Comparison of emission models and top-down/bottom-up approaches on the estimation of vehicular emissions, and its use in air quality modeling in the city of Manizales

Abstract

Vehicular emissions are the main contributors to air pollution in urban centers, causing adverse health outcomes; therefore, an accurate estimation of the emission fluxes from this source is vital. In this work, three vehicular emission inventories were developed using two emission models (IVE and COPERT) and two estimation approaches (Top-down vs Bottom.up). The inventories were developed for the same region (Manizales) and base year (2017), and were validated through air quality simulations using the WRF-Chem model. For that purpose, WRF-Chem predictions of CO were compared against ground measurements in order to identify which inventory led to the most accurate representation of pollution patterns in the city The emission inventories exhibited noticeable differences. Using the COPERT/Top-down inventory as reference, CO emissions with the COPERT/bottom-up and IVE/Top-down inventory were 44% and 118% higher, respectively. The discrepancies between the COPERT/Top-down and COPERT/Bottom-up inventories were caused because the later approach led to higher activity factors; hence, increasing emissions. On the other hand, the differences between the COPERT/Top-down and IVE/Top-down inventories were generated due to the different approaches to relate emission factors with activity factors of each emission model. COPERT relates emission factors with the average traveling speed, while IVE considers the vehicle specific power, which accounts not only for average speed, but for acceleration, vehicle weight, road slopes among other variables. WRF-Chem predictions exhibited a strong sensitivity to the emission inventories. Although all the scenarios led to underestimations of CO, the IVE/Top-down emission inventory allows a reduction of mean bias and mean gross error; hence, providing the most accurate results. On the contrary, the worst model performance was obtained with the COPERT/Bottom-up emission inventory. These results indicated that IVE emission model is more representative of the actual vehicle operating conditions in the city of Manizales.