[Home]  [Master Table of Contents]

Meeting Minutes
Facility Investigations, Closure Projects,
and Well Installations, August 10, 1999

Date: 10 August 1999

Time: 9:30 A.M. - 5:00 P.M.

Place: Camp Stanley Storage Activity, Boerne, Texas

Subject: CSSA Well Placement and Modeling Tasks, RL74, RL83, DO23

Attendees:

Minutes prepared by Julie Burdey and Scott Pearson, Parsons ES

The meeting commenced at 9:30 AM at the CSSA conference room. The objectives of the meeting (see Agenda) were to 1) define an appropriate groundwater modeling approach for CSSA; 2) identify locations for new planned wells; 3) determine well construction specifications; and 4) identify wells in which transducers will be installed.

An overview of the wells currently in place, a history environmental work conducted at CSSA, and the extent and nature of the contaminant plume was given by Chris Beal and Rene Hefner. A brief description of the geologic/hydrogeologic conditions was described by Rene Hefner (Figure 1). This included borehole geophysical and TV camera results, outcrop observations, and literature on the regional

setting. The type and faulting and known displacements were also discussed. The dip of the faults are unknown, but have been assumed to be nearly vertical with the downthrown blocks typically towards the south. Currently, the displacement along the northern fault zone is thought to be slight (up to 30 feet), while boring logs indicate the southern fault may have displacement upwards of 100 feet.

The current well system at CSSA was identified by former/current use and zone(s) of production by Rene Hefner and Chris Beal. The major producing zones of the Middle Trinity Aquifer are the Lower Glen Rose (LGR) and Cow Creek (CC) Limestones. It is unknown if the Bexar Shale (BS) is a transmissive unit in the area. Rene said that his past research has indicated that the Bexar Shale is considered a confining unit. The highest occurrence of TCE has been detected in Well 16, which is an open borehole across both the Lower Glen Rose and Cow Creek Limestones.

A brief description of SWMU and AOC investigations followed the introduction statements. Past work has included surface and subsurface sampling, electromagnetic (EM) and ground penetrating radar (GPR), and soil gas sampling. Seismic data has been attained for the northern cantonment area from Blackhawk Geosciences, Inc., and has been used to define the area of faulting to the north in conjunction with graduate studies research (Hefner and Waterreus) and outcrop mapping by Parsons ES.

Soil gas sampling over a large area encompassing Well 16 has shown that PCE is the major constituent detected in the near surface soils. Two locations southeast of Well 16 have been identified (SWMUs B-3 and O-1) as areas with the highest concentrations. Whereas most detections of PCE/TCE are below 10 mg/l over most of the surveyed area, both B-3 and O-1 have concentrations above 1,000 mg/l. These two SWMUs are considered the most likely source area candidates. SWMU B-2, north of Well 16, may also be a contributing factor, but investigative results to this date have been inconclusive.

From a historical perspective, concentrations of contaminants in groundwater fluctuate often, and seem to be tied to precipitation events and/or fluctuating water levels. Only wells D and 16 have contaminants in excess of action levels. A weather station and transducer have been established and monitored at Well 16 for several years. The aquifer response to precipitation events is dramatic and quick. Water levels may fluctuate over 100 feet during the course of the year. Only Well 4 is known to have ever gone dry. [Note: Well D was dry during September and December 1999 sampling events.]

Camp Stanley intends to initiate a groundwater modeling effort under RL74 to better understand the overall hydrogeologic system and the processes which affect the fate and transport of the contaminants. Three priorities of the model function were identified by Rene Hefner and Jo Jean Mullen as:

A list of data needs or gaps were identified by the meeting attendees as:

Based on the goals of the modeling effort and the amount of data that will be available for the model, Steve Rossello, Jim Williams, Roger Peebles, and Joe Hailer recommended that models consistent with the USGS Modflow engines be considered. Roger Peebles also discussed the prospect of using a finite element fractured flow analysis, but the data necessary to initiate that type of model is currently unavailable at CSSA. As the data set grows, a fracture analysis model could later be implemented. Based on these decisions, four model processing programs based on the USGS MODFLOW program were considered:

The advantages and disadvantages were weighed by the modeling portion of the meeting attendees. Of the four models available, only Argus was viewed not to be an appropriate selection for the project goals. Consideration was given to Visual Modflow since its relative ease of use would more easily allow all users at AFCEE, CSSA, and Parsons ES to manipulate the working model. However, not all parties will use the program, and other packages will better meet the project goals. Groundwater Vistas has one of the better calibration tools packages and is fairly easy to use. However, its standard output is two-dimensional and requires a third-party program to generate three-dimensional graphics.

The remaining package is GMS, which was developed by the U.S. Army, and is available through both the government and commercial vendors. The GMS package has a strong GIS interface and excellent visual capability which includes both standard graphics and ray tracing techniques. The package should be available free of charge to CSSA and its approved contractors. Its disadvantages are that overlays are difficult when rendering ray-traced graphics, training and support is not as readily available, and of all the packages, it has the highest learning curve.

From this discussion the use of the GMS software was deemed most appropriate because of its availability and affiliation with the U.S. Army and its strong capabilities with GIS interfacing and graphics. The limited usability of this model to the general practitioner (such as Brian Murphy) was noted; however, Brian indicated that he would rely on his contractors to actually run the model.

Following the lunch period, the discussion turned to the installation and placement of 15 monitoring wells scheduled to be drilled. Twelve wells are currently scheduled to be drilled in an arrangement of four clusters, with the remaining three wells to be installed as Lower Glen Rose wells only. Jo Jean Mullen stated that she was prepared to modify the DO23 contract to convert the three single wells into an additional well cluster if deemed necessary.

Scott Pearson identified preliminary locations for potential drilling sites on a poster-sized map of CSSA. As the discussion ensued, a "spirited" debate began regarding the end use of the drilling data. The well locations could be optimized for data regarding groundwater flow (physical characteristics), or they could be optimized for obtaining data regarding the contamination plume (chemical characteristics. Some attendees argued that well placement should be dispersed to evaluate the geology/hydrogeology of the watershed, and to implement perimeter monitoring where CSSA boundaries and potential preferential flowpaths (e.g., fault zones) intersect. The assertion was that a better understanding of the entire groundwater flow system would help predict where contaminants may migrate from the known affected area. However, such a scenario would require that some wells be sacrificed from the contaminant investigation to satisfy these criteria.

Jo Jean Mullen argued that the EPA has clearly indicated that the contaminant plume needs to be well defined and the question of whether off-site receptors may be affected must be answered. She added that the subsurface investigation needs to be focused primarily at where the contaminants are now to estimate plume transport and velocity. Jo Jean emphasized the possibility of TCE/PCE in the Cow Creek formation at Well 16 warrants more investigation within the plume area. Jo Jean directed that a compromise between the two approaches must be made.

Steve Rossello and Jim Williams pointed out that for a modeling effort to be effective, wells both inside and outside the fault system would be necessary. Steve Rossello also emphasized that groundwater should be addressed in areas of higher topographic relief to identify if these areas were significant groundwater recharge areas. It was recognized that additional work will be required to fill the data gaps left with the current work scope.

As a result of the debate, locations for four cluster wells and three single LGR wells were tentatively identified at the close of the meeting. Locations of possible future perimeter monitoring wells were not identified during the meeting. The driving forces of the well placement were three-fold: contaminant detection in the affected areas, observation points both within and outside of the northern fault zone, and evaluation of topographic expression on the water. It was stated that the number one priority is to determine if the plume is leaving the base, potentially threatening potable water sources. Although modeling is important and will be key to answering the potable water threat question, it shouldn’t drive the locations of the wells. It was stated that chasing the plume is the most important task now.

Joe Hailer postulated that the installation of cased cluster wells within a highly fractured fault zone may be an exercise in futility since groundwater flow may primarily be vertical within the fault zone. In such a case, prolific vertical fracturing would essentially render any confining properties of the Bexar Shale as void. Rene Hefner noted that the nature and degree of faulting is not known at this time, but a highly brecciated zone is not expected. Steve Rossello thought that there was valuable information to be gained from cluster wells installed within a faulted area so that the effects of horizontal flow disruption could be directly compared to cluster well data collected from outside of the faulted zone.

Scott Pearson voiced concerns about the proposed well intake design. Although the EPA asked that the basal portion of the geologic units be monitored, Scott thought that there was a significant risk to installing screened wells with short intakes (10-foot screens) such that the portion of the formation open to the borehole may not freely yield water (e.g., dolomitic or massive limestone beds). It was the opinion of some that this approach was limiting considering that contaminated zones (especially in the Lower Glen Rose) may be sealed off from the monitored portion of the formation. Scott Pearson proposed that the clusters well should be constructed as cased open boreholes similar to wells MW1 and MW2 (as shown in Figure 3). This would ensure that the wells were not dry, and that the well was free to receive potentially contaminated water from any water-bearing portion of that unit. Roger Peebles and Joe Hailer agreed, and further added that if the Cow Creek was under confining pressure with an upward vertical gradient, then monitoring of the upper portion of the unit would be more appropriate.

A brief discussion regarding the well identification nomenclature was also discussed. The Lower Glen Rose wells will be named MW3, MW4, and MW5. The cluster wells shall be designated CW-1 through CW-4. Individual wells within a cluster will be identified by unique suffix to identify the formational zone (e.g., LGR, BS, CC or [U]pper, [M]iddle, and [L]ower). The identification will be less then ten characters in length to satisfy IRPIMS requirements.

CSSA has been collecting weekly groundwater elevations from all monitoring wells. They have recently purchased four transducers for the RL83 effort. Upon well installation, the transducers will be installed at locations CW-3 and CW-4 to begin monitoring. Throughout the modeling study, the transducers will likely be moved to other locations for comprehensive coverage.

In terms of collecting additional hydraulic data, Scott Pearson suggested "mini" drawdown tests may be feasible during the development of each new well. Currently, an eight-hour development period is scheduled for each well with a 10 gpm submersible pump. The upper zones could be monitored with the transducers during the development of the deeper wells to measure any observable hydraulic connection. In general, the attendees thought it was a good idea, but were concerned that short test time may only yield limited information, and that 10 gpm discharge may not be enough to produce any visible effects. Parsons ES will check into the feasibility of such a test. If appropriate, Parsons ES may make a request for modification to the RL83 contract to maximize use of the development data.

As a future action, it was agreed that a pumping test in the vicinity of CW-3 would be a good idea. It was also agreed that it would be useful to have two pumping tests, one in the fault zone and one outside the fault zone. This would be included in future task orders as well as any additional data gaps identified by the preliminary results of the model.

Roger Peebles suggested the use of diffusion samplers (baggies) to provide qualitative data regarding chlorinated hydrocarbon concentrations at different depths with the water column of the existing wells at CSSA. Roger indicated that he would provide additional information regarding these samplers to Chris Beal. [Note: CSSA installed a diffusion sampler at Well 16 after the meeting. The sampler was removed for analysis in mid-December 1999.]

Joe Hailer suggested monitoring for natural attenuation parameters for chlorinated hydrocarbons and indicated he would look into appropriate analyses and parameters.