Hidden Danger on the Slopes: Why Earthquake Direction Could Decide Whether Hillside Buildings Survive

Hillside buildings have long carried a reputation for seismic vulnerability, but a new study suggests that engineers may have been underestimating the true nature of the threat. Research published in the Bulletin of Earthquake Engineering by Md Ghousul Ansari and Sekhar Chandra Dutta of the Indian Institute of Technology (ISM) Dhanbad, together with Sumit Kumar of Government Engineering College Palamu, reveals that two factors often treated separately in earthquake engineering, the bidirectional character of ground shaking and the angle at which seismic waves strike a structure, combine to make buildings on sloping ground far more fragile than conventional analyses imply. The findings arrive at a moment when rapid urbanization in mountainous regions across the world is pushing more construction onto steep terrain, making the results both timely and potentially consequential for building codes.

The research team focused on three configurations of a five-storey reinforced concrete building. The first was a regular building standing on flat land, serving as the baseline for comparison. The second was a step-back building, in which floors are progressively shortened on one side as the structure climbs the slope, creating a vertically irregular profile. The third was a step-back setback building, which adds a further geometric complication: not only do floors shorten toward the hill, but columns at the downhill side are of vastly different heights, a configuration common in towns built along Himalayan and other mountain slopes. These three systems allowed the researchers to isolate how each type of geometric irregularity alters seismic demand.

At the heart of the study lies a demanding computational campaign. The authors subjected their building models to nonlinear time-history analyses, meaning the structures were simulated through the full recorded shaking of real earthquakes, with the model free to crack, yield, and redistribute forces just as a real reinforced concrete frame would. Twenty far-field ground motions drawn from the publicly available PEER NGA database supplied the excitation. Each record was applied first in a single horizontal direction and then in both horizontal directions simultaneously, reflecting the reality that earthquakes almost never shake the ground along a single clean axis. The team also varied the record scale factor, systematically scaling the intensity of each motion to probe how demand grows as shaking strengthens.

The headline metric was the maximum inter-storey drift ratio, essentially how far one floor displaces sideways relative to the floor beneath it. Drift is a trusted proxy for structural damage: excessive drift means cracked columns, yielded reinforcement, and, in the worst cases, collapse of individual storeys. When the buildings were shaken in one direction only, the flat-land building produced maximum drift ratios ranging from 0.39 to 1.35 percent across the scaled records. The step-back setback building ranged from 0.48 to 1.67 percent, and the step-back building, the most geometrically compromised of the three, ranged from 0.74 to 2.23 percent, already signaling its heightened vulnerability.

The striking result emerged when shaking was applied in two directions at once. Under bidirectional excitation, the drift demands climbed sharply: the flat-land building’s range rose to 0.70 to 2.13 percent, the step-back setback building to 0.86 to 2.33 percent, and the step-back building to 0.93 to 2.79 percent. In other words, simply accounting for the second horizontal component of ground motion, something many simplified analyses still ignore, pushed every configuration toward substantially larger deformations. The physical explanation lies in bidirectional interaction: as columns yield in one direction, their capacity to resist forces in the perpendicular direction degrades, so the two components of shaking effectively erode each other’s resistance and demand accumulates in ways a unidirectional analysis cannot capture.

But the study did not stop there. The direction in which a ground-motion record is oriented relative to the building, known as the incidence angle, was varied systematically. This matters because earthquake records contain two horizontal components, and rotating the record changes which component aligns with the building’s strong and weak axes. For regular flat-land buildings, this rotational sensitivity exists but is comparatively modest. For the hillside configurations, the effect was dramatic. At critical incidence angles, the normalized drift demand of the hillside buildings reached roughly 2.0 to 2.8 times the response of the flat-land building. The step-back building proved especially direction-sensitive, its worst-case angles exploiting the mismatch between the short, stiff columns on the uphill side and the tall, flexible columns at the bottom of the slope.

To convert these deterministic results into a probabilistic picture of risk, the team turned to incremental dynamic analysis, a method in which each ground motion is scaled through a series of increasing intensities until the structure approaches its limits. From this family of analyses they constructed fragility curves, which express the probability that a building will exceed a given damage state as a function of ground-motion intensity, here measured by peak ground acceleration. The damage thresholds were drawn from the drift limits specified in FEMA 356:2000, the widely used American prestandard for seismic rehabilitation, allowing the curves to distinguish between levels of damage from minor cracking through to more severe structural compromise.

The fragility analysis delivered a nuanced verdict. Across the board, the step-back configuration generally emerged as the most seismically vulnerable of the three, consistent with its larger drift demands under both unidirectional and bidirectional shaking. Yet the researchers found something more unsettling: the relative ranking among the flat-land, step-back, and step-back setback buildings is not fixed. It shifts with the seismic incidence angle. A building that appears safer than its neighbor when the ground motion strikes from one orientation can appear more dangerous when the same record is rotated. For code writers and practitioners, this undermines any single-scenario assessment and suggests that hillside structures should be evaluated across a spectrum of incidence angles rather than at one assumed orientation.

For a rapidly urbanizing mountain world, the implications are hard to ignore. Cities such as those sprawling across the Indian Himalayas, the Andes, the Mediterranean hillsides of Turkey and Italy, and countless other sloping terrains host millions of reinforced concrete frames whose geometry resembles the step-back and step-back setback prototypes studied here. Previous research has already documented elevated collapse risk for hillside buildings in the Himalayan belt, and recent studies have proposed new descriptors of the coupled irregularities these structures exhibit. The present study adds a critical layer: even a thorough fragility assessment may significantly understate risk if it considers only one direction of shaking or one angle of incidence. The authors’ demonstration that record-to-record variability, illustrated by plotting mean responses together with mean plus and minus one standard deviation for selected cases, further complicates the picture, reminding engineers that individual earthquakes can deviate substantially from average predictions. The practical message is sobering but actionable: as hillside construction accelerates, seismic evaluation of buildings on slopes should routinely incorporate bidirectional excitation and sweep across incidence angles, because on sloping ground, the direction an earthquake comes from may matter almost as much as its strength.

Subject of Research: Seismic response and fragility of reinforced concrete hillside buildings under unidirectional and bidirectional ground motions

Article Title: Seismic response of hillside buildings under unidirectional and bidirectional ground motions with varying incidence angles

Article References: Ansari, M. G., Dutta, S. C., & Kumar, S. (2026). Seismic response of hillside buildings under unidirectional and bidirectional ground motions with varying incidence angles. Bulletin of Earthquake Engineering. https://doi.org/10.1007/s10518-026-02720-5

Image Credits: AI Generated

DOI: 10.1007/s10518-026-02720-5

Keywords: hillside buildings, seismic response, bidirectional ground motion, incidence angle, inter-storey drift, fragility analysis, incremental dynamic analysis, nonlinear time-history analysis, vertical irregularity, step-back setback, reinforced concrete frames, earthquake engineering