Global assessment of bike-sharing sensitivity and resilience to extreme weather

Source abstract

As climate extremes intensify, assessing how bike-sharing systems respond and recover is vital for urban resilience. This study develops a comparative framework using a time-series Transformer model and scenario analysis to evaluate demand under extreme heat, cold, heavy rain, and snowfall, using hourly trip and weather data from 14 cities worldwide (2023–2025). The sensitivity analysis shows that extreme heat reduces demand by over 40% in hotter cities but remains stable in temperate ones, while extreme cold and snowfall cause the strongest suppression, often exceeding 80%. To capture resilience, we simulate recovery curves after rainfall and snowfall shocks: North American systems typically rebound within 6–7 hours, whereas Montreal and Seoul show more prolonged disruptions, while Oslo and New York recover faster with winter operations and flatter demand profiles. These findings highlight how climatic exposure and operational readiness shape resilience and provide evidence to support climate-adaptive management and long-term system planning.