• Citer

Valorizing granulometric information for geospatial modeling and source apportionment of heavy metals in road dust in Brussels

par Mora, Adrien
Président du jury Debaste, Frédéric
Promoteur Verbanck, Michel
Co-Promoteur Bogaert, Patrick
Publication Non publié, 2024-06-25

Mémoire

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

The study of urban road dust pollution and heavy metal contamination is a major scientificinterest as it meets health and environmental concerns. The current research explores thevalorization of high-resolution granulometric information for the characterization, thegeospatial modeling and the source apportionment of road dust in Brussels.The performance of a multivariate regression model predicting the geospatial distribution ofroad dust heavy metal concentrations in BCR depending on urban conditions presents limitedperformance (determination coefficient ≈ 0.5). The present study shows that this limitation isnot related to granulometric phenomena. Higher heavy metal concentrations are observed forroad dust samples with a higher median particle size, which is unexpected as fine particles aregenerally considered to be more contaminated with heavy metals. An effect of solid-to-soliddilution of road dust by roadside soil cannot be highlighted by comparing granulometricdistances.Source apportionment by Positive Matrix Factorization (PMF) allows to restructure a sampledataset along a combination of modeled factors. The inclusion of high-resolution particle sizedistribution into a source apportionment approach by Positive Matrix Factorization does notlead to convergent solutions. Granulometric information has to be grouped into a limitednumber of size bins to obtain valid solutions. The optimal number of PMF factors isdetermined as being 4, the inclusion of granulometric information does not allow thedetection of an additional number of factors. Two factors contributed to the majority of theheavy metal loads, presenting respective associations between the metals Cr-Cu-Ni andPb-Zn-Ba. The profile of the factors does not allow their identification as specificenvironmental sources. It is unsure whether these factors correspond to complexarrangements of environmental sources or to mere mathematical constructions.