Dept. of Environmental Science
Oregon Graduate Center
The chemical element balance receptor model is explored as a method of determining the major contributors to the fine and total suspended particulate in Portland, Oregon. Its major weaknesses in previous applications are evaluated and an attempt is made to rectify them. Aerosol source chemical compositions from original source tests are reported, evaluated and compared with results from other source tests. A new least squares fitting procedure which incorporates errors in the source compositions as well as the errors in the ambient chemical concentrations is derived. The validity of this "effective variance" least squares calculation method is verified in a series of simulation studies comparing it to the ordinary weighted least squares calculations used in previous element balance studies. These simulations show that the effective variance least squares fit is superior in all respects. A method for identifying specific sources within a source type grouping by averaging source contributions within distinct surface windflow patterns is proposed. To test these refinements in the chemical element balance, filter samples of Portland's fine (<2 Âµm) and total suspended particulate were taken at six locations, two background, one industrial, one urban, one urban/industrial and one residential. Ninety-four days were sampled between July, 1977 and April, 1978. Three or six sequential lo-vol filter samples and one hi-vol sample of fine and total suspended particulate were taken at each site. Each day sampled was classified into a meteorological regime based on surface windflow direction and speed. A 32 day subset of samples was selected to represent each regime in rough proportion to its frequency of occurrence throughout the year. 1300+ lo-vol filters were subjected to x-ray fluorescence, instrumental neutron activation and ion chromatographic analysis for NO[superscript -] [subscript 3], SO[superscript -2] [subscript 4], F[superscript -], Na, Mg,Al , Si, S, Cl, K, Ca, Ti, V, Cr , Mn, Fe, Ni , Cu, Zn, Br, Pb, Se, As, Cd, and Ba. Volatilizable and non-volatilizable carbon concentrations were quantified on hi-vol filters. The data set was validated by interlaboratory and intermethod comparisons. Aerosol source contributions to each specimen were calculated using the effective variance least squares fit implemented by an interactive computer program, *CALCEB. This routine allows the operator to call up any set of data and to attempt fits with various sources and concentrations. The chemical element balance receptor model was found to be successful in estimating major source impacts for control strategy purposes. An average of 40-48% of the total suspended particulate loading in the Portland airshed can be ascribed to geological material, presumably the result of vehicular resuspension. 7-13% of the total suspended particulate is attributed to vegetative burning (i.e. fireplace, woodstove, slash or field) , but this figure is somewhat suspect because of the uncertainties in the chemical composition of emissions from this source type. The average contributions show 1-9% of the TSP coming from automobile exhaust and 2-7% due to the-intrusion of marine background aerosol. Industrial point sources average less than 1% in their contributions; it is probable that these are upper limits. 8-14% of the total suspended particulate is due to volatilizable carbon, 3-7% is due to sulfate and 3-7% is due to nitrate which cannot be accounted for by the primary source types and could be due to secondary formation. Ninety percent of the total suspended particulate is accounted for by the balances, on the average. The chemical element balance is compared to other receptor models and a program of future research in the areas of source characterization, receptor model development, and ambient aerosol study procedures is presented.
Watson, John George, "Chemical element balance receptor model methodology for assessing the sources of fine and total suspended particulate matter in Portland, Oregon" (1979). Scholar Archive. 317.