With no seismometers in the area, an international team used video from two security cameras to calculate how violently the ground moved in Caraballeda on 24 June. Their method could be crucial to understanding how to design for future safety worldwide.
By Dr Shamim Quadir (Senior Communications Officer), Published
Dr Miguel Bravo-Haro, participated in an international study that has reconstructed, for the first time, the ground motion that struck Caraballeda, La Guaria State, Venezuela, during the magnitude 7.5 earthquake of 24 June this year.
The study was led from Universidad de Chile by the EASER research group (Evolution Assessment of Seismic Risk), an ANID-funded Anillo project on seismic risk, with Universidad de Concepción (Chile), City St George's, University of London (UK), Stanford University (USA) and the Venezuelan company Inelectra.
Dr Bravo-Haro, Senior Lecturer in Structural Engineering at the Department of Engineering, School of Science & Technology at City St George's, is a co-author of the research, published in The Seismic Record, a journal of the Seismological Society of America. His contribution was the development of a (bi-directional sliding block) numerical model and the (inversion) algorithm that enabled the reconstruction of the ground motion during the earthquake from CCTV security camera footage.
An earthquake with no recordings
On 24 June, a shallow earthquake of magnitude 7.5 ruptured the San Sebastián and Boconó faults on Venezuela's northern coast. The rupture passed almost directly beneath the communities of La Guaira state. Hundreds of buildings collapsed and many more were damaged, with it being one of the most violent earthquakes in decades to strike a densely populated area.
Yet no seismic instrument recorded how hard the ground actually shook where the damage was greatest, and this is not the first time. When the 1967 Caracas earthquake damaged many of the same neighbourhoods, it also went unrecorded: the few instruments of the era failed or had never been properly installed - an accelerograph in Caracas, for example, had no battery. For decades scientists had to infer the intensity of shaking indirectly, from damage. Almost sixty years later, La Guaira still has no network capable of measuring an earthquake of this size close to the fault.
Why this measurement matters
Large, shallow earthquakes that rupture within a few kilometres of a city are the events engineers fear most. They can produce a strong, long “pulse” of ground motion that subjects buildings to demands far beyond those considered in their design. They are also the least recorded in the world. NGA-West2 is the name of the international database engineers use to build models and produce design codes to protect buildings against earthquake induced ground motions. It shows only two recordings of earthquakes larger than magnitude 7.4 had been made within 5 km of the fault before 2023, with the 2023 Türkiye earthquakes adding 20 more.
A traffic post and a lamp post as seismographs
With no conventional records, the researchers turned to two ordinary security cameras mounted on a commercial building in Caraballeda, which happened to be filming the street during the earthquake. In the footage, two everyday objects respond to the shaking: a plastic traffic delineator post that slides across the pavement, and an 11-metre lamp post that sways from side to side.
Neither is a measuring instrument, but the physics of how each responds to ground motion is well understood. The team tracked the movement of each object frame by frame, then worked backwards to calculate the ground motion that must have made the post slide and the lamp post sway exactly as seen. To make the calculation reliable, members of the team visited the site, surveyed the objects, and measured on the spot the friction between the post's base and the pavement.
Two features make the result robust. First, the two objects respond to the same ground motion in very different ways, one sliding and the other vibrating. Second, the ground motion was calculated from the sliding post alone and then checked against the lamp post, which played no part in the calculation of the ground-motion. That the result explains the movement of both objects gives high confidence in the estimate. Four analysts tracked the post independently, and the final result lies within the range of their traces. To the authors' knowledge, it is the first video-based reconstruction of ground motion to be verified independently in this way.
What they found
The ground at Caraballeda received an extraordinarily strong “pulse”, lasting less than one second and directed mainly north–south, roughly perpendicular to the fault. The peak ground velocity reached about roughly two metres per second. For context, according to the NGA-West2 database, only two recordings from the 1999 Chi-Chi earthquake in Taiwan exceed this value. It is larger than the strongest record from the 1995 Kobe earthquake in Japan, and larger than any recorded during the highly destructive 2023 Türkiye earthquakes.
The intensity also exceeds what the models used by engineers worldwide predict, even after correcting them for the effects of being close to the fault, and it is on average almost twice what the Venezuelan seismic design code specifies for an extreme earthquake across the range of vibration periods typical of mid- and high-rise buildings in the area.
The same cameras captured eight buildings beginning to collapse at the same moment, just as that pulse ended. The authors believe it is the first time video has shown that exact instant.
Why it matters beyond Venezuela
The impact of this study is that in practice, this is the only public, quantitative estimate of how the ground moved in one of the areas worst hit by the earthquake. It is essential information for understanding why buildings failed, for updating seismic hazard maps and Venezuela's design code, and for deciding how to repair and rebuild, all of them urgent tasks. And because security cameras are now almost everywhere, the method can be applied in other earthquake-prone cities that have no instruments: ordinary surveillance video can become a reliable and useful engineering tool.
Reflecting on the international collaboration, Dr Bravo Haro said:
About the study
Read the full article, Ground-Motion Intensity in Caraballeda During the 24 June 2026 Mw 7.5 Venezuela Earthquake Estimated from CCTV Video in the journal, The Seismic Record:
About EASER
EASER (Evolution Assessment of Seismic Risk) is a research group funded by Chile's national research agency ANID through its Anillo programme. It brings together researchers from Universidad de Chile, Universidad de Concepción and other institutions to study how seismic risk evolves and how to reduce it.