FORECASTING EARTHQUAKES Seismic warnings: the Icelandic laboratory
On 27 October 1998, the geophysicist Stuart Crampin of Edinburgh University (UK) sent a stark warning by e-mail to Ragnar Stefánsson, senior scientific officer at the Icelandic Meteorological Office (Vedur) and supervisor of the country’s seismological research and monitoring network: “An earthquake of major magnitude has an 80 in 100 probability of occurring within the next three months,” he cautioned.
As a contributor to the Prenlab European project, a team effort on which researchers from ten EU countries worked alongside their Icelandic colleagues, Crampin based his warning on field observations in the South Iceland Lowlands (SIL) zone, a seismic region lying south of the capital Reykjavik and the subject of particular scrutiny. He had noted a progressive and significant slowing down of the phenomenon known as ‘shear-wave splitting’ due to a quite sustained wave of low-magnitude micro earthquakes. In geological terms, this indicator expresses variations in the alignment of minute fissures in the rock that forms the Earth’s crust due to the effect of internal tectonic strain.
A successful race against time The scientists working on the project located the probable epicentre of the expected earthquake along the 50km line running between two of the SIL seismological observation stations. Two days later, on 29 October 1998, the forecast became more precise, narrowing the time scale to sometime in the following month. At the Vedur’s request, the national civil protection committee put the essentially rural population in the danger zone on maximum alert. On 10 November, the warning was further refined, indicating that the earthquake could exceed five on the Richter scale.
The anticipated earthquake occurred at 10.38 a.m. on 13 November. Its epicentre was 2km from one of the observation stations and its magnitude was 5.1.
This remarkable textbook case of an early earthquake warning was made possible by Iceland’s particular context. The island, straddling the oceanic ridge on the boundary between the Eurasian and North-East Atlantic plates, is without doubt one of the most earthquake-prone places on the planet. Faced with the continuous presence of this kind of risk, Iceland is packed full of observation devices (seismographs, Global Positioning System [GPS] sensors carefully recording the slightest earth movements, boreholes, etc.) producing data that are carefully studied by the very best specialists. The island is a genuine natural geophysics laboratory (hence Prenlab, the project’s name). Its civil protection measures are also well developed.
Seismic warning map, with a scale of 20x20km, produced by the Icelandic Meteorological Office (Vedur) and transmitted to the Icelandic civil protection authorities on the evening of 19 June 2000. The red dots show the location of small shocks recorded during the previous two days and the thin black lines illustrate known fault lines. The green borders delimit the two areas where an earthquake was regarded as an imminent probability. On 21 June 2000, 26 hours after this document was produced, an earthquake of 6.1 magnitude produced the fault highlighted by the thick red line.
Grounds for concern Participants in the Prenlab project, which completed its work in 2001, were also able to monitor live a second remarkable double seismic episode. In this same region of south western Iceland, just after midnight on 17 June 2000, there came a new and more violent earthquake, this time measuring 6.6 on the Richter scale.
Immediately after the first tremor, scientists on the spot deployed all their resources to analyse the many facets of this earthquake and the aftershocks that followed. Their observations quickly gave cause for alarm. During the night of 19 June, that is 80 hours after the first earthquake, they warned the authorities that a second earthquake of about the same magnitude could strike at any time and provided quite a precise map of the likely location (see map). As predicted, the next night it struck.
Since 2001, the fruitful European co-operation of the Prenlab project has continued under the Prepared project. The latter’s mission is to study all the data recorded before, during and after the June 2000 earthquakes and to learn as many lessons as possible from them. All the micro-shocks recorded in advance of the two earthquakes are being scrutinised carefully, as is the information obtained from sensors placed at drilling points, changes in radon concentrations in borehole water, variations in the shear-wave splittings, etc.
"One should not imagine it will be possible to have clear and reliable forecasting in anything like the near future,” stresses Ragnar Stefánsson. “The two successful experiences during the Prenlab project certainly prove it is possible, but subject to three conditions. You need excellent knowledge of a zone’s tectonic conditions, an understanding of the processes at work in the underlying section of the Earth’s crust, and an effective and highly developed system of geophysical observation that functions in real time. Viewed in this light, the further European research carried out by the Prepared project certainly has a scientific significance that goes beyond the Icelandic laboratory. It benefits the geophysics community worldwide.”
When speaking of the future, the words forecast and prediction differ in terms of probability. The first, backed up by an array of reliable, cross checked indicators that can be modelled, often concerns the near future, as in weekly weather forecasts. Seismologists, never quick to presume anything when ...
The question of precursors
Indicators known as precursors can provide a basis for more immediate forecasts. One example of this is the slowing of micro-seismic waves, illustrated in the case of the Prenlab project – but this phenomenon has only been observed in a very limited number of earthquakes. The study of radon concentrations ...
When speaking of the future, the words forecast and prediction differ in terms of probability. The first, backed up by an array of reliable, cross checked indicators that can be modelled, often concerns the near future, as in weekly weather forecasts. Seismologists, never quick to presume anything when it comes to the unfathomable mysteries of the underground world, prefer the latter term. With few exceptions, the present state of knowledge only allows them to make long-term or, at best, medium-term predictions.
These are based on a large array of data. First, there is the increasingly precise map of faults that correspond to the boundaries between the Earth’s tectonic plates. Then there is the painstaking analysis of past earthquakes. Given a more or less constant tectonic pressure, rocks of a certain kind are presumed to experience cyclical ruptures of a comparable magnitude. If this does not occur, one speaks of a ‘seismic gap’, i.e. a growing probability of an earthquake due to the unrelieved build-up of stress. Finally, there are the new technologies of geophysical surveillance (see the case of Iceland), such as the increasingly numerous seismographic stations, probes, satellite pictures and GPS that measure movements on the edges of faults down to the last centimetre.
The question of precursors
Indicators known as precursors can provide a basis for more immediate forecasts. One example of this is the slowing of micro-seismic waves, illustrated in the case of the Prenlab project – but this phenomenon has only been observed in a very limited number of earthquakes. The study of radon concentrations in underground water in an active fault has become a classic precursor, although such an increase in radioactive gas can be linked to other causes, particularly climatic. Measurement of water levels in wells and the presence of bubbles are other indicators.
In the1980s, Greek researchers discovered the pre-seismic existence of electric currents circulating in the quartzite subsoil. However, in addition to the fact that this rock is very particular to certain zones, the validity of their modelling has been much discussed. Reliable studies have confirmed the reality of changes in animal behaviour in the moments immediately preceding an earthquake.
The use of these precursors is, therefore, a very delicate matter. However, there is one model case. During the 1960s, China, a country prone to violent earthquakes, launched a mass campaign designed to alert the population to these precursors. This proved very successful and led to the early warning of the powerful earthquake, measuring 7.3 on the Richter scale, that struck the Haicheng region on 4 February 1975. Unfortunately, just one year after this unique experiment, the Tangshan quake, probably the most murderous on record, hit the country without warning and left hundreds of thousands dead.