Avaliação do uso de diferentes modelos receptores para determinação da contribuição das fontes de partículas totais em suspensão

Abstract

The source apportionment of particulate matter has been a subject of growing interest in the scientific community to assist in the management of air quality. The receptor models are used for identification and quantification of the sources responsible for the contributions from the mass concentration of the contaminant in the receptor. This work evaluates the results of two worldwide disseminate receptor models: chemical mass balance (CMB) and positive matrix factorization (PMF) for TSP data obtained in Enseada do Suá Monitoring Station in Vitoria -ES and sedimentary particles data obtained in the Metropolitan Region of Vitória (RGV) in order to investigate the difficulties in the use of each model as well as its advantages and disadvantages. The CMB model has the disadvantage of requiring the sources profiles, determined experimentally in samples with the aid of sources and inventories of local sources to be applied and also has limitations when the sources involved are similar or collinear. The PMF model does not require the sources profiles, but has the disadvantage of needing large amount of sample concentration of the contaminant in the receiver. The results showed, based on performance measures that the two models were considered good to reproduce the data from the receptors with acceptable settings. However, different results were adjusted, stating the difficulty of working with collinear sources. The CMB model used about 20 sources to fit the applications of sedimentary particles data (based on the inventory of sources), but many of these sources were combined for best results. For TSP data, the CMB model used only six sources and got great settings, while the PMF model found only five types of sources (factors), but these results were considered representative for the study region, and are in bankruptcy with the sources used in CMB model. It was found that in the PMF model some species that are not characteristic of certain factors (sources) appear in the results, which makes the identification of sources even more complex, requiring extensive knowledge about the composition of numerous sources. The wind direction analysis helped in identifying the sources in the CMB model results making the results more realistic with physical meaningful in specific episodes demonstrated in the work.