[Home] [Master Table of Contents]

[Table of Contents] [Next Section]

Well Installation Report - Wells CS-MW3 through CS-MW10, August 2003

Section 4 - Results and Interpretation

Results of geophysical logging, packer tests, groundwater and soil sample analysis, and fracture core analysis are provided in this section.

4.1 - Geophysical Logs

A total of 17 logging runs were performed by GeoCam to collect borehole geophysical data for the 15 new wells that were installed. The four geophysical parameters collected were SP, gamma ray, caliper, and electrical resistivity. Table 4.1 lists all the geophysical runs completed during the drilling effort. Normally, the logging run only covered the formation that was to be screened. A complete composite log was then constructed by merging all data from the individual logging runs. The different lengths of the logging tools account for slight variations in total depth measured for a particular run. The geophysical logs for all wells installed under RL83 can be found in Appendix B (not available electronically). Geophysical logs for the wells installed under DO23 can be found in the DO23 Well Installation Report.

4.2 - Injection Packer Tests

A total of 38 packer tests (seven for DO23 and 31 for RL83) were conducted in selected stratigraphic zones at eight drilling locations (CS-MW3 through CS-MW10). The generalized strategy was to perform two or more tests per formation (LGR, BS, CC) that was to be monitored at that location. Approximately 30 percent of the tests performed yielded �no flow� conditions that were normally terminated within the first 5 minutes of the test. The following sections describe the location tests and presents the calculated values of relative permeability. Results of the testing are presented in Appendix C.

4.2.1   Pack Test Summaries

4.2.1.1   CS-MW3-LGR

In corehole CS-MW3, three packer tests were performed within 5.5 feet intervals in the LGR on February 4, 2001. The interval between 401.5 and 407 feet bgs was selected because it was a zone of high resistivity and low gamma counts. It was also being considered for the screened interval of CS-MW3-LGR. A second interval (304.5 to 310 feet bgs) was tested to determine the permeability of an upper borehole interval of high resistivity. However, this test resulted in a �no flow� condition. Another resistive zone between 266.5 and 272 feet bgs was selected, which also resulted in a �no flow� condition.

Table 4.1 - Summary of Geophysical Logs

4.2.1.2   CS-MW4-LGR

Two packer tests were performed within 5.5-feet intervals in the LGR Limestone at corehole CS-MW4 on February 7, 2001. The 304.5 to 310 feet bgs interval was selected to coincide with the proposed screen interval and the zone with the highest resistivity value. The second tested interval (213.5 to 219 ft bgs) was chosen to coincide with a high resistivity value and because it was in a zone of wackestones and grainstones exhibiting some secondary porosity features.

4.2.1.3   CS-MW5-LGR

At corehole CS-MW5, two packer tests were performed within 5.5 feet intervals in the LGR Limestone on February 2, 2001. The 419.5 to 425 ft bgs interval was selected to determine the permeability of a secondary porosity features (e.g., vugs) with a high resistivity value that coincided with the top of the proposed screened interval. The second tested interval (319.5.5 to 325 ft bgs) was chosen to coincide with a high resistivity value in zone of fossiliferous limestones with some large vugs. This interval corresponds with the 304.5 to 310 feet bgs interval tested at the CS-MW3 location. Both tests were unable to inject water into the formation.

4.2.1.4   CS-MW6 Cluster

At the CS-MW6 cluster, which consists of three wells, seven packer tests were performed within 5.5 foot intervals. Two tests each were performed within the LGR and the CC Limestones, and three tests were conducted within the BS. The tests were performed in the corehole of each formation.

Two tests were performed within 5.5-foot intervals within the LGR on February 21, 2001. The 340 to 345.5 feet bgs interval was selected to coincide with the proposed screen interval, and also an interval with high resistivity value and low gamma count. The second tested interval (196.5 to 202 feet bgs) was chosen to investigate the potential for uphole groundwater contribution from a small interval with secondary porosity features and high resistivity.

Three packer tests were performed within the BS on March 15, 2001. The 409.5 to 415 foot bgs interval corresponds with the mid-point of the proposed screen interval, and also corresponded with the depth of a geotechnical sample submitted for permeability testing. The second test interval (389.5 to 395 feet bgs) was in a zone of lower gamma counts relative to its bounding layers. The third test (377 to 382.5 feet bgs) characterizes the top of the BS. All three tests resulted in �no flow� conditions.

Two injection tests were performed within the proposed screened interval portion of the CC on April 8, 2001. The 470 to 475.5 feet bgs test represents the lower portion of the screened interval, while the 448 to 453.5 foot bgs interval coincides within a zone of increased resistivity at the top of the screened interval.

4.2.1.5   CS-MW7 Cluster

The CS-MW7 consists of two wells: CS-MW7-LGR and CS-MW7-CC. Two packer tests were performed at 5.5-foot intervals within the LGR corehole on February 19, 2001. Two additional tests were performed at 12.1-foot intervals within the CC corehole on June 28, 2001. A larger interval was tested because the packer assembly had been modified to include an integrated submersible pump to allow for collection of discrete groundwater samples. The check valve within the pump was removed to allow unrestricted flow during the injection tests. The modification allowed for a zone to be sampled prior to the injection test, creating an efficient and cost-effective method of obtaining two types of data from one setting of the packer system.

In the LGR, the 324.5 to 330 foot bgs packer test interval coincides with the upper portion of the proposed well screen interval, increased relative resistivity, and a lower gamma count. A second packer test was performed at a depth equal to a perched water level between 68 and 73.5 feet bgs. For the CC tests, the 440 to 452.1 feet bgs was equivalent to the proposed screened interval. The upper portion of CC was tested between the 420 and 432.1 foot bgs interval in a grain-supported limestone that was noted as a water-producing interval.

4.2.1.6   CS-MW8 Cluster

At cluster CS-MW8, which consists of an LGR and CC well pair, two tests were performed at 10.5-foot intervals within the LGR on May 7, 2001; and two tests were performed at 10.2-foot intervals within the CC Formation on June 5, 2001. The injection tests utilized the dual-purpose packer configuration described in CS-MW7 (Section 4.2.1.5). The difference in length between the packer intervals represents a slight variation in connecting hardware used for each lithologic zone.

The lower portion of the proposed LGR screened interval was tested between 346.5 and 357 feet bgs, where the maximum resistivity value was measured. A water-producing vuggy grainstone, based on lithologic observation, was selected as the second interval (305.5 to 316 feet bgs) to test. Two tests were performed across both the upper (434.8 to 445 feet bgs) and lower (449.8 to 460 feet bgs) sections of the proposed screened interval for the CC Limestone well.

4.2.1.7   CS-MW9 Cluster

The CS-MW9 cluster consists of three wells: CS-MW9-LGR, CS-MW9-BS, and CS-MW9-CC. Between November 22, 2000 and January 9, 2001, seven packer tests were performed within 5.5-foot intervals in these wells. Two tests each were conducted within the LGR Limestone and the BS, and three packer tests were conducted within the CC Limestone. All the tests were performed in the CC Limestone corehole.

Within the LGR Limestone, the 294 to 299.5 foot bgs interval was chosen because the resistivity value was high in this interval and the gamma count was low, and because this interval crossed the upper portion of the proposed screened interval. The second test (153 to 158.5 feet bgs) in the LGR Limestone was selected from a water-bearing interval that also had a high resistivity, low gamma count, an increased borehole caliper, and was logged as being very vuggy with core loss. As described in Section 4.2.2, this zone was the most permeable zone tested during the field effort.

Two test zones (360 to 365.5 feet bgs and 370 to 375.5 feet bgs) within the BS were chosen based on relatively lower gamma counts in comparison to the upper and lower portions of the lithologic unit. Three tests were performed in the CC section of the corehole. The deepest test (445.5 to 451 feet bgs), which corresponded to the lower portion of the proposed screened interval, resulted in a �no flow� condition. The packer was reset to coincide with the upper portion of the proposed screen interval (424.5 o 430 feet bgs), which was in a zone of high resistivity and low gamma count. A third test was conducted between 386.5 to 392 feet bgs, in a zone which the core barrel �dropped� about 1 foot during coring operations.

4.2.1.8   CS-MW10 Cluster

At cluster CS-MW10, which consists of wells CS-MW10-LGR and CS-MW10-CC, nine packer tests were performed within 10.4�foot intervals. The injection tests utilized the dual-purpose packer configuration described in CS-MW7 (Section 4.2.1.5). Four tests were conducted within the LGR, two tests were conducted within the BS, and three tests were conducted within the CC Limestone. All the tests were performed in the CC corehole.

On August 24, 2001, four injection tests were performed within the LGR. Two intervals (307 to 317.4 feet bgs and 320 to 330.4 feet bgs) were selected to test the permeability of fault-type fractures noted during coring; however, each of these attempts resulted in a �no flow� condition. A test performed in the 356 to 366.4 feet bgs vuggy interval coincides with high resistivity and low gamma counts above the proposed screened interval. The remaining test (379 to 389.4 feet bgs) targeted the high resistivity zones located within the middle of the proposed screened interval.

Two injection tests were performed in the BS on September 13, 2001. A �no flow� condition was encountered between 416 and 426.4 feet bgs. The second test investigated an interval of small �core barrel drops� reported by the driller between the interval of 436 and 446.4 feet bgs.

Three tests were conducted in the CC Limestone on September 13,feet 2001. Water was injected into a fracture noted during logging between the interval of 460feet and 470.4 feet bgs. A high resistivity value located within the middle of the proposed screened interval was tested between 478 and 488.4 feet bgs. A final test between 504 feet and 514.4 feet bgs was conducted below the main water-bearing unit to determine the relative permeability of the lower section of the CC Limestone.

4.2.2   Data Reduction and Hydraulic Conductivity

To perform the evaluations, some basic assumptions regarding the nature of the aquifer were made. The underlying assumption for these injection tests was that they were performed within the zone of saturation. Historical reports (ES, 1993) indicate that the Middle Trinity Aquifer (e.g., the LGR and the CC Limestones) is primarily under an unconfined water table condition.

While this may hold true on a regional scale, the continual acquisition of data indicates that locally confined conditions can occur, especially in the case of the CC Limestone where the BS acts as a confining bed. In the vicinity of CSSA, the LGR Limestone is more apt to behave as an unconfined unit due to the lack of a truly confining upper layer (Upper Glen Rose [UGR]). However, at the scale for which the testing was conducted, groundwater intersected in the corehole can easily behave as if under confined conditions. This can occur since the secondary porosity features encountered at depth (e.g., karst and fractures) may ultimately interconnect with higher-elevation voids under an unconfined condition. In other words, confinement may occur locally, while the Middle Trinity Aquifer operates regionally as an unconfined aquifer.

With those points being made, the site hydrogeology and testing conditions best matches the analytical method for Zone 3 (below water table in the saturated zone) as described in Section 2.9.3. Even for localized confinement, the method for Zone 3 is still applicable. The mechanics of the Zone 3 equation deal with gravitational head for a water table aquifer, but also apply to the pressure head for a confined aquifer. For purposes of this analysis, there is no distinction between confined and unconfined heads in the calculation of the distance to the water table (h₁) term.

A total of 19 tests were conducted within the LGR, while 7 tests and 12 tests were completed in the BS and CC Formations, respectively. The equation presented in Section 2.9.3 was used to calculate the values of K from the injection tests performed. Table 4.2 lists the resultant K values, and Appendix C contains the computation sheets.

Summary statistics by hydrologic unit are presented in Table 4.3. Of the 38 tests attempted, 11 resulted in a �no flow� condition. With respect to the testing methodology, an impermeable condition was encountered in each of the hydrologic units, with the highest percentage of �no flow� conditions occurring in the BS. While it is understood that these geologic materials possess some coefficient of permeability, for the purposes of this report, those �no flow� field tests are reported with a null value (0).

Table 4.2 - Hydraulic Conductivities derived from Injection Packer Tests

Table 4.3 - Statistical Summary of Injection Packer Tests

Within the CC, BS, and LGR Formations, a total of 1, 4, and 6 �no flow� intervals were encountered, respectively. The normalized data set indicates that the LGR will encounter a �no flow� condition nearly four times more often than the CC Limestone. Likewise, the BS is seven times more likely to yield a �no flow� condition than the CC Limestone. Of course, this data set is skewed by preferential selection criteria used to choose the test zones.

Including the �no flow� tests, K values ranged from some degree of impermeability (0 ft/sec) in all three formations (LGR, BS, and CC) to 5.03 x 10^-5 ft/sec in the LGR. When the entire test population is normalized to relative permeability, the data shows that the LGR and CC formations are 7.5 and 11 times more permeable than the BS, respectively.

From another point of view, when the 11 �no flow� zones are removed from the data set, the least permeable test conducted occurred in the LGR (68-73.5 feet bgs) with a hydraulic conductivity of 3.47 x 10^-8 ft/sec. Within this data subset, the median and average formational permeabilities are implicitly increased as would be expected. The averaged LGR and CC hydraulic conductivities are greater than those permeable sections of the BS by a factor of 4.7 and 5.4, respectively.

For completeness, the injection tests that occurred within the eventual screened intervals of the wells were evaluated. However, it must recognized that the results of this evaluation are skewed since the packer interval only represents a sometimes small portion of the actual water-producing section within the screened interval. In this case, a well test is a more reliable method of determining K of the screened interval. With that being said, the summary statistics indicate that the K of the LGR wells are twice that determined for the BS wells. Likewise, analysis indicates CC well screen intervals are more permeable than the BS wells by a factor of 6.

As Appendix C indicates, only a couple of tests exhibited low injection pressure (less than 21 psi) and elevated injection rates. The remaining intervals generally had relatively consistent injection pressure that varied from 30 to 50 psi with diminishing rate of water loss with time.

According to the Handbook of Hydrology (Maidment, 1993) and with respect to the geologic terrain, the average LGR and CC Formations are typified by lower-permeable karstified limestone, while the BS falls more closely toward a carbonate mud permeability (Figure 4.1). For comparison, an average range for the Edwards Limestone hydraulic conductivity (Maclay, 1995) is included in the diagram to illustrate the stages of karstic development between the Middle Trinity and Edwards Aquifers. Another interpretation of the hydraulic conductivity measurements of the LGR and CC intervals is that they are indicative of a fractured flow regime, and those of the BS interval are suggestive of matrix flow.

Figure 4.1   Relative Hydraulic Conductivity Comparison (all values in ft/sec)

4.2.3   Geotechnical Lab Testing

As an independent study, geotechnical samples were submitted by CSSA to evaluate the permeability and absorption properties of selected recovered core from the respective members of the Middle Trinity Aquifer. Results of this testing was provided by CSSA for inclusion into this report. Data reports regarding this testing can be found in Appendix C.

Three discrete rock samples originating from target depths within either the LGR Limestone, BS, or CC Limestone were submitted for analysis to Raba Kistner Consultants, Inc. (R-KCI) for permeability analysis in Spring 2001. The permeability analysis followed ASTM method D5084, which used regular tap water as the permeant fluid. The parameters quantified by this analysis included the initial moisture content, bulk density (unit dry weight), and vertical permeability. Table 4.4 summarizes findings of the analysis. In general, moisture content of the samples were low (less than 13 percent) and the bulk density of the CC Limestone was less than that of the LGR and BS. The permeability results correspond to the findings of the injection packer tests which indicate that the CC matrix is the most permeable and the BS matrix is the least permeable.

Results of the geotechnical analysis were within one magnitude of difference with those positive test results associated with the injection packer tests. In particular, the LGR geotechnical sample was 4.3 times more permeable than the average LGR injection test. Likewise, the CC geotechnical sample was 9.6 times more permeable than the average CC injection test. However, the BS geotechnical sample was 3.1 times less permeable than the average BS injection test. Differences in the results are likely attributable to small sample population of the geotechnical testing and the significantly lesser interval distance (0.5-foot compared to a 10-or 12-foot interval) between the testing methodologies. All things being considered, results within an order of magnitude could be considered as similar.

Table 4.4 - Geotechnical Permeability Analysis

CSSA also independently conducted matrix absorption tests using a modified version of ASTM C127. The intention of the analysis was to quantify absorptive properties of the rock matrix with respect to a solvent. The test involved weighing the oven-dried core sample (A), soaking the core in a permeant fluid (PCE) for 24 hours, and weighing the saturated sample (B). The absorption potential is given as a percentage in terms of saturated weight divided by weight of the oven-dried sample weight as shown by the following equation.

Absorption % = {(B-A)/A} x 100

CSSA submitted three vadose zone samples (UGR) from the Building 90 area which typified the shallow surface conditions to which solvent contamination had been released. Three samples were obtained from CS-MW6-LGR located near the northwest corner of Building 90. Twenty-four-hour absorption rates ranged from 0 percent to 0.7 percent. The 26.5 to 27-foot sample matrix decomposed while in solution; therefore, the test could not be continued. The 48-hour absorption measurement showed between 0.02 percent and 0.7 percent absorption. Table 4.5 shows the results of the testing, and further data can be located within Appendix C.

Table 4.5 - Solvent Absorption Testing

4.3 - Groundwater Sampling

4.3.1   Discrete Sampling During Well Installation

A total 40 discrete interval groundwater samples were collected at well locations CS-MW7, CS-MW8, and CS-MW10 using those methods described in Section 2.13.2. The discrete samples were used to vertically profile the contaminant characteristics of perched water intervals, the LGR Limestone, and the CC Limestone. During the drilling of a well cluster, discrete samples were collected from either one or both of the wells comprising a well pair. Specifically, comparison samples from the same strata of the LGR and CC wells within a cluster were collected at the CS-MW7 and CS-MW8 locations. After those comparisons indicated that substantial differences in concentration were not evident within the well pair, that procedure was discontinued at cluster CS-MW10.

The discrete interval sampling program was initiated in April 2001 while drilling at CS-MW8. The procedure and approach were refined and tailored through implementation at CS-MW7 and CS-MW10. In the beginning, a small portion of the analytical work was performed by Cooperheat in San Antonio, TX. The bulk of the contaminant screening was carried out by DHL Analytical in Round Rock, TX. Sample results were generally received within 24 hours of collection, and ultimately dictated the progression and construction of the wells.

All samples were analyzed for target VOCs which included MEK, cis-1,2-DCE, trans-1,2-DCE, PCE, TCE, and toluene. In addition, acetone, isopropyl alcohol, and toluene, were identified early on in the drilling program as tentatively identified compounds (TIC), and were subsequently added to the target list. Isopropyl alcohol is the primary ingredient of the foaming agent occasionally used during the drilling, and acetone is a by-product of the isopropyl alcohol degradation.

4.3.1.1   CS-MW7

A total of 10 groundwater samples were collected during the process of coring and drilling CS-MW7-LGR and CS-MW7-CC. Eight samples originated from the LGR Limestone, and the remaining two samples were obtained from the CC Limestone. Two sets of data (e.g., four samples) were stratigraphic duplicate samples between the LGR and CC wells.

Within the LGR section of the aquifer, only PCE and TCE were reported above MDLs. Detected concentrations of PCE ranged between 3.17 �g/L to 10.2 �g/L in six samples (Table 4.6). Detected concentrations of TCE ranged from 1.3 �g/L to 8.31 �g/L in a differing set of six samples. The results indicate that either PCE and/or TCE are present throughout the upper 300-foot section of the LGR within the Middle Trinity Aquifer.

Table 4.6 - CS-MW7 Discrete Groundwater Sampling Results

In the CC portion of the Middle Trinity Aquifer, only acetone was reported above the MDL. As described above, acetone is a degradation by-product of the drilling foam, and therefore, its detection at the low concentrations noted above is not considered to be site-related. Figure 4.2 is a graphical representation of contaminants occurring within the Middle Trinity Aquifer at well cluster CS-MW7.

4.3.1.2   CS-MW8

A total of 21 groundwater samples were collected during coring and drilling activities at CS-MW8-LGR and CS-MW8-CC. Nineteen samples originated from the LGR Limestone, and the remaining two samples were obtained from the CC Limestone. Six sets of data (e.g., 12 samples) were stratigraphic duplicate samples between the LGR and CC wells.

All target analytes, except trans-1,2-DCE, were reported above MDLs within the LGR section of the Middle Trinity Aquifer (Table 4.7). The CS-MW8 cluster is the location at which the discrete sampling program began, and the results from this cluster were used to develop the list of target analytes for the other wells. One occurrence of MEK was reported in a perched water-bearing zone at 53 feet bgs. Acetone was reported in that sample also, in addition to another sample collected at 166 feet bgs. The occurrence of acetone at this cluster is more significant because it was detected prior to the use of any injected foaming agent. The compounds cis-1,2-DCE, PCE, and TCE were reported in samples ranging in depth from 83 feet bgs to 300 feet bgs. Detected concentrations of cis-1.2-DCE ranged between 0.18 �g/L to 0.57 �g/L in six samples. Detected concentrations of PCE ranged between 3.19 �g/L to 57 �g/L in 14 samples. Detected concentrations of TCE ranged from 2.6 �g/L to 20.5 �g/L in 12 samples which also had detectable concentrations of PCE. Toluene was also reported in three samples, with concentrations ranging between 2.32 �g/L and 14.2 �g/L. No contaminants were reported beyond the 300-foot depth.

Results indicate that either PCE and/or TCE are present throughout the upper 300-foot section of the LGR within the Middle Trinity Aquifer. In general, PCE levels appear to decrease with depth while TCE levels increase. The screening data indicate that contaminants in excess of corresponding MCLs are present in the lower-yielding elevations of the LGR. Conversely, concentrations of the compounds are not detectable in the higher-yielding portions of the LGR. Presumably, the apparent loss of contaminant concentration is due to dilution.

In the CC portion of the Middle Trinity Aquifer, no contaminants were reported above the MDL at the CS-MW8 cluster location. Figure 4.3 is a graphical representation of contaminants occurring within the Middle Trinity Aquifer at well cluster CS-MW8.

Table 4.7 - CS-MW8 Discrete Groundwater Sampling Results

4.3.1.3   CS-MW10

A total of nine groundwater samples were collected during the process of coring and drilling CS-MW10-CC. Six samples originated from LGR Limestone, one sample originated from the BS, and the remaining two samples were obtained from the CC Limestone. Since prior tests at CS-MW7 and CS-MW8 clusters indicated similar results between stratigraphic depth duplicates, duplicate samples were not collected from CS-MW10-LGR.

Within the LGR section of the aquifer, only PCE and toluene were reported above MDLs. Detected concentrations of PCE ranged between 0.62 �g/L to 1.51 �g/L in four samples (Table 4.8). Detected concentrations of toluene ranged from 0.49 �g/L to 0.69 �g/L in two samples. Acetone (45 �g/L) was reported within the BS at a depth of 447 feet. However, the sample was visibly soapy, hence the acetone is considered likely to be a result of the foaming agent.

In the CC portion of the Middle Trinity Aquifer, no contaminants were reported above the MDL at the CS-MW10 cluster location. Figure 4.4 is a graphical representation of contaminants occurring within the Middle Trinity Aquifer at well cluster CS-MW10.

Table 4.8 - CS-MW10 Discrete Groundwater Sampling Results

4.3.2   Off-post Impact Monitoring

Due to the proximity of the MW-10 cluster to off-site private wells LS-2, LS-3, LS-6, and LS-7, concerns were raised as to the possible impact drilling of the cluster might have on water quality and contaminant levels in the off-site wells. To monitor the possible impact in these wells, off-site sampling was conducted prior to, during, and following drilling and installation of the MW-10 cluster. The analytical data for these sampling events are summarized in Table 4.9 and Figure 4.5.

All wells showed a decrease in PCE concentrations in samples collected during drilling. However, PCE levels increased 3 days after the MW-10 wells were completed. The last round of samples taken as part of the off-site monitoring were taken about 2 weeks after the MW-10 cluster was completed. Three of the four wells showed decreased PCE concentration, but LS-6 showed a slight increase from 6.7 �g/L to 10.0 �g/L.

4.3.3   Major Ion Data

Between June and December 2001, all DO23 and RL83 wells were sampled and analyzed for major cations and anions. The purpose of this sampling was to evaluate differences, if any, in groundwater chemistry between the LGR Limestone, the BS, and the CC Limestone. Analytes evaluated include calcium (Ca⁺²), sodium (Na⁺), potassium (K⁺), magnesium (Mg⁺²), chloride (Cl^-), carbonate (CO₃^-2), bicarbonate (HCO₃^-), and sulfate (SO₄^-2). These analytical data was used to create Stiff pattern diagrams and a Piper diagram.

Stiff pattern diagrams show the concentrations of cations and anions using three parallel horizontal axes. The cations are shown in milliequivalents per liter on the left of the vertical axis, and the anions are plotted on the right. The horizontal scale increases away from the vertical axis, so the larger the area of the pattern, the more mineralized the groundwater. The resulting polygon can be used to graphically characterize the ionic concentrations of a certain groundwater sample and possibly delineate hydrogeochemical facies within an aquifer (i.e., CC, BS, and LGR).

In Table 4.10 and Table 4.11, the major ion data are shown with the Stiff pattern of each well plotted below each sample. Some visible trends can be seen in the Stiff pattern diagrams when they are grouped with respect to the formation screened in each well.

The major trends in the Stiff patterns are:

Sulfate concentrations are typically two to 10 times greater in the BS and CC wells compared to the LGR wells.

On average, calcium concentrations are lower in the BS and CC wells with the lowest concentrations being in the BS wells

Sodium, potassium, and magnesium are all more abundant in the CC and the BS wells.

The data in Table 4.10 and Table 4.11 were used to generate a Piper diagram to illustrate differences in groundwater chemistry between the three hydrostratigraphic units. Piper diagrams consist of three components: two trilinear diagrams along the bottom and one diamond-shaped diagram in the middle, as shown in Figure 4.6. The trilinear diagrams illustrate the relative concentrations of cations (left) and anions (right) in each sample. For both trilinear diagrams, the chemical data are normalized with only the components shown on that trilinear diagram. Each corner represents 100 percent of the labeled component, with the percentage of that component decreasing away from the corner along the labeled side. Also, all data that plots inside the diagram contain all three components, whereas data that lie along the lines of the triangle contain two components, but none of the third component. On the example trilinear diagram shown in Figure 4.6, the regions where one cation or anion is predominant (greater than 50 percent) are indicated.

The diamond field is designed to show both anion and cation groups. For each sample, a line is projected from its point in the cation and anion trilinear diagrams into the diamond region, and the data point is plotted where the lines intersect. The diamond field is used to illustrate where a particular sample will fall into one of four groundwater types:

	Ca+2 + Mg+2 + SO4-2 + Cl-;
	Ca+2 + Mg+2 + CO3-2 + HCO3-;
	Na+ + K+ + SO4-2 + Cl-; and
	Na+ + K+ + CO3-2 + HCO3-.

Groundwater type generally refers to those cations and anions that constitute more than half of the total on a milliequivalent basis. Therefore, the above examples represent groundwater that would be in equivalent in percentages of all cations and anions, however, this is rarely the case. Many times, groundwater will be dominated by one cation and one anion (e.g., calcium-bicarbonate type water).

As shown in Figure 4.7, the anions present in groundwater collected from all three hydrostratigraphic units mostly consisted of HCO3- molecules, with lesser amounts of CO3-2in the BS and CC groundwater. Also, a greater percentage of SO4-2 is present in the BS and CC than within the LGR. Whereas the LGR samples fall within a closely-spaced cluster, the BS and CC samples diverge from this clustering because of additional SO4-2 mineralization.

Formational clustering by cations is not as distinct as with the anions in Figure 4.7. For nearly all of the LGR samples, K+ and Na+ are less than 20 percent of reported cations. Most of the LGR samples (five of eight) have Ca+2 as the major (greater than 50 percent) cation component. The remaining three LGR samples have increased proportions of Mg+2, Na+ or K+. In this data set, CS-MW3-LGR is aberrant because the major cation (nearly 80 percent) is Mg+2, not Ca+2. Both BS wells group closely in Na+ and K+ portions of the trilinear diagram. These two samples, as well as two CC samples (CS-MW6-CC and CS-MW10-CC) are very low in Ca+2 and high in Na+ and K+ with respect to the bulk of the samples. The remaining three CC samples (CS-MW7-CC, CS-MW8-CC, and CS-MW9-CC) are tightly clustered at 38 percent Ca+2, 45 percent Mg+2, and 17 percent Na+ and K+.

When the cations and anions are plotted together in the diamond-shaped field of Figure 4.7 and based on data given in Table 4.10 and Table 4.11, it is evident that all the samples plot chiefly as HCO3- type groundwater. All the LGR samples plot out as a Ca+2-Mg+2-HCO3- type groundwater. Additionally, with the exception of CS-MW6-LGR, all the LGR samples are grouped in a closely spaced cluster. The BS samples differ in character from the LGR samples because of the absence of Ca+2, and the proportional increase in Na+ and K+ (Na+-K+-HCO3- type groundwater).

There appears to be three types of CC groundwater that are not necessarily explained by their geography or relative position to other wells. The majority of the CC samples (3 of 5) plot out as a Ca+2-Mg+2-HCO3- type groundwater. However, they are distinct from the LGR samples because of increased SO4-2 content. The remaining two CC samples are classified as a either a Na+-K+-HCO3- type groundwater (CS-MW6-CC) or K+-CO3-2 type groundwater (CS-MW10-CC) because of the proportional lack of Ca+2 and Mg+2.

The diagrams clearly show some distinction between formational members of the Middle Trinity Aquifer. As expected from a limestone aquifer, the water source is primarily HCO3- in nature. Water from the LGR portion of the aquifer seems to be very similar throughout the well locations. Within the LGR, the major difference noted is the relative proportions of Ca+2 versus Mg+2. The differences in relative cation abundance do not appear to be a function of down dip mineralization, and are perhaps more attributable to geographic areas of localized dolomitization or some structural control. The same can be said for the CC samples, except the changes in cation abundance are inversely proportional concentrations of Na+ and K+ in lieu of Ca+2 and Mg+2. Also, increased SO4-2 content in the BS and CC may be attributable to the dissolution of anhydrite within those units.

The dissimilarity of groundwater suggests that the three formational members of the Middle Trinity Aquifer are not pervasively in direct hydraulic communication across the post. However, it is likely that localized communication exists in structurally significant areas (faulting planes), and may explain some of the outlying data points. The halide mineralization (Na+ and K+) of the BS and portions of the CC probably represent depositional environment characteristics that are common to both units.

As the data sets increase with the addition of more wells, further differentiation and possibly some statistical analysis can be quantified.

4.4 - Soil and Rock Sampling

Soil and rock samples were collected from each of the corehole locations. The target sampling intervals were surface soils, LGR, BS, and the CC Limestone. Samples obtained as part of the DO23 wells were submitted to O�Brien & Gere Laboratories. Likewise, samples collected from RL83 wells were submitted to APPL Laboratories for similar analyses. The following represents a discussion of detections in each of those stratigraphic zones. The data are tabulated in Table 4.12, Table 4.13, Table 4.14, and Table 4.15, respectively. Full analytical data for each sample collected are presented in Appendix E.

4.4.1   Surface Soil Samples

Five soil samples (including one duplicate) were collected between January and February 2001 (Table 4.12). For metallic constituents in surface samples, only cadmium was reported above the established background concentration (the AFCEE reporting limit [RL]) of 0.1 mg/kg) in three sample submittals. Trace amounts of methylene chloride, toluene, and TCE were reported at two sample locations above the laboratory MDL, but less than the AFCEE RL (F-flagged data).

4.4.2   Glen Rose Samples

Sixteen rock samples (including one duplicate) were collected from the LGR between November 2000 and August 2001 (Table 4.13). Each sample was analyzed for metals and VOCs.

Inorganic constituents were reported above site-specific Glen Rose background levels in 10 of 16 samples. These exceedances included barium (one sample), chromium (one sample), nickel (three samples), zinc (five samples), arsenic (one sample), cadmium (four samples), and mercury (one sample). However, it should be noted that the current LGR background metals levels were calculated based on samples collected near the surface (up to 20 feet deep). A source of metals contamination at the depths observed in the cluster well samples is unlikely; and the concentrations detected during the cluster well drilling are likely representative of background conditions.

As with the surface samples, trace amounts of methylene chloride, toluene, and TCE were reported at four sample locations above the laboratory MDL, but less than the AFCEE RL (F-flagged data).

4.4.3   Bexar Shale Samples

Eight rock samples (including one duplicate) were collected from the BS between November 2000 and September 2001. All samples were analyzed for metals and VOCs.

Inorganic constituents were reported above Texas-specific median background concentrations in seven of eight samples. These values were included as qualitative comparison from the TRRP since no site-specific concentrations were determined for this subsurface unit. Inorganic compounds in excess of the median background concentration included chromium (five samples), nickel (five samples), zinc (one sample), arsenic (one sample), cadmium (five samples), and mercury (one sample). However, since there is no known source of metals contamination at these locations, these concentrations are considered to be indicative of background conditions. Trace amounts of either bromobenzene, 1,4-dichlorobenzene, naphthalene, 1,2,3-trichlorobenzene, and 1,2,4-trichlorobenzene were reported at two sample locations above the laboratory MDL, but less than the AFCEE RL (F-flagged data).

4.4.4   Cow Creek and Hammett Shale Samples

A total of seven samples were obtained from the CC Limestone interval, and two samples (including one duplicate) from the Hammett Shale. The contact between these units is conformable and gradational, and therefore have been combined for this discussion.

Inorganic constituents were reported above Texas-specific median background concentrations in five of nine samples. These values were included as qualitative comparison from the TRRP since no site-specific concentrations were determined for these subsurface units. Inorganic compounds in excess of the median background concentration included nickel (two Hammett Shale samples), arsenic (one sample), cadmium (three CC samples), and mercury (one CC sample). As described above, because there is no known source of metals contamination at these depths or locations, these concentrations are considered to represent background conditions.

Trace amounts of either methylene chloride, naphthalene and toluene were reported at three locations above the laboratory MDL, but less than the AFCEE RL (F-flagged data). Additionally, methylene chloride and toluene were reported above the RL within the CC at a depth of 495 feet bgs at location CS-MW10-CC.

4.5 - Fractured Core Analysis

A total of 80 feet of 2�-inch diameter conventional drill core recovered from four pilot holes drilled during the RL83 program were submitted to Core Labs for fracture analysis. Table 4.14 summarizes the core recovery in each pilot hole submitted for analysis. The summary report is included as Appendix F.

Table 4.14 - Core Intervals Submitted for Fracture Analysis

Twenty natural fractures were identified in the subject cores, of which 18 are �orientable.� The remaining two fractures are termed �unmeasureable,� which means they cannot be traced from one side of the core to the other to obtain an orientation. The fractures are a mixture of broken (5), closed (3), vuggy (1 � [Appendix F - Photos E & F]), mineralized (6) and broken/mineralized (5). The latter include fractures that break the core into two or more pieces, and which have some mineralization on the fracture surfaces. Calcite is the cementing mineral in all cases (Appendix F - Photo C). Some broken fractures show slickensides on the fracture surfaces, indicating that they are shear fractures (Appendix F - Photos A & G). Some of the closed fractures are stained reddish-orange, suggesting that iron-bearing minerals have precipitated from groundwater moving along them (Appendix F - Photos B & D).

Fracture dip angle varies between 25� and 68�, with an average of 50�. Cores from the CS-MW10-CC well show consistently low dip angles (25� to 58�). Fracture strike cannot be determined because the cores were not oriented. However, where two or more fractures are present, a single fracture strike trend is typically observed (Appendix F - Figures 7, 13, & 22). An exception to this is noted in core interval three of the CS-MW7-CC core, which shows possibly two strike trends (Appendix F - Figure 10).

The CS-MW6-LGR cores contain a single broken fracture, and this occurs in the upper or shallower interval. The CS-MW10-CC cores contain seven mineralized and broken/mineralized fractures, and these also occur in the shallower of the two intervals. The other wells contain fractures in both the shallower and deeper core intervals.

[Next Section]