Near-fault ground motions drive directional seismic response in concrete buildings

When the Mw 7.7 Pazarcık earthquake tore through southeastern Türkiye on 6 February 2023, it did more than level buildings — it handed engineers an unprecedented natural experiment in how the direction of ground shaking can determine whether a structure stands or falls. A new study published in the Bulletin of Earthquake Engineering now shows that for reinforced concrete buildings in near-fault zones, the worst-case direction of shaking is not simply the direction of the strongest seismic pulse, but the product of a subtle and sometimes counterintuitive interaction between the ground motion itself and the internal architecture of the building. The finding carries significant implications for how seismic performance is assessed in cities built directly atop active fault systems.

The research team — Emre Gani and Ali Sari of Istanbul Technical University, Ibrahim O. Dedeoglu of Batman University, Osman E. Ozbulut of the University of Virginia, and Musa Yetkin of Firat University — conducted nonlinear time-history analyses of a representative reinforced concrete building drawn from the building stock of Antakya, a city devastated in the February 2023 earthquake sequence. Their focal case is the TK-3124 strong-motion record, a recording that has drawn intense scientific attention because it contains a pronounced velocity pulse and exhibits exceptionally high spectral accelerations across a broad range of vibration periods. Such pulse-like motions are the signature of near-fault ground shaking, where the forward-rupturing front of an earthquake concentrates seismic energy into a single, coherent long-period surge.

Velocity pulses matter because they strike buildings in a fundamentally different way than the quasi-random, broadband shaking recorded farther from faults. When the period of a velocity pulse overlaps with the natural vibration period of a structure, the pulse can drive the building into large, one-sided displacement excursions — a phenomenon long associated with the concentrated inelastic demands and residual deformations observed after major earthquakes. The TK-3124 record, with its spectral energy spread across a wide period band, is particularly dangerous in this respect: it can simultaneously punish long-period buildings with large displacements and short-period buildings with high accelerations, leaving few structural typologies untouched.

What distinguishes the new work, however, is its explicit treatment of directionality. Ground motions recorded near faults are strongly polarized: the horizontal energy is not distributed evenly in all compass directions, but is concentrated along particular orientations controlled by the geometry of the fault rupture. Most engineering practice implicitly ignores this by rotating records to a standard orientation or by using orientation-independent intensity measures. The researchers instead rotated the record through the full range of incident angles and applied it to their building model in each orientation, tracking how damage demands shifted as the loading direction changed. The results were unambiguous: directionality significantly influences both the magnitude of structural response and the spatial distribution of damage, and its effects become even more pronounced when the motion is pulse-like.

Perhaps the study’s most striking conclusion is that the critical direction — the orientation that produces the most severe structural demands — does not necessarily coincide with the direction of maximum pulse intensity. Instead, it emerges from the interaction between the ground-motion characteristics and the configuration of the building’s lateral load-resisting system. A frame with strong axes in one orientation and weak axes in another will respond most severely when the polarized motion aligns unfavorably with those axes, not simply when the strongest horizontal component arrives along a particular compass bearing. This means that two structurally similar buildings on the same city block, oriented differently with respect to the fault, can experience profoundly different damage from the very same earthquake.

The team’s analytical approach centered on detailed nonlinear modeling of the representative RC building, capturing the post-yield behavior of beams, columns and joints as the structure was pushed through the full range of loading directions. By comparing deformation demands, story drifts and damage patterns across orientations, they were able to map how inelasticity migrated through the structure as the incident angle changed. Under pulse-like excitation, the velocity pulse component contributed substantially to increased inelastic demands through spectral amplification in the long-period range, concentrating damage in particular stories and particular plan locations depending on orientation — a finding that helps explain the starkly uneven damage patterns documented across Antakya after the 2023 earthquake.

The researchers also explored how structural modifications alter the picture. Adding perimeter beams and adding shear walls — two of the most common strengthening strategies for deficient reinforced concrete frames — were found to influence seismic performance by changing the structure’s stiffness, redistributing load-transfer mechanisms, and reshaping the spatial pattern of damage throughout the system. Shear walls in particular can dramatically shift where energy is dissipated and where cracking and crushing concentrate. But the study adds an important caveat: these modifications shorten the building’s natural period, and a stiffer structure can actually attract higher acceleration demands in certain cases. Strengthening, in other words, is not a guaranteed improvement in every ground-motion scenario; its benefit depends on where the modified structure’s period sits relative to the spectral content of the expected shaking.

This caveat resonates with the broader lessons emerging from the 2023 Kahramanmaraş earthquake doublet. The Mw 7.7 Pazarcık event and the Mw 7.6 Elbistan event that followed it hours later ruptured segments of the East Anatolian Fault with remarkable speed, with evidence of intermittent supershear rupture — crack propagation faster than the shear-wave velocity — which sharpens and intensifies near-fault pulses even further. Antakya, sitting in a sedimentary basin that further amplified long-period motion, suffered catastrophic damage despite lying tens of kilometers from the surface rupture in places. Previous work by members of the same team had already linked the severity of destruction in Antakya to the combination of pulse-like motions, supershear rupture and basin effects; the new study extends that narrative from the ground surface into the buildings themselves.

The implications for engineering practice are considerable. Modern seismic codes, including Türkiye’s current building code, generally assess structural performance using ground motions applied along the principal axes of the building, effectively assuming that shaking direction is a secondary concern. The new results argue that in near-fault regions this assumption can be dangerously optimistic. Realistic seismic performance assessment, the authors conclude, requires the integrated consideration of directionality effects, pulse-like ground motions and structural system characteristics — treating these not as independent factors to be accounted for separately, but as interacting elements of a single problem. A building that passes a code check under axially applied motions could still face critical demands under an obliquely oriented pulse.

For cities like Antakya — and for the many urban centers worldwide that sit near major strike-slip faults, from the San Andreas to the North Anatolian Fault — the message is sobering but actionable. Retrofit decisions should account for the orientation of a building relative to the dominant fault and the expected polarization of near-fault pulses. Strengthening strategies that merely stiffen a structure may shift rather than resolve its vulnerability if the modified period aligns with amplified spectral content. And in the assessment of existing buildings, the critical direction should be identified explicitly, rather than assumed to follow the strongest recorded component. The wreckage of 2023 demonstrated that near-fault shaking is directional; the new research shows exactly how that directionality propagates through concrete, steel and rebar into the pattern of survival and collapse on the ground.

The study also contributes methodologically to a growing body of literature on orientation-dependent seismic demand. Recent work on the 2023 Kahramanmaraş doublet has documented strong directionality and polarization of response spectral ordinates across the affected region, and researchers have proposed modified ground-motion models to estimate orientation-dependent spectra in strike-slip earthquakes. By coupling these seismological insights with detailed structural analysis of an actual, representative building, the new study bridges a gap that has long separated earthquake scientists from structural engineers: the ground motion community has known that pulses are directional, and the structural community has known that buildings have weak axes, but quantifying the collision of the two in a real damaged urban environment has remained rare.

As Türkiye embarks on the enormous task of assessing and retrofitting hundreds of thousands of buildings in the earthquake zone, findings of this kind offer a more refined lens than simple intensity-based triage. The worst building is not always the one closest to the fault or the one struck by the largest recorded acceleration; it may be the one whose weak direction happens to face the pulse. Understanding that geometry — and designing against it — could mean the difference between a building that cracks and one that collapses the next time the East Anatolian Fault lets go.