Testing and application of diffusion samplers to identify temporal trends in volatile-organic compounds

Abstract

Methods for groundwater sampling have evolved over time. This evolution has been driven by changing theories on how to obtain representative aquifer water samples. Passive sampling is a fairly recent method that relies on the natural flushing capacity of a well to obtain representative samples. The use of diffusion samplers is one method of passive sampling and works well under certain conditions. As part of a 2-year study to determine the temporal variability and trends in concentrations of volatile organic compounds (VOC's) in a large plume (0.5 mi2 area) of contaminated ground water in a glacial-drift aquifer, results of VOC analyses of samples collected with diffusion bag samplers were compared with those of samples collected with other types of samplers. The area of study is the primary source area of the large VOC plume and is located adjacent to a a river that losses flow and recharges the aquifer. The concentrations of VOC's, primarily tetrachloroethylene (PCE), trichloroethylene (TCE), and cw-l,2-dichloroethene (cw-l,2DCE), in samples collected with diffusion samplers show a strong positive linear correlation (root-mean square error of 0.94 and above) with concentrations from purged samples following low-flow sampling procedures. A total of 20 coupled diffusion and peristaltic-pump samples were collected from 7 wells completed in high-permeability glacial-drift. The mean concentration of PCE in the diffusion samples was 1,152 parts per billion (ppb) and the mean from the peristalticpump samples was 1,119 ppb. The standard deviations also were similar. The mean concentrations of TCE were slightly higher in diffusion samples (89.2 ppb) than peristaltic-pump samples (75.4 ppb). The mean concentration of cw-l,2DCE in diffusion samples (95.0 ppb) was virtually identical to the mean in peristaltic-pump samples. Although VOC concentrations changed dramatically at several wells over the sampled period, trends in VOC's detected using diffusion samplers corresponded with trends in VOC's from other lowflow sampling methods. For example, at two wells where coupled diffusion and peristaltic-pump samples were collected, VOC concentrations varied by a half order of magnitude over a two-month period. Although the diffusion sampler was installed and left in the well for the entire period, VOC concentrations in the diffusion sampler at the time of retrieval generally matched those in the instantaneous samples collected with the peristaltic pump on the same day, suggesting relatively rapid equilibration of the diffusion sampler to VOC concentrations in the well. The use of diffusion samplers allowed for the understanding of contaminant transport conditions at the study site because it allowed for an increase in the frequency of sampling without an associated increase in labor cost. For example, spatially variable declines in PCE concentrations were identified over the two-year study that are related to spatial variations in sediment lithology and the location of the plume within the groundwater flow system. Wells screened ...

Effects of a remedial system and its operation on volatile organic compound-contaminated ground water, Operable Unit 1, Savage Municipal Well Superfund Site, Milford, New Hampshire, 1998-2004

“…To better quantify hydrologic conditions for the periods shown in figures 4 and 5, water levels were evaluated based on water-level statistics from long-term (from 1962) monitoring well MI-18 in OU2 ( fig. 1) Precipitation effects on ground water in OU1 were exam ined in Brayton and Harte (2001). Findings from that study showed that the land-surface crown inside the barrier area decreases infiltration of precipitation and recharge to ground water.…”

“…Brayton and Harte (2001) found that recharge rates inside the barrier area decreased by as much as a half from pre-construction periods because of the addition of a semiconical land-surface crown during barrier construc tion. Analysis of additional data subsequent to that reported in Brayton and Harte (2001) indicates that a recharge rate of 2.45 gal/min also may be appropriate (about a 30-percent reduction from outside the barrier). Therefore, simulated ranges in direct recharge varied from 1.75 to 2.45 gal/min.…”

“…Ground-water-flow conditions for pre-barrier wall construction, but prior to operation of the remedial system, were discussed in Brayton and Harte (2001). Post-construction ground-water-flow patterns due to the barrier produce an area of focused recharge (river loss) from the Souhegan River to the aquifer in OU1 near well P-2 ( fig.…”

“…Direct recharge inside the barrier from infiltration of pre cipitation was estimated at a yearly rate of 1.75 gal/min (Bray ton and Harte, 2001). This rate is 50 percent less than the rate for outside the barrier.…”

“…Previous studies by the USGS of the study site include construction and calibration of two ground-water-flow models (Harte and others, 1999;Harte and Mack, 1992), evaluation of effects of ground-water withdrawals on advective transport of contaminated ground waters (Harte and Willey, 1997), analysis of PCE trends for pre-and post-barrier conditions (Harte and others, 2001), and description of results of a monitoring pro gram of continuous water levels from water years -2003(Brayton and Harte, 2001Harte, 2005). Furthermore, solute transport of PCE was investigated and a transport model cali brated for the OU1 and OU2 areas (Harte, 2004).…”