Source apportionment of sediment PAHs in the Pearl River Delta region (China) using nonnegative matrix factorization analysis with effective weighted variance solution

Considering the advantages and limitations of a single receptor model, in this study, a combined technique of nonnegative matrix factorization analysis with effective weighted variance solution (NMF-EWV) was proposed for source apportionment. Utilizing NMF, major linear independent factor loadings with nonnegative elements were extracted to identify potential pollution sources. Then, these physical reasonable factor loadings were regarded as source profiles to apportion contributions using effective weighted variance solutions. Evaluation results indicated that the NMF-EWV method reproduced the source profiles well, and got a reasonable apportionment results for the synthetic dataset. The methodology of the NMF-EWV was also applied to recognize sources and apportion the contributions of polycyclic aromatic hydrocarbons (PAHs) collected from freshwater and marine sediments in the Pearl River Delta (PRD) region which is one of the most industrialized and economically significant regions of China. Apportionment results showed that traffic tunnel made the largest contribution (46.49%) for the freshwater PAH sediments in the PRD, followed by coal residential source (29.61%), power plant (13.45%) and gasoline engine (10.45%). For the marine sediments, traffic tunnel was also apportioned as the largest source (57.61%), followed by power plant (22.86%), gasoline engine (17.71%) and coal residential source (1.82%). Traffic-related sources were the predominant reasons for PAH pollution in that region.

► A combined source apportionment method of NMF-EWV was studied.
► The NMF-EWV method was evaluated with synthetic dataset.
► The sources of PAHs in sediments from the Pearl River Delta (China) have been apportioned.