Managing Cascading Failures in Interdependent Urban Infrastructure: A Systematic Review of Propagation Dynamics, Spatial Bottlenecks, and Crisis Governance

Background Climate-driven failures in one urban infrastructure system often spread to others, yet evidence on interrupting them is split across propagation, critical-component and coordination studies. We asked how these cascades can be interrupted by linking how they propagate, where they can be stopped and who coordinates the response there. Methods Following PRISMA 2020, we searched Scopus and IEEE Xplore (24 September 2026), with OpenAlex citation searching (25 September 2026), for primary studies of interdependent urban infrastructure under extreme weather. Two independent reviewers (Reviewer 1 and Reviewer 2) screened, extracted and appraised studies under pre-established eligibility criteria; discrepancies were resolved through independent adjudication by a third senior reviewer (Adjudicator). Methodological quality was appraised using MMAT and a structured five-domain modeling checklist; synthesis followed SWiM with GRADE-CERQual. Results Of 6,434 records, 92 studies were included; 69 were simulations and 82 examined flooding. Power was the most modeled initiating system (30 of 89 propagation studies); road or access terms appeared in 48 of 89 pathways; in the two examined closely, the disruption was shared flood exposure rather than functional failure propagation (not coded). Bottlenecks were located mainly by simulation ranking and geographic overlay; at most 22 of 84 checked them against an observed event (17 strictly), and 7 treated institutional interfaces as bottlenecks. Just 21 linked all three dimensions. None of 26 governance studies reported an observed comparison of arrangements. Governance findings reflect the review’s operational criteria requiring explicit multi-infrastructure dependencies under climate hazards, highlighting that institutional response arrangements are rarely integrated into physical cascade models. Conclusions Retained studies locate where an urban cascade might be interrupted, not who interrupts it or whether that works, a gap partly reflecting separate analysis of assets and organizations. Modeling institutional interfaces as bottlenecks and evaluating real arrangements against observed events would narrow it.