Source-driven variability of particulate matter oxidative potential at urban and rural coastal sites in the northwestern Mediterranean

Health effects associated with particulate matter (PM) exposure are closely linked to the ability of particles to induce the formation of reactive oxygen species (ROS) and trigger oxidative stress. Accordingly, the oxidative potential (OP) of PM is considered a more health-relevant toxicity metric than mass concentration alone. However, in densely populated and ecologically sensitive areas in the northwestern Mediterranean, the main sources contributing to OP remain poorly constrained, particularly regarding differences between urban and rural environments.

This sutdy systematically evaluated the OP of total suspended particulate matter (TSP) at an urban coastal site (Endoume) and a rural coastal site (Banyuls) in the region. OP was quantified using the dithiothreitol (DTT) assay and evaluated the contributions of primary emission sources and secondary formation to OP. Chemical tracers, dual carbon isotopes (¹³C & ¹⁴C), and positive matrix factorization (PMF) were used to apportion the main local sources.

The results reveal pronounced differences in both magnitude and source contribution to OP between urban and rural coastal aerosols. The annual mean organic-carbon-normalized DTT activity (DTTm) at Endoume was 20.0 ± 9.1 pmol min⁻¹ μg⁻¹ and 17.0 ± 5.3 pmol min⁻¹ μg⁻¹ at Banyuls (Mann–Whitney U test, p = 0.06). The annual mean volume-normalized OP (DTTv) was comparable at both sites (≈ 0.04 ± 0.02 pmol min⁻¹ m⁻³, Mann–Whitney U test, p < 0.05).

At Endoume, DTTv showed strong positive correlations with traffic- and fossil-fuel-combustion-related (FF) tracer metals (Pb, Cu, Zn) and elemental carbon (ρ ≈ 0.60–0.66, p < 0.01), together with pronounced seasonal variability. In spring, OP was primarily controlled by traffic and industrial emission (ρ > 0.75); in summer, ship emission emerged as the dominant driver (V–DTTv: ρ > 0.9, p < 0.01); while in autumn and winter, the contribution from biomass burning (BB) increased substantially (DTTv–nssK⁺: ρ = 0.74, p < 0.01). In contrast, the OP at Banyuls was dominated by traffic emission in spring (Zn–DTTv: ρ > 0.7, p < 0.01), whereas BB and ship emission jointly influenced OP in summer (V–Ni: ρ ≈ 0.7, p < 0.01). Additionally, dust and sea salt contributed significantly to OP at both sites, with a more pronounced influence at Banyuls (nss-Ca²⁺–DTTv: ρ ≈ 0.85, p < 0.01). Carbon isotope analysis showed that autumn samples at both sites exhibited lower OC_NF and DTTv values, indicating that the influence of FF on OP may be more pronounced.

PMF results further show that at Banyuls, traffic emission and BB together accounted for approximately 25% of OPv, with natural dust contributing about 14%, whereas at Endoume, industrial emissions (25%), BB (20%), and traffic emission (19%) were the major contributors to OPv. For OPm, industrial emission dominated at Endoume, while natural sources such as sea salt and dust were the primary contributors at Banyuls; secondary formation processes contributed substantially to OPm at both sites. Overall, this study demonstrates strong spatial and seasonal source dependence of PM oxidative toxicity in the northwestern Mediterranean coastal region, providing important constraints for health-oriented air pollution assessments.