Updates on the M7.7 Myanmar earthquake

An unusually long, possibly supershear rupture

ဤစာကို မြန်မာဘာသာဖြင့် (Google မှ အလိုအလျောက် ဘာသာပြန်ဆိုရန်) ကိုဖတ်ရန် ဤနေရာကိုနှိပ်ပါ။

Citation: Bradley, K., Hubbard, J., 2025. Updates on the M7.7 Myanmar earthquake. Earthquake Insights, https://doi.org/10.62481/9e49eb4a

Earthquake Insights is an ad-free newsletter written by two independent earthquake scientists. Our posts are written for a general audience, with some advanced science thrown in! To get these posts delivered by email, become a free subscriber. If you would like to support our work here, please also consider a paid subscription. If you have recently lost your job or are a student without ability to pay, let us know and we will upgrade your subscription to “paid” at no cost.

This post is only possible because scientists around the world are rapidly working to produce results and sharing those results with the world via social media. In particular, we would like to acknowledge Dr. Zhe Jia, Dr. Jascha Polet, Dr. Hongjun Si, and Dr. Dun Wang; their colleagues; the scientists at the USGS Geologic Hazards Science Center; and the scientists at the Thai Earthquake Observation Division, whose data and results are shown and discussed below, or who alerted us to new results.

Please leave a comment below if you have any new results to share, or any additional insights.

News about the M7.7 earthquake on March 28 in Myanmar continues to become available. As expected, the death toll has climbed — currently standing at >1,700. This likely represents a major underestimate, given the severity of the disaster and the limited capacity for response. The USGS PAGER currently estimates a 67% chance of >10,000 fatalities. Many images showing widespread destruction have emerged. In addition to impacts on buildings, damage has occurred to airports, roads, and bridges, including collapse of the control tower at Naypyitaw airport, severely hindering transportation and rescue efforts.

After destructive events, there is an apparent tension between the need for immediate humanitarian assistance, and the desire to quickly understand what happened and what we can now learn about the Earth. However, these are not contradictory needs. Every large earthquake helps to fill in our understanding of the basic processes that create hazard. We think that the more that scientists and the general public can share a clearer vision of the realms of possibility, the more likely it is that we will actually use Earth’s greatest gift — the long periods of quiescence between damaging earthquakes — to improve our resilience and limit the impacts of future earthquakes globally.

In this post, we want to talk about some of the big-picture questions that earthquake scientists are considering right now. We also want to be clear about why these questions are important, for science and for society. Some of these questions will surely be answered within the coming days, and some may take much longer to address.

Here are a few of the big questions earthquake geologists and seismologists are thinking about right now:

1. How long was the fault rupture?

2. How fast did the rupture grow?

3. What are the implications for damage?

Let’s jump in. As always, we are discussing preliminary information that is likely to change. As always, please leave comments to identify errors or suggest improvements, or to alert us to new information.

How long was the fault rupture?

While earthquakes are often shown as points on maps, representing the location where the rupture started (the hypocenter), they actually arise from slip on 2D surfaces within the Earth’s crust. For a small earthquake, a dot is usually an okay representation. However, for a very large earthquake, the rupture might be tens, hundreds, or even over a thousand kilometers long. When considering earthquake impacts, it then becomes important to consider not just the distance from the epicenter, but the distance from the fault surface. A building located hundreds of kilometers away from the hypocenter might still be directly on top of the ruptured fault! (A note: the hypocenter is the location where the earthquake started, in three dimensions. The epicenter is the projection of the hypocenter to the surface, i.e. where it is shown on a map.)

The Sagaing Fault, which ruptured in Myanmar, is well known and fairly well mapped at the surface. However, because there are very few scientific instruments operating in Myanmar, it has proven quite difficult to quickly determine how much of the fault actually ruptured in Friday’s earthquake.

The initial estimates of fault slip were calculated using only seismic readings from far-away seismometers. These data allowed measurement of the magnitude of the earthquake relatively easily. They also provided a constraint on how long it took for the rupture to happen.

The earliest USGS rupture model estimated up to ~5 meters of slip of over a ~200 kilometer span, reaching from ~22.5°N to 20.5°N, over a duration of about 90 seconds. This number was consistent with a rupture propagating through the ground at a typical speed of ~2 kilometers per second (about six times as fast as the speed of sound through air).

Based on past earthquakes, we can generally expect that a M7.7 earthquake would arise from rupture of a ~200 kilometer long section of fault, so this number made a lot of sense. The images below are based on the classic 1994 paper by Donald Wells and Kevin Coppersmith, and show how magnitude relates to rupture length, based on a set of recorded earthquakes.

Figure 1: Empirical relationships between earthquake magnitude and (left) surface rupture length and (right) subsurface length. Annotations to highlight a M7.7 earthquake added by us. Figure 9A and 14A of Wells and Coppersmith (1994).

However, newer seismological observations strongly suggest that the rupture was longer. As of this writing, the model has been updated to ~6.5 meters of maximum slip on a fault ~270 kilometers long, and a number of other preliminary results suggest an even longer rupture (keep reading). The map below shows the USGS slip model next to the trace of the fault in map view. The model includes slip down to ~20 km depth on a fault dipping 75° to the east. Note that the slip model is not consistent with the map of shaking, which is also a USGS product; we address this below.

Figure 2: Map of estimated shaking (colors) and reported felt shaking (colored squares) from the USGS. Dark gray line shows location of the USGS slip model, which is shown in the offset box. The left side of the slip model at ~4 km depth; the right side is at 20 km depth.

A long, supershear rupture?

We have a reasonably good constraint on the amount of time it took for the rupture to complete. This can be read from seismograms. The longer the shaking and the wider the impulses on the seismogram, the longer the rupture likely was.

Figure 3: Source time function for the Myanmar earthquake, which describes the rate of moment release, relative to the peak moment rate. The event completed within 80-90 seconds. Source: USGS, https://earthquake.usgs.gov/earthquakes/eventpage/us7000pn9s/finite-fault

By combining the rupture length with the total time, we can assess the speed of the rupture. In this case, if the rupture of the Sagaing Fault was much longer than ~250 km, then the average rupture speed must have been faster than usual. For example, if the rupture front traveled 500 kilometers in 90 seconds (hypothetically), then its average velocity would have been 5.5 kilometers per second. (We note that since the rupture can propagate in two directions simultaneously, it is the distance from the epicenter to each end of the rupture that matters. Here, the rupture was mostly southward, but this is important to keep in mind.)

There is a special class of fault ruptures that are called supershear events. During these events, the rupture propagates faster than the velocity of S-waves in the surrounding rock. S-waves, also known as shear waves, are a form of side-to-side shaking that travels through rock much more slowly than the forward-to-backward P-waves (pressure waves). (The strongest shaking typically comes later, from the slower moving surface waves.) So, supershear simply means faster-than-shear-waves.

Much like a supersonic jet creates a sonic boom, a supershear rupture creates a Mach wave: areas in front of the rupture will be simultaneously hit by a wave of superimposed S-waves emanating from different rupture moments.

Figure 4: Diagram showing how a supershear rupture creates a Mach wave. Figure 2ab of Yue et al. (2013). https://doi.org/10.1002/2013JB010594

Supershear earthquakes were theoretically proposed based on laboratory experiments and some scattered field observations, long before they were actually observed. The first earthquake for which supershear was proposed was the 1979 Imperial Valley earthquake in California, which was too small magnitude to be really convincing. Starting with the 1999 M7.6 Izmit earthquake in Turkey, increasingly confident observations have demonstrated that supershear earthquakes really do happen. Since then, supershear has been confidently identified in less than twenty earthquakes, all of them greater than magnitude 6.9, and all on strike-slip faults. About one in six large strike-slip earthquakes (excluding very deep events) seem to have at least a component of supershear — with a bias towards events with good data, suggesting that many events without identified supershear may simply not have good enough data to see it.

There has been a lot of interest in supershear rupture: scientists want to know what physical factors control it, how it starts and ends, and — most critically for society — how might it impact seismic hazard. These questions all bring us back to the recent M7.7 earthquake in Myanmar.

We know that the Myanmar earthquake originated in the north, near Mandalay. To roughly estimate whether a significant part of the fault could have ruptured at supershear velocities, we really need a better sense of how long the total rupture was, especially towards the south. A 200 kilometer rupture could occur in 90 seconds at a little over 2 kilometers per second — generally below the typical S-wave velocities for the crust. S-wave velocities vary depending on rock type, but might typically be ~3-4 kilometers per second.

Fortunately, seismologists have already been working to estimate the rupture length using a couple of advanced techniques. A few early results are available.

Preliminary back-projection results

One way that we can understand what happened in an earthquake is back-projection. This technique uses giant, far-away seismic arrays to track the locations of high-frequency bursts of energy produced during the rupture. The width of these arrays creates a useful aperture, within which the small differences between similar seismic waves can be used to extract detailed information that could not be estimated from only a few stations.

Because these seismic networks are located far away from the earthquake source, the absolute locations of the energy releases are hard to estimate, but their relative locations can be determined pretty reliably. The results from different arrays covering different parts of the globe can also be compared to assess how robust the observations are.

A preliminary back projection study by Dr. Dun Wang, shared by Dr. Hongjun Si, indicates that high frequency energy was radiated from along the Sagaing Fault between 18°N and 23.5°N — a distance of almost 600 kilometers! (Each degree of latitude represents ~111 kilometers, making it easy to convert between latitude and distance for this north-south fault.)

That may be an overestimate, because the study overlays results derived from from three independent seismic networks (Europe, Alaska, and Australia), and some of the data points on the edges are based on only one or another network. Still, the overlap from all networks spans from ~18.5°N to 22.5°N: about 440 kilometers. Most of that distance is to the south of the epicenter.

Figure 5: Back-projection results from three seismic networks, colored by the relevant network. Star shows epicenter of the Myanmar earthquake. Credit: Dr. Dun Wang.

Preliminary sub-event inversion

This result is pretty consistent with a sub-event inversion by Dr. Zhe Jia. A sub-event inversion treats an earthquake as several different energy sources rather than a single idealized source. Because ruptures often evolve in pulses, as the rupture front propagates through onto new parts of a fault, this can help illuminate where and when those different parts ruptured. Jia found that the earthquake can be modeled in two ways. The first option is similar to the initial USGS model: an ~180 km long rupture that occurred at normal rupture speeds (~2 km/s). The second option is a much longer rupture (~350 km), that must have occurred much faster (~5 km/s).

Jia writes:

Only the long supershear model accurately predicts early S-wave arrivals recorded at three separate strong-motion stations (particularly station YGN, see images). In contrast, the subshear scenario underestimates the timing of these arrivals by a wide margin.

Figure 6: Two options for the sub-event inversion. Option 1: a subshear rupture along a fault ~180 km long. Option 2: a supershear rupture along a fault ~350 km long. Option 1 does not fit the S-wave arrival times at three stations (including Yangon, to the south). Source: Dr. Zhe Jia.

Aftershocks

Finally, Dr. Jascha Polet pointed out on Bluesky that Thailand has a publicly accessible online earthquake catalog from their nearby network, which has monitored many aftershocks over the last days. These data are much more reliable than locations from global networks, which cannot generally see small earthquakes in this region. We followed her lead and downloaded the data, plotting a cross section and a time section to see what we could learn.

Figure 7: Mainshock-aftershock sequence for the 2025 Myanmar earthquake. Earthquakes are colored by time, and also projected to the right onto a timeline. A cross-section is shown at the bottom of the image, showing seismicity for a distance of ~400 km south of the mainshock. Note that the magnitudes here are somewhat higher than the USGS magnitudes; the Thai network uses a different scale.

Aftershocks can occur on a fault that has ruptured. They can also occur on nearby fractures and faults in the crust that have been stressed. For large earthquakes, we usually see that aftershocks are most concentrated around the parts of the fault that slipped, forming a sort of halo. This lets us use early aftershock distributions to try to infer the extent of the slip area.

The early aftershocks of the M7.7 earthquake illuminate a north-south length ~500 kilometers long. This suggests that rupture did propagate quite far to the south, perhaps as far as 18°N. This is consistent with the other early results that suggest quite a long rupture.

As also noted by Dr. Polet, there is a ~100 km long region between ~20.5°N and 21.5°N with fewer aftershocks. There are several possible causes for this. Areas of a fault that experienced supershear rupture will commonly present two different symptoms: (1) large rupture at the surface, and (2) a lack of aftershocks. So there may be a hint of supershear rupture behavior in the aftershocks. Alternatively, differences in the geology of the fault, the slip distribution, the distribution of afterslip along the fault, or other factors could cause an aftershock window.

An unusually long rupture

The most recently updated shaking model by the USGS now shows the zone of highest intensity shaking reaching from 18.8°N in the south to 23.1°N in the north — a distance of ~480 kilometers. This is longer than the current USGS slip model, because the shaking model takes into account more types of data, including damage reports and (in this case) the back-projection results above by Dr. Wang. (As far as we know, the sub-event inversion, which came out later, is not currently incorporated.) In contrast, the slip model uses only specific seismic data: according to the USGS website, 29 teleseismic broadband P waveforms, 29 broadband SH waveforms, and 71 long period surface waves.

Figure 8: Map of the Myanmar earthquake mainshock and aftershocks (Thai network). Map also shows faults, colored by type, and contours of slip intensity estimated by the USGS.

It is important to remember that different data products about earthquakes rely on different assumptions in the face of our ignorance about important factors. That is normal — we need to keep our minds open and look at many perspectives to get a better understanding.

From our perspective, it now looks like the rupture was actually quite long. Based on the scaling relations discussed earlier, a 450 kilometer rupture would usually produce a M8.1 earthquake. That difference in magnitude may not seem like much compared to the measured M7.7, but it is actually a factor of ~4 difference in earthquake moment (a value similar to energy release).

One of the basic equations in earthquake science relates fault dimensions and slip to earthquake moment:

\( Moment = rigidity *area * slip\)

If the fault was twice as long as normal, the difference could be made up by a thinner-than usual fault (not extending as deep into the crust), lower average slip than usual, and/or unusually low rigidity. We should eventually have pretty good estimates of the slip area and slip distance from independent observations. At this point, we leave this note simply as a question mark.

Implications for damage

The immediate implication of a long rupture is that many more areas, including the capital city of Naypyitaw south of the epicenter, were likely located very close to the rupture itself. This means that these areas experienced high shaking intensities, as well as direct fault-related deformation. In fact, the current USGS PAGER estimates that nearly 6.5 million people probably experienced either intensity VIII or IX shaking (severe or violent) — nearly twice as many people as in the two huge earthquakes in Turkey in 2023 combined.

Supershear rupture, which seems to be implied for at least part of this long rupture, has additional but less well understood implications.

Dr. Zhe Jia writes:

Supershear earthquakes often produce stronger ground shaking over a broader area, due to high rupture speeds that concentrate seismic energy ahead of the propagating front (a “Mach cone” effect). This means damage can be significantly more severe than expected for a “standard” M7.7 event, potentially impacting communities hundreds of kilometers from the epicenter.

This may have caused exceptionally high shaking in areas south of the epicenter, and may in part explain some of the unusually high shaking observed in Bangkok. This will have to be studies using actual models of the shaking. Note that while Bangkok is located ~1,000 kilometers south of the epicenter, the southern tip of the rupture was quite a lot closer (~600 kilometers).

Of course, it is also possible that damage in Bangkok was exacerbated by site conditions. The amplitudes of seismic waves typically grow as the waves pass through softer sediments, a process known as amplification. In some cases, seismic waves with specific wavelengths can resonate within a subsurface basin. Buildings with the same resonance frequency can then sway more than expected, and become damaged. This process famously happened in 1985 in Mexico City due to an earthquake ~350 kilometers away. We do not yet know if this happened in Bangkok.

The remote effects of large earthquakes can be quite puzzling. On that note, we will end with a quick observation that fell directly out of plotting the Thailand earthquake catalog. The border area between Myanmar and northern Thailand is seismically active. This region, which is about 200 kilometers from the Sagaing Fault, is part of the Shan Plateau. The Shan Plateau is cut across by an unusual array of curved strike-slip faults, which are an intriguing part of the giant India-Asia collision zone, and which are themselves hazardous. We noticed that the border area along the Salween River experienced a sudden uptick in small earthquake activity following the M7.7 earthquake (the dark red circles; see in particular the timeline to the right):

Figure 9: Thailand/Myanmar border region, southeast of the 2025 M7.7 epicenter. Earthquakes are colored by time and also projected to the right onto a timeline. Dark red earthquakes are aftershocks of the M7.7 earthquake.

These new events are small, mostly M2-M3, although one M4.1 did occur. Triggered seismicity is pretty common globally after large earthquakes. This particular cluster of events is fairly mysterious, because it appears to be spread out over a large area. We will hold off on any interpretation for now, which would presumably require better understanding of the giant rupture to the west. If you have any information on the seismicity of this region, or this triggered seismicity, please let us know!

References:

Archuleta, R.J., 1984. A faulting model for the 1979 Imperial Valley earthquake. Journal of Geophysical Research: Solid Earth, 89(B6), pp.4559-4585. https://doi.org/10.1029/JB089iB06p04559

Bouchon, M., Bouin, M.P., Karabulut, H., Toksöz, M.N., Dietrich, M. and Rosakis, A.J., 2001. How fast is rupture during an earthquake? New insights from the 1999 Turkey earthquakes. Geophysical Research Letters, 28(14), pp.2723-2726. https://doi.org/10.1029/2001GL013112

Hubbard, J. and Bradley, K., 2025. Catastrophic M7.7 earthquake caused by rupture of Sagaing Fault in Myanmar. Earthquake Insights, https://doi.org/10.62481/9250a38a

Rosakis, A.J., Samudrala, O. and Coker, D., 1999. Cracks faster than the shear wave speed. Science, 284(5418), pp.1337-1340. https://doi.org/10.1126/science.284.5418.1337

Socquet, A., Hollingsworth, J., Pathier, E. and Bouchon, M., 2019. Evidence of supershear during the 2018 magnitude 7.5 Palu earthquake from space geodesy. Nature Geoscience, 12(3), pp.192-199. https://doi.org/10.1038/s41561-018-0296-0

Wells, D.L. and Coppersmith, K.J., 1994. New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement. Bulletin of the seismological Society of America, 84(4), pp.974-1002. https://doi.org/10.1785/BSSA0840040974

Yue, H., Lay, T., Freymueller, J.T., Ding, K., Rivera, L., Ruppert, N.A. and Koper, K.D., 2013. Supershear rupture of the 5 January 2013 Craig, Alaska (MW 7.5) earthquake. Journal of Geophysical Research: Solid Earth, 118(11), pp.5903-5919. https://doi.org/10.1002/2013JB010594