Typology of U.S. transit systems reveals energy data gap and potential for diesel-based emissions reductions

Abstract We present a typology of 2205 U.S. transit systems, derived from 167 operational indicators in the National Transit Database, to reveal distinct patterns. Using kernel principal component analysis and Gaussian mixture modeling, we identify five distinct transit system types: Metro Bus–Rail , Mid-Sized Urban , Hybrid Multimodal , Urban Specialized , and Diverse Operations . We also discover a gap in diesel consumption reporting (76% of systems), which are predominantly of the Mid-Sized Urban type. To bridge this critical data gap, we apply extreme gradient boosting to predict the diesel consumption for 1674 underreported systems. We then estimate lifecycle emissions, accounting for regional grid carbon intensity. The result highlights significant variability in diesel consumption and emissions across transit types, with Metro Bus–Rail systems contributing the highest emissions. To evaluate decarbonization potential, we simulate fleet electrification and grid decarbonization. Our scenario analyses indicate that up to 82% of 10.27 MMTCO 2 e diesel-based transit emissions can be eliminated through fleet electrification, with further grid decarbonization eliminating the remainder. These findings provide actionable insights for policymakers to prioritize electrification, optimize resource allocation, and support sustainable transit strategies in the U.S.