<p>The routing of material over a distribution network is subject to man-made and natural disruptions, and it is important to understand the network’s spatiotemporal vulnerabilities, i.e., when and where such disruptions will notably affect outcomes. Knowledge of vulnerabilities can inform either protect against disruption or expedite recovery from it to ensure shipments are routed efficiently while meeting any delivery deadlines. This research formulates and examines the bilevel material routing problem, wherein an upper-level problem identifies the respective times and locations for a limited number of fixed-duration attacks on arcs within a distribution network, and a lower-level problem subsequently routes shipments over the network between respective origins and destinations. The defender is presented in the lower-level problem and lexicographically prioritizes minimizing the number of shipments arriving beyond their required arrival time, the penalties for delivering shipments later than desired times, and the shipment weighted distance travelled. The attacker is presented in the upper-level problem and seeks to maximize both the shipment weighted distance travelled and the likelihood the attack strategy is successful. This research leverages a non-dominated sorting genetic algorithm (NSGA-II) to develop high-quality solutions with respect to the attacker’s competing objectives. For a baseline instance of the underlying problem, testing examines the set of the attacker’s Pareto optimal solutions to identify spatiotemporal vulnerabilities over a range of attacker prioritizations. Subsequent testing examines pseudorandom instances of the underlying problem to further investigate the presence of spatiotemporal vulnerabilities among varying routing configurations by the defender to garner additional insights.</p>