Carbon accounting and weather-adaptive electrification dispatch for demand-responsive transit under carbon neutrality

Decarbonizing urban public transport requires not only the replacement of conventional vehicles with cleaner technologies, but also operational strategies that can respond to dynamic demand, service uncertainty, and external disturbances. Although existing studies have extensively examined bus electrification and public transport emission reduction, most focus on fixed-route systems or fleet-level assessment, leaving the order-level carbon performance and adaptive electrification potential of demand-responsive transit (DRT) insufficiently understood. To address this gap, this study develops an integrated analytical framework that combines order-level carbon accounting, weather-sensitive operation modeling, and electrification dispatch simulation. Based on large-scale real-world operational data from an urban DRT system, historical orders are integrated with weather observations, stop locations, and vehicle energy-consumption information. Econometric models are used to identify how weather conditions affect demand, service reliability, and carbon intensity, while scenario simulations are conducted to evaluate alternative electrification allocation strategies. The results show that DRT primarily serves short-distance feeder and intra-district trips, with substantial carbon heterogeneity across orders, vehicle types, operating periods, and service areas. Electric-vehicle orders have 28.7% lower operating-stage carbon intensity per passenger-kilometer than diesel-vehicle orders. Weather disturbances reshape both demand and cancellation risk, highlighting the need for adaptive operational management. In the full-substitution scenario, replacing all diesel-served orders with electric-vehicle services reduces total order-attributed emissions by 24.10%. When electric capacity is constrained, weather-adaptive dispatch outperforms simple demand-based priority allocation. These findings demonstrate that DRT decarbonization depends on the coordinated effects of vehicle electrification, passenger occupancy, demand aggregation, and priority allocation. The proposed framework provides an empirical basis for integrating order-level emission monitoring, demand aggregation, and weather-informed electric-vehicle allocation into low-carbon DRT operations.