<p>Access to mobile cell phone service is an important factor in our response to natural hazards. Mobile phones are crucial tools for facilitating communication between those affected by a disaster and first responders or family members. However, damage to physical network components and power outages can lead to cell phone service outages during and after a natural disaster. To better understand the risk of cell phone service outages, we propose a novel framework to estimate areas within a community most likely to lose cell phone service during a natural hazard. Using only publicly available data on cell network infrastructure, we generate a nominal coverage map with a genetic algorithm, expanding on prior attempts to model cell networks under hazards, which have used entirely synthetic networks with fixed coverage polygons. We apply a hazard event to the generated network using Monte Carlo simulation and measure the resulting outages across our simulated scenarios. We demonstrate this framework using an example of Ann Arbor, MI (USA) under high wind scenarios. This work improves on prior attempts to model cell phone network vulnerability by incorporating real tower and antenna data and by explicitly modeling overlapping service areas rather than using stylized hexagon service polygons. This framework can help decision makers prioritize vulnerable areas for backup power and temporary cell service hotspots during and after hazard events.</p>