Is it finally possible to predict earthquakes?
|
Version 0 of 1. Predicting earthquakes was once thought to be impossible due to the difficulty of calculating the motion of rocky mantle flows. Such flows are caused by high temperatures inside the Earth, the hottest part of which is the iron core. This core heats the bottom of the rocky mantle, which causes it to move slowly in large streams. But thanks to a new algorithm, we now know that it is possible to model these underground streams. While this may be of little solace to the thousands of people affected by the earthquake in Nepal last week, we hope that in the not-too-distant future scientists may be able to provide warnings for at least some similar events, helping to minimise loss of life and wider devastation. At the Delft University of Technology, we are using an algorithm to create a digital model of the North Anatolian Fault, a major geological fault extending westwards from eastern Turkey, across northern Turkey and into the Aegean Sea. The last major earthquake along this fault line occurred in 1999 at Izmit in Turkey – around 17,000 people were killed. The computer-based model seeks to predict where the plates are running together the hardest, and thus where underground stresses are the greatest — an often tell-tale sign of where an earthquake might hit. To do this the model seeks to map underground patterns of activity in the Earth’s mantle across some 100m underground grid points. The model predicts where these stress points will arise through simulating the flow of different rocky mantle flows. The temperature and pressure in the Earth mantle are so high that these rock flows are slow. In essence, the model can ascertain where, as a result of these flows, the plates are likely to run together, and automatically detect these stressed grid points. As good as this computer model is, however, it still needs refinement. This is because the link between earthquakes and underground mantle flows is complex and hard to compute. In practice, this means calculations of such flows involves the construction of very complex mathematical systems made up of millions of pressure and velocity values mapped around the underground grid points. This is a massive exercise and scaling it up – which involves computing over a large expanse of the Earth’s mantle – was previously too difficult. Our algorithm, however, has allowed this scaling up to be achieved using the latest scientific methods and computing power. The model is more accurate and comprehensive, but the project’s complexity has been increased considerably. While others working in the field do not dispute the model we are using there are always others would like to add more details, such as mapping a larger area or looking in even more granularity at factors like friction and viscosity. However, ever greater complexity involves ever more computing time. After finishing our work on the North Anatolian Fault, we intend to try to model the tectonics of the entire Earth. It is a truly ambitious project that will involve perhaps some 1bn grid points in what we call our “fine grid”. To make the computations for these one billion points will require surmounting yet another major hurdle — the parallel computing problem. That is, increasing the number of computers in a system generally means that they work less efficiently. However, our colleagues in Utrecht have already been working with a test-sample of 500 computers and we now believe we have mitigated this problem with our algorithm, but computing the one billion parts is a long-term programme. The hope is that in practice, when key underground points of tension are detected, that this information can be used by governments, NGOs, and communities to better prepare for earthquakes. For instance, governments can mandate stronger building regulations in areas of high stress, or evacuate a given geography. Moreover, the model can also potentially point to earthquake “aftershocks” which have particularly been a problem in Nepal. Despite this breakthrough, however, a journey lies ahead. Computing the 1bn parts in the fine grid is a long-term programme. In order to push forward in the meantime, we are also working on a technique called “coarse grid” acceleration, which only uses a small number of sample points in all of the Earth’s various strata, allowing us to obtain fast, accurate solutions for all of these sample points, leading to considerable savings in computer time. While much more hard work and innovation lies ahead, this new frontier of seismology is genuinely path-breaking and already achieving exciting results. However, as the Nepal earthquake has painfully reminded us all, true success will only be achieved when we reach the stage at which human lives are saved by applying our research in practice. Kees Vuik is a professor of numerical analysis at Delft University of Technology. Join our community of development professionals and humanitarians. Follow@GuardianGDP on Twitter. |