Tragedies like that in Izmit remind us of our limitations. Although we understand the mechanisms in which they have their origins, we are a long way from being able to predict earthquakes in any useful way.
Many scientists distinguish nowadays between earthquake forecasting and earthquake prediction. The former is based on the accurate measurement of small movements of the ground in the vicinity of faults, either by terrestrial instruments or by means of satellites, and on a careful analysis of historical records of seismic occurrences of various kinds. The results are estimates of the long-term probability of tremors of specific magnitudes at certain places, and they are widely regarded as both useful and acceptably reliable.
Earthquake prediction, on the other hand, is concerned with anticipating a major tremor at a certain place anything from hours to months ahead. Despite the considerable optimism that existed a quarter of a century ago, many experts are now of the view that prediction, in this sense, is just not possible.
Great faith was placed in the 1970s on the potential of "empirical precursors" - observed phenomena that were believed to precede a major tremor for reasons not completely understood. It was observed, for example, that water levels in wells might change as fractures opened below the ground. Paradoxically, another symptom of the worst to come was silence - a "seismic quiescence", where seismic activity in an otherwise active region might remain at a low level for several successive years. On the other hand, a series of minor tremors might also be indicative of a major seismic incident to come.
Other techniques of seismic prediction rely on close monitoring of the behaviour of the Earth's crust. There may, for example, be changes in the natural magnetism of rocks, or in their electrical conductivity, or a sudden change in the chemical composition of the local groundwater in the period leading up to a major tremor.
But further research into these matters has revealed that in nearly every case the guidance provided by such criteria is ambiguous. Indeed when some of the apparently successful predictions were checked, they were found to be due to measurement errors or just wishful thinking. Many geologists now would agree that Charles Richter, he of the famous scale, had got it right when he said over 20 years ago: "Journalists and the general public rush to any suggestion of earthquake prediction like hogs towards a full trough. Prediction provides a happy hunting ground for amateurs, cranks and outright publicity-seeking fakers."