Multivariate analysis for assessing sources, and potential risks of polycyclic aromatic hydrocarbons in Lisbon urban soils

Abstract

Urban soils quality may be severely affected by polycyclic aromatic hydrocarbons (PAHs) contamination, as is the case of Lisbon (Portugal). However, to conduct a risk assessment analysis in an urban area can be a very difficult task due to the patchy nature and heterogeneity of these soils. Thus, the present study aims to provide an example on how to perform the first tier of a risk assessment plan in the case of urban soils using a simpler, cost effective, and reliable framework. Thus, a study was conducted in Lisbon to assess the levels of PAH, their potential risks to the environment and human health, and to identify their major sources. Source apportionment was performed by studying PAHs profiles, their relationship with potentially toxic elements, and general characteristics of soil using multivariate statistical methods. Results showed that geostatistical tools are useful for evaluating the spatial distribution and major inputs of PAHs in urban soils, as well as to identify areas of potential concern, showing their usefulness in risk assessment analysis and urban planning. Particularly, the prediction maps obtained allowed for a clear identification of areas with the highest levels of PAHs (close to the airport and in the city center). The high concentrations found in soils from the city center should be a result of long-term accumulation due to diffuse pollution mostly from traffic (through atmospheric emissions, tire debris and fuel exhaust, as well as pavement debris). Indeed, most of the sites sampled in the city center were historical gardens and parks. The calculation of potential risks based on different models showed that there is a high discrepancy among guidelines, and that risks will be extremely associated with the endpoint or parameters used in the different models. Nevertheless, this initial approach based on total levels was useful for identifying areas where a more detailed risk assessment is needed (close to the airport and in the city center). Therefore, the use of prediction maps can be very useful for urban planning, for example, by crossing information obtained with land uses, it is possible to define the most problematic areas (e.g., playgrounds and schools).