Mexico earthquake partially reloads fault that ruptured in devastating 1985 shock

Though unlikely sufficient to trigger another great earthquake soon, about one-quarter of the fault surface that broke in the 1985 earthquake is now closer to failure.

By Ross S. Stein, Ph.D., Temblor, Inc., and Shinji Toda, Ph.D., IRIDeS, Tohoku University

Citation: Stein, R., and Toda, S, 2022, Mexico earthquake partially reloads fault that ruptured in devastating 1985 shock, Temblor,

Este artículo también está disponible en español.

On Sept. 19, 2022, a magnitude-7.6 shock struck the Pacific coast of Mexico between Manzanillo and Zihuantanejo at a shallow depth of 15 kilometers (nine miles). Fortunately, this is not a densely populated area. The quake was nevertheless widely felt, from Puerto Vallarta and Guadalajara to the northwest, Mexico City to the northeast and Acapulco to the southwest, with strong shaking within 100 kilometers (60 miles) of the mainshock.

Not a surprise

Though this earthquake was not forecasted, it wasn’t a surprise. Although earthquakes cannot be predicted, their long-term likelihood can be assessed by balancing the accumulation of strain recorded by GPS against the past century of earthquakes, which release that strain. From that vantage point, one sees that the expected time between magnitude-7.6 quakes (also called the average return time or recurrence interval) is between about 30 and 100 years.

The megathrust surface along which this quake took place is slipping, or moving, at about 50 millimeters per year (about two inches per year). A magnitude-7.6 earthquake moves the fault about 3 meters (10 feet). A return time of about 60 years is also consistent with that simple assessment.

The last large shock in this location was a magnitude-7.5 in 1973 that might have slipped the same patch of fault, or perhaps an adjacent one at greater depth. Thus, at least 50 years elapsed between the last shock and this one — enough time for the fault to rebuild significant stress.

map with cool colors in a linear pattern
Here we use Temblor’s T-GEAR model to display the return time of magnitude-7.6 earthquakes in Mexico. T-GEAR is a long-term model that attempts to capture the past century of behavior by blending strain rate measured by GPS and the release of that strain in earthquakes occurring over the past 117 years. The typical return time of quakes this size is about 100 years.


A larger shock unlikely imminent

Monday’s shock likely does not portend a larger event, because most large shocks are not followed by still larger earthquakes. Nevertheless, the Sept. 19 quake has already been succeeded by a magnitude-6.8 shock on Sept. 22. Earthquake sequences do occur: The Sept. 7, 2017, magnitude-8.2 Tehauntepec shock was followed 12 days later by the magnitude-7.1 Puebla earthquake 600 kilometers (370 miles) to the northwest, and four days after that, a magnitude-6.8 shock struck 250 kilometers (155 miles) to the southeast of the Tehauntepec mainshock.

Did the 2022 earthquake reload the site of the 1985 magnitude-8.0 Michoacán shock?

To answer this key question, we calculated the stress transferred by the magnitude-7.6 shock to the surrounding megathrust surface. Stress dropped profoundly in a central “core” where the fault slipped. We forecast that within about 5 years, aftershocks will cease in this core, and will likely remain shut down for up to a century. But the surrounding corona has now been brought closer to failure, and so this is where we expect vigorous aftershock activity during the next decades. These inferences are based on our study of much larger megathrust earthquakes, greater than or equal to magnitude-9.0, worldwide (Toda and Stein, 2022). Whether these inferences from the largest megathrust earthquakes apply to magnitude-7.6 events is less certain.


Surrounding the core is a “corona” of increased stress, which in this case overlaps with the patch of the megathrust that ruptured in 1985, as shown below. Only about a quarter of the 1985 patch (in its southwest quadrant) has been brought significantly closer to failure. Therefore, although the 1985 patch could re-rupture, this might not be enough stress, or it might not cover a large enough stressed area, to trigger another great earthquake. Interestingly however, this quadrant has already been the site of the magnitude-6.8 aftershock that occurred on Wednesday, Sept. 21, which supports these calculations. Unquestionably, the area between the Sept. 19 epicenter and Zhuatenejo will be a region that the Servicio Sismológico Nacional of Mexico will closely monitor.

map with colorful area in the middle
Here we calculate the stress transferred by the Sept. 19, 2022 magnitude-7.6 earthquake to the surrounding fault patches, using Coulomb 3.3 (Toda et al., 2011). The stress drops at the site of high slip (the ‘core’), and rises in the surrounding annulus (‘corona’), as explored in Toda and Stein (2022). Stress is also transferred to the southwest portion of the rupture surface of the 1985 Michoacán earthquake (Mendoza and Hartzell, 1989), which has already been the site of a magnitude-6.8 aftershock, consistent with these calculations. Gonzales-Huizar et al. (2022) published an independent stress transfer model, in which they also identify a stress increase at the site of the 1985 magnitude-8.0 earthquake.


Our aftershock forecast for the next month

Aftershocks become less frequent — but no smaller — with time, and will probably last for decades. So, we used Temblor’s Realtime Risk model to forecast seismicity over a broad swath of Mexico for the 30 days beginning Sept. 20, 2022 — the day of the mainshock.

The strongest effects are associated with the most recent earthquakes, whose impact will fade over the next decade. Five years after the 2017 magnitude-8.2 Tehauntepec earthquake (Meng et al., 2019), the event is still influencing surrounding seismicity.

We expect between eight and twenty shocks greater than or equal to magnitude-5.0, and one to two shocks greater than or equal to magnitude-6.0, during this period. This number is between five and fifteen times higher than that during the timeframe between 2000 and 2012. Though the greatest concentration of these events is forecast to be close to the site of the 2022 earthquakes, the 2017 Tehuantepec sequence is still causing an elevated rate of seismicity to the southeast of Oaxaca, as its aftershocks are still occurring.

map of central Mexico
Here we use Temblor’s Realtime Risk (Toda and Stein, 2020) model to forecast earthquakes greater than or equal to magnitude-5.0 for the 30 days beginning Sept. 20, 2022. For this calculation, we used the stress transfer from all magnitude-7.0 and greater shocks since 2012, the background seismicity, and earthquake focal mechanisms. We use the Servicio Sismológico Nacional (UNAM) seismic catalog for earthquakes greater than or equal to magnitude-4.0 for the timespan ranging from 2000 to 2012, and Global CMT focal mechanisms.


Bottom Line

This sequence may not be over. A larger coastal earthquake, or one of similar size closer to Mexico City, remains possible. This is why Mexico’s national seismic and geodetic monitoring networks (which were invaluable for our study), its Earthquake Early Warning System and the government’s efforts to make Mexico City’s building stock more resilient in the wake of the damaging 1985 magnitude-8.0 Michoacán and 2017 magnitude-7.1 Puebla, shock, are so important.
[See Comment/Reply figures below in comment section]


Carlos Mendoza, Stephen H. Hartzell; Slip distribution of the 19 September 1985 Michoacan, Mexico, earthquake: Near-source and teleseismic constraints. Bulletin of the Seismological Society of America1989;; 79 (3): 655–669. doi:

Gonzalez-Huizar, H., Pérez-Campos, X., Melgar, D., 2022, Mexico rattled by pair of earthquakes, Temblor,

Meng, L., Huang, H., Xie, Y., Bao, H., & Dominguez, L. A. (2019). Nucleation and kinematic rupture of the 2017 Mw 8.2 Tehuantepec earthquake. Geophysical Research Letters, 46, 3745– 3754.

Toda, S., Stein, R.S. Central shutdown and surrounding activation of aftershocks from megathrust earthquake stress transfer. Nat. Geosci. 15, 494–500 (2022).

Toda, Shinji, Ross S. Stein; Long‐ and Short‐Term Stress Interaction of the 2019 Ridgecrest Sequence and Coulomb‐Based Earthquake Forecasts. Bulletin of the Seismological Society of America 2020; 110 (4): 1765–1780. doi:

Toda, Shinji, Stein, R.S., Sevilgen, Volkan, and Lin, Jian, 2011, Coulomb 3.3 Graphic-rich deformation and stress-change software for earthquake, tectonic, and volcano research and teaching—user guide: U.S. Geological Survey Open-File Report 2011–1060, 63 p., available at