Abstract | Fragility curves are a seismic risk modeler’s bread and butter, relaying the probability of reaching or exceeding each limit state of interest given the ground motion intensity. Yet, as they convey essential information, they also hide assumptions, especially when used to characterize a group of similar or even seemingly identical structures. Chief among them is the concept that the dynamic properties of such structures are invariable and uncorrelated. However, the latter does not necessarily apply to groups of adjacent “identical” spherical pressure vessels, used for storing gaseous products in industrial facilities. The quantity of product contained within a vessel directly affects its dynamic response. Using an ensemble of four “identical” pressure vessels as a case study, a comprehensive set of fragility curves is developed, each corresponding to a different fill ratio of a single vessel. Then, different approaches are explored to combine said fragilities and assess the group of four. These include full-scale Monte Carlo simulation with or without filling level correlation, as well as the computation of a single “law-of-total-variance” fragility curve. The latter approach is decidedly simpler, yet its use cannot be justified without some knowledge of the facility’s operational profile in terms of day-today fluctuation of the filling level, as it can otherwise lead to unconservative and/or biased results.