Why does Taiwan have so many earthquakes?

Two plates, one island

The Earth’s outer shell is broken into plates that move slowly against one another. Most large earthquakes happen where plates meet. Taiwan is one of those places: the Philippine Sea plate lies to the east and south-east, and the continental margin of the Eurasian plate to the west.

Geologists describe the island as the product of an oblique collision, which began within the last few million years, between the Luzon volcanic arc, carried on the Philippine Sea plate, and the edge of the Eurasian continent. Seen on a map, Taiwan lies between two subduction zones that dip in opposite directions, the Manila Trench to the south and the Ryukyu Trench to the north-east, and the collision in between has pushed up the island’s mountain ranges.

The Central Weather Administration (CWA) has measured this movement directly. In 1993 it began building a network of GPS stations to monitor crustal movement caused by the collision of the Eurasian and Philippine Sea plates. How that network fits into Taiwan’s wider earthquake monitoring is explained on Watching the ground move.

How fast is the squeeze?

One of the first complete pictures came from a survey of 131 stations of the Taiwan GPS Network, measured four to six times between 1990 and 1995 by researchers of Academia Sinica’s Institute of Earth Sciences. Relative to Paisha in the Penghu Islands, on the stable Eurasian side, the island of Lanyu (Lan-Hsu) off south-eastern Taiwan was moving at 81.5 ± 1.3 millimetres a year towards an azimuth of 306°, that is, towards the north-west. The authors noted that this is essentially the same direction as the motion of the Philippine Sea plate relative to Eurasia estimated independently.

Eight centimetres a year is fast by geological standards. The motion is not taken up evenly. The same survey found a marked jump in velocities across the Longitudinal Valley in eastern Taiwan, between the Coastal Range and the rest of the island. Where the rock cannot slide smoothly, strain accumulates over years or centuries and is released in earthquakes.

Why Taipei is different

The collision is not the same along the whole island. A study of GPS measurements from 1995 to 2005 in northern Taiwan describes the north as a “postcollisional” mountain belt: it no longer experiences collision and is moving into the regime of the Ryukyu Trench, where the Philippine Sea plate dives beneath the Ryukyu arc. Towards the north-east, near the Yilan Plain at the western tip of the opening Okinawa Trough, the crust is being stretched rather than squeezed.

In that study, the area around Taipei moves only 1.0 to 7.8 millimetres a year relative to the stable margin, far slower than the east of the island. Taipei is not immune, however. Seismologists studying Taipei 101 point out that long-period ground motion from large earthquakes is a main cause of seismic damage to high-rise buildings, and the soft sediments of the Taipei Basin, described on City & Region, affect how the ground there shakes.

Taipei’s own fault

The western edge of the Taipei Basin is marked by the Shanchiao fault, which runs north–south along the margin of the metropolitan area. Unlike the faults that build mountains, it is a normal fault: one side drops down relative to the other, which is how the basin has been subsiding.

Its past earthquakes are recorded underground. Researchers who drilled boreholes along the fault found evidence of repeated sudden subsidence in the early to middle Holocene, with events roughly 11,000, 9,000–9,300 and 8,400–8,600 years ago, and displacements of 2.3 to 4.5 metres per event. Offsets that large imply earthquakes of about magnitude 6.9 to 7.1, and the authors conclude that large earthquakes on this fault are one of the major seismic hazards for metropolitan Taipei.

A separate GPS-based study estimated that a future earthquake on the Chinshan–Shanchiao fault system could reach a moment magnitude of about 6.85, with a possible recurrence interval of about 335 years. The two studies use different methods and their numbers are estimates with large uncertainties; neither says when the next earthquake on the fault will happen.

Where, yes; when, no

Knowing the tectonics tells seismologists where earthquakes are likely and roughly how large they can be. It does not tell them when. The US Geological Survey puts it bluntly: neither it nor any other scientists have ever predicted a major earthquake, and they do not expect to know how in the foreseeable future. A real prediction would have to give the date and time, the location and the magnitude.

What science can offer instead is a probability for a region over a number of years, and, once an earthquake has begun, a few seconds of warning. Taiwan’s earthquake early warning system, which sends alerts to phones after a strong earthquake starts, is described on Watching the ground move. The earthquake that most shaped that system, the 1999 Chi-Chi earthquake, has its own article: The 1999 Chi-Chi earthquake: what happened and what it changed.

Compared with Poland

Poland lies far from any plate boundary, and it shows. According to the Institute of Geophysics of the Polish Academy of Sciences, the most seismically active area of the country is the Carpathians, which is linked to the young age of these mountains, while the East and West European platforms show little natural seismic activity. The strongest earthquake in Polish history is considered to be the one of 3 December 1786, with its epicentre in the Beskid Średni.

The strongest natural earthquake of the instrumental era struck Podhale on 30 November 2004. The institute gives its magnitude as 4.5 on one of its pages and 4.7 on another. Most tremors felt in Poland are not natural at all but induced by mining, in Upper Silesia, the Legnica–Głogów Copper District and around Bełchatów; the strongest mining tremors reached magnitude 4.5 in the copper district in 1977 and 4.6 at Bełchatów in 1980.

Magnitude is logarithmic: one whole unit means about 32 times more energy released. The 1999 Chi-Chi earthquake in Taiwan had a moment magnitude of 7.6, about three units more than Podhale in 2004. Depending on which Podhale figure is used, that is very roughly 20,000 to 45,000 times more energy; the comparison is only approximate, because magnitudes of small and large earthquakes are often measured on different scales.