BOREHOLE-DILUTION TESTS TO MEASURE NATURAL FLUSHING OF SCREENS IN WELLS MONITORED WITH PASSIVE DIFFUSION SAMPLING

Passive diffusion sampling is used to monitor ground-water quality at an increasing number of sites. In this method, natural flow is assumed to flush the monitoring-well screen so that the water inside and outside the screen has similar concentrations of the contaminants of interest. Therefore, contaminant concentrations in water in a diffusion sampler placed in the screen for a sufficient equilibration time should be representative of concentrations in the aquifer. In a study comparing diffusion- and pumped-sampling methods to monitor volatile organic compounds in ground water on Cape Cod, MA, there was good agreement between the methods for most of the 88 wells tested. The degree to which concentrations agreed was repeatable for a given well and, where there was poor agreement, the diffusion-sample concentrations were lower than the pumped-sample concentrations. One possible explanation for the poor agreement for some wells is that the natural flushing rate of these wells is low.

To test this explanation, flushing rates were measured with the borehole-dilution method in eight of the wells – four that had good agreement and four that had poor agreement between diffusion- and pumped samples. The wells have 2- to 5-ft-long screens. In the borehole-dilution method, a saline solution is added to the packer-isolated screened interval to increase the specific conductance of the water in the screen. The conductance is then monitored with a probe placed in the screened interval to measure the rate of dilution by the natural flow through the screen. Ambient ground-water velocities in the Cape Cod sand and gravel aquifer are about 1 to 2 ft/day. The measured bulk velocities across the screens of the eight wells were from 0 to 2.1 ft/day, a range equivalent to flushing rates of zero to 16 screen volumes per day. Higher flushing rates were expected to be related to better agreement between the pumped- and diffusion-sampling methods, but this was the case for only four of the eight wells. This inconsistency may indicate that other factors, such as concentration variation along the well screen in addition to the natural flushing, may affect the extent to which the results from the two methods agree.