Earth science

Slip-sliding away

The side of an oceanic volcano, one of the Hawaiian islands, has been caught sliding towards the sea. The distance concerned was only a few centimetres. But it could be an indicator of a huge landslip to come.

Seeing is believing. For Earth scientists especially, the adage holds considerable weight because the 'seeing' is so rare. The geological record tells us that mountain ranges have been built and then washed down to the sea; that entire ocean basins have opened and closed like a door; and that ice a mile thick blanketed the globe a dozen times over. Scientists believe that these events took place, but still, they are hard to imagine. For most people such incidents might smack more of science fiction than science fact — but then, seeing is believing.

Who, having viewed film of Mount Pinatubo exploding, or the aftermath of a large earthquake, doesn't accept that 'big stuff' really does happen? In their paper on page 1014 of this issue1, Cervelli et al. provide a glimpse of what might end up to be big stuff — the whole side of an oceanic volcano falling into the sea, an event known as flank collapse.

Over time, virtually all oceanic volcanoes grow, become too steep, and slough off flank material. We know this to be true because sonar surveys around most volcanic island chains reveal dozens of old, overlapping debris fields. The Hawaiian islands alone host 70 collapse fields dating from 20 million years ago. Adding up the number of debris fields from all of the ocean's volcanic islands yields the estimate that one flank collapse happens somewhere in the world every 10,000 years on average. Flank collapses are nature's great landslides. Embracing up to 5,000 km3 of rock, they compare to a one-and-a-half-kilometre-thick slice of the state of Rhode Island or the island of Majorca racing sideways for 30 or 60 km. In contrast, when Mount Saint Helens erupted in 1980, only 3 km3 of material blew away.

While flank collapses of oceanic volcanoes are common geologically, none has been caught in action — until now. Cervelli and colleagues' glimpse1 of the action came from a network of 20 continuously recording stations of the global positioning system (GPS) scattered about the southeast slope of Kilauea volcano on Hawaii's big island. In November 2000, over a 36-hour period, these GPS stations witnessed a 20-km-long and 10-km-wide chunk of the southeast flank move seaward at the speed of 6 centimetres per day. For geophysicists accustomed to tectonic motions of a few millimetres per year, a few centimetres per day is like rocket travel.

To soothe any doubting Thomas, Cervelli et al. spend half of the report reviewing the details of their analysis. The effort is certainly thorough enough to dispel any notion that the signal is a fluke or masquerading noise. For me, their map of a dozen GPS displacement arrows (Fig. 1 on page 1015) all pointing out to sea far beyond their error ellipses tells the whole story. What else can they indicate but some early stage of one of those flank collapses that litter the geological record? A 2,000-km3 piece of Hawaii is slip-sliding away.

In terms of predicting a collapse, the authors interpret their observations more cautiously than I do. By means of dislocation modelling, however, they confirm that the observed GPS displacement field could be explained by 10 cm of offset on a shallow dipping surface that lies 4.5 km under their network and that probably extends well out to sea. The offset was a silent earthquake, if you will, on the fault that may eventually detach the whole flank. Thankfully, the November 2000 slide stopped short, but what would it take to dislodge the whole block? Experts believe2,3,4,5 that intense intrusion of the flank by molten magma dikes might provide the extra nudge, especially if the detachment fault is also lubricated by the injection of high-pressure groundwater.

Last year, Simon Day and I published an article6 on the tsunami waves that might be generated by the collapse of another oceanic island volcano, Cumbre Vieja on La Palma in the Canary Islands. Our computer simulations predicted that a tsunami stirred by a 500-km3 landslide there could span the entire Atlantic basin, keeping amplitudes of 10–20 metres. Based on these calculations, if a 2,000-km3 piece of Kilauea ever does push into the sea, it could, under certain conditions, parent a tsunami that will strike much of the Pacific Rim (Fig. 1). Historical time has not seen a tsunami of this scale, but many researchers argue that geological deposits and landform shapes preserve the signature of older ones.

Figure 1: Computer simulation7 of the tsunami waves that might be set off in a collapse of Kilauea's southeast flank.

The simulation assumes that a block 40 km long, 20 km wide and 2,000 m thick (blue, inset) slides for 60 km at 70 m s−1. The tsunami wave fronts are pictured at two-hour intervals from 2 to 14 hours. Red and blue contours are wave elevations and depressions, respectively, and the numbers are sample wave heights in metres. Tsunamis from this collapse would have 1.7 × 1019 J of energy (equivalent to about 4,100 megatons of TNT), and focus slightly towards the southeast. The waves span 280° of arc, sparing only locations to Hawaii's north and west. Tsunamis from volcanic flank collapses can dwarf those generated by earthquakes of any plausible magnitude. The model predicts potentially devastating 30-m waves beaching on the west coast of North America. These decay to 10 m in height by the time of their arrival at the tip of South America.

The implication that volcanic island collapses could raise extensive tsunamis grips both one's imagination and concern. Could potential collapses be monitored and perhaps forecast? Cervelli et al. demonstrate that current GPS technology deployed in networks at 5-km intervals can provide real-time detection of even seismically silent shifts in volcanic edifices: small, silent shifts that may presage a fully fledged flank failure. The world's oceanic volcanoes are stages best not left unwatched. In a few years, dedicated radar satellites may take up volcano guard duty. For now, GPS provides one of the sharpest views.

People should not lose sleep over large but rare natural hazards. They should not run blind either, particularly when a useful eye exists. Until the next volcanic island does collapse we will never know how nature's great landslides play out, but for me, Cervelli and colleagues' article supports the case that seeing is believing.


  1. 1

    Cervelli, P., Segall, P., Johnson, K., Lisowski, M. & Miklius, A. Nature 415, 1014–1018 (2002).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Elsworth, D. & Voight, B. J. Geophys. Res. 100, 6005–6024 (1995).

    ADS  Article  Google Scholar 

  3. 3

    Iverson, R. M. J. Volcanol. Geotherm. Res. 66, 295–308 (1995).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Day, S. J. Geol. Soc. Lond. Spec. Publ. 110, 77–93 (1996).

    ADS  Article  Google Scholar 

  5. 5

    Elsworth, D. & Day, S. J. J. Volcanol. Geotherm. Res. 94, 323–340 (1999).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Ward, S. N & Day, S. Geophys. Res. Lett. 28, 3397–3400 (2001).

    ADS  Article  Google Scholar 

  7. 7

    Ward, S. N. J. Geophys. Res. 106, B6, 11,201–11,216 (2001).

    ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Steven N. Ward.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ward, S. Slip-sliding away. Nature 415, 973–974 (2002).

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.