Collapse misclassification dominates the uncertainty in city-scale earthquake fatality estimates: a validated probabilistic seismic risk model for the residential building stock of Peshawar, Pakistan

Abstract Probabilistic seismic risk estimates for rapidly growing cities in data-scarce regions are increasingly used to set mitigation priorities, yet the fatality component of these estimates is rarely tested. This study develops an updated event-based seismic risk model for the residential building stock of District Peshawar, Pakistan (361,357 buildings; 4.31 million people; USD 7.83 billion replacement value), using the seismic source model of the 2021 Building Code of Pakistan, a census-based exposure model mapped to seven Global Earthquake Model (GEM) building classes, and GEM analytical fragility functions. A hazard-consistent closed-form model is then derived by inverting the event-based aggregate loss curve into an equivalent spectral-acceleration hazard curve. The closed-form model reproduces the event-based average annual loss (AAL), and its 475- and 2,475-year intensities agree with the building-code hazard and 13 published hazard studies. The baseline AAL is USD 20.2 to 22.8 million (0.26 to 0.29% of replacement value), and unreinforced brick and adobe masonry account for 97% of it. Fatality estimates, however, depend critically on how the complete-damage state is translated into casualties. Treating every completely damaged building as collapsed, as is common in practice, yields 412 to 881 deaths per year, which, in a back-analysis of the 2015 Mw 7.5 Hindu Kush earthquake, can exceed the death toll of the entire province. Applying HAZUS collapse fractions, or GEM fatality vulnerability functions, reduces the estimate to 51 to 77 deaths per year (1.2 to 1.8 per 100,000) and reconciles the model with the 2015 observations. A nine-factor sensitivity analysis and a 1,500-run Monte Carlo propagation show that the casualty model, the hazard level and the collapse lethality dominate the uncertainty in fatalities (5 to 95% range 18 to 259 per year), whereas the choice of vulnerability model dominates AAL. Retrofitting single-storey adobe houses is five to six times more cost-effective per life saved than retrofitting brick masonry, although no strategy is economically justified on direct losses alone.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-16
DOI
https://doi.org/10.21203/rs.3.rs-11492895/v1
Primary Topic
Seismic Performance and Analysis
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Collapse misclassification dominates the uncertainty in city-scale earthquake fatality estimates: a validated probabilistic seismic risk model for the residential building stock of Peshawar, Pakistan

Shahid Ullah, Wasim Khan, Huma Aman
3 links
Research Square
Seismic Performance and Analysis
preprint

Collapse misclassification dominates the uncertainty in city-scale earthquake fatality estimates: a validated probabilistic seismic risk model for the residential building stock of Peshawar, Pakistan

Shahid Ullah, Wasim Khan, Huma Aman
preprint en

Abstract

Abstract Probabilistic seismic risk estimates for rapidly growing cities in data-scarce regions are increasingly used to set mitigation priorities, yet the fatality component of these estimates is rarely tested. This study develops an updated event-based seismic risk model for the residential building stock of District Peshawar, Pakistan (361,357 buildings; 4.31 million people; USD 7.83 billion replacement value), using the seismic source model of the 2021 Building Code of Pakistan, a census-based exposure model mapped to seven Global Earthquake Model (GEM) building classes, and GEM analytical fragility functions. A hazard-consistent closed-form model is then derived by inverting the event-based aggregate loss curve into an equivalent spectral-acceleration hazard curve. The closed-form model reproduces the event-based average annual loss (AAL), and its 475- and 2,475-year intensities agree with the building-code hazard and 13 published hazard studies. The baseline AAL is USD 20.2 to 22.8 million (0.26 to 0.29% of replacement value), and unreinforced brick and adobe masonry account for 97% of it. Fatality estimates, however, depend critically on how the complete-damage state is translated into casualties. Treating every completely damaged building as collapsed, as is common in practice, yields 412 to 881 deaths per year, which, in a back-analysis of the 2015 Mw 7.5 Hindu Kush earthquake, can exceed the death toll of the entire province. Applying HAZUS collapse fractions, or GEM fatality vulnerability functions, reduces the estimate to 51 to 77 deaths per year (1.2 to 1.8 per 100,000) and reconciles the model with the 2015 observations. A nine-factor sensitivity analysis and a 1,500-run Monte Carlo propagation show that the casualty model, the hazard level and the collapse lethality dominate the uncertainty in fatalities (5 to 95% range 18 to 259 per year), whereas the choice of vulnerability model dominates AAL. Retrofitting single-storey adobe houses is five to six times more cost-effective per life saved than retrofitting brick masonry, although no strategy is economically justified on direct losses alone.

Research Square
University of Engineering and Technology Peshawar (PK)
Sustainable cities and communities, Good health and well-being
Seismic Performance and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon