Considerations for design of source apportionment studies

Abstract

This report recommends procedures for source and ambient sampling and analysis in source apportionment studies. The recommendations are based on the results of receptor model studies of atmospheric particles in urban areas, especially a recent study of Houston, TX, undertaken as part of the Mathematical and Empirical Receptor Models Workshop (Quail Roost II). The recommendations are presented at three levels of increasing cost and detail of information obtained. Existing mass emissions inventories combined with chemically resolved test data from similar sources (not necessarily in the same locale) can be used to initially estimate the sources of elements present on ambient particles. To aid local users in construction of chemically resolved emission estimates, the U.S. Environmental Protection Agency (EPA) is compiling a library of compositions and size distributions of particulate emissions from major source types. More reliable source characterization can be achieved if the actual sources are tested directly. EPA should develop and publish detailed procedures for source sampling that would be more appropriate for receptor model use than are existing standard methods. Source and ambient sampling should be conducted by similar methods. If possible, particles from sources should be collected in a way that simulates changes that would normally occur before they reach distant receptors (e.g. by diluting and cooling the particles from hot sources). It is recommended that particulate samples be routinely collected in two size fractions by use of virtual impactors and that all samples be subjected, at a minimum, to mass and X-ray fluorescence analyses. Additional measurements are suggested for obtaining more detailed information: neutron activation analysis; X-ray diffraction; automated particle classification by electron microscopy; analyses for classes of organic species, ^(14)C and thermally released carbonaceous species; and real-time observation of several gases during sample collection. Methods for collecting meteorological data in parallel with ambient samples are described, as are methods for incorporating such data into the source identification process.

Files

Additional details