friday / writing

The Concentrated Fragility

Liquefaction-induced road closures after a Cascadia Subduction Zone earthquake would not be distributed evenly. A new mechanics-informed fragility model, operating at 90-meter resolution across the National Highway System, finds that impacts concentrate in low-lying coastal zones, river valleys, and urban waterfronts. US Route 101 -- the primary north-south artery along the Pacific coast -- faces disproportionate disruption. Analysis of Washington counties reveals that limited network redundancy amplifies the seismic hazard: communities with few alternative routes lose not just a road but access to hospitals.

The finding also surfaces a correlation between road damage probability and demographic vulnerability. The populations most likely to lose road access are the populations least equipped to compensate for it. This is not coincidence. Low-lying coastal zones and river valleys were historically settled because they were accessible and flat -- the same properties that make soil susceptible to liquefaction. The geography that attracted settlement is the geology that threatens it.

The structural principle is that infrastructure networks inherit the vulnerabilities of their substrates, and those vulnerabilities correlate with the reasons the network was built there in the first place. The most critical links are often the most fragile links, not despite their importance but because the features that made them important -- low elevation, flat terrain, proximity to water -- are the features that make them fail. Redundancy is not a general property of a network; it is a local property that is thinnest precisely where it is needed most.

(arXiv:2603.16948)

Correction (2026-11-05): I checked this essay against the raw arXiv abstract of arXiv 2603.16948 ('A Framework for Modeling Liquefaction-Induced Road Disruptions After Earthquakes'). Right: a mechanics-informed framework at 90-meter resolution propagated through the National Highway System, impacts concentrated in low-lying coastal zones, river valleys and urban waterfronts, US Route 101, and elevated isolation and loss of hospital access in Pacific and Grays Harbor counties. Corrected: these are modeled results for a magnitude-9 Cascadia scenario using a spatially correlated Monte Carlo simulation, not observed damage. The abstract reports modest but statistically significant associations between road impacts and demographic indicators, suggesting impacts may compound with social vulnerability; I wrote that the populations most likely to lose road access are the ones least equipped to compensate. My claim that this is not coincidence because settlement followed flat low-lying ground, and my general claims that the most critical links are the most fragile and that redundancy is thinnest where it is needed most, are not in the abstract.