An Integrated Framework for VOC Source Apportionment Based on Chemical Transport Modeling and Observation-Constrained Optimization

Volatile organic compounds (VOCs) undergo extensive atmospheric processing, complicating source impact quantification. In this study, we developed an integrated framework for VOC source apportionment by coupling a chemical transport model (CTM) with observation-constrained optimization and applied it to five supersites in the Yangtze River Delta (YRD), China, during May–September 2018. Index of agreement (IOA) was increased by 15–114%, and the root mean square error (RMSE) decreased by 18–71%. Except at the densely vegetated Chenchai Reservoir (CC) site, the cosine similarity between emission and receptor-derived profiles for natural sources ranged from 0.25 to 0.37 at anthropogenically influenced sites, indicating substantial profile divergence driven by complex atmospheric processing. Low-reactivity alkanes were enriched at receptor sites, whereas highly reactive species, such as alkenes and isoprene, were depleted and showed greater temporal variability, reflecting cumulative atmospheric aging. The source apportionment was substantially improved through the markedly better agreement between simulations and observations. Moreover, the receptor-derived source profiles were more environmentally representative, providing a stronger basis for robust source identification by receptor models. The integrated framework may also be extended to PM2.5 source apportionment, even in systems with substantial secondary aerosol contributions.