At 18:04 local time on 24 June 2026, an M7.2 earthquake struck inland in Yaracuy state, near Yumare. Thirty-nine seconds later, an M7.5 mainshock ruptured almost the same patch of crust at 10 km depth, about 16 km southwest of the coastal town of Morón. It was the strongest earthquake in Venezuela since 1900. This is an engineering reading of the event, written with our thoughts for everyone affected.
The first lesson is in the map. Early coverage placed the shaking on the Caribbean coastline, but the United States Geological Survey locates both epicentres inland, on the western San Sebastián fault, roughly 160 km west of Caracas. The mechanism is right-lateral strike-slip on an east-west plane: the signature of the plate-boundary transform system itself, not of a subduction zone.
The 24 June 2026 doublet in its full record
A 100 km-wide boundary, mostly locked
Northern Venezuela is not a subduction margin. It is a roughly 100 km-wide dextral transpressional boundary between the Caribbean and South American plates, taking up about 20 mm per year of relative motion. Most of that budget is carried by the Boconó, San Sebastián and El Pilar fault system, and the San Sebastián fault alone absorbs roughly half of it along the central coast. Its western segments were locked, storing strain, and on 24 June they released about 150 km of it.
The plate boundary and its active faults
The faults are near-vertical and the plates slide horizontally, so the rupture and the strongest shaking run along the coast, not down into the interior. Diagrams that picture Venezuela as a subduction zone get the mechanism, and therefore the loss footprint, wrong.
A 500-year catalogue on a 1,000-year fault
This is where the risk hides. Paleoseismic work on the Boconó fault indicates recurrence of roughly 1,100 to 1,500 years, and the offshore San Sebastián segments are far less constrained. The historical catalogue, by contrast, is only about 500 years long, shorter than a single seismic cycle on these structures. The documented events cluster around Caracas and the El Pilar fault to the east, while the western coast that just ruptured looked deceptively quiet.
| Date | Mag. | Source and location | Depth | Impact |
|---|---|---|---|---|
| 24 Jun 2026 | 7.5 | 16 km SW of Morón, Yaracuy (San Sebastián fault) | ~10 km | This sequence. Strike-slip, ~150 km rupture, maximum intensity MMI IX. |
| 24 Jun 2026 | 7.2 | Near Yumare, Yaracuy (San Sebastián fault) | ~22 km | This sequence. Right-lateral, east-west plane. Foreshock, 39 s earlier. |
| 09 Jul 1997 | 6.9 | Cariaco, El Pilar fault | ~10 km | Cariaco devastated, about 80 lives lost. |
| 29 Jul 1967 | 6.5 | Caracas, offshore Caraballeda | ~25 km | About 280 lives lost, four buildings collapsed in Caracas. |
| 17 Jan 1929 | 6.9 | Cumaná, El Pilar fault | shallow | Cumaná largely destroyed. |
| 29 Oct 1900 | 7.7 | Central coast, offshore (San Narciso) | shallow | Severe damage from Caracas to the coast. Largest of the historical record. |
| 26 Mar 1812 | 7.5 | Caracas, San Sebastián fault | shallow | Caracas and Mérida largely destroyed, with very heavy loss of life. |
When recurrence exceeds the catalogue, a catalogue-driven model systematically under-predicts the hazard. The only defensible path is fault-based: characterise each source by its geometry, its geodetic and geologic slip rate and its maximum magnitude, and treat the declustering and maximum-magnitude choices as first-order judgements, not software defaults.
From engineering to risk intelligence
"Most probabilistic seismic models are calibrated on generic or historical data. Ours is built on a unique, proprietary dataset generated through the design of critical infrastructure — from Panama Canal bridges to metro systems and the high-speed rail between Dubai and Abu Dhabi. This gives us a fundamental advantage: a hazard engine that is not only statistically sound, but informed by real-world engineering at the highest level."
— Carlos Caramés Molero, Founder & Partner, DynamisFrom catalogue to consequence
A record barely 500 years long misses most of the hazard, including a western coast that looked quiet until 24 June. Sismicus, the Xpectral hazard engine, characterises every source by its geometry and slip rate and recovers the strong events the catalogue never captured. Risco then turns that hazard into expected loss on today's exposure: the Caracas, Valencia and Maracay of 2026, far larger than the cities that stood when the boundary last ruptured. For boundaries like this one, with long recurrence, a short record and dense exposure, fault-characterised hazard is not a refinement. It is the difference between seeing the risk in advance and being surprised by it.
The epicentre was here. The catastrophe was 150 km away.
The hazard is only half the loss. The 24 June rupture sent very strong shaking about 150 km east into the coastal metro corridor, where it met the ground conditions and the dense, vulnerable building stock of La Guaira and Caracas. That is where the loss landed, far from the epicentre. Hazard, vulnerability and exposure, read together, are what turn an earthquake into a number.
How the shaking travelled, and where the loss landed
The rupture ran about 150 km east along the coast, not a single point. It carried very strong shaking 150 km from the epicentre into the Caracas and La Guaira corridor.
On La Guaira's soft coastal ground, intensity VII met a vulnerable building stock. The same shaking on engineered structures would not collapse them. Vulnerability is the multiplier.
Loss follows people, not epicentres. About 5.9 million were exposed to very strong shaking along the coastal metro; the loss landed there, not at the fault.
Key questions
Where did the 24 June 2026 Venezuela earthquakes strike?
Inland in Yaracuy state, near Yumare and about 16 km southwest of the coastal town of Morón, on the western San Sebastián fault rather than on the coastline. An M7.2 foreshock was followed 39 seconds later by an M7.5 mainshock at 10 km depth, with a maximum intensity of MMI IX.
Why was such a large earthquake not anticipated by the historical record?
Recurrence on these faults runs roughly 1,100 to 1,500 years, while the historical catalogue is only about 500 years long. The western segments had stayed quiet while documented events clustered around Caracas and the El Pilar fault to the east, so a catalogue-driven model systematically under-predicts the western hazard.
How should a boundary like northern Venezuela be modelled instead?
Source by source, from fault geometry, geodetic and geologic slip rate and maximum magnitude, rather than by counting past events. That hazard is then translated into expected loss on today's exposure, which is far larger than the cities that stood when the boundary last ruptured.
Sources
- USGS — M7.5 and M7.2, 24 June 2026, Yaracuy, Venezuela (event us6000t7zp)
- Audemard, Beck and others — active tectonics of the Boconó, San Sebastián and El Pilar faults
- Historical catalogue — 1812 Caracas, 1900 San Narciso, 1929 Cumaná, 1967 Caracas, 1997 Cariaco
Magnitude and event parameters from USGS; fault geometry on the map is schematic. Historical figures are approximate and vary across published sources. Not a substitute for a formal risk report.