INTERCOMPARISON OF AMBIENT AEROSOL SAMPLERS USED IN WESTERN VISIBILITY AND AIR-QUALITY STUDIES

Several past and ongoing visibility measurement programs in the rural western United States have used ambient aerosol samplers of differing designs. These differences make a direct comparison of aerosol data from these studies difficult. In order to establish the equivalence, bias, and predictability of mass concentrations determined with these samplers, a 4-week intercomparison experiment was conducted at Reno, NV, during May and June 1984. Nineteen samplers belonging to 11 design types were included in this study. Concurrent measurements of particles in two size fractions—FINE (0-2.5 μm) and TOTAL (0-10 or 0-15 μm)—were made. This paper describes the samplers and the intercomparison methodology and provides an assessment of the measurement precision and comparisons of their performance. Sampler performance comparison measures included effective-variance-weighted slopes, intercepts, and correlation coefficients, average ratios, average differences, and distributions of differences. The quantitative results of these measures were then used to assess the equivalence, bias, and predictability of each sampler's measurements relative to the benchmark samplers. Most measured differences in mass concentrations were consistent with known differences and similarities of the size-selective inlets (SSIs) used on the different samplers. The comparisons show that FINE mass concentrations measured by all the samplers are predictable from the values provided by the CYCLONE benchmark sampler, within estimated measurement precision. The SFU, DICHOT, and 2 × 4 are equivalent to the CYCLONE. The SCISAS and WRAQS FINE mass data agreed closely, but are generally slightly lower than the CYCLONE and the SFU data. All FINE mass data are also predictable based on the SCISAS2 data, which are representative of the SCENES program FINE data. All samplers except the CYCLONE and SFU are equivalent to the SCISAS2. There is more variability among TOTAL mass data sets, and in most cases differences are directly related to the sampling effectiveness of different SSIs. All 15- and 10-μm TOTAL mass concentrations, except that of the WRAQS-11, are predictable from SSI-15 and SSI-10 benchmark sampler measurements. All 15-μm TOTAL mass data sets are predictable from the SCISAS2 data. No samplers were equivalent to the SSI-15, and only the WRAQS-2 and SCISAS 1 were equivalent to the SCISAS2. © 1990, American Chemical Society. All rights reserved.