Publication details
Journal: Atmosphere, vol. 17, 971, October 2nd 2026
Doi: doi.org/10.3390/atmos17100971
Arkiv: hdl.handle.net/11250/5594852
Archive: nva.sikt.no/registration/01a1a5f7df11-9d9f7a09-4d70-4271-b03e-0729d43d57b1
Summary:
Real-time single-particle fluorescence instruments are increasingly used for bioaerosol monitoring, yet interpretation of fluorescence-positive particles depends strongly on how the baseline or threshold is defined. This study evaluates deionised water (H2O), sodium chloride (NaCl), and quartz/silica (SiO2) as low-fluorescence references for the SwisensPoleno Jupiter and examines how baseline selection influences threshold-based interpretation. Controlled reference datasets were used to characterise baseline fluorescence distributions, assess particle size effects, and define baseline-specific thresholds using a consistent mean plus three standard deviations (μ+3σ) criterion. Threshold sensitivity was also examined using the empirical 99th percentile (P99) and a robust median absolute deviation (MAD)-based approach. The three reference materials produced distinct baseline responses. NaCl generated the broadest fluorescence ranges across all excitation sources, with μ+3σ upper thresholds of approximately 52.0×10−3, 39.3×10−3, and 38.2×10−3 arbitrary units, compared with 6.6×10−3, 10.1×10−3, and 3.1×10−3 for H2O and 12.1×10−3, 9.1×10−3, and 2.6×10−3 for SiO2. Although NaCl produced larger and more variable particles, its broader response persisted within a shared 10–30 µm size interval. Applying the thresholds to representative test materials showed that baseline choice altered both the proportion of particles exceeding the low-fluorescence background and their excitation–source combination patterns, particularly for near-baseline and moderately fluorescent samples. P99 generally produced results closest to μ+3σ, whereas MAD-based thresholds produced larger changes for several baseline–source combinations, while baseline–material differences remained evident across all three approaches. Summer and winter ambient case studies showed the same baseline dependence. H2O- and SiO2-based thresholds produced broadly similar fluorescence exceedance structures, whereas NaCl-based thresholds retained fewer weakly fluorescent particles. These findings show that both baseline material and threshold definition should be reported explicitly and that the resulting categories should be interpreted as operational fluorescence exceedance patterns rather than definitive biological particle classifications.