Rapid seismic screening of non-engineered URM buildings in liquefaction-prone areas

Rapid visual screening is essential for prioritising interventions in densely populated, data-scarce areas. However, this process is often complicated by the combined influence of structural vulnerabilities and geotechnical hazards. This study proposes an engineering-driven framework for rapid susceptibility screening that combines k-means clustering with parameter weighting derived from the analytic hierarchy process (AHP). This framework was applied to 322 non-engineered unreinforced masonry (URM) buildings in a liquefaction-prone area of Cali, Colombia. The identified susceptibility levels were evaluated against a calibrated macroseismic model, using the vulnerability index and mean damage grades for a 475-year return period. The inclusion of AHP weightings improved the agreement between susceptibility classification and benchmark damage estimates from 42% under standard unweighted k-means to 74% under the AHP-weighted framework, showing clear improvements across most damage categories. To support rapid urban prioritisation, structural susceptibility is integrated with simplified potential liquefaction impact profiles, enabling traffic-light-based screening classification. Results indicate that 57% of the buildings fall into the highest priority categories. Consistency analyses also show that approximately 81% of buildings maintain the same susceptibility class in the Monte Carlo simulations, confirming high stability under parameter uncertainty. The proposed framework is consistent with expected seismic behaviour and aligns well with EMS-98 vulnerability classes, highlighting the influence of liquefaction effects on urban prioritisation. The integration of structural and geotechnical interactions within the proposed framework provides a practical and adaptable tool for seismic prioritisation planning in data-scarce areas.