Effects of Expressway Geometric Alignment on Passenger-Car CO2 Emissions and Development of a Segment-Level Prediction Model

Carbon dioxide (CO2) emissions from passenger cars constitute a substantial share of global road transport emissions, yet the impact of expressway geometric alignment on real-world CO2 emissions remains insufficiently quantified, particularly with respect to the combined effects of alignment parameters and the need for macro-scale prediction. Using On-Board Diagnostics field tests on a typical expressway, this study evaluates the effects of six canonical alignment parameters: horizontal curve radius, horizontal curve length, vertical curve radius, vertical curve length, grade, and grade length. A random forest regressor identifies grade as the dominant predictor, accounting for 59.7% of the total feature importance, approximately six times the value for grade length. Conversely, horizontal and vertical curve parameters exhibit minor effects on emissions. Based on these findings, a prediction model is developed, integrating a steady-state emission term derived from standard certification cycles with a grade-based geometric correction term. Unlike existing models reliant on second-by-second trajectory data, the proposed model requires only segment-level alignment inputs. Independent field validation using a portable emissions measurement system on another expressway, with the test route segmented into 88 grade-defined sections, demonstrates high segment-level accuracy, with overall relative errors of 10.41% and 8.95% on opposing directional lines, respectively. This work provides a computationally efficient tool for low-carbon expressway design evaluation and operational carbon accounting.