An interesting article by Matt Kaplan in the 17 February of the New Scientist is about the millennia old claims that many animals, including, dogs, snakes, elephants, can sense an earthquake sometimes days before it strikes.
Animals have frequently been claimed to demonstrate unusual behaviors prior to earthquakes, including agitation and attempts to escape, if they were captive. Apparently in China, according to the article, they now even keep and monitor snakes for strange behaviors in hopes of predicting earthquakes.
One problem with all the existing claims is that they have all been anecdotal and when more scientific data are taken into consideration, nothing concrete emerges. For example, the article mentions that some villagers in Thailand told the story that on 26 December 2004 their buffalo that had been grazing on a beach had gotten agitated and headed for a nearby hill minutes before the devastating tsunami came (the villagers who followed the buffalo survived). But at the same time in Sri Lanka, a WWF scientist who was monitoring the movements of radio-collared elephants did not notice any unusual activity in one herd that was 100 m away from a beach before the tsunami arrived. All the elephants did was move behind a sand dune after they saw the tsunami approaching.
Animals' quake-predicting abilities will become more respectable only if someone can predict an earthquake using animal behavior well-before the earthquake happens; post-predictions do not count. The inherent difficulty associated with predicting earthquakes stems from the fact that an earthquake is a complex event that results from the interactions of a large number of factors. This was summarized by Geller et al.1 as follows:
Whether any particular small earthquake grows into a large earthquake depends on a myriad of fine details of physical conditions throughout a large volume, not just in the immediate vicinity of the fault. This highly sensitive nonlinear dependence of earthquake rupture on unknown initial conditions severely limits predictability. The prediction of individual large earthquakes would require the unlikely capability of knowing all of these details with great accuracy. [citations omitted]
Along these lines, the New Scientist notes that "the activity associated with different earthquakes is highly variable". That being the case, I don't see much hope for the reliable use of animals to predict earthquakes.
1Geller et al. 1997. Earthquakes cannot be predicted. Science 275:1616.