When I finished my civil engineering training, Colombia’s NSR-10 — the Reglamento Colombiano de Construcción Sismo Resistente, the country’s seismic-resistant construction code — was no longer new. It had been in force for years and was simply the framework under which every structure in the country got designed. But working under a code isn’t the same as understanding why it’s organized the way it is, or why — even though it’s been called “NSR-10” since 2010 — it’s still the subject of update decrees, technical errata, and standing commissions actively revising it. Designing under this code, including a project like the Teatrino — a reinforced-concrete amphitheater that required checking in-plan torsion and several vibration modes — made it clear to me that NSR-10 isn’t a static document. It’s a living body that absorbs every real earthquake, every new microzonation study, and every advance in international seismic engineering.

What it is and how it’s organized

NSR-10 was born after the 1999 Armenia earthquake, which exposed serious gaps in the previous code (Ley 400 of 1997 / NSR-98), particularly around unreinforced masonry and the lack of real enforcement of technical supervision on site. It replaced NSR-98 through Decreto 926 of 2010, and since then it’s been structured into eleven titles — A through K — each with a distinct scope: Title A sets the general seismic-resistant design requirements (seismic hazard zones, design spectra, use groups); Title C covers structural concrete; Title E, structural masonry; Title G, steel structures; Title H, geotechnical studies; and so on through Title K, which regulates complementary requirements like utility networks and equipment.

This modular structure is deliberate: it lets a single title — say, Title A’s seismic hazard map — get updated without rewriting the entire code. It’s, in a sense, the same design principle I value in any well-built system, whether it’s a concrete structure or a data pipeline: modules with clear responsibilities that can be versioned independently.

Why periodic revision matters

Seismic hazard isn’t a fixed number — it gets recalculated through probabilistic seismic hazard studies that incorporate new instrumental seismicity, updated attenuation models, and finer municipal microzonation. Bogotá, Medellín, and Cali, for instance, each have their own microzonation studies that refine locally what Title A’s national map proposes in general terms. Every destructive real earthquake, inside or outside Colombia, leaves lessons that eventually filter into the code: the 2017 Mexico City earthquake reinforced the discussion around torsional irregularity in mid-rise buildings; the 2023 Turkey-Syria earthquakes reopened the international debate on construction-site supervision quality — something NSR-10 already requires explicitly under Title I, but which in practice remains the weakest link in the chain.

Understanding how earthquakes are generated — plate mechanics, released energy, magnitude versus surface-felt intensity — is, in fact, the underlying input for any serious revision of the design spectrum: the code translates that physics into a number a structural engineer can plug directly into analysis.

Practical application in structural design

Day to day, NSR-10 feels less like a reference book and more like a sequence of mandatory decisions:

  1. Classify the project by use group (I through IV), which sets an importance coefficient and demands higher seismic performance for essential buildings like hospitals.
  2. Define the site’s seismic hazard zone (high, intermediate, or low) per Title A and any local microzonation study.
  3. Build the elastic design spectrum and, depending on the chosen structural system, apply the corresponding energy-dissipation capacity coefficient (R).
  4. Check for plan and vertical irregularities — torsion, soft story, diaphragm discontinuity — since every detected irregularity demands a more rigorous analysis, not just a penalty factor.
  5. Document the technical supervision required under Title I, which in practice is as decisive for a structure’s real-world performance as the calculation itself.

That last point is, to me, the code’s most underrated one: a flawless design on paper is worthless if the structure isn’t built with the exact seismic detailing — anchorages, development lengths, confinement in critical zones — the calculation assumed.

Where the code is headed

Today I work more with data than with drawings, but the underlying logic hasn’t changed: just as a machine learning model gets recalibrated when new data arrives, NSR-10 gets recalibrated when new seismic evidence arrives. The technical commissions discussing its update — under Colombia’s Comisión Asesora Permanente para el Régimen de Construcciones Sismo Resistentes — now work with more instrumental data than ever, fed by accelerograph networks and increasingly refined structural engineering studies. That intersection between growing empirical evidence and normative regulation is, at its core, the same discipline I apply today analyzing data: no model, seismic or statistical, stays still in the face of new evidence.