Brunswick Aquifer study : reports for Bryan County, Chatham County, Ebenezer Bend (Glynn County), Effingham County, and Glynn County / prepared for the Georgia Environmental Protection Division, Georgia Geologic Survey ; prepared by Golder Associates

Golder Associates Inc.
3730 Chamblee Tucker Road Atlanta, GA 30341 Telephone: (770) 496-1893 Fax: (770) 934-9476

REPORT ON BRUNSWICK AQUIFER STUDY
BRYAN COUNTY
Prepared for: Georgia Environmental Protection Division
Georgia Geologic Survey Room 400
19 Martin Luther King, Jr. Drive S.W. Atlanta, Georgia 30334
Prepared by: Golder Associates Inc. 3730 Chamblee Tucker Road
Atlanta, GA 30341

DISTRIBUTION:
1 Copy GAEPD 1 Copy USGS 1 Copy Golder Associates Inc.
January 2003

GOLDER ASSOCIATES INC.

003-3020

GA Environmental Protection Division

Dr. William McLemore

- i -

TABLE OF CONTENTS

January 2003 003-3020

Table of Contents ....................................................................................................................i

SECTION

PAGE

1.0 INTRODUCTION AND BACKGROUND INFORMATION .................................................................. 1 1.1 Introduction..................................................................................................................................................... 1 1.2 Regional Hydrogeology............................................................................................................................... 2
2.0 AQUIFER ASSESSMENT FIELD PROGRAM BRYAN COUNTY.................................................. 5 2.1 Drilling and Well Installation....................................................................................................................... 5 2.2 Aquifer Testing............................................................................................................................................... 7
3.0 AQUIFER TEST ANALYSES ...................................................................................................................... 10 3.1 Data Compilation and Correction.............................................................................................................. 10 3.2 Test Analysis Software and Application.................................................................................................. 11 3.3 Standard Curve Matching Analyses.......................................................................................................... 12 3.4 Pressure Derivative Analyses..................................................................................................................... 12 3.5 Aquifer Test Analysis Results.................................................................................................................... 13
4.0 AQUIFER YIELD ANALYSIS..................................................................................................................... 15 5.0 SUMMARY AND CONCLUSIONS ........................................................................................................... 17

FIGURE 1 FIGURE 2 FIGURE 3 FIGURE 4 FIGURE 5 FIGURE 6 FIGURE 7 FIGURE 8 FIGURE 9
FIGURE 10
FIGURE 11 FIGURE 12

FIGURES

In Order Following
Page 18

Generalized Correlation of Geologic and Hydrogeologic Units Bryan County Site Location Bryan County Site Layout Generalized Boring Log and Geophysical Log BB-1 Generalized Boring Log and Geophysical Log BB-3 Bryan County Cross Section Bryan County Pumping Test Pumping Test Analysis - Theis Curve Match Log Analysis of Pumping Well Pumping Phase Upper Brunswick Test Bryan County Site Log Analysis of Observation Well Pumping Phase Upper Brunswick Test Bryan County Site Normalized Transmissivity Plot Upper Brunswick Test Bryan County Site Simulated Drawdown Contours Bryan County Site

APPENDICES
APPENDIX A Well Construction Data and Boring Logs TABLE A-1 Summary Well Construction Data Table FIGURE A-1 Boring Log BB-1 FIGURE A-2 Boring Log BB-2 FIGURE A-3 Boring Log BB-3

APPENDIX B Pumping Test Information and Analyses TABLE B-1 Summary of Aquifer Test Results Table FIGURE B-1 Pumping Test Plot FIGURE B-2 Barometric Pressure During Pumping Test

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- ii -

January 2003 003-3020

FIGURE B-3 Time vs. Drawdown Plot FIGURE B-4 Theis Curve Match FIGURE B-5 Hantush and Jacob Leakage Plot FIGURE B-6 Theis Recovery Plot for Observation Well FIGURE B-7 Theis Recovery Plot for Pumping Well FIGURE B-8 Cooper & Jacob Time-Drawdown Plot FIGURE B-9 Pumping Test Data Report for Pumping Well FIGURE B-10 Pumping Test Data Report for Observation Well

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 1 -

January 2003 003-3020

1.0 INTRODUCTION AND BACKGROUND INFORMATION

1.1 Introduction Extensive use of the Upper Floridan Aquifer system has caused cones of depression in the potentiometric surface along the Georgia Coastline (particularly around the communities of Savannah and Brunswick) and salt water intrusion has become a significant concern. Consequently, the Georgia Environmental Protection Division (EPD) has developed an Interim Strategy for Managing Salt Water Intrusion in the Upper Floridan Aquifer of Georgia. Key components of the strategy are to conduct research that will lead to development of measures to stop salt water intrusion and to develop a long term management plan to protect the Upper Floridan Aquifer. Consistent with these components, the Brunswick Aquifer system is being considered as an alternative water source to the Upper Floridan aquifer. Potential use of the Brunswick aquifers is of particular interest in Glynn County, Chatham County, and portions of Bryan and Effingham counties where a cap on further groundwater withdrawals from the Upper Floridan Aquifer has been imposed.

The majority of the Miocene related work completed to date has focused on identifying formation sequences using existing corehole data and geophysical logs (Clarke et. al., 1990). Limited work has been performed defining the hydraulic characteristics of the Brunswick aquifers (Steele and McDowell, 1998). Therefore, an investigation was conducted to characterize the Brunswick aquifer system in these areas. The main objectives were to determine both the potential of the Brunswick aquifers to provide sustainable sources of potable water supply and to what extent these aquifers could be developed without adversely affecting the Floridan aquifer system. The investigation included installation of test wells and observation wells at four sites in Bryan, Chatham, Effingham and Glynn Counties, and performance of pumping tests at each site. The resultant aquifer test data were analyzed and evaluated to provide estimates of specific aquifer parameters and to determine the water-bearing capacity of the aquifer(s) at each location.

This report presents the methodology used and subsequent results of the Brunswick Aquifer investigation in Bryan County, Georgia. The investigation included a single 72-hour aquifer test performed on the Lower Brunswick aquifer beginning on November 14, 2000. Information derived from this investigation can be used by local governments, commercial developers and industry for their water-supply planning. Specifically, the investigation provides information as to whether the Brunswick aquifers offer an alternative source of water to the Upper Floridan Aquifer in the capped areas. The investigation also provides critical input parameters for regional

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 2 -

January 2003 003-3020

groundwater flow and solute transport models currently being developed by the United States Geological Survey (USGS) and others.

1.2 Regional Hydrogeology The groundwater resources of Coastal Georgia consist of three primary aquifer systems: the Surficial Aquifer system, the Brunswick Aquifer system (comprised of the lower surficial aquifer and the upper and lower Brunswick aquifers) and the Floridan Aquifer system (comprised of the Upper and Lower Floridan aquifers). The Upper Floridan aquifer is the principle source of groundwater in the region. The surficial aquifers, the upper and lower Brunswick, and the Lower Floridan are considered secondary aquifers. However, due to saltwater intrusion into the Floridan aquifer, the water bearing potential and impacts of development of the upper and lower Brunswick aquifers are being evaluated. A brief description of each of these aquifer systems is presented below. A generalized correlation of geologic and hydrogeologic units in Coastal Georgia is shown on Figure 1.

The surficial aquifer consists of interlayered sand, clay, and thin limestone beds of Pliocene to Pleistocene age. The lower surficial aquifer is not always present, but where found includes the uppermost sand unit of Miocene age deposits. The aquifer is generally unconfined, but can be locally confined above by thick clays. This is the case in Glynn County, where as much as 100 feet of lower surficial sands and gravels were observed below a clay confining unit separating the overlying unconfined surficial aquifer near Anguilla. The aquifer ranges from 65 feet to greater than 380 feet thick and is generally confined below by dense phosphatic limestone or dolomite, and phosphatic silty clay. Wells screened within this aquifer yield in the range of 2 to greater than 180 gallons per minute (gpm). The surficial aquifer is utilized primarily for small domestic uses, such as lawn irrigation and rural drinking water supplies.

The upper and lower Brunswick aquifers are formed by two Miocene depositional sequences. Each sequence forms a geologic unit consisting of a basal carbonate layer, a middle clay layer, and an upper sand layer. The upper sand layers represent the upper and lower Brunswick aquifers (Clarke, Hacke, and Peck, 1990). Limestone within the basal carbonate layers also contribute water to the lower Brunswick in some areas.

The upper Brunswick aquifer consists of poorly sorted, fine to coarse, slightly phosphatic and dolomitic quartz sand. The aquifer ranges in thickness from less than 20 feet in Effingham and

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 3 -

January 2003 003-3020

Chatham counties to 150 feet in Wayne County (Clarke et. al., 1990). In general, the thickness of the upper Brunswick increases toward the south-south east along the Georgia coastline. Typically the upper Brunswick aquifer is separated from the overlying surficial aquifer by a confining unit consisting of phosphatic silty/sandy clay or limestone. Because few individual wells have tapped into the upper Brunswick there is limited hydrologic data available for this aquifer. As of 1990 there had been no documented pumping tests performed on the upper Brunswick.

The lower Brunswick aquifer also consists of poorly sorted, fine to coarse, slightly phosphatic and dolomitic, quartz sand in most of coastal Georgia, however, in Glynn County, the aquifer can be comprised of sandy limestone. This unit is absent in the Savannah, GA area because it has been eroded away or was never deposited. The thickness of the lower Brunswick ranges from about 36 feet to 70 feet (Clarke et. al., 1990). As with the upper Brunswick, the lower Brunswick aquifer thickens to the south and is thickest in the Glynn County area. A confining layer consisting of silty clay and phosphatic limestone or dolomite separates the upper and lower Brunswick aquifers.

The lower Brunswick aquifer is absent throughout most of Chatham and Effingham counties and occurs sporadically throughout Bryan County. Analyses of pumping test data from test wells in Bryan County wells yielded transmissivity values ranging from 90 ft.2/day to 2700 ft.2/day (Jordan, Jones and Goulding, 1999). Well yields from the lower Brunswic k in Glynn County can be as much as 1000 gpm. Transmissivity values range from less than 1400 ft2/day to 4700 ft2/day and hydraulic conductivity values range from 20 ft/day to 60 ft/day (Clarke et. al., 1990). Several of the wells installed in the lower Brunswick aquifer in Glynn County are flowing artesian wells. Based on these data, the lower Brunswick aquifer in Glynn County is expected to have greater groundwater production potential relative to the other three counties.

The underlying Floridan aquifer system consists of interbedded clastics and marl in the updip area (upper coastal plain) and massive limestones and dolostones in the downdip area (near-coastal region). The aquifer system consists of several formations that range in age from Eocene to Oligocene. Principal units that comprise the system are the Oldsmar and Avon Park Formations and the Ocala, Santee, and Suwannee Limestones. In the Brunswick area, rocks of Paleocene age (Cedar Keys Formation) and the Late Cretaceous age are also part of the system.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 4 -

January 2003 003-3020

The aquifer system developed from this continuous carbonate sequence is hydraulically connected in varying degrees. Zones of low permeability and aerial continuity exist within the Floridan system and separate it into two permeable, water-bearing zones; the Upper Floridan aquifer and the Lower Floridan aquifer. Semi-confining silty clay and phosphatic limestone or dolomite units from the Miocene deposits separate the Floridan from the lower Brunswick.

Hydrographs from existing wells screened in the upper and lower Brunswick aquifers exhibit water level fluctuations and trends similar to one another, indicating a hydraulic connection between the aquifers. Likewise, the water levels fluctuate in response to seasonal climatic changes and regional pumping from the Upper Floridan aquifer, indicating some degree of communication with overlying and underlying aquifers. The degree of interconnection is largely dependent upon the thickness of the confining layers separating the aquifers, their hydraulic conductivity, and the presence or absence of discontinuities with higher transmitting properties compared to the surrounding material.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 5 -

January 2003 003-3020

2.0 AQUIFER ASSESSMENT FIELD PROGRAM BRYAN COUNTY

The Bryan County Site (Site) selected by EPD is located on undeve loped land near the town of Belfast off Belfast-Keller Road, approximately two miles west of Georgia Highway 144 and two miles east of Interstate I-95 (Figure 2). The general Site layout is shown on Figure 3. The aquifer assessment program at this Site consisted of the following activities:

Drilling and installation of a test well and corresponding observation well in the upper Brunswick aquifer;
Performance of a 72-hour pumping test in the test well to obtain data to adequately characterize the water-bearing properties of the aquifer;
Analysis of pumping test data using analytical methods to calculate specific aquifer parameters such as transmissivity, storativity, hydraulic conductivity, etc. in the vicinity of the wells;
Based on parameters derived from aquifer tests, forward modeling was performed using numerical simulation techniques to assess the sustainable yield of the aquifers; and
Drilling, geophysical logging, and subsequent abandonment of a lower Brunswick aquifer pilot hole.

2.1 Drilling and Well Installation A 4-inch diameter pilot hole was advanced using mud rotary techniques into the Miocene deposits down to approximately 305 feet below ground surface (ft.bgs). The cuttings were logged during drilling by a qualified geologist and provide the basis for the lithologic logs presented in Appendix A. At completion of the drilling, geophysical surveys were conducted and included resistivity and gamma logs (Appendix A). Geophysical marker horizons depicted on gamma logs within the Miocene deposits were used to identify the depositional sequences within the Miocene strata (Clarke et. al., 1990). These data, along with resistivity data, were used to identify the location of the upper Brunswick aquifer and select the appropriate screen intervals for the pumping and observation wells.

The upper Brunswick aquifer consisted of brown and black, well rounded to sub-rounded, fine to coarse phosphatic sand with some olive green clay and quartz sand, shell fragments and limestone. The sandy limestone unit was encountered from approximately 230 to 247 ft.bgs, and is confined above and below by a dark to light olive green and silty clay. A generalized boring log describing the major lithologic units encountered and relative geophysical surveys is shown on Figure 4.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 6 -

January 2003 003-3020

Once the target screen interval was selected, the pilot hole was plugged to the base of the upper Brunswick aquifer such that communication with the lower Brunswick aquifer would not occur. The pilot hole was then reamed with an 11.88-inch diameter bit to approximately 50 ft. bgs and 10-inch diameter surface casing was installed. The remainder of the pilot hole was then reamed. Six inch diameter, schedule 40 PVC well riser pipe and 40-feet of 20-slot PVC screen were installed to a total depth of 275 ft.bgs, with a five foot sump at the bottom of the well and a screened interval of 230 - 270 ft.bgs. A sand pack was then installed from approximately 218 to 280 ft.bgs. Bentonite chips were used to fill the annular space to approximately 40 ft.bgs. Sand was added to the annular space to the surface. This well was labeled BB-1 (USGS Grid Number 35N069).

A 4-inch diameter Schedule 40 PVC observation well with the same screened interval was then installed approximately 100 feet from the pumping well using the same approach described above. The same lithology was encountered in the boring for this well. This well is referred to as BB-2 (USGS Grid Number 35N070). The wells were developed by repeatedly surging and airlifting until clear water was produced and specific conductance measurements had stabilized.

Based on the results from the geophysical survey, it was initially interpreted by both USGS and Golder that the pilot hole had been advanced to the top of the Upper Floridan Aquifer and that the lower Brunswick aquifer was not present at this location. However, additional information from other coastal drilling investigations indicated that the presence of early Miocene and late Oligocene-aged limestone units might have been misinterpreted as the Upper Floridan Aquifer in both the geophysical survey and the drill cuttings. Given these additional data, a second pilot hole (BB-3) was installed approximately 50 feet from the original pilot hole (BB-1) during June 2002. A 6.75-inch diameter drill bit was advanced using mud rotary techniques through the Miocene deposits down to approximately 420 ft. bgs. The cuttings were logged during drilling by a qualified geologist and provide the basis for the lithologic log presented in Appendix A.

At completion of the drilling, lithologic descriptions were reviewed and geophysical surveys (including resistivity and gamma logs) were conducted (Appendix A). Geophysical marker horizons depicted on gamma logs within the Miocene deposits were used to identify the depositional sequences within the Miocene strata (Clarke et. al., 1990). These data, along with resistivity data, indicated that the lower Brunswick aquifer was present at this location, and consisted of light and dark green well-rounded to sub-rounded, fine to coarse, quartz sand and

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 7 -

January 2003 003-3020

clay with some phosphatic sand and little sandy limestone. The unit was encountered from approximately 302 to 345 ft.bgs and is confined above by a dark-green silty clay and is semiconfined below by a white and dark green partially lithified limestone with some silty clay. However, there does not appear to be a significant confining unit between the lower Brunswick and Upper Floridan aquifers at this location. The top of the upper Floridan aquifer, based on cuttings and correlation with surrounding wells, was encountered at approximately 418 ft.bgs.

Upon review of the geophysical survey data by both USGS and Golder, it was determined that although the lower Brunswick aquifer was present at this location, the water-bearing potential of the aquifer did not appear to be significant based on the unit thickness and the resistivity signature. Well installation and testing were not included in the scope of work for this second pilot hole investigation. Therefore, the pilot hole was backfilled to the surface with a bentonitecement grout such that communication between all of the penetrated aquifers would not occur. This pilot hole was labeled BB-3.

Generalized boring logs describing the major lithologic units encountered and relative geophysical surveys are shown on Figures 4 and 5. More detailed boring logs for the wells are included in Appendix A, as well as a summary table of well construction data for all wells installed during the Brunswick aquifer investigation. A generalized cross-section originally presented by Clarke et. al. (1990) and modified by Golder to incorporate these new test wells is shown on Figure 6.

2.2 Aquifer Testing The main objective of the aquifer test program was to derive accurate hydraulic parameters to characterize the upper Brunswick aquifer at the Site. No test wells were installed within the lower Brunswick aquifer at this location. The aquifer testing program for the upper Brunswick aquifer was consistent with the specifications set forth in the EPD's Scope of Work Document and sufficient to derive values for hydraulic conductivity (K), transmissivity (T), storativity (S), leakage (b), specific capacity (Q/s), and the anticipated safe yield for the aquifer. To increase the reliability of the calculated aquifer parameters, the wellbore storage and skin influenced portion of the data was identified and distinguished from the undisturbed formation properties. In addition to the derivation of standard aquifer parameter values, the testing and analysis program provided data that allowed Golder to determine:

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 8 -

January 2003 003-3020

the degree of heterogeneity from changes in unit thickness (or variation in aquifer properties) away from the well or from no-flow boundary effects; and
the presence or lack of recharge from vertical infiltration from overlying and underlying units (or possibly from hydraulic connection to surface water bodies).

Meeting these objectives required a good test design using appropriate equipment and high resolution analysis techniques. The testing equipment included the following:
A properly sized pump to adequately stress the aquifer and create sufficient drawdown in the observation well;
An adequate power source (generator) to supply uninterrupted power to the pumps;
Flow measurement device to accurately measure pumping rate (Q), which was maintained within +/- 5% of the desired Q;
Pressure transducers to measure water levels inside the pumping well and in the observation well and at the surface to measure barometric pressure fluctuations;
Dataloggers to define measurement intervals and record transducer readings during the tests; and
Water level indicators to take manual measurements of water levels and verify transducer readings.
A constant rate pumping test was conducted at the test site to provide data for analysis of aquifer parameters and estimates of sustainable yields in the upper Brunswick. Prior to the pumping test, water levels were monitored in the pumping well to determine static water levels and assess ambient potentiometric trends. Furthermore, barometric pressure readings at the surface were monitored to assist in analyzing the transducer data. These data were corrected accordingly prior to analysis as discussed in Section 3.1.

Pumping well BB-1 was pumped at a constant rate of approximately 15 gpm for a period of 72 hours, with a total resultant drawdown of approximately 57 feet and a specific capacity of 0.26 gpm/foot. The flow rate was closely controlled using a butterfly valve and flow meter assembly, which allowed for small flow adjustments to be made to compensate for head loss in the well column and subsequent flow rate loss over the course of the test period. The monitoring frequency of water level measurements in the wells was adjusted throughout the test to obtain an adequate number of early, middle and late time data points for analysis. The dataloggers were

GOLDER ASSOCIATES INC.

GA Environmental Protection Division

Dr. William McLemore

- 9 -

January 2003 003-3020

synchronized in time and programmed to collect water levels on a log frequency scale. The transducer data were backed up with manual water level measurements collected periodically throughout the test. Data were plotted during the course of the test to detect any anomalies and to indicate when equilibrium conditions were reached. These data are presented in both written and graphical form in Appendix B.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 10 -

January 2003 003-3020

3.0 AQUIFER TEST ANALYSES

Methods for analyzing the aquifer test data were developed to provide appropriate aquifer parameter estimations in order to facilitate an accurate and comprehensive evaluation of the water-bearing characteristics of the upper Brunswick aquifer at the Bryan County Site. The general approach used in performing these analyses was to:

Correct the aquifer test data (if warranted) for barometric fluctuations and significant trends;
Perform standard curve matching analyses assuming a homogeneous model that is infinite in aerial extent;
Perform derivative analyses (using more complicated flow models if necessary) to match the entire test data;
Identify aquifer boundary conditions and geometry (e.g. narrowing, thickening, recharge boundary, etc.) based on the pressure derivative analyses; and
Develop a basic three-dimensional numerical flow model using MODFLOW to estimate the anticipated yield of the Brunswick Aquifer(s) in the vicinity of each Site.

3.1 Data Compilation and Correction The pumping test data were compiled into a series of time versus drawdown tables and graphical plots (Appendix B). Raw pressure head data were converted to water level below top of casing values. Significant potentiometric trends were not apparent in the pre-test ambient water level data.

Barometric effects were identified and factored into the analyses using the following methodology. The initial barometric pressure at the start of each test was measured using a barometric pressure transducer at the surface. The relative changes in barometric pressure throughout the test period were recorded at the same frequency as the water level data. The relative change in pressure (positive or negative) was converted into consistent units and subtracted from the raw water level data, resulting in a very similar curve to that produced by the original data, as shown on Figure 7. As a test case, the non-corrected and corrected curves were both analyzed and were in agreement. The small diurnal changes (less than 0.2 feet) manifested in the background data were considered insignificant given the total drawdown in the pumping and observation wells.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 11 -

January 2003 003-3020

3.2 Test Analysis Software and Application The data were analyzed using three separate software packages. Waterloo Hydrogeologic, Inc.'s AquiferTest software package was used to perform initial parameter estimates with conventional curve-matching solution techniques such as Theis (1935) and Cooper-Jacob (1946). These results were used to finalize the governing assumptions (confined, leaky confined, etc.) by examining the quality of fit of the data to the standard type curves.

The selection of software packages for pressure derivative analyses was dependent on the observed formation response. In general, the tests that showed homogeneous formation response or small changes in hydraulic properties away from the borehole were best suited for Golder's FLOWDIM. Test data that showed more complicated formation response and/or boundary effects or significant changes in pumping rates were analyzed with INTERPRET/2 of Scientific Software Intercomp Ltd. (SSI). A brief description of the software packages is presented below.

FLOWDIM (version 2.14) is a software package developed by Golder for aquifer test analysis that has been verified in assessment of potential nuclear repository investigations. All test types may be analyzed including constant rate, constant head and slug phases. Both homogeneous and composite flow models may be used to interpret the data. Flow geometry may also be matched to infer the local connectivity of a fracture network or changes in the thickness of the aquifer away from the borehole. This package also includes the derivative of pressure (i.e., rate of pressure change) with respect to the natural logarithm of time that has shown to significantly improve the diagnostic and quantitative analysis of slug and constant-rate pumping tests (Spane and Wurstner, 1993).

INTERPRET/2 is an interactive program that uses a constant rate solution to provide optimized hydraulic parameters for a wide range of potential reservoir models. This program incorporates pressure derivative data to improve the recognition of the formation response beyond the influence of skin effects. The slope of the derivative data is characteristic of flow regimes at different times. This characteristic in the data can be used to select an appropriate model and obtain transmissivity values from the correct radial flow portion of the data. The analysis is iterated between the deconvolution plot and Ramey A, B and C type curves, which emphasize different time regimes. A good match on all plots confirms the reliability in the derived parameters.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 12 -

January 2003 003-3020

Some of the features of INTERPRET/2 include extensive superposition of constant rate events, nonlinear regression and multi-event rate normalized plots. Multi-event plots allow the relevant phases to be presented on single plots to evaluate for consistency of the formation response throughout the test. Additionally, it can accommodate changing wellbore storage and skin between the test periods.

3.3 Standard Curve Matching Analyses As discussed previously, Waterloo Hydrogeologic, Inc.'s AquiferTest software package was used to perform initial parameter estimates with conventional solution techniques such as Theis (1935) and Cooper-Jacob (1946). A homogeneous formation model was used to match the data assuming a finite wellbore for the inner boundary and infinite in lateral extent outer boundary condition. Emphasis was placed on matching the middle time data to derive estimates for transmissivity and storativity.

Based on the lithology encountered at the Bryan County Site, the upper Brunswick aquifer was assumed to be confined. Subsequent analysis of the quality of fit of the observation well drawdown data to the standard Theis curve strengthened this assumption, as no indication of leakage was evident. Figure 8 displays the fit of the drawdown data to the standard Theis curve.

The observation well data were further analyzed using the Cooper-Jacob Time-Drawdown method. Recovery data was also analyzed using the semi-log Theis Recovery method in both the pumping well and the observation well. Similar transmissivity values were derived from these analyses and are in general agreement with the resultant transmissivity from the Theis analysis. These analyses are shown in Appendix B.

3.4 Pressure Derivative Analyses The general pressure derivative analysis approach is summarized below: Select the most accurate input parameters; Diagnose the flow model based on derivative data and constrain the number of possible
solutions; As the solution is non-unique, determine most appropriate flow model based on geology and
other available information; Perform analysis to determine hydraulic parameters; Compute unit parameters by dividing by the saturated thickness; and

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 13 -

January 2003 003-3020

Compare the pumping and observation zone derivative data on a normalized transmissivity plot to evaluate for the consistency of the formation response.

Selection of the appropriate flow model is best accomplished by log analysis using the derivative of the pressure change plotted against time. Use of the derivative increases the resolution in identifying the flow model and greatly improves parameter reliability (Spane and Wurstner, 1993).

Parameters (T, S, etc.) are derived from type curve matching after the selection of the most appropriate model. To increase the reliability of the calculated aquifer parameters, the skin influenced portion of the data was identified and distinguished from the undisturbed formation properties. Also, to the extent practical, the degree of heterogeneity from changes in unit thickness or variation in properties away from the well were evaluated. This was accomplished using pressure derivative analysis methods (Figure 9). The data were also evaluated to determine possible effects from recharge boundaries and from leakage from overlying and underlying units.

The test performed in Bryan County did not exhibit significant leakage or indicate the presence of recharge boundaries within close proximity to the test site. The pressure derivative analysis did indicate some degree of heterogeneity within the aquifer (i.e., increasing transmissivity away from the pumping well) as shown on Figure 10. This is attributed to thickening of the aquifer unit.

As a consistency check, data from some of the pumping tests were analyzed with both software packages (FLOWDIM and INTERPRET/2). The derived transmissivity values were in agreement, indicating that FLOWDIM results are verifiable using a commercially available software package. Figures 9, 10, and 11 present the results of the pressure derivative analyses performed on both the pumping and observation wells.

3.5 Aquifer Test Analysis Results The resultant aquifer parameters derived from the pumping test analyses are summarized below. It is important to note that the transmissivity and storativity values presented below are estimates. Transmissivity values are considered to be accurate well below a factor of two, and are derived from the period of radial flow (between the wellbore storage and skin effects in the early time data and outer boundary effects apparent in the later time data). This period of flow manifests as

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 14 -

January 2003 003-3020

a semi-log straight line (or leveling off) in the derivative data in log-log coordinates, and is therefore easily recognizable. Storativity is a time-dependent parameter and a good match to the data confirms the reliability of the results, with uncertainty considered to be a factor of two or less.

TABLE 1 SUMMARY OF BRYAN COUNTY SITE AQUIFER CHARACTERISTICS

AQUIFER PROPERTIES

UPPER BRUNSWICK AQUIFER

Formation Model Geometry
Inner Boundary

confined, no apparent leakage
homogeneous composite (increasing thickness away from the pumping well)
wellbore storage and skin

Outer Boundary

infinite in lateral extent

Significant recharge and/or discharge boundaries Transmissivity (T)

None 86 - 95 ft.2/day

Aquifer Thickness (b)

10 feet*

Hydraulic Conductivity (K) Storativity (S)
Specific Storage (Ss)

8.6 9.5 ft./day 3.7 x 10-5
3.7 x 10-6 feet -1

Specific Capacity (Q/s)

0.26 gpm/foot

Notes: *Based on the assumption that the majority of water is produced in the 230 to 240 ft.bgs portion of the screened interval. Note that large screened intervals were installed in an attempt to produce as mu ch water as possible. However, for accurate parameter estimation, only the portion of the screened interval that indicated production based on geophysical information was used as the aquifer thickness.

A summary table including data derived from pumping tests performed at all of the Brunswick aquifer investigation sites is included in Appendix B.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 15 -

January 2003 003-3020

4.0 AQUIFER YIELD ANALYS IS

Information obtained from the preliminary Aquifer Assessment Reports, the drilling program, and the aquifer test analyses were used to estimate the anticipated yield for each aquifer tested. A simplified conceptual model incorporating the aquifer parameters derived from this investigation was used to generate a numerical flow model of the upper Brunswick aquifer at the Site using the USGS's groundwater flow code, MODFLOW (McDonald and Harbaugh, 1988). MODFLOW was developed by the USGS for analysis of groundwater flow systems. The model uses a finitedifference approach to approximate the analytical solution of the partial-differential gr oundwater flow equation. MODFLOW is generally considered to be the most widely used and accepted numerical model available.

A two-dimensional model of the aquifer was constructed. Model set-up included the creation of a suitable model grid to represent the aquifer within the area of interest. A telescoping grid (45 by 45 cells) was used to provide a greater level of detail within the Site area. Cell widths were set to incrementally increase towards the perimeter of the grid. This large model grid was required to ensure that the simulated drawdown was not influenced by the model boundaries.

Model input parameters were based on data collected from each site during the testing program. The parameters derived from the aquifer tests are represented for conditions within the radius of influence. Radius of influence is directly dependent on duration of pumping. For the predictive modeling, the radius of influence will extend beyond that which was examined by the aquifer tests. The results of the numerical model are based on the assumption that hydraulic properties and aquifer thickness do not significantly change with distance from the pumping well, and therefore provide rough estimation of drawdown effects at greater distances away from the Site.

The aquifer was modeled as a confined, homogeneous system with infinite lateral extent. Boundaries were simulated using Dirichlet conditions located outside the portion of the model impacted by drawdown. The model was calibrated by simulating the pumping test performed at the Site and matching the model simulated drawdown to drawdown observed in the monitoring well during the pumping test. Based on this calibration, predictive simulations were performed to determine the extent that the aquifer could be developed as a water supply source.

The model was then simulated to generate calculated water levels (heads) over the pumping period. The amount of drawdown caused by the pumping well is determined by comparing the

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 16 -

January 2003 003-3020

simulated heads at a given time to the static heads at the start of the simulation. The resulting drawdown contours represent a cone of depression around the pumping well that expands for a period of time as pumping continues. The cone of depression is dependent on the diffusivity of the aquifer (transmissivity/storativity) and will continue to enlarge until a recharge or no-flow boundary is reached.

The locations of the five-foot and ten-foot drawdown contours relative to the pumping well after one year and ten years of continuous withdrawal were used as evaluation targets. Furthermore, since these data will be used towards a preliminary evaluation of the potential for aquifer development, additional limiting simulation criteria were used in the sustainable yield simulations. These included an assumed well efficiency of 75% and a maximum allowable drawdown of 100 feet (a conservative target representing approximately 50% of the maximum available drawdown at the Site) in the pumping well after ten years of continuous pumping. These limitations provide a relative basis for aquifer development evaluations, and result in conservative estimates of the potential impact from aquifer development.

The calibrated model was simulated under transient conditions with continuous withdrawal over one year and ten year time periods. Given the criteria discussed previously, the model results indicate that the maximum sustainable pumping rate for the upper Brunswick Aquifer at the Bryan County Site is approximately 22 gpm.

This pumping rate resulted in ten feet of drawdown approximately 9,650 feet away from the pumping well and five feet of drawdown 21,000 feet away from the pumping well after continuous pumping over one year. Furthermore, the simulated model results after ten years of continuous pumping show approximately ten feet of drawdown 30,700 feet away and five feet of drawdown approximately 66,500 feet away. Figure 12 shows the distribution of the simulated drawdown contours in the vicinity of the Site. These contours are generated under the assumptions that the aquifer properties are consistent and no significant boundaries are encountered throughout the radius of influence.

This model should be considered a preliminary evaluation tool and is not intended to simulate all aspects of the Brunswick Aquifer System. As such, the calibration and sensitivity analyses performed on the model was limited. However, for the purposes of this investigation, the model provides reasonable estimates of potential drawdown scenarios.

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 17 -

January 2003 003-3020

5.0 SUMMARY AND CONCLUSIONS

Golder performed a drilling investigation and subsequent aquifer testing to determine the potential viability of water resource development of the Brunswick Aquifers near Belfast in Bryan County, Georgia.

The upper Brunswick aquifer consisted of brown and black, well rounded to sub-rounded, fine to coarse phosphatic sand with some olive green clay and quartz sand, shell fragments and limestone. The sandy limestone unit was encountered from approximately 230 to 240 ft.bgs., and is confined above and below by a dark to light olive green and silty clay. The static water level in the upper Brunswick aquifer at this Site is approximately 20 ft. bgs.

The lower Brunswick aquifer consisted of light and dark green well-rounded to sub-rounded, fine to coarse, quartz sand and clay with some phosphatic sand and little sandy limestone. The unit was encountered from approximately 302 to 345 ft.bgs and is confined above by a dark-green silty clay and is semi-confined below by a white and dark green partially lithified limestone with some silty clay. However, there does not appear to be a significant confining unit between the lower Brunswick and Upper Floridan aquifers at this location. The top of the upper Floridan aquifer, based on cuttings and correlation with surroundin g wells, was encountered at approximately 418 ft.bgs.

A pumping well and observation well were installed into the upper Brunswick aquifer. A 72-hour pumping test was completed using a pumping rate of approximately 15 gpm. Analysis of the pumping test data using pressure derivative techniques yielded a transmissivity value of 86 to 95 ft.2/day, a storativity value of 3.7 x 10-5, and indicated a general increase in transmissivity away from the pumping well that is attributed to thickening of the aquifer unit. The range for hydraulic conductivity values (based on a saturated thickness of 10 feet) is between 8.6 - 9.5 ft./day. The specific capacity of 0.26 gpm/foot was calculated based on a pumping rate of 15 gpm and drawdown of 57 feet at the end of 72 hours of pumping. The total head available for pumping is approximately 200 feet based on a static water level of approximately 20 ft.bgs and the top of the unit at 230 ft.bgs.

The long-term performance of the upper Brunswick aquifer was evaluated using numerical methods using MODFLOW. Aquifer parameter values derived from the pumping test were used as model input. Criteria were established such that an estimate of the impact of potential aquifer

GOLDER ASSOCIATES INC.

GA Environmental Protection Division Dr. William McLemore

- 18 -

January 2003 003-3020

development could be assessed. Based on these criteria, the maximum sustainable pumping rate was estimated to be about 22 gpm. The results of this investigation indicate that the upper Brunswick Aquifer at the Bryan County Site appears to be a potentially viable groundwater supply alternative for small domestic users, but is not likely a viable supply for light industrial or small development uses.

No test wells were installed in the lower Brunswick aquifer at this location. However, based on the geophysical signature of the unit, this aquifer is likely more productive than the upper Brunswick aquifer at this location and may be a viable groundwater supply for light industrial or small development uses. The lack of a significant confining unit between the lower Brunswick and Upper Floridan aquifers at this location suggests that vertical leakage between these aquifers may occur.

GOLDER ASSOCIATES INC.

FIGURES

Source:Weems&Edwards,2001

TITLE

GeneralizedCorrelationofGeologicandHydrogeologicUnits

Atlanta, Georgia

CLIENT/PROJECT

DRAWN

GAEPD/MioceneAquiferStudy/GA

CHECKED REVIEWED

JES DATE
SCALE
FILENO.

01-May-2002 JOBNO.
DWGNO.
N/A
SUBTITLE
120.cdr

003-3020.001
REV.NO.
FIGURENO.
1

BB-1 Approx. 100 ft.
BB-3

BB-2

TOBELFAST

LEGEND
UpperBrunswickPumpingWell UpperBrunswickObservationWell LowerBrunswickBoring

CLIENT/PROJECT

EPD/MIOCENE/GA

DRAWN

CHECKED
JES

REVIEWED

BELFAST-KELLER ROAD

TITLE

BRYANCOUNTYSITELAYOUT

Atlanta, Georgia

DATE

SCALE

FILENO.

JOBNO.

DWGNO./REVNO.

FIGURE

04-May-01

NottoScale

003-3020

003-3020

b_site.cdr

3

31U008

200 100
0 -100 -200 -300

USNEDDIFIMFEENRTESNTAINADTEMDIOPCOESNTE-MUIONCITEANE
MIOCENE UNIT B MIOCENE UNIT C PRE-MIOCENE

ELEVATION (FT MSL)

-400

-500

-600

31U008

-700 -800 -900

BULLOCH

32R002

34R032

BRYAN

34P01394P014 35P108

CHATHAM

35N350N35025 36N011 BB-3

37N00637N004

LEGEND

GEOPHYSICAL LOG - GAMMA RADIATION

Adapted from Clarke et al, 1990. Modified by Golder Associates based on project results and USGS archive data.

EPD/MIOCENE/GA

32R002 34R032 34P019 34P014 35P108 35N035 BB-3 35N025 36N011 37N006 37N004

ELEVATION (FT MSL) 35N035 BB-3 35N025

MARKER A MARKER B MARKER C MARKER D
100 0
-100 -200 -300 -400 -500
BRYAN COUNTY CROSS-SECTION

0

10 11/19/2000
20

Begin: 11/26/2000 11:00am

30

40

50

60

70

80

90 0

200000

400000

600000 Time (sec)

CLIENT/PROJECT

GAEPD/Miocene/GA

DRAWN

CHECKED
JES

REVIEWED

DATE

SCALE

10-August-2001

BB2 With Barometric Correction BB1 Pumping Well
BB2 Observation Well
BB1 With Barometric Correction

800000

End: 11/29/2000 11:00pm

1000000

1200000

TITLE

BryanCountyPumpingTest

Atlanta, Georgia

FILENO.

JOBNO.

DWGNO./REVNO.

FIGURE

NA

003-3020

003-3020

bryan_6.cdr

7

CLIENT/PROJECT

GAEPD/Miocene/GA

DRAWN

CHECKED
JES

REVIEWED

DATE

SCALE

10-August-2001

TITLE
Atlanta, Georgia

LogAnalysisof PumpingWell PumpingPhase UpperBrunswickTest BryanCountySite

FILENO.

JOBNO.

DWGNO./REVNO.

FIGURE

NA

003-3020

003-3020

bryan_8.cdr

9

CLIENT/PROJECT

GAEPD/Miocene/GA

DRAWN

CHECKED
JES

REVIEWED

DATE

SCALE

10-August-2001

TITLE

LogAnalysisofObservationWellPumpingPhase

UpperBrunswickTest

Atlanta, Georgia

BryanCountySite

FILENO.

JOBNO.

DWGNO./REVNO.

FIGURE

NA

003-3020

003-3020

bryan_9.cdr

10

CLIENT/PROJECT

GAEPD/Miocene/GA

DRAWN

CHECKED
JES

REVIEWED

DATE

SCALE

10-August-2001

TITLE

Atlanta, Georgia

TransmissivityNormalizedPlotfor UpperBrunswickTestattheBryanCountySite

FILENO.

JOBNO.

DWGNO./REVNO.

FIGURE

NA

003-3020

003-3020

bryan_10.cdr

11

APPENDIX A
WELL CONSTRUCTION DATA AND BORING LOGS

APPENDIX B PUMPING TEST INFORMATION AND ANALYSES