Hydrogeologic effects can confound sampling.

Last month, in Part 1 of this series, it was shown that traditional sampling techniques can adversely affect the quality of collected samples due to the entrainment of artificially generated turbidity. The following explains how hydrogeologic effects also can occur and confound our understanding of the true contaminant distributions and concentrations, and illustrates these effects with simple graphical examples.

Mixing of contaminated with uncontaminated waters in both the subsurface and in a monitoring well can occur when purging and sampling are improperly performed. This problem increases with longer well screens, higher pump rates, bailing and when variable stratigraphy exists across the screened interval.

Long well screens tend to average the contaminant concentration over the vertical screen dimension, i. e. over the screen length. This is because traditional sampling techniques simultaneously pull water from all zones, both contaminated and uncontaminated, that the screen intersects (Figure 1). This may be satisfactory if the desired result is a concentration integrated over a fairly large volume of the aquifer. It yields little information, however, about plume thickness or contaminant concentration gradients within the actual plume.

This problem can be worsened by stratigraphic variations that result in zones of high natural ground water flow layered with less permeable zones across the screened interval. In this situation, the water is preferentially transferred into the well casing from the higher permeability flow zones, whether or not these are the zones of maximum contamination (Figure 2).

High pump rates also can yield false plume locations and higher-than-actual contaminant concentrations, especially when combined with the above long-screen scenarios. Figures 1 and 2 illustrate that wells with long screens can provide incorrect, low contaminant values but a potential overestimation of plume thickness. Figure 3 shows the potential for high pump rates to pull contaminated water into a zone where there previously was none. The figure depicts this occurrence via high-permeability sand, but fractured rock also could be very susceptible to such errors. This scenario results in overestimation of the natural extent of the plume as well as spreading of the contamination into uncontaminated zones. Although horizontally induced migration is depicted, field data has shown that vertically induced migration through the aquifer also is possible.

Ideally, the screened interval should encompass only the sampling zone of interest, i.e. no more than the maximum vertical integration length that allows adequate characterization of the plume location and contaminant concentrations. The determination of what constitutes an adequate degree of characterization depends, of course, on the goals of the monitoring program. In general, pump rates should be set to attain only waters from the zones of interest and minimize induced hydrodynamic effects.

Figure 4 depicts such a sampling scheme. In this nested well scheme, there are separate sampling points, with no sandpacks, and the aquifer has been allowed to collapse around the sampling point. These can be installed with small diameter augers or with the use of drive-point technologies. These types of monitoring points probably are best sampled through low-flow purging and sampling techniques, with the monitoring of purging parameters such as turbidity, Eh and DO. It also is possible to sample such monitoring wells using passive, or unpurged sampling, depending on the natural ground water flow to maintain purged conditions. ND