Monte Carlo Simulation of Earthquake Damage to Mexico City:A Cadastral Exposure Model with Voronoi-Weighted Spatio-Temporal Point Processes
This paper presents a Monte Carlo simulation model that integrates cadastral property data, ETAS seismic source clustering, and site-specific ground motion amplification to estimate future earthquake-induced monetary losses in Mexico City, revealing an average annual loss of 2.26 billion MXN and a 1-in-100-year loss of 40 billion MXN.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Mexico City sits on a geological trap. Built atop the drained bed of an ancient lake, the city rests on layers of soft, water-saturated clay. When seismic waves travel through the hard rock of the surrounding mountains, they move quickly and with little effect. But when those same waves hit the soft lake sediments beneath the city, they slow down and swell in amplitude, much like a jelly wobbling on a plate. This phenomenon, known as site amplification, turns distant, moderate earthquakes into devastating local catastrophes. The city has felt this power twice in recent decades: first in 1985, when a massive quake hundreds of kilometers away caused widespread collapse, and again in 2017, when a smaller, closer event repeated the destruction. Because the ground motion behaves so differently here than anywhere else, standard ways of predicting earthquake damage fail. To understand the true financial risk facing this megacity, researchers needed a model that respected the unique physics of the lake bed and the specific value of the buildings sitting upon it.
A team of researchers at the Universidad Iberoamericana has built such a model, creating a detailed simulation of what future earthquakes could cost Mexico City. Instead of relying on broad economic estimates or population counts, they used the city's official land registry, known as the cadastral register. This government database lists the assessed value of every single parcel of land and the structures on it, providing a granular map of exactly what is at risk. The researchers combined this financial map with a sophisticated computer model of how earthquakes happen. They analyzed decades of seismic records to identify two distinct types of earthquake sources that threaten the region: shallow events near the coast where tectonic plates slide past each other, and deeper events occurring within a subducting plate far beneath the surface. By treating these as separate sources, the model can generate thousands of synthetic earthquake scenarios, each with a specific location, time, and magnitude.
The core of the study involves running these scenarios through a virtual version of the city. For every simulated earthquake, the model calculates how much the ground would shake at the city's location, accounting for the distance from the epicenter and the massive amplification caused by the lake sediments. It then translates that shaking into physical damage for every building in the registry, converting structural harm into monetary loss. To ensure the model reflects reality, the researchers calibrated it against the 2017 earthquake. They did not simply adjust the numbers to match the final bill; instead, they tested the model against five different independent facts from that event: the total money lost, the number of buildings that collapsed, the number of severely damaged structures, the geographic area covered by damage, and the age of the buildings that fell. This rigorous check confirmed that the model accurately reproduces the complex patterns of real-world destruction.
The results of this massive simulation, which ran through two hundred million years of hypothetical time, reveal a wide range of possible outcomes. The average annual loss is estimated at 2.26 billion Mexican pesos, but this average masks the extreme nature of the risk. In a typical year, the city might see almost no damage at all. However, the model predicts that a once-in-a-century event could inflict 40 billion pesos in damage. The uncertainty in these numbers is significant, driven largely by how much the soft ground amplifies the shaking. The researchers found that the potential loss could be as low as 14 billion pesos or as high as 439 billion pesos for a single century-scale event, depending on the specific behavior of the soil. This wide gap is not a failure of the model but a reflection of the current lack of precise data on the city's soil layers.
The study also uncovered a hidden layer of economic reality regarding how damage is reported. When the model was calibrated to the 2017 event, it showed that the official government tally of 16 billion pesos in losses captured only a fraction of the actual physical damage. The model suggests that the true economic cost of that earthquake was likely closer to 44 billion pesos, with the difference consisting of minor structural damage that owners repaired without reporting it to authorities. This finding implies that relying solely on official reports underestimates the true financial burden of seismic risk. The researchers emphasize that while their simulation provides a robust framework for understanding the city's vulnerability, the most critical next step is to obtain detailed soil maps and official construction values to narrow the wide range of possible losses. Until then, the simulation stands as the most precise estimate available, painting a picture of a city where the ground beneath it remains the single most powerful variable in its future safety.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.