One of the crowning monumental achievements of the ancient Mediterranean world is Emperor Justinian’s church of Haghia Sofia in what was once Constantinople and is now Istanbul. This magnificent domed structure was severely damaged by earthquakes in AD 557, 989 and 1346. Today, Haghia Sophia is in the firing line again.

The great church, and the enormous modern city that surrounds it, lies along one of the Mediterranean’s most dangerous earthquake faults. Turkey’s North Anatolian fault is Europe’s answer to the San Andreas Fault in California. This 900-kilometre crack in the ground cuts across northern Turkey from east to west, allowing the interior of Anatolia to move westwards with respect to Europe.

Over the last 60 years, successive strands of this fault line have ruptured in large earthquakes, each event releasing stress on one part of the fault and passing it down the line to the next strand. The result is that, rather like a set of falling dominoes, the ruptures of the North Anatolian fault have moved steadily westwards – in 1939, 1942, 1943, 1944, 1957 and 1967 – from the comparatively sparsely populated parts of eastern Turkey to the industrial heartland of the north-west. Then, in August and November 1999, two of the strands just east of Istanbul that had yet to break ruptured in earthquakes that left over 35,000 dead, destroyed 15,000 buildings and cost $10–25 billion in damage.

Earthquake geologists are convinced that the quakes have now added stress to the last remaining significant unruptured strand, the section of fault that lies in the Marmara Sea, directly offshore of Istanbul. If the pattern of the past continues, then Istanbul, a conurbation of some ten million people, is due for a direct seismic hit in the coming years or decades.

If ‘earthquake storms’ can occur over decades, might they also explain apparent sudden bursts of devastation that affected the ancient world in a similar short period of time? Most controversial is the theory that an earthquake storm may have been responsible for the abrupt physical and political collapse of Aegean Bronze Age world around 1200 BC. Some geologists and archaeologists point out that most of the ancient cities that fell at that time lie along the plate-boundary of the eastern Mediterranean and show signs of destruction typical of earthquakes. It supports a view that a storm of earthquakes successively ‘unzipped’ the plate boundary, so weakening the cities along the way that they were left vulnerable militarily, inviting attacks from opportunistic neighbours.

Earthquakes have frequently been used by historians and archaeologists as convenient explanations for cataclysmic destructions and abandonments. But earthquakes rarely wipe out entire cities, let alone entire regions. More often, seismic shocks leave cities as jumbles of ruined, damaged and intact buildings, encouraging their inhabitants not to flee but to stay and rebuild their houses and livelihoods. It is a pattern that we see in modern earthquake disasters, and there is little sign that human nature was any different in the past. In general, the intensity of seismic effects weakens away from their source, ensuring that individual earthquakes are unlikely to wreak regional devastation. Typically, seismic disasters strike down one or two neighbouring settlements, but those a few tens of kilometres away are virtually unscathed. The size, or magnitude, of the earthquake determines the extent of the ‘fallout zone’. The 1999 Turkish earthquakes struck only 80 kilometres from the centre of Istanbul, yet the death toll in the metropolis was comparatively light, approximately a thousand people. Nevertheless a direct hit on the city next time round may well produce a seismic catastrophe on a scale far greater than anything witnessed by the ancient world.