Relationship between residential dust-lead loading and dust-lead concentration across multiple North American datasets

摘要

Dust-lead monitoring studies typically report values as either dust-lead loadings (mu g/ft(2)) or as dust-lead concentrations (mu g/g). This study presents an approach for estimating dust-lead concentrations, when only dust-lead loading values are available. A literature search identified five large studies (200 homes) that contained original data for both dust-lead loading and dust-lead concentration, and could be used to develop an empirically-based loading to concentration (LTC) model. The R-2 values, standard error of the regression, slope, and intercept were improved when these original data sets were refined and pooled together. An additional thirty-two studies had summary statistics available for both dust-lead loading and dust-lead concentration, which enabled an independent evaluation of the LTC model. Despite differences in study design, sampling method, and analytical procedures, the LTC model made consistent predictions across studies. Reported central-tendency summary statistics from the independent evaluation data sets showed that measured central tendency dust-lead concentrations from these studies were similar to predicted dust-lead concentrations using the LTC model. Across 142 paired dust-lead loading and dust-lead concentration values from the evaluation datasets, slightly more than 90% of estimated dust-lead concentration values fell within a 90% prediction interval. These results indicate that, when only dust-lead loading values are available, the LTC model provides a reasonable approach for estimating dust-lead concentrations. A limitation of this approach is the greater uncertainty at high dust-lead loadings, caused by variability in total dust loadings (g/ft(2)) that result in one dust-lead loading value being linked to a range of dust-lead concentration values.