A HYBRID APPROACH FOR SOURCE APPORTIONMENT OF ATMOSPHERIC POLLUTANTS IN THE NORTHEASTERN UNITED STATES

Deterministic modeling has been the traditional method employed to estimate contributions of specific sources or source regions to the pollutant levels at receptor locations. Deterministic models simulate atmospheric motions and the associated physical and chemical processes. These models rely on detailed source emissions inventories and meteorolgical descriptions. Receptor modeling has recently emerged as a means of assessing the contributions of specific sources or regions to the measured pollutants. Receptor models use empirical evidence, such as elemental composition at a measurement location, to deduce the sources of the pollutants. While certain applications may prohibit the use of one of the modeling approaches, the greatest understanding of the sources of the atmospheric pollutants may be obtained from their complementary use.; In this study, meteorological observations were utilized together with aerosol and gaseous data taken during a comprehensive field study in an attempt to understand the nature and sources of the atmospheric acidity measured in the Northeast United States. The main data base used in this study was from the 1983 Allegheny-Laurel Experiment. Examination of the detailed chemical and elemental analyses revealed that the atmosphere was decidedly acidic, and the aerosol consisted essentially of {dollar}{lcub}rm Hsb2 SOsbsp{lcub}4{rcub}{lcub}2-{rcub}{rcub}{dollar} in various degrees of neutralization. This acidic aerosol was largely associated with transported coal combustion products arriving at the Allegheny and Laurel sites from west. Local source contributions, primarily from large coal-fired power plants, were estimated to comprise about half of the measured HNO{dollar}sb 3{dollar}(g) and SO{dollar}sb 2{dollar}(g).; Efforts to identify and quantify the air pollution sources contributing to the fine and coarse particle mass were hampered by meteorological uncertainty. Fine mode particles were observed to shift into the coarse mode under prolonged periods of atmospheric saturation. Receptor modeling estimates of source contributions under these saturated conditions were highly uncertain. The CMB model ultimately failed to account for half of the fine particle mass.; A Hybrid modeling approach was utilized to quantify the regional contributions to observed trace element and {dollar}{lcub}rm SO{rcub}sbsp{lcub}4{rcub}{lcub}2-{rcub}{dollar} concentrations. Distinct regional contribution differences were observed for certain trace elements and elemental ratios. Results from this study indicate that one can determine regional contributions to ambient concentrations using only observed data.