The Fair-Weather Lane: Capacity Credit and Grid-Style Market Design for Urban Air Mobility

Urban air mobility (UAM) is often promised as a way to fly over road congestion. This paper argues that the promise fails at exactly the moments it matters most. Small electric aircraft have almost no ability to store a queue in the air, and they are grounded by the same weather that degrades roads. In electricity terms, UAM is an intermittent resource. Its capacity credit, the capacity a planner can count on during stress periods, is far below its nominal capacity. We test this idea with ten years (2016–2025) of hourly reanalysis weather for six cities: Seoul, Tokyo, New York, London, Los Angeles and Dubai. We apply illustrative operating minima for wind gusts, precipitation and snowfall. These measures cannot capture low visibility or thunderstorms, so the availabilities reported are upper bounds. Under the baseline rule, weekday peak-hour availability ranges from 84% to 97%. In wet peak hours, when road capacity is known to fall, it drops to 66–84%. Grounding is 1.8–8.6 times more likely in wet peak hours than in dry ones. Even wind alone grounds aircraft 1.3–3.1 times more often in wet hours, so the coupling is not an artefact of the rain threshold. Every city also sees "stranded-return days", when the morning peak is flyable but the evening peak is not. There are about 5 to 33 such weekdays a year under the baseline rule, and about 19 in Seoul. Three implications follow. Transport appraisal should give UAM little or no credit against road or rail capacity. Operators should pay for the ground backup that society keeps on their behalf and guarantee passengers a way home. And because the sky, like the grid, must balance a non-storable capacity in real time, its allocation should borrow from four decades of electricity-market design. That means an independent slot operator, nodal and time-varying prices, firm contracts for scheduled mass vehicles, a real-time market for personal flights, and reserved capacity for emergencies. Working paper (not peer-reviewed). A Korean abstract is included at the end of the PDF. All numbers and figures are reproducible with fetch_weather.py and analyze.py (weather data: Open-Meteo / ERA5, CC BY 4.0).