As vapor intrusion has become an increasing concern at corrective action sites, the USEPA and a number of state regulatory agencies have recently issued guidance documents covering the collection of subsurface gas, and above-ground air samples during vapor intrusion investigations. Most of these guidance documents do not provide specific recommendations concerning the number of each type of sample required to evaluate the presence or absence of a vapor intrusion impact and, in fact, the number of samples needed to evaluate vapor intrusion depends on both the spatial and temporal variability in VOC concentrations in the environmental media sampled. Through the collection of a large number of samples focused around individual buildings at two study sites, we have been able to evaluate the spatial and temporal variability in VOC concentration in groundwater, well headspace, soil gas, sub-slab, ambient air, and indoor air samples. At both test sites, we observed higher spatial variability in subsurface gas samples than in above-ground air samples. This indicates that a larger number of spatially-separated samples are required to accurately characterize average VOC concentrations in these media. For example, six subslab samples would be required to
estimate the mean VOC concentration with an accuracy of +/- 67% while only one indoor air sample or two ambient air samples would be required to achieve the same level of accuracy.
For all media, short-term temporal variability (i.e., a time scale of days) was low, with 65% of paired samples showing a relative percent difference (RPD) of less than 30%, the standard quality assurance objective for duplicate samples. Longer-term temporal variability (i.e., a time scale of months) was significantly higher than short-term variability and was similar in magnitude to the observed spatial variability for subsurface gas samples. These findings indicate that collecting multiple samples from the same sampling point over a period of days provides little additional information concerning average VOC concentrations. However, multiple sample events spaced over a period of months do provide an improved understanding of average VOC concentrations over time at the sampling points.