Fuel supply disruptions caused by natural disasters can cascade rapidly through regional distribution networks, yet most emergency response models focus on last-mile delivery and overlook the upstream middle-mile segment that connects refineries to distribution terminals. This gap leaves emergency management organizations without tools to evaluate how refinery outages, pipeline failures, or disruptions to waterborne fuel imports via marine terminals propagate through the network. Using Southern California as a case study, this project extends an existing Operational Flow Capacity (OFC) framework into the middle mile through a discrete-event simulation, built on open-source Python tools, that quantifies fuel flows and identifies throughput bottlenecks under specified disruption conditions. The model was tested across three scenarios: a refinery fire, a major earthquake, and a 2-week disruption occurring together with the imminent decommissioning of the Phillips 66 refinery. In this last scenario, the combined effect reduced overall network throughput to 68% of its normal value over the duration of the disruption; applying a fuel-blending policy could partially offset this impact, raising throughput to 79.6%. More broadly, the results reveal that the network is resilient to short, localized disruptions, but that large, multi-asset failures can isolate inland cities like Las Vegas and Phoenix, with recovery taking nearly three months. Based on these findings, the project recommends pre-authorizing bundled emergency regulatory waivers, establishing minimum safety stock requirements at distribution terminals, and expanding pipeline capacity through targeted federal financial incentives; measures that, if adopted, could reduce supply shortfalls, shorten regional recovery timelines, and maintain critical fuel access during major disaster events. While Southern California serves as the primary case study, the simulation framework is designed to be transferable to other regional fuel networks, requiring only locally calibrated input parameters, making it a broadly applicable tool for emergency planners across the country.
