Damage to structures by pyroclastic flows and surges, inferred from nuclear weapons effects

In order to define the risk from explosive eruptions, one must constrain both the probability of explosive events and the effects, or consequences, of those events. This paper focuses on the effects of pyroclastic flows and surges (here termed ‘pyroclastic density currents', or PDCs) on buildings, infrastructure elements, and to some extent on vehicles. PDCs impart a lateral force to such structures in the form of dynamic pressure, which depends on the bulk density of the PDC (which in turn depends mainly on particle concentration) and its velocity. For reasonable ranges of particle concentration (10⁻³ to 0.5) and velocities (10 to 300 m/s), dynamic pressure on the upstream face of a structure ranges from 0.1 kPa to 10⁴ kPa. Lateral loads ranging up to about 100 kPa were produced during nuclear weapons tests in the 1940s and 1950s that were designed to study the effects of such loading on a variety of structures for civil defense and emergency response purposes in the event of nuclear war. Although considerable simplifications are involved, the data from these weapon tests provide useful analog information for understanding the effects of PDCs. I reviewed data from the nuclear tests, describing the expected damage from different loadings. Tables are provided that define the response of different structural elements (e.g., windows, framing, walls) and whole structures to loading in probabilistic terms, which in principle account for variations in construction quality, orientation, and other factors. Finally, damage documented from historical eruptions at Mt. Lamington (1951), Herculaneum (AD 79 Vesuvius eruption), and St. Pierre (1902 Mt. Pelee eruption) is reviewed. Damage patterns, combined with estimates of velocity, provide an independent estimate of particle concentration in the PDCs. Details of structural damage should be recorded and mapped around future eruptions in order to help refine this aspect of consequence analysis. Another fruitful approach would be to combine numerical simulations of eruption scenarios, which can produce simulated maps of dynamic pressure, with GIS-based data on structures for a given region; the result would be predictions of consequences that could be used for planning and emergency response training.