SCIENCE: FINDING FAULTS

Earthquakes are terrifyingly unpredictable, but plotting their past strikes may one day help to give early warning of the quakes of the future. Sanjida O'Connell reports

The earth's surface, always uneasy, is prone to slip along jagged breaks known as fault planes, which extend up to 20km below the ground. When this occurs, huge amounts of energy are released and the result is an earthquake. Accurate instrumental records of earthquakes only go back as far as the 1960s but archaeologists and geologists have now begun to chart their history over the past 3,000 years, as they attempt to extend their knowledge of one of nature's most terrifying and unpredictable phenomena.

"It is like a detective story," says Professor Nicholas Ambraseys, from the Civil Engineering Department at Imperial College in London. Professor Ambraseys pieces together where and when earthquakes might have struck and how powerful they were using a variety of clues. Often he starts by unearthing ancient documents. The most useful date from Roman times and detail payments made to restore earthquake-damaged buildings such as temples.

The Jordan rift, which runs through Jericho, is the only place in the world with historical records extending back 10,000 years. In 31BC, a fault plane that carves its way across Jericho caused an earthquake which destroyed the town of Qumran (where the Dead Sea scrolls were found). The devastation was recorded by the Jewish historian Josephus: "At this time it was that the fight happened at Actium between Octavius Caesar and Anthony in the seventh reign of Herod and it was also that there was an earthquake in Judea, such a one as has not happened at any other time and which brought a great destruction to the cattle in that country. About 10,000 men also perished" (Antiquities of the Jews, Book XV, chapter five, verse two).

In some cases, there are no historical leads to follow. But digging a trench through a fault plane can reveal the record of an earthquake's passing. A layer of sand marked with turbulent patterns as if cigarette smoke had solidified is clear evidence. The extraordinary energy released during a quake turns the earth to liquid; human artifacts and grains of pollen are caught up in its flow. This layer then solidifies, but the artifacts and pollen can be carbon-dated, allowing scientists to pin-point when the earthquake took place.

Professor Paul Hancock of Bristol University uses another recently discovered investigative technique. He and his colleague Dr Erhan Altunel, a Turkish geologist, have been examining ancient Aegean sites. The ways columns have fallen and buildings have cracked provide clues which indicate where previously undiscovered fault planes lie. One of the main sites Professor Hancock and Dr Altunel studied is the Turkish town of Hierapolis, now called Pamukkale; by charting the damage to the town they have been able to record its earthquake history. The geologists' research has shown that Hierapolis lies on a jagged fault plane which stretches from the north- west to the south-east. In this region the earth's surface has been pulling apart for 3,000 years. The first earthquake probably hit the town in 60BC and created what is known as a fault scarp: during the quake one part of the earth's crust slides past another either horizontally, or as at Hierapolis, vertically, creating a wall of rock. At Hierapolis the scarp is at a 60 degree incline. A carving of Diana, Mercury, and the local deity Hierapolis, which dates back to Roman times and was probably carved in the second century AD, is suspended three metres above the scarp. Since no-one would have been able to reach the stone to carve the gods, it must have been uplifted with the fault plane during an earthquake, says Professor Hancock.

It is also possible to date earthquakes by the approximate age of artifacts that have been displaced or broken. A petrified man-made water channel runs under the town's Roman fort. The fault plane also passes directly blow the building and both it and the channel have been cracked by quakes. The channel has been off-set (moved relative to the original line of the walls) twice as much as the fort, so there must have been two earthquakes caused by the same fault plane. The first occurred in Roman times (which cracked the channel), the second after the fort had been built (cracking the channel again and also the fort). It is likely that this one was in the 13th century as a coin from that era was found trapped between fallen bricks.

Another technique used to identify the passing of a quake is to examine limestone deposits. There are a number of carbonate-rich hot-water springs in the area; the carbonates gradually precipitate out as a kind of limestone known as travertine. Local water channels have grown walls of travertine - there are several in Hierapolis that reached a height of 10m before being spectacularly ruptured. Sheets of travertine have also built up on a Roman sarcophagus overturned by an earthquake. By working out the rate of deposition, it has been possible to reveal when the sarcophagus was toppled.

Scientists have to be wary of blaming earthquakes too precipitately, though. The tombs in two of the necropolises in Hierapolis are damaged and, at first, scientists thought the destruction was caused by an earthquake. However, a closer examination revealed that the lids had been rotated by 30 degrees, although the bases were still intact - more likely to indicate grave-robbers.

The path of the earthquake alters as it passes through ancient relics, hitting them at right angles because the buildings are rigid in comparison to the soil they are built on. Often this can give a false picture of where the fault plane is and in what direction it lies. Fortunately, in Hierapolis it is possible to see the fault line - where the fault plane has intersected the earth's surface - before it hits the town. One of the quakes knocked over several columns - all lined up in parallel where they fell - by a sacred pool (now in the centre of the Pamukkale Motel) before passing through a Byzantine basilica.

Professor Hancock was able to surmise that Hierapolis was the epicentre of the fault - in other words, the heart of the earthquakes were directly beneath the town. He believes that there were three major quakes in the area, one around AD60 (which created the scarp), one in the sixth century and one in the 13th century, with several smaller shakes in between. The distance by which relics have been displaced and the damage they suffered indicate that the AD60 quake had an intensity of IX on the modified Mercalli scale. The scale extends from I to XII and is a measure of the energy released from the epicentre of the quake. An intensity of IX would cause shaking that would "cause houses to fall down, destroy their foundations, cause cracks in the ground, and make sand and mud boil as water comes up from below the earth's surface". Looking back at the intervals between the quakes, Professor Hancock estimates that the town will be hit by a large earthquake every 500 to 1000 years.

A 500-year warning hardly seems sufficient for the residents. "Prediction is totally useless. If you predict an earthquake, you have to take the consequences. Who wants that responsibility?" asks Ambraseys. No one has yet managed to predict when and where an earthquake will next appear, but by at least showing where the fault plane lies, local people have some indication that they are living in an earthquake zone. Professor Hancock adds, "An analogy would be if a foreigner visited Britain in November. If he didn't take precautions - by wearing warm clothes - we'd consider that imprudent." !