Apportioning variability of polycyclic aromatic hydrocarbons (PAHs) and chlordanes in indoor and outdoor environments.

Measurements of semi-volatile organic compounds (SVOCs) in air are subject to substantial variability and uncertainty. This study apportions total variance of polycyclic aromatic hydrocarbons (PAHs) and chlordanes to variability and uncertainty components. Concentrations of PAHs and chlordanes were measured inside and outside of 116 residences in three large cities in the U.S. during 1999-2000. Total variance was apportioned to between-city, between-tract, between-residence, and seasonal variation, as well as measurement uncertainty using variance component analysis and log-transformed data for frequently detected compounds. Outdoors, seasonal variation was the greatest portion (44-67%) of total variance, and city effects were significant (19-24%). Indoors, seasonality dominated variability of PAH measurements (>50%). Gas-phase PAHs varied more within city than between cities; particulate-phase PAHs varied significantly between cities but were largely homogeneous within cities. Gas-phase chlordanes showed larger intra-urban variation (63%) than seasonal variation (18%). Measurement uncertainty was generally below 10% with a few exceptions occurring at very low concentrations. Results indicate a need to collect multiple-season samples to account for the large temporal variation between seasons. Samples from centrally located monitoring stations could be representative of ambient SVOCs. Variance component analysis is useful to weigh influential factors in SVOC concentrations, identify and apportion sources, evaluate method performance, and design effective monitoring programs.