Temporal and spatial variation in residential soil metal concentrations: implications for exposure assessments.

Understanding temporal and spatial variation in soil chemicals is critical in exposure assessments. We measured eight metals in subsamples, duplicates (~0.3 m), and repeat soil samples taken 1-6 years after initial sampling (~5 m). We estimated variance components (VCs) of metal concentrations using nested analyses accounting for sampling area, land use and soil type, and calculated coefficients of variation (CVs) for repeat sample pairs. Total variance for all metals was similar, but VCs were proportioned differently by metal and sample type. Spatial variation explained the majority of variance in duplicate samples. CVs of metal concentrations were not significantly different over the long time interval, but repeat samples had larger VCs for unexplained error. Sampling area and land use were important for Ba and Mn, and Pb and Hg, respectively. Results suggest metals are stable over long times and suitable for exposure assessments, but that individual metal behavior should be considered.