka, W. Nature 427, 827–829 (2004).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Detrick, R. S. & Crough, S. T. J. Geophys. Res. 83, 1236–1244 (1978).

    ADS  Article  Google Scholar 

  5. 5

    Olson, P. in Magma Transport and Storage (ed. Ryan, M. P.) 33–51 (Wiley, New York, 1990).

    Google Scholar 

  6. 6

    Ribe, N. M. & Christensen, U. R. J. Geophys. Res. 99, 669–682 (1994).

    ADS  Article  Google Scholar 

  7. 7

    Moore, W. B., Schubert, G. & Tackley, P. J. Science 279, 1008–1011 (1998).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Ribe, N. M. & Christensen, U. R. Earth Planet. Sci. Lett. 171, 517–531 (1999).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Sleep, N. H. J. Geophys. Res. 99, 9327–9343 (1994).

    ADS  Article  Google Scholar 

  10. 10

    Von Herzen, R. P. et al. J. Geophys. Res. 94, 13783–13799 (1989).

    ADS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ribe, N. Through thick and thin. Nature 427, 793–795 (2004).

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.