A recent study concludes that using direct push is, indeed, a viable option.

Monitoring well installation using direct-push technology is potentially useful and cost effective, and the use of direct-push technology is increasing. Many state regulatory agencies are hesitant to make decisions using data generated with direct-push instruments.

BP Corp. North America Inc. and the Underground Storage Tank (UST) program of the U.S. Environmental Protection Agency (EPA) conducted a study to determine whether measurements of ground water parameters obtained using direct-push wells are comparable to those obtained from conventional monitoring wells. The researchers expect that their work will help assuage some of these concerns and provide a sound basis for further work.

The objective of the project was to determine whether direct-push wells yield results that are comparable to conventional monitoring wells for water level measurements, ground water chemical concentrations, hydraulic conductivity and natural attenuation (geochemistry) parameters. Conventional wells versus direct-push wells were compared at four retail fuel stations with dissolved-phase hydrocarbon plumes. Sites were chosen using existing conventional wells to screen the geologies and to choose three wells at each site that exhibited a range of concentrations. Two sites were located in Georgia and two were located in Ohio - a wide range of conditions was studied.

About the Wells

Conventional monitoring wells were either 2-inch or 4-inch diameter wells installed in accordance with state-approved methods prior to the study. Construction consisted of PVC casing connected to 10- or 15-foot long PVC screen intersecting the water table. They were installed with 4- or 8-inch diameter boreholes and the annular space was filled with a filter pack, a bentonite seal and grout. For each of three conventional wells at a site, a direct-push well, screened over basically the same interval, was installed 2.5 feet to the west. The screen of the direct-push well was set over basically the same interval as the conventional well to ensure that the sample would be obtained from similar geologic and hydraulic conditions as the conventional well - provided the wells are properly developed and the screens are not clogged. Overall, there were 12 clusters and 24 wells (12 conventional; 12 direct-push).

The direct-push wells were installed with a Geoprobe direct-push apparatus. During drilling, soil samples were taken for grain-size analysis. Continuous soil samples were collected from the saturated zone that corresponded to the screened interval of the conventional well.

After soil sampling was completed and the boring reached the proper depth, the Geoprobe rods were extracted from the bore. The direct-push well, which consisted of a 1-inch diameter, schedule 80 PVC screen and riser pipe, was properly assembled and inserted into the bore. The diameter of the bore was approximately 0.2 inches larger than the outer diameter of the well screen and riser, and no filter pack, bentonite or grout was used in the annulus. The upper 2 feet of all direct-push wells were sealed at the surface with bentonite and neat cement grout. The surface finish of each of the direct-push wells consisted of a flush-mounted, 5-inch diameter, traffic-rated vault set in a 1-foot-square concrete pad. After they were installed, measuring points on the direct-push wells and all existing monitoring wells at the site were surveyed to a commondatum. Each direct-push well was developed by purging.

There was no filter pack with the direct-push wells because this type of apparatus was not available at the time of installation. However, this may be an advantage because filter packs may not be needed. A filter pack in wells is designed to do two things. It increases the effective hydraulic diameter of a well and it retains most of the formation material, thereby filtering fines from the well. Specifically for environmental monitoring wells, the filter pack is designed to exclude the entrance of fine silts, sands and clays into a monitoring well. Therefore, the only effect the filter pack should have on analytical analyses from environmental monitoring wells is that sediment and formation fines are minimized. If a filter pack is used in the construction of an environmental monitoring well, it will affect the hydraulic diameter of the well, and it may influence the results of hydraulic conductivity tests - such as slug tests - if the filter pack is not properly accounted for in the calculations. There should be no real difference in the analytical results for dissolved phase constituents.

During the study, the wells were sampled four times, resulting in 768 analytical data values for MTBE, BTEX constituents, total BTEX, naphthalene and total suspended solids (TSS). Water levels were measured on each of the wells prior to sampling. Additional data included duplicate samples. The data was collected for a year. Hydraulic conductivities were measured twice using rising head tests. Two of the sites also were sampled for geochemistry parameters: dissolved oxygen, pH, carbon dioxide, alkalinity, ferrous iron, total iron, nitrate, sulfate and methane.

The following conclusions can be drawn:

Ground water levels measured in conventional versus direct-push monitoring wells are nearly identical.

• For MTBE measurements, there is no difference between the concentrations measured in samples from direct-push and conventional monitoring wells.

• For BTEX measurements, there is no difference between the concentrations measured in samples from direct-push and conventional monitoring wells across three of the sites. For one site, the measurements were consistently and significantly higher in samples obtained from the direct-push wells, suggesting a systematic error. Subsequent analysis and sampling suggests that the screen or borehole may have become contaminated during installation of the direct-push well.

• The mean hydraulic conductivity from the conventional wells is 4.4 times greater than from the direct-push wells, also suggesting a systematic error or problem.

• The concentrations of TSS were significantly higher in samples from direct-push wells than those from conventional wells, which likely results from the lack of a filter pack and possibly incomplete well development.

• The naphthalene concentrations exhibited slightly higher concentrations in samples from direct-push wells than those from conventional wells, but the result was not consistent across all sites, and there was considerable spatial variability.

• The consistently lower hydraulic conductivity and higher TSS concentrations in the direct-push wells, the variability in naphthalene concentrations and possibly the difference in BTEX concentrations at one of the sites are believed to be due to poor well development of the direct-push wells.

• For geochemical parameters indicative of natural attenuation (dissolved oxygen, carbon dioxide, ferrous iron, nitrate, methane, alkalinity and sulfate), the statistical analysis indicates that there is no difference in concentrations measured in samples obtained from direct-push and conventional monitoring wells. The caveat is that there is only a small amount of data and it exhibits some variability.

Analysis of the data suggests that, provided the wells are properly developed, there is good reason to believe that all measurements obtained from direct-push monitoring wells are equivalent to those obtained from conventional monitoring wells.

Lessons Learned

The study was well planned and the procedures were well documented and carried out. Future studies would do well to adopt most of what was conducted in this study. There are five lessons learned:

1. Direct-push wells need to be developed properly. The direct-push wells in this study were developed by simple purging of the well. Other methods such as using a surge block should be considered.

2. The statistical analysis was more complex than that applied in many previous monitoring well studies that used non-parametric methods. The complex statistical analysis, making use of transformation, linear models and normal statistics, not only enhanced the usefulness of the small data set, but also highlighted questionable data. This demonstrated its usefulness, and its application should be considered in future studies.

3. Hydraulic conductivities often varied by more than a factor of two when comparing calculated results for the same well but on different dates. This was not noted at the time, and care should be taken to determine if well parameters may have changed and affected the results.

4. The fraction of organic carbon in the soil should be measured in order to check whether some concentrations might correlate with TSS.

5. The sampling and analytical methods used to determine the concentrations of the geochemistry parameters should be reviewed to determine the extent that they could contribute to the variability in the data. Perhaps new methods should be used in future studies.
ND