Quantifying role of source variations on PM2.5-bound toxic components under climate change: measurement at multiple sites during 2018-2022 in a Chinese megacity

Understanding the response of PM_2.5-bound toxic components to source variations under climate change is crucial for public health protection. However, the lack of long-term and multi-site observational data of toxic components limits such efforts. Here, we conducted a five-year PM_2.5 measurement (2018-2022) at 10 sites across a Chinese megacity, analyzing 15 polycyclic aromatic hydrocarbons (PAHs), 6 organophosphate esters (OPEs), and 9 potentially toxic elements (PTEs). Using explainable machine learning, we found that source variations from particle matter mass reduction under climate change can impact PM_2.5-bound toxic components. Meteorological factors like extreme heat days and max temperature impact most toxic components, while geographic, socioeconomic, and anthropogenic factors mainly affect PTEs, especially Cu. We also designed 10 extreme heat and source variation scenarios to predict the response of toxic components. When comparing scenario 2-1 (source variation without temperature change) with scenario 2-2 and 2-3 (the same source variation but higher temperatures), many PM_2.5-bound organics and As show higher reduction rates under climate change, highlighting the need to focus more on gas-phase organics and products of atmospheric process. Benzo[b]fluoranthene (BbF) is most sensitive to traffic source reductions, and Cu, Mn, Zn and Fe are more sensitive to industrial source reductions.