Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4434721 | Atmospheric Pollution Research | 2016 | 11 Pages |
A reliable identification of sources and their relative time dependent contributions to ambient aerosol load is an important air pollution research problem. Given the inherent complexity of contributing sources in urban/mega-cities, an appropriate statistical investigation is needed to characterize sources and to understand their timeline trend profiles. Daily average ambient particulate matter (PM) loads, PM10 (aerodynamic diameter <10 μm) and coarser particulate matter (CPM: aerodynamic diameter >10 μm) were collected once a week over 4 years at a receptor site in Delhi. The samples were analyzed to quantify the presence of 17 marker elements. Time series data of PM loads, and that of associated marker elements was subjected to Positive Matrix Factorization (PMF) to identify sources and to quantify their contributions to each PM fraction with reference to the associated marker elements. The resolved time series data of each contributing source was further subjected to Ensemble Empirical Mode Decomposition (EEMD) analysis to extract their timeline trend profile over four years in CPM and PM10 load. Three sources contributed to the CPM load: anthropogenic (15%), primary crustal (59%), and fine crustal dust (26%). Four sources contributed to the PM10 load: coarser grain crustal material (9%), fine grain crustal material (12%), industrial and vehicular emissions (23%), and wind assisted transport and re-suspension of surface dust (56%). The timeline trend of sources contributions to CPM and PM10 displayed a non-linearity. The unique composite-PM10 source contributed maximum to the ambient PM10 load. Distinct underlying processes of this source involved convective re-suspension and city-wide cleaning associated upliftment of surface deposits back into the ambient environment.