De Econometrist neemt een statistische kijk op de wereld.
Ever since we, humans, started to ruin this planet, we have had to deal with earth’s reciprocation that came in the form of earthquakes. Fortunately, over this long period of time, we got better and better at constructing houses and other types of buildings that could withstand these tremors, up to some extent. However, a hut made out of timber is rebuilt quicker than a skyscraper and, during an earthquake, one probably resides rather in the first than in the latter. Even with present-day architectural technology, there remain incidents claiming hundreds or even thousands of lives. It therefore comes as no surprise that every so often an earthquake and its aftermath make it to global news. Being able to forecast weather, sales or even predict the position of Jupiter in a hundred years, why can’t we figure out when an earthquake will strike?
Humans have been recording earthquakes for around 4,000 years. The deadliest documented earthquake occurred in 1556 in China, which is estimated to have killed 830,000 people. The Valdivia earthquake (Chile, 1960) was measured to be of the highest magnitude ever recorded, 9.5 on the Richter scale, claiming “only” 5,700 people. More recent destructive earthquakes struck Indonesia, China and Japan, in 2004, 2009 and 2011, respectively. Taking a closer look, the greatest part of all documented earthquakes happened in the same regions. When looking at a list of all significant earthquakes in the period of 1900 to 2017, one can clearly see a pattern of some sort:
The orbs indicating earthquakes of certain magnitudes clearly form a system of lines around the earth. Why does this pattern exist?
The pattern in the above picture suggests that the surface of the earth is divided into multiple pieces. This turns out to be exactly the case: the earth’s crust is separated into multiple so-called tectonic plates, which float on a molten part of the earth’s mantle. Taking a look at a map of the earth dissevered into its tectonic plates, it becomes clear that the orbs indicating earthquakes all lie on top of the boundaries of two or more plates.
When familiar with the cause of an earthquake, this resemblance is extremely sensible.
Tectonic plates constantly move away from, towards and parallel to each other. The movements only range from 1 to 20 centimeters a year, but they bring about enormous forces. When they move parallel to each other, or when one plate shifts underneath the other, this does not always go smoothly. When one plate gets stuck behind the other, tension builds up, and eventually, when released, an earthquake will occur:
To measure the movement of such tectonic plates can be difficult, but it is workable. Predicting whenever an earthquake will strike a certain area is much more complicated, however. As many, hard to measure factors are involved with the development of an earthquake, this is an extremely complex task. Think of the diverse densities of landmasses, the changing temperature underneath the plates or the gravitational forces, which differ during planetary cycles. All these factors have an enormous impact on the exact timing of an earthquake.
In the past, scientists have tried to analyse all data concerning previous earthquakes and succeeded to find some sort of pattern. This was to be expected, as many important factors are cyclical. The problem is, however, the accuracy of the predictions. Of what value is a prediction stating that an earthquake will probably strike somewhere in the eastern part of Japan within the next 40 years? For such a prediction to be of any use, is has to be in a small enough time frame so that people can take action, for example to evacuate the area.
Another great problem is brought about by the cost of such safety measures. One can imagine that evacuating a whole metropol and preparing it for a coming earthquake is extremely costly. Thousands or even millions of people have to leave the area and have to be accommodated for several days, if not weeks. A wrong prediction can make for large unnecessary expenditures. Also when the probability of an earthquake occurring is very small, say 3%, should those safety measures be taken at all?
Fortunately, technology may once again bring an answer to the problem. As primary seismic waves are not very hard to record, even most types of smartphones are capable of doing so. With a word-wide network of connected devices able to detect seismic activity, a useful dataset can be made up. The data acquired by cellphones won’t be enough to alarm directly for a coming quake, but will certainly be of tremendous help to analyse geological data. Nasa’s Quakesim software, for example, which identifies regions at risk, would profit substantially from such a crowdfunded data system. Ofcourse, this would only be possible if people are willing to exchange some of their privacy for the lives of others.
Dit artikel is geschreven door Pieter Dilg