Source apportionment with uncertainty estimates of fine particulate matter in Ostrava, Czech Republic using Positive Matrix Factorization

Özet

A 14-week investigation during a warm and cold seasons was conducted to improve understanding of airpollution sources that might be impacting air quality in Ostrava, the Czech Republic. Fine particulatematter (PM2.5) samples were collected in consecutive 12-h day and night increments during spring andfall 2012 sampling campaigns. Sampling sites were strategically located to evaluate conditions in closeproximity of a large steel works industrial complex, as well as away from direct influence of theindustrial complex. These samples were analyzed for metals and other elements, organic and elemental(black) carbon, and polycyclic aromatic hydrocarbons (PAHs). The PM2.5 samples were supplementedwith pollutant gases and meteorological parameters. We applied the EPA PMF v5.1 model with uncertaintyestimate features to the Ostrava data set. Using the model's bootstrapping procedure and otherconsiderations, six factors were determined to provide the optimum solution. Each model run consistedof 100 iterations to ensure that the solution represents a global minimum. The resulting factors wereidentified as representing coal (power plants), mixed Cl, crustal, industrial 1 (alkali metals and PAHs),industrial 2 (transition metals), and home heat/transportation. The home heating source is thought to belargely domestic boilers burning low quality fuels such as lignite, wood, and domestic waste.Transportation-related combustion emissions could not be resolved as a separate factor. Uncertaintyestimates support the general conclusion that the factors identified as representing coal power and homeheat/transportation dominate the percent contribution to fine mass. Apportionment of regulated individualspecies is also presented.