From real-time to long-term source apportionment of PM 10 using high-time-resolution measurements of aerosol physical properties: methodology and example application at an urban background site (Aosta, Italy)

Identifying aerosol sources is essential for designing effective air quality policies. This study introduces a novel PM10 source apportionment approach – RASPBERRY (Real-time Aerosol Source apportionment using Physics- Based Experimental data and multivaRiate factoR analYsis) – based on the analysis of aerosol physical properties, namely particle size distributions in the accumulation and coarse modes (diameter in the range 0.18–18 μm) and spectrally resolved light absorption (at 7 wavelengths in the range 370–950 nm). The availability of such measurements at high temporal resolution, down to a few minutes, enables aerosol mass source apportionment from real time to long-term scales. To demonstrate the implementation of RASPBERRY, we apply the method to a 5-year hourly dataset (2020–2024) from an urban background site in the north-western Italian Alps, combining observations from a cost-effective optical particle counter (Palas Fidas 200) and an aethalometer (Magee Scientific AE33). RASPBERRY identifies six source factors contributing to PM10: traffic (9 %), biomass burning (10 %), two secondary aerosol modes (condensation, 23 %, and droplet, 16 %), desert dust (21 %), and local dust resuspension (21 %). Hourly resolved RASPBERRY estimates show strong agreement with traditional chemical source apportionment techniques when aggregated to daily resolution to match that of the chemical analyses. Further validation is provided through comparisons with ground-based remote sensing (lidar-ceilometers, sun photometers) and modelling tools (Validated ReAnalysis ensemble from the Copernicus Atmosphere Monitoring Service, CAMS). Selected realtime applications are also presented, including emergency surveillance during accidental events and the rapid identification of regional transport of secondary particles, as well as long-range transport of desert dust and Canadian wildfire smoke. The effective variance least squares (EVLS) method is additionally implemented within RASPBERRY as an enhanced variant (RASPBERRY+EVLS), enabling full propagation of uncertainties associated with both the source profiles and the measurements. Although demonstrated at a single site, RASPBERRY is readily transferable to international air quality networks engaged in aerosol mass source apportionment, as it relies on optical instruments commonly employed by regulatory authorities and environmental protection agencies. The RASPBERRY and RASPBERRY+EVLS codes and the dataset described in this paper can be freely accessed