Receptor modeling of volatile organic compounds. 1. Emission inventory and validation

Air samples, speciated for 23 organics (23-NMOC) by gas chromatography, were collected on 15 days between July 1 and September 9, 1987, at three sites in Chicago, 20--70 km apart. Four-hour samples were collected simultaneously at all three locations, along with continuous measures of NO/NO[sub 2] and an additional 20-h sample at one site. The 23-NMOC concentration pattern was evaluated for each of 55 samples using a chemical mass balance (CMB) receptor model and eight source categories: vehicle exhaust, gasoline vapor, petroleum refineries, vapor degreasing, architectural coatings, graphic arts, dry cleaning, and wastewater treatment. The largest contribution was from vehicles (33% of total VOC). Average predictions agreed well with the emission inventory with the exception of petroleum refineries, which were over 5 times the inventory values. Comparison of CMB results with wind trajectories and refinery locations supported this conclusion. Atmospheric reactivity was not a problem for source allocation for the typical 3--5-h time periods between release and sampling. 37 refs., 5 figs., 4 tabs.