Last Thursday, Ed Nissen from the Geophysics department crossed Kafadar to deliver a talk on earthquakes to the geologists of Berthoud Hall. Nissen’s research focuses broadly on faulting and LIDAR applications. (LIDAR is a type of satellite mapping that penetrates through ground cover, such as trees, to give detailed bare-Earth topography.) He spoke about the Zagros Mountains, which range across the entire western border of Iran, specifically focusing on an island at the southernmost end of the range called Qeshm Island.
Nissen began by differentiating oceanic plate-boundary tectonics from intracontinental tectonics. Most world tectonism, meaning any deformation of the Earth’s crust, including folding and faulting, but referring mainly in this case to seismic slip along faults, occurs along the boundaries of the plates which make up the Earth’s crust. If a map of measured earthquakes is overlaid on a map of the plates, seismic activity clusters tightly where one plate meets another. The majority of this activity is in subduction zones, where an oceanic plate is being overridden by a continental plate. Regions like this cause the largest earthquakes, including the one that instigated the 2004 Indonesian tsunami, the quake responsible for the 2011 Fukushima nuclear meltdown, and the single most powerful earthquake ever recorded, a magnitude 9.5, which occurred in Chile in 1960.
Where two continental plates hit each other, however, the character of seismicity is different. These earthquakes do not cluster along the plate boundary, as in a subduction zone or along an oceanic spreading ridge. Instead, they make a diffuse cloud, sometimes occurring thousands of kilometers from the collision zone. Furthermore, while most of the seismogenic (that is, earthquake-causing) faults in a subduction zone are normal or reverse, steep faults with vertical slip, the faults in a continent-continent collision zone can be of any kind. Continental earthquakes also tend to be of smaller magnitude than subduction zone quakes, but the death toll and damage resulting from continental quakes, especially when the effects of tsunamis are ignored, and only deaths and damage caused directly by the earthquake itself considered, is greater by far than that resulting from oceanic-boundary quakes.
The main locus for continental earthquakes on this planet is a collisional band known as the Alpine-Himalayan Earthquake Belt, which ranges from Turkey to Nepal. Here, the Eurasian plate in the north contacts the African, Arabian, and Indian plates to the south. Iran itself comprises the collision zone between Arabia and Eurasia. The country’s mountainous topography is an analogue for its seismicity, as most continental faulting is related to mountain-building. Indeed, the modern mountain-building now going on in this area can tell a lot about the formation of ancient, now extinct mountain belts, such as the Appalachians and even the Front Range. Iran’s mountains were formed starting around 20 million years ago (MYA), when the Arabian plate, having separated from the African plate about 200 MYA, struck Eurasia. This mountain range is far from done growing, as can be seen in the extreme tectonic activity of the area. Short of Nepal, Kyrgyzstan, and the Phillipines, the country has the highest uniform risk of earthquake in all Asia; global risk assessment places it the third most vulnerable to natural disaster of any country in the world. In other words, nowhere in Iran is safe. Iran held the earthquake with the highest percentage of deaths in the area affected, when 85% of the population of the city of Tabas were killed in 1978. Over 120,000 people have been killed by earthquakes in Iran in just the last century. In the 9th century, two earthquakes killed more than double that number; and the city of Tehran, which has a current population of 12 million, has already been destroyed four times by earthquakes. Therefore, research into the causes and nature of continental quakes, especially in this region, is of the greatest imperative.
The Zagros mountains, where Nissen did his research, have the densest concentration of earthquakes of any mountain range in the world. Yet these quakes can account for only 20% of the crustal shortening observed in the region. This means that the rest of the shortening must be taken up by other means, either aseismic slip (movement along a fault which does not result in an earthquake) or ductile deformation (folding of the crust, rather than breaking). Determining the answer to this mystery is no easy prospect. For one thing, unlike in most other places with major seismic activity, the quakes here do not produce surface ruptures, only minor bedding-plane slip, rather than causing a visible fault scarp, the only surface expression of seismicity is in the form of small extension cracks on the ground. This means that the local faults are blind, or buried beneath the ground surface, and makes the determination of their exact nature quite difficult. Further complicating matters is the lack of data available: only four seismic reflection profiles of the region have ever been released to the public by the Iranian national oil company, meaning there is essentially no data available on the nature of the subsurface, which seismic reflections image. The result is that multiple interpretations abound, with little agreement and no good constraints.
Nissen used a satellite technology called InSAR to measure the phase shift in the ground caused by a magnitude 6.0 earthquake on Qeshm Island, in the SFB. InSAR, which has centimeter-scale resolution and an error of only a single centimeter, produces a 3D image called an interferogram, showing how the ground has shifted up or down in response to a quake. An interferogram with a bullseye pattern, such as Nissen and his colleagues found on Qeshm, indicates a buried fault. The wavelength of the bullseye rings allows determination of the depth of the fault, and detailed modelling allows the researchers to make a good estimate of the fault’s orientation. Following these analyses, Nissen discovered something unexpected: the underground rupture which cause the Qeshm earthquake was centered much shallower than anticipated, at a depth of about 5-8 km, putting it within the competent group of rocks. From this, he inferred the group must contain a weak unit along which slip could have occurred, such as a marlstone.
In order to constrain the depth of the fault, Nissen’s team set up a microseismic measurement network in the area to keep track of aftershocks. They were unable to complete this set-up until about a month after the main quake, so only small aftershocks were registered; the largest ones would have occurred shortly after the original earthquake. The measurement network used the temporal separation between the arrival of the two waves created by each aftershock, P, or pressure waves, which move faster, and S, or shear waves, which move slower, to pinpoint the quake’s source and ultimately create a model of the underground structure of the crust. Because aftershock focii tend to cluster around the original rupture, Nissen was surprised to find the aftershocks measured by his network were concentrated far deeper than the original quake, about 10-20 km depth.
The measured slip in the competent group, Nissen reasoned, had been triggered by a much deeper, aseismic crustal rupture. He decided to reexamine his original depth calculations for the main quake, just to be certain the shallow result was not the product of human error. To do this, he used teleseismic data, or the measurements of the waves produced by the quake as collected from distant seismometers. Using a process called teleseismic body-waveform modelling, he was able to constrain the depth of the Qeshm earthquake to between 6 and 10 km depth, a result that agreed with the InSAR data. Finally, he compared the depths calculated for the main quake and the aftershocks against the single available seismic reflection profile for this region. This placed the aftershocks directly in the midst of the Hormuz salt. Because salt flows in response to pressure, it does not fault, therefore, it could not be the source for an earthquake. Yet the Hormuz contains many large blocks of entrenched Cambrian rocks, and flowing of the salt caused by a large earthquake could break up these kilmeter-scale inclusions, thus inducing the observed microquakes. Nissen did one final analysis to see if this hypothesis was valid.
Generally, the magnitude of an earthquake is directly related to the size of the rupture zone that causes it, though the rupture zone may be smaller than the fault itself. As the measured aftershocks ranged in magnitude from 1.0-4.0, Nissen was able to apply this relationship and derive a maximum rupture size of about 1 kilometer. This matches the observed sizes of entrained rock blocks within the Hormuz, thus supporting Nissen’s hypothesis as to the source of the aftershocks. An earthquake the size of the Qeshm quake, on the other hand, a magnitude 6.0, gives a rupture size of about 10 kilometers. This is smaller than the total thickness of sediments here (~15 kilometers), which means that the seismic portion of the quake did not involve the full thickness of the crust. This suggests that the Hormuz salt serves as a sort of decoupling agent, a fault rupture cannot propagate through the salt, so an earthquake occurs either in the units above or below the salt, but not both. Because the rupture’s size is thus limited, the Hormuz also serves to limit the maximum magnitude of an earthquake in this area. Indeed, when only the rock units above the salt are considered, the area’s seismicity accounts for most of the local crustal shortening. Though the area is seismically active down to 20 kilometers depth, 75% of this activity occurs in the shallow crust (meaning the sub-salt units are not involved) and only two seismic events have been of magnitude greater than 6.0. Both of these large quakes occurred in association with steep, asymmetrical folds, suggesting that the salt has been locally evacuated from the subsurface. Without the salt as a buffer, rupture size is uninhibited, allowing for much larger earthquakes. If the salt distribution throughout Iran can be better mapped, the earthquake risk can be determined with much better resolution.