@article{fdi:010095433,
  title = {{M}achine-learning-driven reconstruction of organic aerosol sources across dense monitoring networks in {E}urope},
  author = {{J}ouanny, {A}. and {U}padhyay, {A}. and {J}iang, {J}. {H}. and {V}asilakos, {P}. and {V}ia, {M}. and {C}heng, {Y}. and {F}lueckiger, {B}. and {U}zu, {G}a{\¨e}lle and {J}affrezo, {J}. {L}. and {V}oiron, {C}{\'e}line and {F}avez, {O}. and {C}hebaicheb, {H}. and {B}ourin, {A}. and {F}ont, {A}. and {R}iffault, {V}. and {F}reney, {E}. and {M}archand, {N}. and {C}hazeau, {B}. and {C}onil, {S}. and {P}etit, {J}. {E}. and de la {R}osa, {J}. {D}. and de la {C}ampa, {A}. {S}. and {N}avarro, {D}. {S}. {R}. and {C}astillo, {S}. and {A}lastuey, {A}. and {Q}uerol, {X}. and {R}eche, {C}. and {M}inguillon, {M}. {C}. and {M}aasikmets, {M}. and {K}eernik, {H}. and {G}iardi, {F}. and {C}olombi, {C}. and {C}uccia, {E}. and {G}ilardoni, {S}. and {R}inaldi, {M}. and {P}aglione, {M}. and {P}oluzzi, {V}. and {M}assabo, {D}. and {B}elis, {C}. and {G}range, {S}. and {H}ueglin, {C}. and {C}anonaco, {F}. and {T}obler, {A}. and {T}imonen, {H}. {J}. and {A}urela, {M}. and {E}hn, {M}. and {S}tavroulas, {I}. and {B}ougiatioti, {A}. and {E}leftheriadis, {K}. and {G}ini, {M}. {I}. and {Z}ografou, {O}. and {M}anousakas, {M}. {I}. and {C}hen, {G}. {I}. and {G}reen, {D}. {C}. and {P}okorna, {P}. and {V}odicka, {P}. and {L}hotka, {R}. and {S}chwarz, {J}. and {S}chemmel, {A}. and {A}tabakhsh, {S}. and {H}errmann, {H}. and {P}oulain, {L}. and {F}lentje, {H}. and {H}eikkinen, {L}. and {K}umar, {V}. and van der {G}on, {H}. {A}. {D}. and {A}as, {W}. and {P}latt, {S}. {M}. and {Y}ttri, {K}. {E}. and {S}alma, {I}. and {V}asanits, {A}. and {B}ergmans, {B}. and {S}osedova, {Y}. and {N}ecki, {J}. and {O}vadnevaite, {J}. and {L}in, {C}. {S}. and {P}auraite, {J}. and {P}ikridas, {M}. and {S}ciare, {J}. and {V}asilescu, {J}. and {B}elegante, {L}. and {A}lves, {C}. and {S}lowik, {J}. {G}. and {P}robst-{H}ensch, {N}. and {V}ienneau, {D}. and {P}revot, {A}. {S}. {H}. and {M}edbouhi, {A}. {A}. and {B}anos, {D}. {T}. and de {H}oogh, {K}. and {D}aellenbach, {K}. {R}. and {K}rymova, {E}. and {E}l {H}addad, {I}.},
  editor = {},
  language = {{ENG}},
  abstract = {{F}ine particulate matter ({PM}) poses a major threat to public health, with organic aerosol ({OA}) being a key component. {M}ajor {OA} sources, hydrocarbon-like {OA} ({HOA}), biomass burning {OA} ({BBOA}), and oxygenated {OA} ({OOA}), have distinct health and environmental impacts. {H}owever, {OA} source apportionment via positive matrix factorization ({PMF}) applied to aerosol mass spectrometry ({AMS}) or aerosol chemical speciation monitoring ({ACSM}) data is costly and limited to a few supersites, leaving over 80% of {OA} data uncategorized in global monitoring networks. {T}o address this gap, we trained machine learning models to predict {HOA}, {BBOA}, and {OOA} using limited {OA} source apportionment data and widely available organic carbon ({OC}) measurements across {E}urope (2010-2019). {O}ur best performing model expanded the {OA} source data set 4-fold, yielding 85 000 daily apportionment values across 180 sites. {R}esults show that {HOA} and {BBOA} peak in winter, particularly in urban areas, while {OOA}, consistently the dominant fraction, is more regionally distributed with less seasonal variability. {T}his study provides a significantly expanded {OA} source data set, enabling better identification of pollution hotspots and supporting high-resolution exposure assessments.},
  keywords = {source apportionment ; machine learning ; deeplearning ; {E}urope data set ; spatial-temporalanalysis ; air quality ; organic aerosols ; {EUROPE}},
  booktitle = {},
  journal = {{E}nvironmental {S}cience and {T}echnology {L}etters},
  volume = {12},
  numero = {11},
  pages = {1523--1531},
  ISSN = {2328-8930},
  year = {2025},
  DOI = {10.1021/acs.estlett.5c00771},
  URL = {https://www.documentation.ird.fr/hor/fdi:010095433},
}