The holy grail of earthquake prediction
Wouldn’t it be wonderful if we could predict damaging earthquakes and be forewarned about their time, size and location days or hours before they occur—like we predict cyclones? The tragic Nepal earthquake that occurred not far from the Indian border on 25 April gives further impetus to that question.
Scientists these days have a fairly good idea of how and why earthquakes occur. They have been able to make sophisticated models of earthquake generation along the fault lines from the study of seismic waves. The grand unifying theory in Earth sciences, called plate tectonicsm, provides the theoretical underpinnings of earthquake generation. This theory tells us that the Earth’s top solid crust is divided into various tectonic plates, and that they are all in relative motion as they move at less than a snail’s pace over underlying semi-molten material. While some of them grind past each other, or go under, there are areas where they are colliding to form huge mountains—like the Himalayas. The borders of these opposing plates are locales where tremendous strain accumulates over long periods and, finally, the stresses overwhelm the strength of the rocks, generating quakes that result in releasing the stored energy outward through seismic waves.
Thousands of small earthquakes occur every day—the question is why some of them evolve into huge earthquakes. To reach the holy grail of earthquake prediction, we need to first understand what transforms a small offset at the source into a real big rupture—like what happened in Japan in 2012 or recently in Nepal. We do not know this because our understanding of such processes is limited as of today.
All earthquakes nucleate at depths of tens of kilometres below the surface of the Earth and we do not have enough understanding of the properties of rocks, including the pressure, temperature and fluids and the attendant complexities at those depths. Nor is it possible to observe closely the earthquake nucleation processes through the huge thickness of opaque solid rock, unlike the above-surface atmospheric processes that can be monitored.
Predicting the exact time and size of an earthquake, thus, remains elusive.
But we have made progress in finding the probability of an earthquake occurring in a region over a span of decades, and how much of stress is building up in a particular region. In some places, we also know how often big earthquakes occur and the overall budget of accumulated versus used-up slip in earlier earthquakes. We can call this “long-term forecast” and scientists have been able make such forecasts of earthquake vulnerability of some parts of the world, like the Himalayan arc.
The combined use of satellite-based measurements and modelling of crustal deformation and fault-specific geological studies of earthquake recurrence have made such forecasts possible, at least in some regions.
Satellite-based radar imaging of surface deformation before and after an earthquake is a new tool that is likely to aid earthquake forecasting. Knowing the recurrence period or pattern of earthquakes on the fault lines will also contribute to long-term forecasting of earthquakes. So the current state of knowledge might tell you which is the most likely place of an earthquake disaster, but the precise timing of it is beyond our present capabilities.
A million-dollar question is whether any precursory signals can be captured years, days or hours before a big earthquake. For, ultimately, the success of earthquake prediction science depends on resolving this question.
To understand this issue further, we need to go back to the first and only successful case of an earthquake prediction—the 4 February 1975 Haicheng earthquake in north China. The story unfolded during the murky days of China’s Cultural Revolution and the exact details came out much later, after American seismologists visited the affected areas and consulted the people involved in this exercise to bring out the facts.
They reached the conclusion that what happened in the 1975 Haicheng earthquake was a true case of a successful prediction. The leader of the American fact-finding team, Kelin Wang, describes this as “a blend of confusion, empirical analysis, intuitive judgment and good luck”. The precursory signals included land level changes in the nearby areas that prompted an initial long-term earthquake warning. But a short-term warning was later given based on the increased foreshock activity. The location of the quake was also identified. Taking the cue from the warnings, local residents began leaving the area and when the 7.3 magnitude earthquake eventually occurred at 7:36pm, it killed about 1,000 people.
But the euphoria of this apparent successful prediction was short-lived. On 28 July the next year, a 7.6 earthquake that occurred in a place called Tangshan could not be predicted as there were no precursory foreshocks or landlevel changes, and 250,000 people lost their lives.
Earthquakes behave differently and, essentially, they reflect the complexities of the source regions.
But the quest goes on. Some groups of scientists think that during the preparatory stages of earthquakes, microcracking occurs in the source regions and it generates electromagnetic emission in various frequencies. So the idea is to catch this signal before the earthquake strikes a particular area. Some researchers expect to see some sort of ionospheric perturbations before the earthquake strikes and they hope satellite-borne experiments might be able to establish the applicability of this method.
A practical way to deal with earthquakes is to generate scenarios of site-specific ground motion in response to expected maximum near-field earthquake magnitudes. Once we have that information, a prudent option is to develop an earthquake-resistant built environment. As a robust response to earthquake hazard, this is the best option currently available to us.
The writer is a professor of geodynamics at Jawaharlal Nehru Centre for Advanced Scientific Research, Bengaluru.