Uncovering the Uncertainties and Variability of Travel Time and Fuel Consumption Using Floating Car Data: A Case Study in Wuhan

Understanding the probabilistic behavior of both travel time and fuel consumption is critical for improving urban traffic efficiency and developing reliable urban transportation systems. While previous studies have extensively examined travel time distributions, the joint distributional patterns of fuel consumption remain underexplored, limiting the effectiveness of green routing strategies. A novel data-driven framework, integrating map-matching, second-by-second trajectory interpolation, and microscopic fuel consumption estimation (CMEM) is developed for jointly analyzing the distributions and spatiotemporal variability characteristics of link-level travel time and fuel consumption using floating car data (FCD) in Wuhan, China. The results reveal that hourly travel time follows a Lognormal distribution for over 80% of road links, while fuel consumption is best represented by a Gamma distribution in more than 30% of links. Peak-hour analyses (08:00 and 18:00) indicate substantial increases in both the mean and 90th-percentile values of the two metrics, reflecting heightened uncertainty during congestion. Coefficients of variation are consistently higher on weekdays than on weekends, emphasizing increased variability in urban mobility. Hot spot analysis further shows that areas with dense traffic signals and commercial activities tend to form clusters of high travel time and fuel consumption. These findings highlight the necessity of incorporating reliability and energy efficiency considerations into urban traffic management and eco-routing navigation systems.