📄 earth science

Tectonic loading rate controls magnitude–rupture-length scaling

This study demonstrates that tectonic loading rate significantly controls earthquake magnitude by influencing stress drop and rupture area, revealing that slowly loaded faults can produce earthquakes up to six times larger than fast-loading faults of the same rupture length, a critical factor that should be incorporated into seismic hazard scaling laws.

Original authors: Eyüp Sopacı, Yann Klinger, Luca Dal Zilio

Published 2026-10-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Eyüp Sopacı, Yann Klinger, Luca Dal Zilio

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Earthquakes are the sudden release of energy stored in the Earth's crust, a process that occurs when tectonic plates grind against one another and eventually snap. To understand how dangerous a specific fault line might be, scientists have long relied on a simple rule of thumb: the longer the crack that breaks, the bigger the earthquake. This relationship allows researchers to estimate the size of a disaster by measuring the length of a scar on the ground left by a past event. However, this rule has always had a frustrating flaw. Two faults of identical length can produce earthquakes of vastly different sizes, with the energy released differing by as much as six times. For decades, this inconsistency was treated as random noise, an unavoidable uncertainty in predicting the worst-case scenario for our cities.

A new study challenges the idea that this variation is random. Researchers have discovered that the speed at which the Earth's crust moves is the missing key to the puzzle. By analyzing a global catalog of major earthquakes and running detailed computer simulations of how faults behave over thousands of years, the team found that slowly moving faults produce significantly larger earthquakes than fast-moving ones, even when they break for the same distance. This finding suggests that the standard rules used to assess earthquake danger are systematically underestimating the potential size of quakes in stable continental interiors, such as the vast, slow-moving regions of central Asia.

The investigation began by looking at the great earthquakes of the 20th century in Mongolia. These massive events, some reaching a magnitude of 8.4, ruptured faults that move at a glacial pace of less than one millimeter per year. In contrast, the famous San Andreas Fault in California moves at a rate of about 30 millimeters per year, yet its major earthquakes are often smaller than expected for their rupture length. When the researchers plotted the data from fifty historical earthquakes against their rupture lengths, a clear pattern emerged. Faults moving slower than five millimeters per year consistently produced earthquakes with higher magnitudes than the global average. The difference was substantial: for a long rupture, the magnitude could be half a unit higher, which translates to a sixfold increase in the energy released.

To understand why this happens, the scientists turned to physics-based simulations that mimic the life cycle of a fault. They modeled two distinct environments: the cold, rigid interior of the Mongolian continent and the warmer, more active boundary of the San Andreas Fault. The simulations revealed two complementary reasons for the size difference. First, the slow movement in Mongolia means that the time between earthquakes is much longer. This extended period allows the fault to "heal" and strengthen significantly, building up a much higher level of stress before it finally breaks. When it does snap, the release of this stored energy is far more violent, resulting in a larger slip and a bigger earthquake.

Second, the temperature of the rock plays a critical role. The cold lithosphere beneath Mongolia allows the brittle, earthquake-generating layer of the crust to extend much deeper—down to about 24 kilometers—compared to the warmer San Andreas Fault, where this layer thins out at roughly 14 kilometers. A deeper layer means there is a larger surface area of rock available to rupture. When a fault breaks, it does not just slide along a line; it tears through a three-dimensional volume of rock. The combination of a deeper rupture zone and the extra strength gained from long periods of healing allows the slow-moving faults to generate earthquakes that are much larger than their length alone would suggest.

The researchers tested this idea by running three-dimensional computer models of earthquake cycles. In these simulations, a fault moving at the slow Mongolian rate produced a rupture that was not only deeper but also released about six times more energy than a fault moving at the fast San Andreas rate, even when both broke for the same length. The models confirmed that the slow loading rate is the primary driver, allowing the fault to accumulate a state of high frictional strength that is simply not possible on rapidly moving boundaries. While the deeper rupture zone contributes to the size, the extended healing time is the dominant factor, acting like a slow, powerful spring that is allowed to wind up for centuries before releasing its tension.

This discovery has profound implications for how we assess seismic risk. Current hazard models often use a single, universal formula to convert fault length into earthquake magnitude, a method that works reasonably well for fast-moving plate boundaries but fails for the slow, stable interiors of continents. By ignoring the speed of tectonic loading, these models systematically underestimate the maximum possible earthquake size for faults in places like Mongolia, the Tibetan Plateau, and other continental interiors. The study suggests that for these slow-moving faults, the largest possible earthquake could be significantly larger than previously thought.

The findings do not just apply to the extreme cases of Mongolia. The research indicates that the relationship between loading speed and earthquake size is a continuous spectrum. As the speed of fault movement decreases, the potential size of the earthquake increases. This means that hazard assessments for any fault should consider its specific loading rate. For the slowest faults, which may have recurrence intervals of thousands of years, the potential for a massive, high-magnitude event is real and measurable. By incorporating the speed of tectonic movement into the scaling laws, scientists can refine their predictions, offering a more accurate picture of the danger posed by the Earth's slowest-moving cracks. The study concludes that the size of an earthquake is not just a matter of how long the fault is, but also a reflection of how long it has been waiting to break.

Technical Summary: Tectonic Loading Rate Controls Magnitude–Rupture-Length Scaling

Problem Statement
Seismic hazard assessments rely on empirical scaling laws that convert mapped fault rupture length into the expected magnitude of an earthquake. However, observations reveal significant scatter in these relationships: earthquakes with nearly identical rupture lengths can differ in moment magnitude ( $M_w$ ) by 0.5 units or more, corresponding to a sixfold difference in seismic moment. This variability is often treated as irreducible noise. A specific anomaly exists in the Mongolia–Altai region, where intraplate faults with extremely slow tectonic loading rates (0.06–0.3 mm yr $^{-1}$ ) have produced $M_w \ge 8$ earthquakes with rupture lengths comparable to, or smaller than, those of fast-loading plate-boundary faults like the San Andreas Fault (30 mm yr $^{-1}$ ), yet with significantly larger magnitudes. The physical mechanism controlling this offset remains unresolved, raising the question of whether standard scaling relations systematically underestimate the maximum credible earthquake size on ultra-slow faults.

Methodology
The authors employed a multi-faceted approach combining observational data analysis, physics-based numerical simulations, and analytical modeling:

Global Catalog Compilation: A dataset of 50 historical and recent $M_w \ge 6$ strike-slip earthquakes was compiled from diverse tectonic settings (California, Turkey, Iran, China, Japan, New Zealand, Alaska, and Mongolia). Each event was characterized by rupture length, mean coseismic slip, and tectonic slip rate ( $V_p$ ).
Thermal and Rheological Framework: The study contrasted two end-member settings: the cold, slowly deforming continental lithosphere of Mongolia (geothermal gradient $\sim 18^\circ$ C km $^{-1}$ ) and the warm, rapidly deforming San Andreas Fault system ( $\sim 35^\circ$ C km $^{-1}$ ). One-dimensional yield-strength envelopes were constructed to determine the depth of the brittle-ductile transition (BDT), which defines the base of the seismogenic zone.
Numerical Simulations:
- 2D Simulations: Used to isolate the effect of loading rate on interseismic healing, stress drop, and slip using an aging law rate-and-state friction formulation. Loading rates varied from $10^{-2}$ to 60 mm yr $^{-1}$ .
- 3D Simulations: Performed using the PyQuake3D code to model full earthquake cycles, capturing the combined effects of frictional healing and seismogenic width (determined by thermal structure) on rupture area and seismic moment.
Analytical Modeling: A reduced rate-and-state model was derived to analytically link loading rate, state variable evolution (healing), and peak-to-residual stress drop. This included sensitivity analyses to distinguish the contributions of frictional healing versus seismogenic width.

Key Results

Systematic Offset: The observational catalog reveals that faults with slip rates below 5 mm yr $^{-1}$ define a systematically higher magnitude–length trend compared to faster faults. For long ruptures, this offset approaches 0.5 magnitude units. The four major Mongolia–Altai earthquakes consistently plot above the standard empirical curve, while San Andreas events plot below it.
Mechanisms of Enhancement: The study identifies two complementary mechanisms by which slow loading increases seismic moment for a fixed rupture length:
1. Frictional Healing: Slow loading extends the interseismic interval, allowing the fault state variable to evolve to higher values. This results in greater frictional strength, a larger peak-to-residual stress drop, and increased coseismic slip.
2. Seismogenic Width: Slowly deforming continental interiors are underlain by colder lithosphere, which pushes the brittle-ductile transition to greater depths (e.g., 24 km in Mongolia vs. 14 km in San Andreas). This creates a thicker seismogenic layer, increasing the potential rupture area ( $A$ ) for a given strike length.
Quantitative Impact: 3D simulations show that for equivalent rupture lengths, the cold/slow model produces approximately twice the seismic moment of the hot/fast model. This is driven by a combination of larger slip (16 m vs. 13 m) and a rupture area 1.7 times larger.
Dominance of Healing: Sensitivity analysis indicates that for continental strike-slip faults with widths $W \gtrsim 3$ km, the contribution of interseismic healing to stress drop exceeds the geometric contribution of seismogenic width by an order of magnitude or more (reaching a factor of ~50 for Mongolian parameters).
Analytical Scaling: The analytical model demonstrates that the maximum state variable scales approximately as $\theta_s \sim V_p^{-1} \ln(V_p^{-1})$ . Consequently, a two-order-of-magnitude decrease in loading rate can produce a several-megapascal increase in stress drop.

Significance and Claims
The paper argues that tectonic loading rate is a measurable, physical control on maximum earthquake size that must be incorporated into magnitude–rupture-length scaling laws. The authors claim that:

Standard scaling relations, often calibrated on active plate boundaries, systematically underestimate the magnitude, slip, and rupture area of earthquakes on ultra-slow, cold intraplate faults.
The "scatter" in magnitude–length relations is not random but physically explicable, driven by the loading history and thermal structure of the fault.
For seismic hazard assessment, particularly in probabilistic models where rupture length is converted to magnitude (or vice versa), ignoring loading rate leads to biases. Specifically, it may underestimate the maximum credible magnitude for slow faults and the associated ground motion at near-fault sites.
The study provides a physical basis for weighting scaling-relation logic tree branches by tectonic setting, rather than applying a single universal curve. The magnitude difference between end-members (fast vs. slow) is approximately 0.5 units, a critical factor for low-probability, high-consequence design scenarios.

The authors conclude that while their models simplify natural complexity, the direction of the effect is robust: longer interseismic intervals allow faults to restrengthen, and cold lithosphere provides a larger area for stress release, resulting in larger earthquakes for a given rupture length.

More like this