An Improved Method for Estimating in Situ Sampling Rates of Nonpolar Passive Samplers

Abstract

The quality of passive sampling methods for measuring concentrations of dissolved hydrophobic contaminants relies on accurate knowledge of in situ sampling rates. In currently used methods for estimating these sampling rates from the dissipation rates of performance reference compounds (PRCs), the PRCs that show either insignificant or complete dissipation are ignored. We explored the merits of nonlinear least-squares (NLS) methods for estimating sampling rates, aiming to retain the information stored in PRC data that is neglected in the traditional methods. To this end, we examined the error structure of weighted NLS, unweighted NLS, and the traditional methods, using model simulations. The results show that sampling rates are best estimated using unweighted NLS. Uncertainties in the sampler-water partition coefficients may result in biased estimates that only weakly depend on the number of PRCs being used. The major advantage of unweighted NLS over the traditional method is that sampling rate estimates and uncertainties are available where the traditional method fails, and that the variability of sampling rate estimates is smaller.