Re-Interpretation of Metal(Loid) Concentrations in Urban Soils of Two Different Land Uses by Positive Matrix Factorisation

Positive matrix factorisation (PMF) is a receptor model, which is frequently used in environmental geochemistry for metal(loid)s' source apportionment in various matrices. In this study, concentration data for As, Cu, Co, Cr, Fe, Mn, Ni, Pb, Sb, V and Zn in urban soils of two land uses (playground and roadside soils) from the city of Bratislava (Slovak Republic) were reinterpreted using the PMF model to quantify the contributions of individual sources of metal(loid)s in soils. Brake and tyre abrasion (27.0%), mixed traffic/industrial sources (14.3%), and natural origin (58.7%) were identified for roadside soils, while mixed traffic sources (28.4%), Cu-specific source (24.6%), and natural origin (47.0%) contributed to the metal(loid) concentrations in playground soils. Factor contributions revealed prevailing non-exhaust sources dominated by Cu, Sb and Zn over mixed traffic/industrial sources in roadside soils, following the world trend. Strong negative correlations of the 206Pb/207Pb isotopic ratio with Factor 1 in both groups of soils, representing traffic and industrial sources, are an evidence of the correct assignment of the obtained factor profiles to the actual sources of metal(loid)s in soils. There were significant differences in the proportion of mixed traffic sources in playground soils among the five urban districts, while the largest proportion of this source was identified in the city centre with the longest urbanisation and industrial history. Error estimation methods (displacement, bootstrap, displacement with bootstrap) proved that the data fit the model well and there was a minimal rotational ambiguity and random errors in the base model run.