Improved BVOC estimation accuracy by integrating high-resolution land cover and species-specific emission factor in a subtropical city.

Biogenic volatile organic compounds (BVOCs), key precursors of near-surface ozone (O3), remain highly uncertain because coarse-resolution land cover data and plant functional type (PFT)-based emission factors (EFs) cannot capture fragmented vegetation and interspecific variability. Here, we developed a 10 m species-specific BVOC emission framework for Xiamen, China, by integrating satellite data with ground-based plant surveys. A localized BVOC EF database covering 115 tree species was established using field measurements of 14 dominant species and literature data. High-resolution mapping better resolved fragmented urban vegetation, increasing urban contributions from 1.71% at 500 m to 5.58% at 10 m. The species-specific framework estimated higher annual BVOC emissions (35.03 Gg) than the 500 m (23.96 Gg) and 10 m (31.29 Gg) PFT-based approaches, while correcting the underestimation of monoterpene emissions in pine-dominated forests caused by uniform PFT-based EFs. Monte Carlo simulations showed markedly reduced uncertainty, with the coefficient of variation decreasing to 1.69. O3 formation potential were concentrated in mountainous forests, where Pinus massoniana accounted for 28.26%-34.59% of total emissions. This study highlights the importance of integrating high-resolution and species-level data to improve BVOC estimates and support targeted O3 mitigation strategies.