This post is a follow up to my comments on the 7.2 magnitude earthquake that hit Northern Japan in June 2008. Of late, I have been thinking deeply on how to detect natural disasters by using Computational Intelligence techniques. In the course of my thoughts, I recalled that I had read something about many animals fleeing the tsunami hit zones several hours before the disaster struck. After a short Google search, I discovered many articles on this issue, the best of them all being the Ecologist.

The March 2005 edition of  the Ecologist reported that many animals escaped the great Asian tsunami on Boxing Day, 2004. It is said that elephants in Sri Lanka and Sumatra moved to high ground before the giant waves struck. In Thailand, a heard of buffalo were grazing by the beach, then they suddenly and started running up the hill.Some bewildered villagers followed them uphill and were also saved. There are lots of stories about these strange pre-tsunami behavior of animals in this region.

Rupert Sheldrake, the author of the Ecologist of  March 2005, further pointed out that the strange behavior of animals before the Asian tsunami was not a new phenomenon. He says  that historically animals also seem to know when other kinds of calamities are about to strike. He says that there is a large body of evidence of unusual animal behavior before earthquakes. Many wild and domesticated animals generally behave in fearful, anxious, or unusual ways several hours or even days before the earthquakes struck. He concludes his article by saying that by paying attention to unusual animal behavior, earthquake warning systems can be developed. Pet owners would be involved strange behavior of their animals to a central control unit where a computer system would analyze the places of origin of the incoming calls. If there were an unusual number of calls, it would sound an alarm, and display on a map the places from which the calls were coming.He further says that the cost of such a system would be far much cheaper than the amount of money being invested in earthquake and tsunami research currently.

It is now more than three years since Rupert wrote his article. I, therefore, would like to find out if any research institution, University or government department has ever tried to implement a computer system as suggested by Rupert? In my view, it would be very easy to implement such a system. I, however, think that such a system would just serve as a starting point. There is a need for we, scientists, to learn from animals. We must find out the reason why they feel uncomfortable whenever there is an impending natural disaster. As a Christian, I know that God can communicate to some people, by way of prophecy, about an impending disaster. But in this case, our focus is on animals which know by instinct that a disaster is about to strike. In the field of Computational Intelligence, we attempt to mimic nature for problem solving. I, therefore, would like to inform my fellow computational intelligence researchers that a new area has been added to our research endevours.I look forward to seeing some joutnal/conference papers entitled  Mimicking animal behaviour for detecting natural disasters in the near future.