RESUMO
The rapid progress of intelligent transportation systems (ITS) has enabled the development of a highly spatiotemporally resolved vehicular VOC emission inventory. However, up to this point, the emission factors applied in vehicular VOC emission inventories worldwide are either independent of driving conditions or estimated by emission models, resulting in significant bias. In this study, by using the speed-dependent VOC emission factor measured online from a typical fleet in Guangzhou and collecting multiple sources of ITS data, we developed, for the first time, a link-level dynamic vehicular VOC emission inventory. The results reveal that the emission factors for vehicles at speeds higher than 50 km/h are only around 30 % of those at 5-20 km/h. Consequently, the total vehicular VOC emission in Guangzhou is estimated to be 16.19 kt in 2019, around 40 % lower than the estimates by the static emission inventory using the average emission factor during a short transient driving (STD) cycle. This discrepancy is mainly due to the much lower average speed of the STD cycle (20 km/h) compared to the average traffic speed on the road network (36 km/h). The discrepancy in VOC emissions was even higher for highways, with the static emission factors being 75-93 % higher than the speed-dependent ones. Such a large discrepancy underscores the necessity of applying localised speed-dependent emission factors to improve the estimation accuracy of vehicular VOC emissions. This study provides more accurate insights for policymakers in formulating targeted strategies to reduce vehicular VOC emissions and mitigate their contributions to ozone and PM2.5 pollution in urban areas.