The event
On 24 June, two mainshock-class earthquakes, M7.2 then M7.5, ruptured the same stretch of the Venezuelan coast 39 seconds apart. Thirty-nine seconds is not enough time to inspect, evacuate, red-tag or repair anything. The second rupture struck structures that were still moving, already cracked, with stiffness and strength degraded by the first.
Our thoughts are with everyone affected. What follows is an engineering reflection, not a comment on the response. We covered the event itself, and why the historical record never saw it coming, in a separate case study.
Why it breaks the curve
A standard fragility curve answers one question: given a level of shaking, how likely is an undamaged structure to reach a given damage state. Damage, though, is path-dependent. After the first shock a building carries:
- A longer natural period. It has lost stiffness, so it responds to the ground differently than the intact model assumes.
- Lower, capped strength. Part of its capacity is spent and does not come back between shocks.
- Residual drift. It is no longer standing straight; its starting point has shifted.
- Consumed deformation capacity. Every cycle of the first shock used up part of the budget the second one needs.
Every one of those moves the collapse fragility to the left. So the second shock does not read off the intact curve: it reads off a degraded one. A single-event model, run twice independently, misses this entirely, and it under-counts collapses and large losses precisely in the tail that sizes your capital.
State-dependent fragility: why a 39-second doublet breaks single-event curves
The fix
The remedy is not a fudge factor. It is to make fragility state-dependent: a family of curves conditioned on the damage the structure is already carrying, not just on the shaking. That means running the sequence through the structural model as a sequence, so residual deformation and accumulated deterioration carry across from the first shock into the second.
- Back-to-back analysis. Mainshock and aftershock records run through the structural model in one pass, with no reset in between. Residual drift, stiffness loss and strength loss all carry across.
- Degrading behaviour, element by element. Coupling beams and walls calibrated with energy-based deterioration, so the second shock starts from an already deteriorated structure.
- Intensity read where the damaged building responds. The second shock is evaluated at the structure's elongated period, not the intact one, because the building that receives it is not the building that was designed.
- Real sequence pairs. Kumamoto 2016 and Ridgecrest 2019 as doublet analogues, plus Christchurch, L'Aquila and Emilia.
- Sequence-aware frequency. The second shock enters the model through how often aftershocks actually follow this kind of rupture, and where, never as an independent draw.
This is the same degrading-stiffness, energy-based deterioration Dynamis calibrates in performance-based design: the cyclic response of coupling beams and walls is exactly where the second shock does its work. It is the engineering that sits behind how we build fragility, and it is why sequences belong in the vulnerability conversation, not in a footnote.
"Fragility curves today are built on simplified assumptions. Ours are built on 15 years of designing structures that must not fail. By pairing performance-based seismic design with AI-assisted evidence collection, we transform fragility from generic to engineering-grade."
Carlos Caramés Molero, Founder & Partner, DynamisSingle-event fragility is fine for the body of the loss distribution. Sequences live in the tail, and the tail is the part you are paid to get right.
Key questions
Why do standard fragility curves under-estimate losses in earthquake sequences?
A standard fragility curve is conditioned on an undamaged structure: it answers how likely an intact building is to reach a damage state at a given level of shaking. In a sequence, the second shock strikes a building that has already lost stiffness and strength and may be permanently leaning. Its collapse fragility has shifted, so reading the intact curve twice under-counts collapses and large losses, precisely in the tail of the loss distribution.
How much can a first shock raise the collapse probability in the second one?
Schematically, at the same level of shaking, the probability of collapse of a frame that has been through a strong first shock can climb from around 5% to around 58%, an order-of-magnitude shift. The exact numbers depend on the structure and the records, but the direction and the scale of the effect are well documented in real sequences such as Kumamoto 2016 and Christchurch 2011.
What is a state-dependent fragility curve?
A family of fragility curves conditioned not only on the level of shaking but on the damage state the structure is already in. It is generated by running mainshock and aftershock records through the structural model back to back, without resetting it, so residual deformation and degraded stiffness and strength carry across into the second shock.
Sources
- USGS, preliminary event parameters for the 24 June 2026 Venezuela doublet (M7.2 and M7.5, Yaracuy)
- Audemard and others (2000), regional active-fault context for northern Venezuela
- Documented mainshock-aftershock sequences: Kumamoto 2016, Ridgecrest 2019, Darfield-Christchurch 2010-2011, L'Aquila 2009, Emilia 2012
Event parameters are USGS preliminary values. The figure and the quoted probabilities are schematic and illustrative, not calibrated to a specific asset. Not a substitute for a formal risk report.