On March 28, 2025, the crust in north-central Myanmar ruptured for over 500 kilometers, a rupture length so unprecedented that it challenged decades of assumptions about seismic hazard. The magnitude 7.7 earthquake on the Sagaing Fault not only destroyed villages, killing thousands of people, but also provided an unusual, high-resolution case study now changing the way scientists conceptualize fault segmentation, earthquake cycles, and the boundaries of strike–slip fault behavior.
The Sagaing Fault, a north–south strike–slip fault system approximately 1,400 kilometers in length across Myanmar, has been likened to California’s San Andreas Fault for years by geometry and tectonic setting. Both faults are straight, mature plate boundaries for horizontal motion between blocks of crust. Prior to this occurrence, hazard models hypothecated by the seismic gap hypothesis estimated a rupture of approximately 300 kilometers along a segment that had been locked since 1839. What happened instead was that the 2025 earthquake ruptured along that gap and extended into neighboring segments most recently broken in 1930 and 1946, finally covering ~510 kilometers.
Researchers from Caltech, with postdoctoral scholar Solène Antoine at the lead in the lab of Jean-Philippe Avouac, used sophisticated satellite geodesy to quantify the deformation. With optical image correlation from Sentinel-2 and other missions, in addition to radar amplitude correlation, they determined the horizontal displacements even where radar interferometry does not work because of decorrelation and lack of sensitivity to north–south motion. Analysis showed an extraordinarily homogeneous slip distribution averaging 3.3 meters along the rupture, with no shallow-slip deficit a characteristic very seldom observed in great continental strike–slip earthquakes. The rupture propagated at supershear velocity, outrunning the shear waves it generated, a phenomenon facilitated by the Sagaing Fault’s smooth, mature geometry.
Such findings directly challenge the concept of fixed fault segmentation. Historically, hazard assessments treated fault sections as quasi-independent, with recurrence intervals tied to slip deficits. But the Myanmar event demonstrated that old strike–slip faults can break across several “segments” during a single event, releasing more than the deficit accumulated since the previous rupture. As Avouac summarized, “The study shows that future earthquakes might not simply repeat past known earthquakes. Successive ruptures of a given fault, even as simple as the Sagaing or the San Andreas faults, can be very different and can release even more than the deficit of slip since the last event.”
To test the physics underlying this behavior, the team performed quasidynamic earthquake cycle simulations on a simplified, nonplanar model of the Sagaing Fault. These models generated sequences with unregular segmentation, clustering, and variable magnitude, including events like the 2025 rupture. The simulations indicate that Mw > 7.5 earthquakes on this fault recur at unregular intervals averaging 141 years, with a standard deviation of approximately 40 years. This heterogeneity highlights the limitations of time-independent statistical models, which infer probabilities over constant-sized windows without regard to the timing, location, and size of the previous slip.
The Myanmar findings are consistent with insights from other tectonic regimes. In subduction settings, decades of geodetic and seismic observations have shown that large ruptures can bridge multiple asperities and low-coupling zones in a manner that is not captured by simple recurrence models. In the same way, strike–slip faults such as the San Andreas have formed previous ruptures e.g., the 1906 San Francisco earthquake that extended across assumed segment boundaries. The Sagaing event offers us a contemporary, data-sufficient example of this effect with surface displacement maps, offset data, and finite slip models now in the public domain for model testing.
For the purpose of hazard estimation, the consequences are significant. Current U.S. Geological Survey models for the San Andreas Fault estimate a 60% chance of a magnitude 6.7 or greater earthquake near Los Angeles in the next 30 years, based largely on statistical recurrence. But the Myanmar case suggests that such models may underestimate the potential for multi-segment, high-magnitude ruptures. Physics-based, time-dependent models incorporating stress redistribution, fault geometry, and observed slip histories offer a path toward more realistic forecasts. As shown in Myanmar, combining high-resolution geodetic imaging and dynamic rupture simulations has the potential to expose behaviors hidden from historical records alone.
The 2025 Sagaing rupture is both a humanitarian disaster and a scientific milestone. Through an integration of state-of-the-art remote sensing with state-of-the-art modeling, scientists have mapped a fault larger than its assumed limits, providing a cautionary example for other global major strike–slip systems. For hazard planners and geoscientists, the warning is unambiguous: the next “big one” won’t necessarily resemble the previous one.
