The press announced in April that the USGS and other western US geoscience institutes had issues the first ever comprehensive earthquake forecast for California (see http://www.scec.org/ucerf/) , but it was cautiously phrased in terms of probabilities of destructive magnitudes (>6.7) over the next 30 years. That might be fine and dandy for administrators and civil engineers, but not so good for anyone who becomes a victim at the precise time this or that Californian fault ‘goes off’. People world-wide have rarely chosen where to live based on knowledge of geological risks; indeed most threatened communities have little choice, for many reasons. What would be useful is being warned that a devastating earthquake is definitely due where one lives, and it will happen sometime in the next few days or weeks. Even an hour’s warning will save many lives. But no geological survey will commit itself to that kind of pronouncement, except perhaps some of the many surveys in China. The fact that all kinds of phenomena, such as nervousness among animals, rising water levels in wells and so-on have been shown to occur shortly before many big earthquakes has prompted a kind of ‘barefoot’ monitoring that is officially co-ordinated in some parts of China. It is said that lives have been saved on a number or recent occasions.

It is easy for western scientists to make the analogy with homeopathy, and pooh-pooh such methodology. Also, there has been a succession of observations from space that could prove useful, such as ‘earth lights’ and magnetic-field fluctuations that accompany some seismic events (see Remote signs of earthquakes in EPN August 2003, Early warning of earthquakes in EPN December 2005). The latest odd, but conceivably useful connection is an association of unusual cloud formations with earthquakes in Iran (Guo, G. & Wang, B. 2008. Cloud anomaly before Iran earthquake. International Journal of Remote Sensing, v. 29, p. 1921-1928). The authors, from Nanyang Normal University in China, scrutinised free, hourly images from the geostationary Meteosat-5 satellite covering the whole of Iran, where seismicity is concentrated on a single large zone of deformation that trends NW-SE through the Zagros mountains. On several dates they found cloud formations parallel to the fault zone. Between 60 to 70 days later large eathquakes took place along the fault, including the highly destructive Bam earthquake of 26 December 2003. Indeed, a noticeable thermal anomaly in clouds directly above Bam occurred 5 days before the disaster.

How often do tsunamis occur?

Fortunately, truly destructive tsunamis on the scale of that of 26 December 2004 are rare events. So much so that nobody has a clear idea of their average frequency at different exposed shorelines; a vital statistic for risk analysis. Tsunamis produce high energy marine deposits, but unless they are preserved in accessible locations their incidence would be difficult to estimate, and they may be confused with tempestites generated by hurricanes. One characteristic of tsunamis is that they are waves that affect the entire ocean volume, unlike wind waves whose effects are restricted to a few tens to hundred of metres, which can create unique features. Canadian, US and Omani sedimentologists have examined a sediment deposited in Oman by a recorded tsunami generated by a large earthquake off Pakistan in 1945 and have discovered one such signature (Donato, S.V et al 2008. Identifying tsunami .deposits using bivalve shell taphonomy.  Geology, v. 36, p. 199-202). The deposit, a coquina rich in bivalve shells, contains an unusually high proportion of still-articulated shells, suggesting that living animals were ripped from the seabed and then flung into a lagoon. Along with oddities in fragmentation of other shells and the sheer size and extent of the coquina, this feature seems to be characteristic of tsunamites. Features in the Oman example closely match those in another on the eastern shore of the Mediterranean Sea in Israel.