Earthquakes: the landslides of the deep

By Carly Faber, researcher, UiT

Carly is a rock detective who hunts for evidence of ancient earthquakes around the world. She completed her PhD in Tromsø and a postdoc at McGill University in Montreal. In this article, Carly provides some background about earthquakes, and then goes on to tell us how she identifies ancient earthquakes in rocks. This tool is also applied to landslides. Next time you are hiking in the Norwegian mountains, you may just be stepping on signs of an old earthquake or landslide!

Like landslides, earthquakes are dangerous natural phenomenon that cause loss of human life and costly damage to infrastructure. And like coastal landslides, earthquakes can cause tsunamis, which compound the seriousness of their effects. Earthquakes can also trigger landslides and rockfalls when they happen in mountainous areas. Humans have been recording earthquakes for nearly 4000 years, and the deadliest earthquake in recorded history occurred in 1556 in Shaanxi, China where it is estimated that around 830 000 people died. The magnitude was around 8. The highest ever recorded magnitude for an earthquake was 9.5 for the 1960 Valdivia earthquake in Chile (Kanamori, 2006). There are several methods to measure earthquakes, and the most common scale used to estimate the magnitude of an earthquake today is the moment magnitude scale (MMS). The MMS measures shaking during an earthquake. The MMS replaces the outdated Richter scale, which cannot adequately measure earthquakes over magnitude 8. The MMS is a log scale, which means that at magnitude 9.5, the Valdivia earthquake shook around 32 times more than the magnitude 8 earthquake at Shaanxi. Norway’s most significant earthquake happened in 1904, in the Oslofjord area. It had a magnitude 5.4, and while there were no casualties, damage was caused to buildings.

Will Norway see an earthquake like the one depicted in “Skjelvet”?

What is an earthquake?

When the Earth’s tectonic plates move, they cause stress to build up in rocks. This happens most along plate boundaries where the tectonic plates move past each other, such as the San Andreas Fault in California, USA.  An earthquake occurs due to build-up of these stresses on a fault that is locked (i.e. not moving). When stresses can no longer be sustained, the sides of the fault slip past each other (failure) and this releases the stresses (like letting go of a loaded spring), causing the generation of seismic waves; an earthquake. It is still not possible to predict earthquakes. If the fault on which an earthquake occurs is one that reaches the surface, then it is possible to see a fault scarp where one side is uplifted relative to the other. If this happens underwater, then the uplifted side displaces water and this generates a tsunami wave. We are lucky in Norway in that we are not situated on a plate boundary, and therefore are not at high risk of significant earthquakes. Small intraplate earthquakes (earthquakes that happen in the middle of a tectonic plate, away from plate boundaries) can occur, but are rare. The Oslofjord earthquake of 1904 is probably associated with small movement of faults in the extinct Oslo Rift Zone, and minor earthquakes happen in the area every year but are generally too weak to be felt. The risk for a large earthquake in the area is low.

How do we study earthquakes?

Earthquakes have traditionally been studied using geophysical methods, such as measuring seismic waves, and by surface monitoring, such as physical surface mapping, or by satellite or airborne monitoring of the movement of the Earth’s surface (e.g. GPS and LIDAR). One of the major problems that earthquake scientists face is that we cannot directly observe an active fault at depth. There are some drilling programs, that have managed to drill through active faults (e.g. SAFOD; San Andreas Fault Observatory at Depth (, but drilling allows only for the study of shallower parts of faults, and they limit study to a very small area of the fault. Geologists therefore study the nearly instantaneous processes that happen on a fault during an earthquake by using lab experiments that recreate earthquake conditions (sliding two blocks of rock against each other in opposite directions at speeds of around 1 m/s), or by interpreting the earthquake processes from ancient faults that have previously experienced earthquakes, but which are no longer active. These ancient faults are now visible at the Earth’s surface due to erosion and tectonic uplift, and they offer “snapshots” of processes that they preserve when they formed or deformed. The trouble is that faults deform over long periods of time, sometimes by creeping (slow non-earthquake sliding), and with deformation from younger earthquakes destroying evidence of older ones. And it is therefore not straightforward to walk up to a fault and to easily be able to recognize direct evidence for an earthquake, nor is it easy to decipher the physical earthquake processes.

Hunting for earthquakes in rocks

The only accepted direct evidence for an earthquake on a fault is the presence of a rock type called pseudotachylyte. Pseudotachylyte forms because the two sides of the fault move so rapidly past one another that the heat generated by the friction causes the rocks to melt. This small amount of melt then cools and solidifies rapidly when the fault sliding ceases. The presence of melt on the fault surface has an effect such that it assists the two sides of the fault in sliding past one another (a type of weakening effect), and it helps to facilitate continued fault sliding. In rocks, pseudotachylytes are found as veins on fault surfaces (e.g. blue arrows on the figure below), and they are often accompanied by injection veins (red arrows on figure below), which are small offshoots of melt injected into the walls of the fault. Injection veins are evidence that the pseudotachylyte was once a melt that could flow. Finding pseudotachylites in the rock record is rare; much rarer than what we expect given how frequent earthquakes are. And it is still unclear whether pseudotachylytes simply do not form in all earthquakes, or if they just do not preserve well in most rocks. In Norway, examples of pseudotachylytes can be found in Lofoten (throughout the islands, but commonly around Nusfjord; e.g. Steltenpohl et al., 2006), western Norway (around Bergen; e.g. Austrheim & Boundy, 1994), and just east of Jotunheim National Park (around Espedal; Dietrichson, 1953).

A pseudotachylyte vein (blue arrows) with its accompanying injection veins (red arrows) from the Fort Foster brittle zone in the Norumbega Fault in coastal Maine, USA. This pseudotachylyte vein is the fossil of an earthquake that most likely happened between 280 and 300 million years ago when North America and Africa were connected as the supercontinent Pangaea.

Since pseudotachylytes are solidified frictional melt, they can be found in other places where melt has been created by frictional heating, for example meteorite impact structures, and even in the basal sliding surface of large landslides. Below is a picture of a pseudotachylyte layer, complete with injection veins, within the basal sliding surface of the Markagunt gravity slide in southwestern Utah, USA. This gravity slide (about 1700-2000 km3 of material) is one of the largest known gravity slides on Earth. It occurred around 21-22 million years ago when part of the Miocene-aged Marysvale volcanic field collapsed. The material is thought to have slid in a single event for more than 30 km over the Miocene land surface (Hacker et al., 2014). The sliding surface in the pictures below separates the dark, brecciated volcanic rocks of the slide (above), from lighter-coloured sandstone (below). This spectacular example shows that pseudotachylytes can also be found in landslides, although they are extremely rare.

The Markagunt gravity slide. Both pictures show the sliding surface (white arrow) containing a thick pseudotachylyte vein (blue arrows), with injection veins (red arrows).

Geological processes are often those that happen over long periods of time. Earthquakes and landslides are catastrophic processes that happen rapidly, but which are affected by more long-term geological processes that continue in between these catastrophic events. Our ability to study these processes hinges on our ability to recognize and describe historical examples. Earthquake-related pseudotachylytes are usually small and difficult to recognize, and next time you step on thin, black, fine-grained, remarkably planar rock vein in the mountains, you may just be stepping on a fossil earthquake.



Austrheim, H., and Boundy, T.M. (1994). Pseudotachylytes generated during seismic faulting and eclogitization of the deep crust. Science, 265, 82-83.

Dietrichson, B. (1953). Pseudotachylit fra de kaledonske skyvesoner i Jotunheimens forgårder, Gudbrandsdalen, og deres dannelsesbetingelser. NGU 184, 23-70.

Hacker, D.B., Biek, R.F., and Rowley, P.D. (2014). Catastrophic emplacement of the gigantic Markagunt gravity slide, southwest Utah (USA): Implications for hazards associated with sector collapse of volcanic fields. Geology, 42, 943-946.

Kanamori, H. (2006). Lessons from the 2004 Sumatra-Andaman earthquake. Phil. Trans. R. Soc. A, 364, 1927-1945.

Steltenpohl, M.G., Kassos, G., and Andresen, A. (2006). Retrograded eclogite-facies pseudotachylytes as deep.crustal paleoseismic faults within continental basement of Lofoten, north Norway. Geosphere, 2, 61-72.

Leave a Reply

Your email address will not be published. Required fields are marked *