Blog

 
The Annapurna range in the Nepal Himalaya. Click for full size.

On April 25, 2015 a major earthquake (Magnitude M_w = 7.8) occurred in central Nepal with an epicenter roughly halfway between the Nepalese capital city Kathmandu and the popular tourist town of Pokhara. On May 12, 2015 a second major earthquake (not an aftershock) occurred a little more than 100 km east, between Kathmandu and the highest mountain in the world, Mt. Everest. This earthquake, with a magnitude of M_w = 7.3, was in many ways similar to the earlier event. With two large earthquakes in the same region in such a short time period, questions naturally arise such as:

• Are these two earthquakes related, and if so, how?
• Are these earthquakes in some way unique, or are they typical for a setting like the Himalaya?
• Why did they occur so closely together, in space and in time?
• Will there be additional large earthquakes? Are there other hazards?

This post aims to provide some insight into earthquakes in the Himalaya and the questions above.

The earthquakes

Before we can consider talking about the questions above, we need to look at both recent earthquakes in a bit more detail. We can start by looking at a map of the locations of the earthquakes that occurred on April 25 and May 12.

 
Topographic map of the region surrounding Nepal with the location of the recent large earthquake epicenters (red, blue stars). Relative convergence between India and Tibet (heavy black arrows) occurs at a rate of ~45 mm/y with most of the ~20 mm/y of convergence in the Himalaya taken up across the Main Frontal Thrust fault. The thin black line marked A-A′ is the location of the simplified cross sections below. The shaded red and blue boxes mark the approximate extent of fault slip during the April 25 and May 12 earthquakes, respectively. Elevation data is ETOPO1 (Amante and Eakins, 2009), fault locations are from HimaTibetMap v1.1 (Taylor and Yin, 2009; Styron et al., 2010) and the map was created using GMT (Wessel et al., 2013). Click for full size.

On the map above, we can see two important things. First, the earlier earthquake on April 25, 2015 occurred west of the earthquake on May 12, 2015. This alone might not seem important, but the fault slip calculations indicate slip on the fault that produced the earthquake began near the mapped epicenter and propagated eastward along a section of the fault ~100 km wide. The estimated maximum amount of fault slip along any section is roughly 3 meters. This is part of the reason that Kathmandu experienced such significant shaking. So, important item #1: Eastward propagation of fault slip. Second, both earthquake epicenters roughly the same distance from the Main Frontal Thrust fault, which is the main active fault accommodating convergence between India and Tibet in the Himalaya today. The Main Frontal Thrust connects at depth to the Main Himalayan Thrust fault, the main fault system uplifting the Himalaya via fault slip atop the Indian craton. This relationship between the Main Frontal Thrust and Main Himalayan Thrust is clearly shown on the simplified tectonic cross-section below.

 
Simplified cross-section through the Himalaya showing the main active faults. The motion of India northeastward into Eurasia results in the Indian craton being pushed downward beneath the Himalaya and uplift of the Himalayan mountains (heavy black arrows). The Main Himalayan Thrust fault system exists between the down-going Indian craton and the Himalaya, and it is known as the Main Frontal Thrust where it intersects the surface of the Earth ~60 km south of Kathmandu. The locations of the recent Nepal earthquakes (red, blue stars) have been projected onto this cross-section to indicate their approximate position relative to the Main Himalayan Thrust. The heavy red and dashed blue lines are the approximate extent of fault slip during the April 25 and May 12 earthquakes, respectively. Cross-section A-A′ location shown on the map above. Cross-section simplified from Avouac, 2003. Click for full size.

The earthquake that occurred on April 25 took place at a depth of ~10 km and the available data suggest fault slip was along a fault dipping gently to the north at ~11°. The location of this earthquake appears to fall directly on the Main Himalayan Thrust fault, and the dip angle of fault motion is consistent with this location as well, suggesting that earthquake is the result of fault slip on the Main Himalayan Thrust. The second earthquake occurred slightly further north, with respect to the Main Frontal Thrust, at an estimated depth of ~16 km and a similar calculation for fault slip on a fault dipping northward at ~9°. This position, though not directly on the Main Himalayan Thrust in the cross-section above, is clearly quite close and within the uncertainty of the position of the Main Himalayan Thrust at depth. The dip angle is slightly shallower than expected based on the cross section, but it is important to recognise that the geometry of the Main Himalayan Thrust may vary from east to west. Regardless, this earthquake appears to have also occurred on the Main Himalayan Thrust, or an associated fault. So, important point #2: Two earthquakes appear to have occurred on the same fault system. Before we address how these important points relate, let's have a look at the active fault systems in the Himalaya and past earthquake locations.

The Nepal earthquakes in the context of Himalayan tectonics

The Himalaya have formed as a result of convergence between the Indian craton and Eurasia, with collision of the continents occurring roughly 50 million years ago. Over that time, a number of major fault systems have been active as the mountain system has grown and evolved. The cross-section below shows a time evolution of the activity of major fault systems in the Himalaya.

Simplified cross-section through the Himalaya showing the main active faults over that past 23 million years. The Main Central Thrust became active around 23 Ma and was mainly active up until ~12 Ma, when deformation stepped south onto the Main Boundary Thrust. That structure was active until around 3 Ma, whereafter the Main Frontal Thrust became active. All three major fault systems connect at depth to the Main Himalayan Thrust, present thrust front in the Himalaya. Cross-section A-A′ location shown on the map above. Cross-section simplified from Avouac, 2003. Click for full size.

As you can see, the major thrust fault systems that have been active throughout the recent history of the Himalaya have each acted to accommodate the relative convergence between India and Eurasia. In this simplified view of Himalayan tectonics, faulting has progressed southward, with each new fault system connecting to the Main Himalayan Thrust at depth.

At present, the Main Frontal Thrust accommodates ~20 mm/y of convergence across the Himalaya. A convergence rate of 20 mm/y might not seem like much, but if we allow the relative motion between India and Tibet to accumulate over 100-300 years (a rough approximation of the time between large earthquakes), we would expect to see something like 2-6 meters of fault slip if the accumulated stress is completely released. Of course, it is not likely the stress is completely released when fault slip occurs, but the point stands that the gradual convergence will produce large earthquakes from time to time. These recent two earthquakes are just part of the ongoing evolution of the Himalaya and the natural result of tectonic plate convergence.

Moderate and larger earthquakes in the Himalaya since 1976

As we have seen above for the recent Nepal earthquakes, when earthquakes occur in regions with enough seismometers, scientists can calculate the orientation of the transmission of the seismic energy (the earthquake). In other words, we're able to figure out the orientation of the fault where the earthquake occurred, which is quite useful, of course. The figure below shows the typical way in which these fault plane solutions are displayed for earthquakes of magnitude M_w = 5 and greater since 1976. If you're looking for some information about how to interpret the symbols, you may want to have a look at the IRIS page on focal mechanisms.

 
Topographic map of the region surrounding Nepal with the locations of focal mechanisms calculated for earthquakes of magnitude M_w = 5 and greater since 1976. Focal mechanisms are colored by depth, and the majority of Himalayan earthquakes clearly occur at depths of 11-20 km. The two larger focal mechanisms in black are from the recent Nepal earthquakes. Elevation data is ETOPO1 (Amante and Eakins, 2009), fault locations are from HimaTibetMap v1.1 (Taylor and Yin, 2009; Styron et al., 2010), focal mechanism data are from the Global Centroid-Moment-Tensor (CMT) Project (Dziewonski et al., 1981; Ekström et al., 2012) and the map was created using GMT (Wessel et al., 2013). Click for full size.

Looking at the earthquake fault plane solutions, a few things are clear. First, most of the fault plane solutions suggest thrust-sense fault motion on shallowly dipping fault planes that strike roughly parallel to the Himalayan front. If this doesn't make any sense, perhaps having another look at the IRIS page on focal mechanisms would be a good idea. To say this another way, most of the earthquakes shown above are smaller than magnitude 7, but otherwise similar to the two recent Nepal earthquakes. They occur with fault motion consistent with slip on the Main Himalayan Thrust (or other associated thrust faults) with an orientation that suggests the fault dip angles are shallow 5-20°, again consistent with the geometry of the Main Himalayan Thrust. This appears to be the norm for earthquakes of magnitude M_w ≥ 5 in the Himalaya, as we might expect in this tectonic setting. Second, we can observe that most of these earthquakes occur at depths of 11-20 km. These depths are both within the range of the depths of the Main Himalayan Thrust across much of the southern Himalaya, and clearly within the temperature range in the crust corresponding to the seismogenic zone, the part of the Earth's crust where most earthquakes occur. Generally speaking, this corresponds to regions with temperatures of ~300°C or below in quartz-rich continental crust. The picture gets more complicated when considering earthquakes of smaller magnitude or looking at those within the Tibetan Plateau, but for the larger earthquakes in the Himalaya it appears most are related to thrust faulting on relatively shallow faults.

Earthquakes consistent with numerical model predictions

As mentioned above, the fault plane solutions for fault slip in the Himalaya generally suggest slip on thrust faults dipping gently north with thrust motion roughly perpendicular to the mountain front. This observation is true both for the historical earthquakes since 1976, and for both recent major earthquakes in Nepal. This observation is also consistent with the fault motion that is predicted in 3D mechanical numerical models of mountain systems similar to the Himalaya. An example of this type of model is shown in the figure below.

 
3D mechanical model of the western half of a Himalaya-like mountain system with neighboring plateau. The upper panel shows an example 3D velocity calculation with active shear zones (model equivalents to faults) indicated with blue to pink colors. The lower panel shows the model location of the Main Frontal Thrust and Main Himalayan Thrust (MHT), as well as the northward motion of India and resulting uplift of the Himalaya. These models are similar to those described in detail in Whipp et al. (2014). Click for full size.

As we can see in the model calculation above, the relative convergence between India and Tibet is accommodated by the Main Himalayan Thrust (MHT), which dips gently northward beneath the model equivalent of the Himalaya. The bright pink color of this shear zone indicates it is the main active structure (or fault system) in the model, where earthquakes would be expected to occur. In this case, the model is designed with a weak region beneath the Himalaya, where the shear zone develops dynamically. Details about this type of model are available in Whipp et al. (2014).

Looking down on the top of a similar model, we can see several additional interesting features.

 
Planform (map) view of calculated velocities and shear strain rates for the western half of a Himalaya-like mountain system. These model results are based on Whipp et al. (2014). Click for full size.

First, the velocity vectors in the lower part of the figure show the northward movement of India relative to Tibet and the wide, curved pink region is the model equivalent of the Main Frontal Thrust. On both the left and right sides of the model, the convergence velocity decreases from 20 mm/y to zero across the Main Frontal Thrust, clearly indicating it is the only active fault in those regions. In this reference frame it may be hard to see, but this reduction in velocity also predicts that any fault slip would be perpendicular to those sections of the mountain front. In the central region, where the mountain front is oblique to the incoming motion of India, the velocity vectors do not decrease to zero across the Main Frontal Thrust, but instead a component of velocity parallel to the mountain front remains within the model equivalent of the Himalaya. This also may be difficult to envision, but if you consider that the incoming Indian velocity vector has both a component parallel and perpendicular to the mountain front, the residual velocity in the model Himalaya results from a decrease in the component perpendicular to the front. In other words, here too we expect thrust motion to occur perpendicular to the mountain front. This is particularly nice to observe because it is consistent with the observations from both major Nepal earthquakes and the majority of the fault plane solutions shown in the map above. Again, if you're interested in more details about these models, you should check out Whipp et al. (2014).

The final model comparison we'll consider has to do with the region across which earthquakes are observed in the Himalaya.

 
Calculated velocity vectors and temperatures for a model cross-section in the Himalaya of Bhutan. The 300°C isotherm is indicated by the heavy black line and the region at temperatures below 300°C is likely entirely in the seismogenic zone, capable of producing earthquakes. As can be clearly seen, nearly all of the southern part of the Himalaya and the Main Himalayan Thrust fault reside within the seismogenic zone. The thermal model is from Coutand et al. (2014). Click for full size.

Above, we see a thermal model calculation for a cross-section through the Himalaya of Bhutan. Bhutan is 200-300 kilometers east of the recent earthquake locations in Nepal, but the general thermal model calculations above are still likely representative of the central Himalaya. What we can see is that much of the Main Himalayan Thrust down to 15-20 km depth is at temperatures below 300°C, and likely entirely within the part of the crust where earthquakes would be expected (the seismogenic zone). This means that fault motion along most of the shallowly dipping sections of the Main Himalayan Thrust would be predicted to produce earthquakes like those recently observed in Nepal. Again, we see consistency between the observed locations of earthquakes and their fault plane solutions (mostly between 11-20 km depth) and the predictions from thermal models in the Himalaya. The picture is becoming clear...

As a quick note, the model above is from Coutand et al. (2014) and was part of a study to try to determine the geometry of the Main Himalayan Thrust at depth using mineral dating of surface rocks. If you're curious, I urge you to have a look at the paper.

Returning to the questions - What can we say?

At this point we can address some of the questions posed at the start of this post.

• Are these two earthquakes related, and if so, how? 
  It seems the answer is yes. Though it is hard to be certain, it seems reasonable to suggest that the two recent Nepal earthquakes are related. The first earthquake propagated eastward from its epicenter about 100 km and the area where the propagation stopped is right around the location of the second earthquake. If we consider that both earthquakes appear to have occurred on the same fault system (the Main Himalayan Thrust), it is likely that the region that experienced the second large earthquake could "feel" the effects of the earlier fault slip. At the location where fault slip propagation stopped, it is quite likely that stresses on the fault segment that had not slipped were increased, putting this region of the fault under greater stress and an increased likelihood of reaching the stresses required for fault slip to occur. It appears that the fault stresses in the region of the second earthquake must have been just below the critical stress needed for fault slip after the first earthquake, resulting in slip only two weeks afterward. This kind of earthquake activity is somewhat similar to what was observed in the great Sumatran earthquakes of 2004 and 2005.
• Are these earthquakes in some way unique, or are they typical for a setting like the Himalaya? 
  They're both quite big, but otherwise pretty typical. The Himalaya have formed along the tectonic plate margin between the Indian and Eurasian plates, which currently converge at a rate of ~45 mm/y. The Main Himalayan Thrust fault is the current main active fault beneath the Himalaya and appears to be the location of both recent Nepal earthquakes. Fault motion and the resulting earthquakes of this type are predicted by both mechanical models of Himalaya-like mountains and the temperatures in the crust in the Himalaya.
• Why did they occur so closely together, in space and in time? 
  This is a good question. At this point, it seems that the earlier April earthquake may have released stress on part of the Main Himalayan Thrust fault, but increased stresses to the east, where the second earthquake occurred. This is a logical prediction for two earthquakes apparently on the same fault system and in close proximity to one another. In effect, one earthquake may have loaded the Main Himalayan Thrust locally and resulted in the second soon afterward. As noted above, this type of activity is not entirely unusual on large thrust faults (see question 1), and if we can take anything positive away from how these two earthquakes occurred, it is that we can be grateful that there were two separate earthquakes and not a single larger, continuous rupture that could have produced an earthquake of magnitude M_w > 8.
• Will there be additional large earthquakes? Are there other hazards? 
  Absolutely there will be more large earthquakes in the future in the Himalaya, but the big question is when. I'm not a seismologist, but one of the holy grails of seismology would be the ability to predict earthquake occurrences. In a setting like the Himalaya, earthquakes are part of life, but we still cannot be sure when the next big earthquake will occur. For now, it seems much of the stress on the MHT beneath Kathmandu may have been released, but looking forward, there is no reason to think there won't be another large earthquake in the future. It may not be likely to occur tomorrow, but earthquakes occur at tectonic plate margins and we as a society should do our best to acknowledge this reality and prepare accordingly. Other than the possibility of some significant aftershocks, which should decease over the coming weeks, landslides are likely the largest risk in the region affected by the recent earthquakes. The earthquakes produced a number of significant landslides and may have put some areas at greater risk for future landslides. With the coming monsoon season, it is possible that some of these areas may be affected by large landslides. Furthermore, rivers have been dammed by landslides in some areas, and the landslide dams may fail during the heavy rains of the monsoon, potentially flooding regions downstream. If you're interested, there's a great blog about landslides by Dave Petley, and Marin Clark, Sean Gallen and Nathan Niemi at the University of Michigan have performed a landslide hazard analysis following the recent Nepal earthquakes.