Magma fragmentation in eruptions

Article metrics

Some volcanic eruptions are gentle and effusive, and even (as on Hawaii) support a tourist trade. Others, such as the eruptions of Mount St Helens in 1980 and of Mount Pinatubo in 1991, are violent and terrifying, causing severe damage, loss of life and even temporary climate change. The difference between them stems from the nature of the erupting magma and the way that it breaks up into fine particles and ash by a process called fragmentation.

Papers by Zhang1 and by Marti et al.2 (pages 648 and 650 of this issue) provide new insights into magma behaviour during fragmentation. Zhang describes a new criterion to assess brittle failure of magma. Any brittle material can support stress (force) by undergoing strain (bending or stretching) up to its elastic limit, after which it either bends permanently, or breaks by brittle failure. When the thin walls between bubbles in a magmatic foam are stressed sufficiently during bubble growth, they break by brittle failure, and the bubbles coalesce. Marti et al. delve into the details of the stress state within the magma before and during fragmentation by examining the morphology of pumice from a volcano in the Andes.

It is bubbles generated deep within a volcano that are the driving force behind eruptions. They grow by diffusion of dissolved magmatic gas and decompression on the way up to the Earth's surface3,4,5,6, as described in Box 1. But fluid magmas (even highly viscous ones) should allow bubbles to grow until their internal pressure matches that of the exterior (1 atmosphere) before fragmentation, and thus lead to more gentle fragmentation and less explosive eruptions than are observed.

This apparent discrepancy is reconciled by the theoretical formulation and laboratory observations of Zhang1 and of Marti et al.2. They show that as the viscous material deforms rapidly due to bubble growth, it can no longer be treated as a fluid with a Newtonian viscosity, but assumes the characteristics of a brittle material. In part, brittleness may be exacerbated by cooling of the bubble walls by the latent heat of exsolution of water evaporating from the magma, so that they may even turn to glass in extreme cases7. With brittle behaviour of magmatic foams, fragmentation does not depend so much on the total amount of strain (as for typical solids), but rather on the strain rate.

The key to brittle fragmentation during eruptions is for the bubbles to grow and deform so quickly that there is no time for the magma to ‘relax’ as would a fluid. It thus behaves like a brittle solid that can be broken, rather than bent or stretched. Although one normally considers the total amount of strain (bending or stretching) as a failure criterion for elastic solids, magmas do not behave quite like solids. If a small stress is applied, fluid flow will eventually relieve that stress (relax) as the fluid flows (strains) to accommodate it. But when the stress is large enough, the strain rate to accommodate it is too large for the magma, and it breaks like an elastic solid.

Zhang finds that, under sufficiently high strain rates, a magma can be treated as an elastic solid, and he applies brittle-failure theory to an idealized case of spherical bubbles surrounded by spherical shells of magma. As the bubbles grow and cause thinning of the walls of magma that separate them, the tensile stresses grow in the magma until they surpass its tensile strength, causing fragmentation. As further evidence for the brittle behaviour of magma under high strain rates, Marti et al. find that the morphology of tube pumices (volcanic foams with highly elongated bubbles) records the state of strain at the point of fragmentation. With this strain marker, they demonstrate that the magmatic foam from which the pumice was derived deformed as a fluid until shortly before fragmentation, then sheared in a visco-elastic manner just before fragmenting as an elastic material. They conclude that fragmentation was associated with a great increase in strain rate.

The disruption of a magmatic foam8 marks a fundamental change in the topology of a two-phase magmatic system as the system shifts from a bubbly melt to a gassy spray. A magma containing bubbles is a continuous ‘liquid’ melt with isolated gas bubbles that can communicate only through diffusion of gas through the melt. After fragmentation, however, the system becomes a continuous gas that is carrying tiny droplets or fragments of magma (which do not communicate with each other at all). Further, because of the low viscosity of the gas, the droplets may have a different velocity from that of the gas, and so may be segregated. In some cases, this can lead to nuées ardantes, devastating, high-density ‘avalanches’ of gas and ash particles that can move downslope very quickly and over long distances — as, for instance, one did on Mount Pelée, Guadeloupe, destroying the coastal town of St Pierre and its inhabitants.

With this new understanding of fragmentation, volcanologists can reconcile the violent behaviour of many volcanoes with the characteristics of the erupted magmatic materials, leading us another step closer to predicting the nature of eruptions. As described in Box 1, however, the question of bubble nucleation remains. Once bubbles exist, we know how they grow. Once they grow, we now know how they ‘pop’. But we still don't know how they form in the first place.


  1. 1

    Zhang, Y. Nature 402 648–650 1999).

  2. 2

    Marti, J., Soriano, C. & Dingwell, D. B. Nature 402, 650–653 (1999).

  3. 3

    Alidibirov, M. & Dingwell, D. B. Nature 380, 146–148 (1996).

  4. 4

    Alidibirov, M. & Panov, V. Bull. Volcanol. 59, 481–489 (1998).

  5. 5

    Gardner, J. E. et al. Bull. Volcanol. 58, 144–162 (1996).

  6. 6

    Papale, P. Nature 397, 425–428 (1999).

  7. 7

    Sahagian, D. L. & Proussevitch, A. A. J. Volcanol. Geotherm. Res. 74, 19–38 (1996).

  8. 8

    Proussevitch, A., Sahagian, D. & Kutolin, V. J. Volcanol. Geotherm. Res. 59, 161–178 (1993).

  9. 9

    Toramaru, A. J. Geophys. Res. 100, 1913–1931 (1995).

  10. 10

    Proussevitch, A. & Sahagian, D. J. Geophys. Res. 103, 18223–18251 (1998).

  11. 11

    Proussevitch, A. A. & Sahagian, D. L. J. Geophys. Res. 101, 17447–17456 (1996).

Download references

Author information

Correspondence to Dork Sahagian.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Sahagian, D. Magma fragmentation in eruptions. Nature 402, 589–591 (1999) doi:10.1038/45099

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.