Earthquake Jitters: Why The Quietest Faults May Be The Most Dangerous

Like sleeping giants, dormant faults store energy over millions of years before being rudely awakened by human activity.

Some faults can go tens of millions of years without motion. No tremors, no ground shaking, no historical record of seismic activity. By conventional wisdom, these quiet zones should be among the safest places to drill for natural gas, extract geothermal energy, or store carbon dioxide underground. Except they’re not.

Across parts of the Netherlands and other supposedly quiet regions around the world, earthquakes have been striking with alarming frequency in places where the ground hasn’t budged in geological ages. Homes have cracked. Infrastructure has buckled. Communities that never built for seismic activity have found themselves rattled by tremors spawned by human drilling and extraction.

Scientists have now uncovered why: These sleeping giants have been quietly building strength for millions of years, storing energy like coiled springs. When human activities alter underground pressures, that stored energy can release catastrophically.

A study published in Nature Communications reveals that long-dormant faults gain strength through a process called “healing,” where mineral grains slowly bond together and rock compacts over geological time. After roughly 30 million years of inactivity, even faults that scientists classified as inherently stable can rupture violently. The finding revises a long-held assumption that velocity-strengthening faults can’t host earthquakes.

“Faults could develop significant interface strength, expressed as increase in static frictions by around 0.25, due to ‘healing’ over thousands to millions of years,” the research team from Utrecht University in the Netherlands writes in the paper. That friction increase means substantially more force holding the fault in place and, when it finally gives way, a much larger energy release.

The discovery has major consequences for the energy transition. As societies shift away from fossil fuels, many plans involve using the subsurface for geothermal power, hydrogen storage, and carbon sequestration. Site selection has often focused on avoiding areas with recent earthquake history. But this research shows that tectonic silence isn’t a guarantee of safety, particularly where faults have experienced both long periods of inactivity and fast healing rates. Under these conditions, initial induced events may be larger than anticipated.

Microscopic Changes Build Over Millennia

Fault healing happens at scales invisible to human observation but becomes consequential over deep time. When a fault sits still, pressure causes mineral grains at contact points to grow together. Chemical processes cement the surfaces. Laboratory experiments have documented this over hours and days, showing progressive strengthening as faults remain locked.

Researchers Meng Li, Andre Niemeijer, and Ylona van Dinther asked what happens when healing continues for millions of years. They built numerical models capable of simulating both the glacial pace of geological processes and the violent seconds of earthquake rupture. The team focused on a configuration matching the Groningen gas field, Europe’s largest natural gas reservoir, where earthquakes up to magnitude 3.6 have damaged thousands of homes since the 1990s despite no prior seismic history.

Simulations tracked how fault strength evolves during dormancy and how that strength releases during human-induced pressure changes. The researchers tested healing periods from 10 years to 100 million years, varying a key parameter called “b” that controls healing rate. Higher b values mean faster strength accumulation.

After 30 million years of healing at high rates, simulated faults generated earthquakes reaching slip rates around 1 meter per second when gas extraction reduced underground pressure. The modeled earthquake occurred about 35 years after production started, matching the actual delay observed in Groningen between 1963 and 1991. Simulated stress drops reached approximately 3.0 megapascals, similar to the 2.5 megapascals measured in Groningen’s largest earthquake in 2012.

Laboratory tests on actual Groningen reservoir rocks showed they have velocity-strengthening friction, the type that should resist sudden ruptures. This had puzzled scientists: How could earthquakes occur in these supposedly stable rocks? The answer lies in healing that accumulated since the last fault movement, which geological studies date to 30-65 million years ago.

During that enormous span, the faults weren’t weak and primed for rupture. They were gradually becoming stronger, building up potential energy while the region remained tectonically dormant. Gas extraction then provided the trigger, changing underground stresses enough to overcome the healed strength.

Why First Quakes Hit Hardest

Once a velocity-strengthening fault ruptures and releases its healed strength, it cannot generate another major earthquake on human timescales. After the first event, these faults transition to slow, steady sliding rather than sudden ruptures. They essentially become pressure relief valves, releasing stress gradually instead of explosively.

“Velocity-strengthening faults can no longer host earthquakes, because subsequent slip on human lifetimes is stable,” the researchers explained.

Velocity-weakening faults, the type responsible for most natural earthquakes, behave differently. They can produce recurring earthquakes during ongoing human activity. However, subsequent events pack less punch. Fault strength at failure stays around typical levels rather than elevated healed levels, cutting potential stress drops by half or more.

This creates an unusual hazard pattern. The first induced earthquake on any healed fault poses the greatest danger, regardless of fault type. Later earthquakes carry reduced threat. Ruptured velocity-strengthening segments even act as barriers, blocking rupture propagation and limiting the size of future events on neighboring fault sections.

Predicting Which Faults Will Wake

The research provides a framework for assessing which dormant faults might wake dangerously. Two factors matter most: healing rate and healing duration. Faults need both sufficient time (typically over 100,000 years) and fast healing to become seismogenic.

Applying this framework to real-world cases yields testable predictions. The Slochteren sandstone hosting Groningen’s gas reservoir needs at most 25 million years of healing before earthquakes become possible. Geological evidence shows the actual healing time exceeded this threshold. The overlying Basal Zechstein caprock layer displays weakly velocity-weakening properties, making it seismogenic even without extended healing. Seismological data confirms earthquake sources locate in these predicted layers.

Surrounding clay and shale layers show velocity-strengthening behavior with negligible healing rates. Models predict these cannot host earthquake nucleation regardless of healing time, matching observations that place no earthquake hypocenters in these zones.

Carbonate rocks such as limestone and dolomite, which hold half of global conventional oil reserves and host numerous earthquakes, often show velocity-weakening or weakly velocity-strengthening behavior with high healing rates in laboratory tests. Such properties make them potentially seismogenic, and healing amplifies that potential.

Shale reservoirs hydraulically fractured across North America display varied properties. Clay-rich shales tend toward velocity-strengthening behavior with minimal healing, suggesting lower seismic potential, though variability across formations means case-by-case assessment remains necessary.

Basalts being explored for carbon dioxide storage also show mixed properties depending on temperature and mineral composition. Some studies report velocity-weakening behavior or high healing rates that would elevate seismic risk.

Rethinking What Makes Faults Dangerous

Standard earthquake models typically assume static friction around 0.6. But in the Groningen-tuned simulations, fault friction at failure reached approximately 0.8 after 30 million years of healing. Faults are less critically stressed than models assume, requiring larger stress changes to trigger failure.

Recognizing higher initial strength reconciles why Groningen needed 30 years and tens of megapascals of pressure reduction before earthquakes began. Models using standard strength values would need to invoke different stress conditions or material properties to match observations.

The work also clarifies a key point: Human activities don’t make faults stronger or weaker. Healing elevates fault strength over geologic time, and human-driven stress changes can overcome that stored strength. The faults gain their energy from millions of years of dormancy, not from the drilling itself.

For selecting safer sites, the research points toward several strategies. Favor velocity-strengthening rocks that heal slowly and have experienced relatively recent fault activity. Treat large mapped faults conservatively, since even small induced events can trigger larger earthquakes nearby. Settings with geometrical complexity that promote smaller, more frequent earthquakes might reduce maximum magnitudes, as ruptured velocity-strengthening segments create strong barriers to later ruptures.

Regional tectonic history becomes more important than previously recognized. Velocity-strengthening reservoir rocks that have healed since Neogene times (roughly 2 to 23 million years ago) may harbor larger seismic hazards than currently understood. The longest-quiet zones, often assumed safest, might actually require the most careful evaluation.

Of course, not all dormant faults pose equal risks. But dismissing tectonically quiet regions as inherently safe has proven dangerously wrong.

---