Dynamic restrengthening and fault heterogeneity explain megathrust earthquake complexity

This study demonstrates that the complex rupture dynamics of the 2011 Tohoku-Oki megathrust earthquake, including multiple rupture episodes and large tsunamigenic slip, can be self-consistently explained by the interplay of preexisting fault heterogeneity and rapid coseismic frictional restrengthening.

The Gist

Imagine the Earth's crust as a giant, complex machine where tectonic plates grind against each other. Usually, when these plates get stuck and then suddenly snap free, they cause an earthquake. But the 2011 Tohoku-Oki earthquake in Japan was a "monster" event. It wasn't just a simple snap; it was a chaotic, multi-stage explosion of energy that created a massive tsunami and confused scientists for years.

This paper is like a high-tech detective story. The authors used supercomputers to build a virtual time machine, simulating that earthquake to figure out why it was so weird and complex.

Here is the simple breakdown of their discovery:

1. The Mystery: A "Spaghetti" Earthquake

Most earthquakes are like a clean tear in a piece of paper: it starts at one point and rips smoothly in one direction. The Tohoku earthquake was more like a bowl of spaghetti.

• It started, stopped, and started again: The rupture didn't just go; it jumped, reactivated, and spiraled.
• It went deep and shallow differently: Deep down, the fault moved in short, sharp jerks (like a pulse). Near the surface, it slid smoothly and for a long time (like a crack spreading).
• It reached the ocean floor: It slid all the way to the trench, lifting the ocean floor violently to create a giant tsunami.

Scientists wondered: What physical rules allowed such a chaotic mess to happen?

2. The Ingredients: Two Secret Sauce Factors

The researchers found that you don't need to invent new physics to explain this. You just need the right combination of two existing ingredients:

Ingredient A: The "Rubber Band" Effect (Dynamic Restrengthening)
Imagine a rubber band. When you stretch it fast, it gets hot and weakens (friction drops). But the moment you stop stretching it, it instantly snaps back to being strong again.

• In the earthquake: As the fault slipped, it got super hot and slippery (weakening), allowing it to move fast. But the moment that specific patch of rock stopped moving, it instantly "healed" and became sticky again.
• The Result: This caused the earthquake to "stutter." The slip would stop because the rock got sticky again, but the stress would build up behind it until it forced the rock to slip again in a new burst. This created the "stop-and-start" reactivation we saw.

Ingredient B: The "Bumpy Road" (Fault Heterogeneity)
Imagine driving a car on a road. If the road is perfectly smooth, the car moves in a straight line. But if the road is bumpy with potholes and speed bumps (stress heterogeneity), the car will bounce, swerve, and speed up or slow down unpredictably.

• In the earthquake: The fault line wasn't uniform. Some parts were under more pressure than others.
• The Result: These bumps and stress variations acted like a chaotic conductor, telling the earthquake where to jump next, where to slow down, and where to speed up.

3. The Simulation: The Virtual Time Machine

The authors didn't just guess; they ran a massive 3D simulation on a supercomputer. They fed it:

• The actual shape of the ocean floor and the fault.
• The "Rubber Band" friction rules (fast weakening, fast healing).
• The "Bumpy Road" stress map (based on real data from previous studies).

The Magic Moment:
They didn't tell the computer how the earthquake should behave. They just set the rules and let it run.

• The computer spontaneously created the chaos.
• It naturally produced the "stop-and-start" reactivation.
• It naturally created the mix of deep "pulses" and shallow "cracks."
• It naturally slid all the way to the trench, creating the tsunami.

4. Why This Matters

Before this, scientists thought you needed to manually program "special spots" (asperities) into the model to make the earthquake act weird. This paper says: No, you don't.

The complexity wasn't a glitch; it was a natural feature of how rocks behave when they are hot, fast, and unevenly stressed.

• The Analogy: Think of it like a forest fire. You don't need to tell the fire exactly where to jump. If you have dry wood (stress), wind (slip rate), and uneven terrain (heterogeneity), the fire will naturally create complex, unpredictable patterns.

The Takeaway

The 2011 Tohoku earthquake wasn't a mystery that broke the laws of physics. It was a perfect storm of rapid friction healing (the rubber band snapping back) and uneven stress (the bumpy road).

This discovery is huge for the future. It means we can build better models to predict how future giant earthquakes might behave. We don't need to guess the "special spots"; we just need to understand the physics of how the fault heals and how the stress is distributed. This helps us prepare better for the next big one, potentially saving lives by improving tsunami warnings and building codes.

Technical Summary

Here is a detailed technical summary of the paper "Dynamic restrengthening and fault heterogeneity explain megathrust earthquake complexity" by Wong, Gabriel, and Fan.

1. Problem Statement

Megathrust earthquakes, such as the 2011 $M_W$ 9.0 Tohoku-Oki event, exhibit complex rupture dynamics that challenge current physical models. Key observed complexities include:

• Multiple rupture episodes: Evidence of rupture reactivation at the hypocenter.
• Depth-dependent behavior: A contrast between slow, pulse-like rupture downdip and fast, crack-like rupture updip.
• Large near-trench slip: Massive displacement (>50 m) reaching the trench, despite the shallow region typically being characterized by velocity-strengthening friction (which should inhibit rupture).
• Limited along-strike extent: An unusually confined rupture width for an earthquake of its magnitude.

Previous studies often attributed these features to pre-existing static heterogeneities (e.g., specific frictional asperities or stress patches) or used simplified friction laws. However, the physical mechanisms driving these complexities spontaneously during dynamic rupture evolution, particularly when combining realistic friction laws with data-constrained heterogeneity, remained uncertain.

2. Methodology

The authors employed fully dynamic 3D rupture simulations using the open-source code SeisSol (Arbitrary High-order Derivative Discontinuous Galerkin method).

• Model Geometry: A realistic 3D slab geometry based on the Japan Integrated Velocity Structure Model (JIVSM), incorporating high-resolution topobathymetry. The mesh contained ~50 million elements, resolving seismic waves up to 1.5–2 Hz.
• Friction Law: A fast velocity-weakening rate-and-state friction law was implemented. Crucially, this law includes:
  • Coseismic weakening: Rapid friction reduction at high slip rates (simulating thermal pressurization and flash heating).
  • Dynamic restrengthening: Rapid frictional healing as slip rates decelerate.
  • Depth dependence: Velocity-strengthening behavior in shallow (<9 km) and deep (>45 km) zones, and velocity-weakening in the seismogenic zone (10–45 km).
• Initial Conditions (Prestress):
  • Regional Stress: Constrained by the World Stress Map (principal stress orientation).
  • Heterogeneity: Instead of prescribing arbitrary asperities, the initial stress heterogeneity was derived from the median slip distribution of 32 published finite-fault models of the Tohoku-Oki earthquake. This provided a data-informed, spatially varying initial stress field.
  • Parameters: The study performed a grid search varying the amplitude of stress heterogeneity ($\alpha$) and the regional ambient stress level ($R_0$) to find the preferred model.
• Off-Fault Physics: Included Drucker-Prager visco-elasto-plasticity to account for inelastic deformation in the shallow accretionary wedge.

3. Key Contributions

• Spontaneous Complexity: Demonstrated that complex rupture behaviors (reactivation, mixed rupture styles) can emerge spontaneously from the nonlinear interaction of dynamic friction evolution and data-informed stress heterogeneity, without needing to prescribe specific frictional patches or structural barriers.
• Mechanism of Restrengthening: Identified rapid coseismic frictional restrengthening as the primary driver for repeated rupture reactivation and self-healing pulse-like behavior.
• Unified Explanation: Provided a single physical framework explaining three distinct observations simultaneously: the depth-dependent rupture style, the multiple reactivation episodes, and the large near-trench slip.
• Heterogeneity vs. Friction: Showed that while both stress heterogeneity and frictional heterogeneity can produce complex ruptures, stress heterogeneity combined with dynamic restrengthening is more effective at reproducing the specific "pulse-to-crack" transition and spontaneous arrest observed in Tohoku.

4. Key Results

A. Rupture Reactivation and "Spiraling" Fronts

The preferred model spontaneously reproduced multiple rupture reactivation episodes near the hypocenter.

• Mechanism: As a slip pulse propagates, rapid frictional restrengthening causes the fault to "heal" and stop. However, dynamic stress redistribution concentrates stress behind the healing front, eventually overcoming the local strength and re-nucleating a new pulse.
• Observation: This resulted in "spiraling" rupture fronts and back-propagating phases, consistent with back-projection imaging of high-frequency radiation. The model captured six distinct slip episodes at the hypocenter.

B. Depth-Dependent Rupture Styles

The simulation reproduced the observed dichotomy in rupture styles:

• Downdip (Deep): Characterized by short-duration, self-healing slip pulses. These pulses radiate high-frequency seismic energy but accumulate less total slip. The rapid restrengthening allows the rupture to arrest locally, creating a pulse.
• Updip (Shallow): Characterized by sustained, crack-like rupture. As pulses reach the free surface, they reflect and interact, preventing the formation of healing fronts. This leads to prolonged slip durations (>50 s) and large total slip accumulation.
• Result: This explains why high-frequency radiation is inversely correlated with total slip: high-frequency radiation comes from the deep pulses, while large slip occurs in the shallow crack-like zone.

C. Large Near-Trench Slip

Despite the shallow region having velocity-strengthening friction (which typically stabilizes slip) and off-fault plasticity (which dissipates energy), the model produced ~50 m of slip at the trench.

• Mechanism: Dynamic weakening (thermal pressurization/flash heating) was strong enough to overcome the velocity-strengthening barrier. Additionally, wave-mediated dynamic stressing from the deep rupture and free-surface reflections provided the necessary traction to sustain shallow propagation.
• Significance: This confirms that large tsunamigenic slip can occur even in regions with stabilizing frictional properties if dynamic weakening is sufficiently strong.

D. Spontaneous Rupture Arrest

The model successfully reproduced the spontaneous arrest of the rupture along the strike, limiting the magnitude to $M_W$ 8.97 (close to the observed 9.0).

• Mechanism: Rupture arrest occurred where the available strain energy was insufficient to overcome the fracture energy, driven by the specific distribution of the initial stress heterogeneity.
• Contrast: Models with laterally homogeneous prestress failed to arrest, rupturing the entire fault and producing an unrealistic $M_W$ 9.61.

5. Significance

• Hazard Assessment: The study highlights that dynamic effects (specifically restrengthening and stress redistribution) are fundamental controls on rupture complexity. Future seismic and tsunami hazard assessments must incorporate these dynamic processes rather than relying solely on static asperity models.
• Physics-Based Modeling: It establishes a robust, self-consistent framework for simulating megathrust earthquakes that does not require ad-hoc tuning of frictional patches to match observations.
• Generalizability: The identified mechanisms (dynamic restrengthening driving reactivation and mixed rupture styles) are likely applicable to other subduction zones globally, improving the predictability of rupture behavior and tsunami generation potential.

In conclusion, the paper demonstrates that the complex, multi-faceted nature of the Tohoku-Oki earthquake is an emergent property of the interplay between data-constrained fault heterogeneity and rapidly evolving frictional laws, specifically the cycle of dynamic weakening and restrengthening.